Biophysics For The Life Sciences: Series Editor: Norma Allewell

Biophysics for the Life Sciences
Series Editor: Norma Allewell
For further volumes:

http://www.springer.com/series/10230
Linda O. Narhi
Editor
Biophysics for Therapeutic

Protein Development
Editor
Linda O. Narhi
Research and Development
Amgen, Inc.
Thousand Oaks, CA, USA
ISBN 978-1-4614-4315-5 ISBN 978-1-4614-4316-2 (eBook)

DOI 10.1007/978-1-4614-4316-2
Springer New York Heidelberg Dordrecht London
Library of Congress Control Number: 2013931098
© Springer Science+Business Media New York 2013

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this
publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s
location, in its current version, and permission for use must always be obtained from Springer. Permissions
for use may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to
prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Contents
1 Introduction ............................................................................................. 1
Linda O. Narhi
2 High-Throughput Biophysical Approaches to Therapeutic
Protein Development............................................................................... 7
Feng He, Vladimir I. Razinkov, C. Russell Middaugh,
and Gerald W. Becker
3 Techniques for Higher-Order Structure Determination ..................... 33
James Kranz, Fatma AlAzzam, Atul Saluja, Juraj Svitel,
and Wasfi Al-Azzam
4 Biophysical Techniques for Protein Size Distribution Analysis .......... 83
Ziping Wei and Alla Polozova
5 Qualification of Biophysical Methods for the Analysis
of Protein Therapeutics .......................................................................... 99
Yijia Jiang, Cynthia Li, and John Gabrielson
6 Application of Biophysics to the Early Developability
Assessment of Therapeutic Candidates and Its Application
to Enhance Developability Properties ................................................... 127
Hasige Sathish, Nicolas Angell, David Lowe, Ambarish Shah,
and Steven Bishop
7 Application of Biophysics in Formulation, Process,
and Product Characterization: Selected Case Studies ........................ 147
Satish K. Singh, Qin Zou, Min Huang, and Muralidhara Bilikallahalli
8 Biophysical Analysis in Support of Development of Protein
Pharmaceuticals ...................................................................................... 173
Sreedhara Alavattam, Barthelemy Demeule, Jun Liu, Sandeep Yadav,
Mary Cromwell, and Steven J. Shire
v
vi Contents
9 Case Studies Applying Biophysical Techniques to Better

Characterize Protein Aggregates and Particulates
of Varying Size ......................................................................................... 205
Tingting Wang, Sangeeta B. Joshi, Ozan S. Kumru,
Srivalli Telikepalli, C. Russell Middaugh, and David B. Volkin
10 Investigation of Nonconformance During Protein Therapeutic
Manufacturing......................................................................................... 245
Zai-Qing Wen, Gianni Torraca, Guiyang Li, Xiaolin Cao,
and Chanel K. Yee
11 Higher-Order Structure and Protein Aggregate Characterization
of Protein Therapeutics: Perspectives from Good Manufacturing
Practices and Regulatory Guidance ...................................................... 261
Evi B. Struble, John F. Cipollo, Chava Kimchi-Sarfaty,
Zuben E. Sauana, Jack A. Ragheb, and Ewa Marszal
Index ................................................................................................................. 283

Contributors
Fatma AlAzzam TechnoPharmaSphere LLC, Downingtown, PA, USA

Wasfi Al-Azzam Biopharmaceuticals Development, GlaxoSmithKline, King of
Prussia, PA, USA
Nicolas Angell Process and Product Development, Amgen Inc., Thousand Oaks,
CA, USA
Gerald W. Becker Drug Product Development, Amgen Inc., Seattle, WA, USA
Steven Bishop Formulation Sciences, Biopharmaceutical Development,
MedImmune, Gaithersburg, MD, USA
Xiaolin Cao Department of Process and Product Development, Amgen Inc.,
John F. Cipollo Division of Bacterial, Parasitic and Allergenic Products, Office of
Vaccine Research and Review, CBER, FDA, Bethesda, MD, USA
Mary Cromwell Late Stage Pharmaceutical Development, Genentech, South San
Francisco, CA, USA
Barthelemy Demeule Late Stage Pharmaceutical Development, Genentech, South
San Francisco, CA, USA
John Gabrielson Analytical Sciences, Amgen Inc., Longmont, CO, USA
Feng He Drug Product Development, Amgen Inc., Seattle, WA, USA
Min Huang Pfizer Inc., Andover, MA, USA
Yijia Jiang Process and Product Development, Amgen Inc., Thousand Oaks, CA,
USA
Sangeeta B. Joshi Department of Pharmaceutical Chemistry, Macromolecule and
Vaccine Stabilization Center, University of Kansas, Lawrence, KS, USA
vii
viii Contributors
Chava Kimchi-Sarfaty Division of Hematology, Office of Blood Research and

Review, Center for Biologics Evaluation and Research (CBER), Food and Drug
Administration (FDA), Bethesda, MD, USA
James Kranz Biopharmaceuticals Development, GlaxoSmithKline, King of Prussia,
PA, USA
Ozan S. Kumru Department of Pharmaceutical Chemistry, Macromolecule and
Cynthia Li Process and Product Development, Amgen Inc., Thousand Oaks, CA,
USA
Guiyang Li Department of Process and Product Development, Amgen Inc.,
Jun Liu Late Stage Pharmaceutical Development, Genentech, South San Francisco,
CA, USA
David Lowe MedImmune Ltd, Cambridgeshire, UK
Ewa Marszal FDA, Center for Biologics Evaluation and Research, Bethesda,
MD, USA
C. Russell Middaugh Department of Pharmaceutical Chemistry, Macromolecule
and Vaccine Stabilization Center, University of Kansas, Lawrence, KS, USA
Bilikallahalli Muralidhara MedImmune, Gaithersburg, MD, USA
Linda O. Narhi Process and Product Development, R&D, Amgen Inc., Thousand
Oaks, CA, USA
Alla Polozova Biopharmaceutical Development, MedImmune, Gaithersburg, MD,
USA
Jack A. Ragheb Division of Therapeutic Proteins, Office of Biotechnology
Products, Center for Drug Evaluation and Research, FDA, Bethesda, MD, USA
Vladimir I. Razinkov Drug Product Development, Amgen Inc., Seattle, WA,
USA
Atul Saluja Drug Product Science & Technology, Bristol-Myers Squibb,
New Brunswick, NJ, USA
Hasige Sathish Formulation Sciences, Biopharmaceutical Development,
MedImmune, One Medimmune Way, Gaithersburg, MD, USA
Zuben E. Sauna Division of Hematology, Office of Blood Research and Review,
Center for Biologics Evaluation and Research (CBER), Food and Drug
Ambarish Shah Formulation Sciences, Biopharmaceutical Development,
MedImmune, Gaithersburg, MD, USA
Contributors ix
Steven J. Shire Late Stage Pharmaceutical Development, Genetech, South San

Francisco, CA, USA
Satish K. Singh Pfizer Inc., Chesterfield, MO, USA
Sreedhara Alavattam Late Stage Pharmaceutical Development, Genentech,
South San Francisco, CA, USA
Evi B. Struble Division of Hematology, Office of Blood Research and Review,
Juraj Svitel Process and Product Development, Amgen Inc., Thousand Oaks,
CA, USA
Srivalli Telikepalli Department of Pharmaceutical Chemistry, Macromolecule and
Gianni Torraca Department of Process and Product Development, Amgen Inc.,
David B. Volkin Department of Pharmaceutical Chemistry, Macromolecule and
Tingting Wang Department of Pharmaceutical Chemistry, Macromolecule and
Ziping Wei Analytical Development, Novavax, Rockville, MD, USA
Zai-Qing Wen Department of Process and Product Development, Amgen Inc.,
Sandeep Yadav Late Stage Pharmaceutical Development, Genentech, South San
Francisco, CA, USA
Chanel K. Yee Department of Process and Product Development, Amgen Inc.,
Qin Zou Pfizer Inc., Chesterfield, MO, USA
Chapter 1
Introduction
Linda O. Narhi
The last few decades have seen the evolution of protein therapeutics, from a few
drugs isolated from natural sources (such as insulin) to numerous engineered mol-
ecules that are designed to target specific diseases; these biotherapeutics comprise
an increasingly larger part of the commercial drug market (Walsh 2010; Pavlou and
Belsey 2005). While more difficult to make than the traditional small molecule
medications, proteins have the advantage of having a sustained half-life under phys-
iological conditions, can be targeted to specific sites, and do not have the same
concerns around metabolites generated during clearance in vivo, resulting in fewer
side effects. Initial protein therapeutics were hormones and growth factors that were
designed, using the new tools of molecular biology, to replace the native proteins
that were missing or nonfunctional in certain disease states. As the field and industry
matured, the focus shifted to designing molecules to act as antagonist or agonist
targeted to specific components (such as receptors or ligands) on a particular cell
type. Targeting to receptors commonly found on cancer cells has resulted in very
effective oncology treatments with fewer of the debilitating side effects of the tradi-
tional chemotherapy treatments. Inflammation is another area that has seen the
generation of effective biotherapeutics targeted to turn off the inflammatory response
of the effected part of the body in autoimmune diseases (for instance the joints in
rheumatoid arthritis). In order for proteins to be used as a drug, they must be stable,
retaining both activity and structure, during isolation and storage for up to 2 years,
often at high concentration. These are conditions that favor protein aggregation and
denaturation.
Proteins are complex macromolecules that are polymers of the 20 amino acids
connected through peptide bonds. This primary sequence is then folded into the
L.O. Narhi ()

Process and Product Development, R&D, Amgen Inc., One Amgen Center Drive,
Thousand Oaks, CA 91320, USA
e-mail: [email protected]
L.O. Narhi (ed.), Biophysics for Therapeutic Protein Development, Biophysics 1

for the Life Sciences 4, DOI 10.1007/978-1-4614-4316-2_1,
2 L.O. Narhi
secondary, tertiary, and quaternary structure required for function. The secondary
structure consists of alpha helix, beta sheet, beta turns, or unstructured conforma-
tions which are defined by the configuration of the peptide backbone, while the
tertiary structure is defined as the overall global protein fold or three-dimensional
structure. Because of this higher order structure, a folded protein is stabilized by
interactions between amino acids that are located far away from each other on the
linear amino acid sequence. These can be hydrophobic, van der Waals or charge-
charge interactions, hydrogen bonds, salt bridges, disulfide bonds, etc.
Proteins can be simple single domain globular proteins like the growth factors
and cytokines that were the first biotech drugs or complex multi-domain molecules
such as the antibodies, the active form of which is actually made of several different
polypeptides held together by disulfide bonds between cysteines. Quaternary struc-
ture is the association of more than one polypeptide chain. This includes association
necessary for the formation of a functioning molecule, like the antibodies and
hemoglobin, association resulting in undesirable nonnative species such as hetero-
geneous partially unfolded protein aggregates that can be formed by stressing the
native monomer, and association into the well-defined but nonnative structures
characteristic of amyloids.
The conformation (overall global fold or three-dimensional structure) of a pro-
tein is important for maintaining biological activity and stability during long-term
storage. It can also impact the safety profile and biological consequences of bio-
therapeutics (Dobsen 2004; Wang et al. 2007). Proteins can be unfolded by many
different types of stress, including extremes of pH, high or low ionic strength and
other solution conditions, heat, and mixing of the air-liquid interface in solution. All
of these conditions can be encountered during the manufacturing, storage, and
administration of the protein therapeutic. The amino acid residues comprising the
molecule can also undergo chemical modifications which might or might not affect
other molecular properties.
Proteins have an inherent affinity to self-associate under the appropriate solution
conditions. This reaction in its simplest form is a concentration-dependent thermo-
dynamic equilibrium, but can also proceed as a complex, multistep reaction that
includes some irreversible steps. A very important irreversible reaction is the forma-
tion or shuffling of disulfide bonds between the monomeric subunits which can
occur under appropriate red/ox conditions. These self-associated species, or aggre-
gates, can range in size from dimer to species containing hundreds of monomers.
Protein aggregates themselves are comprised of a heterogeneous population contain-
ing molecules of different sizes, morphologies, and chemical modifications, formed
through different pathways and intermediates, all determined by the protein sequence
and the conditions to which the protein has been exposed (Narhi et al. 2010).
Even subtle changes in conformation can result in changes in the safety and effi-
cacy of the product, including decreased activity, inability to bind to receptors,
changes in the pharmacokinetics and pharmacodynamic profiles, and the potential
for immunogenicity. Biophysical methods are the principle techniques available to
determine if a potential protein product has the appropriate higher order structure
and can maintain the active conformation during manufacturing, storage, and delivery.
1 Introduction 3
Thus, biophysical techniques play a key role during the development of protein
therapeutics.
The development life cycle for pharmaceutical proteins begins with target iden-
tification and demonstration of the biological activity of an engineered protein. The
next steps are identification of lead therapeutic candidates with this desired activity
and selection of recombinant production cell lines. This is followed by process and
formulation development and characterization and finally clinical trials and then, if
successful, commercialization. The launch of a product changes the focus to prod-
uct consistency and lot release assays, exploration of different delivery devices and
therapeutic indications, comparability assessments, and support for product and
process failure investigations.
Attributes that are important to screen during cell line development include the
amount of monomer and aggregate of the target protein present in the cell culture
media prior to purification. This information will contribute to selection of the final
commercial cell line that delivers optimal yield and product quality. For selection of
the product candidate itself, in addition to biological activity, the ideal molecule will
need to be stable to process and storage conditions, including low pH for viral clear-
ance if it is a mammalian cell line-derived molecule or refolding conditions if it is
an Escherichia coli-derived protein in the form of inclusion bodies. The protein
therapeutic also needs to withstand storage in solution at 4–8°C for 2 years, often at
protein concentrations above 100 mg/mL (Mahler et al. 2009). Screening for this
type of stability usually involves stressing the material and then assessing the impact
of the stress on the integrity of the protein, with particular emphasis on protein
aggregation and irreversible unfolding of the native three-dimensional structures.
The requirements for the biophysical assays used to characterize therapeutic pro-
teins vary with stage of development. During the early stages of drug product devel-
opment, there is very little of the therapeutic protein available and little time to
perform the assessments; a relative ranking of the proteins or cell lines is often
sufficient to select the candidate to move forward into development. Therefore,
assays with minimal material requirements and high throughput are especially valu-
able, and categorizing candidates into pass or fail is an acceptable output.
Once the molecule to be developed as a therapeutic has been chosen, the focus
becomes developing the production process and formulation to be used for the com-
mercial product. At this stage material availability is no longer rate limiting, and
more rigorous techniques comparing the higher order structure of the actual mate-
rial obtained during the different processing steps can be used to ensure that the final
product was not irreversibly damaged by the conditions being used for its
manufacture.
During formulation development, the stability of the target protein is assessed in
different buffer compositions, pH, storage conditions, delivery devices, etc. These
studies typically involve the generation of many samples which need to be analyzed
in order to arrive at the optimal formulation conditions. Many of the principles that
apply during candidate selection apply here as well, and some of the same assays
can even be used. Actual long-term stability studies are also performed to demon-
strate that the protein drug in the formulation chosen has the desired shelf-life
4 L.O. Narhi
stability and stability to administration by the chosen delivery method. Determination

of conformational integrity and amount of protein aggregation are important aspects
of these studies, which usually involve more rigorous characterization and quantita-
tive analyses.
As the product moves into production, the emphasis switches from developing
and optimizing conditions to maintaining process/product control; biophysical tech-
niques to follow the protein higher order structure are important parts of compara-
bility studies and are required for obtaining licensure of the drug. In this case the
methods must be shown to be fit for purpose, and the sensitivity of the assays to
detect changes in the product needs to be determined. Another important aspect of
preclinical and clinical development is the monitoring of stability samples, stored
both under accelerated and recommended conditions, for comparability. Biophysical
techniques are also used as tools to help ensure that changes in device, concentra-
tion, and formulation made as different indications or patient populations are added
do not affect the conformation of the biotherapeutics. Techniques which can give
reproducible and accurate results, and where the readout is understood, are most
commonly used at this stage of development rather than the high throughput tests
that were employed in the beginning of the product development life cycle. These
analyses must be sensitive to changes in the protein conformation that can occur if
the protein is exposed to slightly different process or storage conditions, as demon-
strated by samples exposed to conditions outside the normal parameters.
Finally, during commercial production, batches occasionally fail the various lot
release assays; biophysical techniques can be used o help identify root cause and
contribute to the safety assessment of lots. In this case, very often, a single visible
aggregate is being studied and so the methods must have the sensitivity to detect and
analyze a very small amount of protein and provide a positive ID of the material if
possible.
In this volume, organized to follow the product life cycle, the applications of
biophysical analyses during the therapeutic protein development are presented.
The initial chapters describe the underlying theory, strengths, and weaknesses of the
different biophysical techniques commonly used during protein therapeutic devel-
opment, while the later chapters present case studies to illustrate the application of
biophysics to protein therapeutic development. Chapter 2 describes the strategy,
development, and implementation of high throughput methods, while Chap. 3 pro-
vides an overview of the biophysical techniques commonly used for higher order
structure analysis during the development of a biotherapeutic. One of the key prod-
uct quality attributes that are be tracked during protein processing and storage is
protein aggregation. The methods being developed and implemented for particle
size distribution analysis and aggregate characterization are discussed in Chap. 4.
And finally Chap. 5 describes the qualification of biophysical methods for the char-
acterization of protein higher order structure and demonstration of fit for purpose of
these methods. The sensitivity and limits of the methods must be demonstrated in
order for them to be used in licensure filings, during characterization, and for com-
parability studies.
1 Introduction 5
Chapters 6–10 present case studies organized to follow the product life cycle.
Chapter 6 demonstrates the use of higher order structure analysis during discovery
and candidate selection, Chap. 7 and 8 illustrates the use of biophysical techniques
during process and product development. Chapter 9 focuses on characterization of
protein aggregates, one of the key product attributes that have the potential to effect
product safety. Chapter 10 contains several case studies demonstrating how bio-
physical techniques are used to identify particles and sources of product and process
failure investigations. And finally Chap. 11 presents a regulatory perspective on
higher order structure analysis across the life cycle. It is our hope that this volume
will enhance the readers understanding and appreciation for the important role that
biophysics plays in successful therapeutic protein development.
References
Dobsen CM (2004) Principles of protein folding, misfolding, and aggregation. Semin Cell Dev
Biol 15:3–16
Mahler H-C, Friess W, Grauschopf U, Kiese S (2009) Protein aggregation: pathways, induction
factors and analysis. J Pharm Sci 98:2909–2934
Narhi LO, Jiang Y, Deshpande R, Kang S, Shultz J (2010) Approaches to control protein aggrega-
tion during bulk production. In: Wang W, Roberts CJ (eds) Aggregation of therapeutic proteins.
Wiley, Hoboken
Pavlou AK, Belsey MJ (2005) The therapeutic antibody market to 2008. Eur J Pharm Biopharm
59:389–396
Walsh G (2010) Biopharmaceutical benchmarks. Nat Biotechnol 28:917–924
Wang W, Singh S, Zeng DL, King K, Nema S (2007) Antibody structure, instability and formula-
tion. J Pharm Sci 96:1–26
Chapter 2
High-Throughput Biophysical Approaches
to Therapeutic Protein Development
Feng He, Vladimir I. Razinkov, C. Russell Middaugh, and Gerald W. Becker
2.1 Introduction
Protein therapeutic products typically experience many development cycles, in

which decisions are empirically derived concerning the identity, manufacturing pro-
cess, final product presentation, and administration methods. To obtain a quality
therapeutic product, significant resources are spent on the search for appropriate
development parameters throughout a product’s life cycle. Due to the length and
complexity of such development in the pharmaceutical industry, any process that
can reduce the amount of time and resources while still reaching acceptable out-
comes is highly preferred. One prominent area of such improvement is to increase
the throughput capability of existing technologies. The focus of this chapter is on
providing an introduction to the biophysical methods and techniques possessing
high-throughput capabilities. We will especially focus on those techniques that are
frequently used in the development of protein therapeutics. Though the definition of
high throughput can be quite broad, the technologies and methodologies highlighted
in this chapter all possess multi-sample throughput capability and/or automation-
enabled measurements and analysis.
During the development cycle of protein therapeutics, it is critical to understand
the physical integrity and other structural characteristics of the product. In the early
stages of product development, preferred physical properties are often used as
F. He () • V.I. Razinkov • G.W. Becker

Drug Product Development, Amgen Inc., 1201 Amgen Court West,
Seattle, WA 98119, USA
e-mail: [email protected]; [email protected]; [email protected]
C.R. Middaugh
Department of Pharmaceutical Chemistry, University of Kansas,
2030 Becker Dr., Lawrence, KS 66047, USA

8 F. He et al.
criteria to select protein candidates (described in Chap. 6). The selection philosophy
can be based on an established relationship between a particular biophysical
property and product quality, or simply follow the general assumption that better
biophysical characteristics can lead to more stable products if other methods cannot
differentiate the drug candidates. Typically, early stage development is conducted at
small scale employing in vivo and in vitro systems because protein availability is
limited. Thus, methods that consume less material but generate useful information
are the most attractive. Once a protein drug candidate is selected, development is
primarily focused on optimizing the parameters that effectively enable the manufac-
turing, packaging, storage, and delivery of the final product. Before these parame-
ters can be finalized for commercial processes, however, numerous analyses need to
be performed to define the space and limitations of physical conditions that best fit
the product. High-throughput biophysical tools often play an important role in these
efforts by providing a faster readout when the protein product is subjected to a range
of experimental conditions.
Two areas during protein therapeutic development frequently employing high-
throughput biophysical analyses are downstream and formulation development.
The goal of downstream development is to find the most suitable purification pro-
cess so that it can be properly scaled to deliver commercial product. Although the
final optimization usually takes place at scales similar to the commercial settings,
small-scale and high-throughput approaches are often utilized to predict protein
behavior. For example, multiwell plate-based chromatographic techniques com-
bined with high-throughput liquid handling instruments can provide a tremendous
amount of information on protein–resin interaction and therefore help guide purifi-
cation design with very affordable material input (Coffman et al. 2008). Although
the ability of a process to perform adequately at a large scale remains to be derived
empirically, small-scale development is critical because it provides opportunities to
test the potential manufacturing steps while varying physical parameters. In fact,
many regulatory authorities, when evaluating new drug product applications, require
such small-scale data that can demonstrate the robustness of the processes as well
as the edges of failure. Besides purification, the formulation development of protein
therapeutics also requires a large number of experimental trials to select suitable
conditions for final product presentation. The general aim is to find a formulation
that will best retain the physical, chemical, and biological properties of the product
while meeting the desired shelf life requirements. More recently, patient conve-
nience and comfort have become major factors that influence formulation develop-
ment. For instance, product formulations compatible with self-administration and
above-freezing storage are preferred when treating chronic diseases. To derive suit-
able formulations, large arrays of factors are typically screened during the develop-
ment phase for their ability to stabilize the drug product. Since it is well established
that physical instability may negatively impact protein therapeutics, biophysical
techniques offer relevant tools to monitor protein conformation in response to for-
mulation and storage conditions. An important area of focus during formulation
development is the evaluation of protein aggregation propensity. Protein aggregates
are thought to often impose detrimental effects on the therapeutic potency and side
2 High-Throughput Biophysical Approaches to Therapeutic Protein Development 9
effect profile of the drug which may even lead to significant clinical safety concerns
(Jiskoot et al. 2012; Rosenberg 2006). As a result, protein aggregation is fre-
quently used to differentiate formulation candidates. Over the last decade, it has
been well documented in the literature that high-throughput biophysical methods
are very capable of detecting the presence of protein aggregates over a wide size
range (Mach and Arvinte 2011). The feasibility and performance of a protein ther-
apeutic product can also be dependent on other key properties of the molecule. An
emerging example is protein viscosity. To reach the desired bioavailability, protein
drugs delivered via the subcutaneous route typically require high concentrations.
This often leads to high solution viscosity that may cause significant difficulties
during product manufacturing and administration of the protein (Shire et al. 2004).
Traditional analytical methods have limitations in material consumption and
throughput (Jezek et al. 2011), but newly developed biophysical techniques offer
significant advantages, including a reduction of sample volume as well as increases
in throughput capability (He et al. 2010a; Wagner et al. 2012).
In the following sections, the role of high-throughput biophysical analysis in the
development of protein therapeutics is discussed. Technical background for selected
biophysical methods is reviewed as well as their high-throughput utility. In addition,
a general introduction to empirical phase diagram (EPD) is presented as an example
of the application of high-throughput biophysical characterization and data inter-
pretation in the development of protein therapeutics.
2.2 High-Throughput Biophysical Techniques That Can Be

Applied to Protein Therapeutic Development
Due to increasing demands in sensitivity, diverse sample applicability and through-

put, biophysical instrumentation has been transformed significantly over the past
decade to implement multiwell measurements and automation modules. Such
advancement has accelerated the research and development of protein-based thera-
peutic entities by decreasing the resource requirement and widening the experimen-
tal design space. In this section, a brief introduction is provided, concerning the
background and utility of selected high-throughput techniques.
2.2.1 Surface Plasmon Resonance
Characterization of target binding is an essential component during the develop-

ment of protein therapeutics. Since the majority of therapeutic targets bind to spe-
cific targets, the ability of protein therapeutics to interact with their respective
sites of interaction is often used to differentiate among drug candidates. Many
high-throughput biophysical tools are available to characterize the binding con-
stant, Kd, and stoichiometry, (n), between protein product and therapeutic targets.
10 F. He et al.
Protein–protein interactions can be characterized with the help of surface plasmon

resonance (SPR) (Rich and Myszka 2000). This method is based on the measure-
ment of changes in refractive index near a sensor surface caused by the binding or
dissociation of a protein with its target (receptor). SPR instruments evaluate pro-
tein binding in real time without labeling, but surface immobilization of at least
one component is required. SPR methods can be both qualitative and quantitative
(Karlsson 2004). In the standard setup, measurements are performed sequentially
with two flow cells: one that holds the sample and the second a reference solution.
Recent microfluidic and automation technologies have created new opportunities
for development of high-throughput instrumentation based on the SPR principle.
The Biacore A100 biosensor can process multiple samples (1,000 samples/day)
and allows high-quality data sampling. This instrument has been used to develop
an in vitro high-throughput kinase assay (Takeda et al. 2006). Another SPR instru-
ment, the Biacore Flexchip microarray device, has been used for rapid identifica-
tion of high-affinity human antibodies from a phage display screen. Fab fragment
analysis with surface plasmon resonance microarrays in a high-throughput format
permits the determination of kinetic constants for 96 different Fab fragments in a
single experiment (Wassaf et al. 2006). High-throughput antibody affinity charac-
terization accelerates early discovery of lead candidates. The ProteOn™ XPR36
multiplexed SPR instrument from Bio-Rad Laboratories (Hercules, CA) (Bravman
et al. 2006) employs microfluidics integrated into 6 × 6 interaction array. The
ProteOn™ XPR36 has also been used for determination of antibody affinity
(Bravman et al. 2006). Plexera® Bioscience (Woodinville, WA) offers a
PlexArray™ HT instrument which can test thousands of protein interactions in
only 30 min. This system is based on a high-density array with the capacity to
evaluate more than 1,000 spots with a spot size as small as 100 µm. Microarray
technology has been widely applied in SPR-based settings to miniaturize spot
size and sample volume to increase throughput (Otsuki and Ishikawa 2010).
A wide variety of other instruments based on SPR and related phenomena are
available as well.
2.2.2 Liquid Chromatography
Liquid chromatography (LC) is perhaps the most heavily used tool in the field of
biotechnology. LC methods are the main components for the purification and analy-
sis of protein therapeutics during their development cycle (Ahrer and Jungbauer
2006; Andrew and Titus 2001). Understandably, most LC technologies have adopted
the high-throughput scheme by automating and streamlining the sampling
mechanism. The principle of protein LC is based on the nature of protein–resin
interactions in a given liquid mobile phase. An initial major step towards high-
throughput LC technology is the implementation of high-performance liquid chro-
matography (HPLC) (Swadesh 2001). This technique uses a fast flow rate under
high enclosed pressure, instead of gravity force, to drive the mobile phase through
the column. This permits faster sampling throughput and improves the resolution of
sample partition. A recent advancement has been categorized as ultra-performance
liquid chromatography (UPLC) where smaller diameter particles and subsequently
higher pressures enable even faster separation times (Xiang et al. 2006). UPLC
technology has clearly shown its advantage in reversed phase (RP) chromatography
and has become the primary separation method for mass spectrometry analysis
(Stackhouse et al. 2011; Szapacs et al. 2010).
Another application of high-throughput LC technology is the development of
protein purification methods. The goal is to obtain the largest amount of the thera-
peutic protein in its active form while reaching the highest purity possible. The
range of such methods is currently extremely wide, and the specific purification
steps often need to be derived empirically. Primary methods include affinity, size-
exclusion, ion-exchange, and hydrophobic interaction chromatographies. Multiwell
plates or miniaturized columns are frequently used to screen a large number of
parameters including type of resin, protein loading and elution conditions, as well
as efficiency for the product of interest (Fahrner et al. 2001). The quantity of column
resin and protein material needed is often measured in microliters, and the experi-
mental steps are simply executed by gravity, centrifugal force, or a pump. This type
of high-throughput approach offers an opportunity to evaluate as many variables as
possible early in development with a minimum of time and protein. The outcome of
the screening can often help to select product-specific processes for further develop-
ment and may even allow companies to bypass intellectual property restricted com-
mon practices. The latest improvement in this field highlights the use of automated
liquid handling systems coupled to positive displacement liquid transfer technology
(Susanto et al. 2008; Wiendahl et al. 2008). Such a system has greater similarity to
large-scale chromatographic instruments used in manufacturing by generating a
pressurized liquid flow through the microcolumns. The results obtained using such
high-throughput techniques are believed to be more representative of a protein’s
behavior during large batch purification.
2.2.3 Light Absorption Spectroscopy
Protein concentration measurements are essential during purification and formulation

development, and high-throughput compatible UV spectroscopy is a well-utilized
tool (Zhao et al. 2010). Besides protein concentration measurements, second
derivative UV spectroscopy has been frequently used to characterize protein ter-
tiary structure (Kueltzo et al. 2003; Mach and Middaugh 2011). Many absorption
spectrometers are compatible with a multiwell plate format and/or automation
modules. Typically, light passes through the plate vertically while the light source
and detector move from sample to sample. Alternatively, the plate itself can be
moved. Since the signal from proteins is typically strong, UV detection and absor-
bance integration are generally quick, allowing fast analysis of a large number of
samples. The disadvantages of this technique are also very well understood.
12 F. He et al.
The path length of the vertical light absorbance is poorly controlled and highly
dependent on the amount of sample present in the well. In addition, proteins with
high extinction coefficients, such as monoclonal antibodies, can easily saturate the
light detector. New adjustable path length spectrophotometers have recently
become available, although their throughput is not as high and light scattering fre-
quently complicates the use of this method.
Compared to UV absorbance, high-throughput optical density (OD), or turbid-
ity, assessment is more widely applied to the protein therapeutic development.
Turbidity usually refers to the obscuration of a sample at wavelengths near the
visible light range where proteins in solution do not manifest significant absor-
bance. The most widely employed wavelength range for this purpose is 350–
400 nm, which avoids any specific absorbance arising from amino acid side chains
or common color pigments. Turbidity is proportional to the amount of light
blocked or scattered by the solution components. In a protein sample, aggregates
and precipitates are known to give rise to solution turbidity as measured by OD,
making it a quick method to assess the quality of protein samples with respect to
the presence of protein aggregates. High-throughput turbidity assessment is fre-
quently used during protein formulation development to evaluate a large number
of samples that are put on storage or under environmental stresses (Zhao et al.
2010; Capelle et al. 2009). Though quantitative determination of protein degrada-
tion is not usually possible with turbidity measurements, the information obtained
is sufficient to discriminate or rank order formulations. Turbidity assessment is
generally noninvasive and can be applied using a variety of spectroscopic instru-
mentation and sample cells, including microtiter plate readers and pharmaceuti-
cally relevant containers. The latter provide a unique opportunity to assess sample
quality of protein therapeutics in their actual storage units, and permit a real-time
monitoring of aggregation during the manufacturing and distribution of a protein
commercial product.
Another common tool for protein analysis is circular dichroism (CD). CD is a
technique that measures the difference in absorbance of left- and right-handed cir-
cularly polarized light. CD is generally divided into near-UV (250–350 nm) and
far-UV (190–250 nm) measurements (Li et al. 2011). Near-UV CD is sensitive to
the tertiary structure of proteins, due to the presence of optically active chromo-
phores including the aromatic amino acid side chains and disulfide bonds. Far-UV
CD, on the other hand, is used to study the secondary structure of proteins. Alpha
helix, beta sheet, turns, and disordered structure all display unique CD spectra.
Expanding the high-throughput capabilities of CD instruments is achieved either by
increasing the number of cuvettes that an instrument can employ or the use of autos-
amplers. In addition, it has recently become possible to obtain data from the near-
and far-UV regions in a single scan. CD is typically more time-consuming than
other light absorbance measurements and can be significantly affected by both light
scattering and absorption flattening phenomena. In order to minimize the interfer-
ence, high-concentration protein samples are often measured via the use of short
path length (µm range) cells (Harn et al. 2007).
2.2.4 Vibrational Spectroscopy
Vibrational spectroscopy includes Raman, Fourier transform infrared (FTIR), and

near-infrared (NIR) spectroscopy. With respect to protein therapeutics development,
various Raman techniques and FTIR are commonly used to analyze protein second-
ary structure, while NIR is often applied to the analysis of the sample components
other than protein, such as organic compounds (Siebert and Hildebrandt 2008).
Because most chemical materials have identifiable spectral patterns, these tech-
niques are also frequently used for raw material and forensics analysis (described in
Chap. 10). Another common advantage of these techniques with protein therapeu-
tics development is that they can be used to analyze lyophilized protein samples
directly, a task at which many other biophysical techniques fail. Near-IR instrumen-
tation has successfully incorporated an automated sample presentation unit that per-
mits higher throughput analysis. The high-throughput development of FTIR,
however, is challenging due to geometrical problems. The only currently available
high-throughput option for FTIR involves the drying of liquid samples, which might
produce structural changes in the protein. Novel solutions to this problem are known
to be under investigation. In contrast, incorporation of multiwell plate and microar-
ray technologies instrumentation has been demonstrated with Raman spectroscopy
(Anquetil et al. 2003). The advantage of Raman-based technologies is that they are
typically compatible with a wide range of transparent container types and suffer less
from interference by water. The major disadvantage of Raman spectroscopy is that
signals from proteins are normally weak. Resonance and surface-enhanced Raman
technologies can partially overcome this problem but suffer from limited applicabil-
ity for a variety of reasons. In summary, vibrational spectroscopic techniques pos-
sess unique abilities to provide quality and conformational information on protein
products, including lyophilized samples. High-throughput applications of these
techniques, however, clearly require further development.
2.2.5 Fluorescence Spectroscopy
The aromatic side chains in proteins serve as fluorophores which emit photons at
higher wavelengths once excited with UV range light. The excitation and emission
profile is highly dependent on the polarity of the side chain environment and, therefore,
is sensitive to the local tertiary structure of the protein (Lakowicz 2006). This fluores-
cent property is referred to as protein intrinsic fluorescence. Since most large proteins
have widely distributed aromatic amino acids, the overall intrinsic fluorescence pro-
vides a good measure of protein folding. Upon conformational changes, the intrinsic
fluorescence emission peak generally shifts in wavelength. If present, tryptophan (Trp)
emission dominates the fluorescence of proteins with tyrosine (Tyr) contributing indi-
rectly. This method is often applied to protein formulation studies to detect changes
in protein conformation as a result of stresses such as temperature, pH, and solute.
14 F. He et al.
The high-throughput capability of fluorescence techniques also typically relies on a

multi-sample cuvette holder, a multiwell plate compartment, or an autosampler.
Generally, fluorescence intensity is detected at a 90o angle to the light source to mini-
mize scattering. In the case of multiwell plates, fluorescence signal is usually acquired
via a top or bottom reading method. Because of the noninvasive nature of the intrinsic
fluorescence measurements, it can be applied to analyze protein conformation in sam-
ples pulled from long-term studies. The utility of intrinsic fluorescence can be extended
by fluorescence lifetime and anisotropy measurements, through which further informa-
tion concerning protein conformation is obtained (Fowler et al. 2002; Owicki 2000).
Use of front surface geometry permits the fluorescence from highly scattering samples
to also be obtained. Recently, a high-throughput, microtiter plate-based fluorometer
(the Avacta Optim 1000) has been developed, which measures fluorescence spectra and
light scattering simultaneously. This instrument is capable of performing rapid thermal
melts with a resultant overall increase in throughput of better than a factor of 10.
Useful fluorescence signals can also be obtained from extrinsic fluorescent
probes such as small molecule dyes. There is a wide variety of fluorescent dyes
commercially available for protein research and development (Hawe et al. 2008a).
Different dyes can be employed to reveal different properties of protein solutions.
For example, hydrophobic dyes are typically used to probe the presence of apolar
sites on a protein. Anilinonaphthalene sulfate (ANS)-based dyes are the most com-
mon hydrophobic probe choices. An increase in surface hydrophobicity is believed
to be an indication of protein unfolding. This usually results in enhancement of fluo-
rescence intensity by the associated dye. Another widely used dye is SYPRO
Orange, a probe originally developed as a gel stain. Recently, SYPRO Orange has
been used to detect protein aggregation in monoclonal antibody product samples
and has been shown to display specificity to structurally perturbed protein aggre-
gates (He et al. 2010b; Mach et al. 2011). Similarly, a number of other dyes have
been reported to be sensitive to the aggregated protein species (Hawe et al. 2010a;
Hawe et al. 2010b). SYPRO Orange has also been employed to study protein unfold-
ing (described in Sect. 2.7). Other extrinsic probes have been employed to measure
solution viscosity (Schäfer and Schmidt 2006; Haidekker et al. 2005; Kung and
Reed 1989). Though the application of these viscosity-sensitive dyes has been
extensively demonstrated in biologically relevant samples, their utility in probing
therapeutic protein viscosity remains to be fully tested. One major concern is that
these dyes can interact with proteins in high-viscosity samples, which typically cor-
relates with high protein concentration (Hawe et al. 2010a). In addition to probing
the protein itself, other extrinsic fluorescent tools are available to study formulation
buffers. For instance, 1-N-phenylnaphthylamine (NPN) and 9-anthryldiazomethane
(ADAM) have been shown to detect polysorbate, a common surfactant used in pro-
tein formulation, and its degraded form in solution (Khossravi et al. 2002).
Fluorescence spectroscopy is also compatible with HPLC instrumentation as an
on-line detection method to monitor elution (Hawe et al. 2008b), offering yet
another high-throughput analytical option. The combination of chromatography
and fluorescence is particularly attractive, since real-time information can be
obtained on specific protein species as they fractionate and elute from the column.
2.2.6 Isothermal Titration Calorimetry
Isothermal titration calorimetry (ITC) is another sensitive method to evaluate

interactions in solution (Torres et al. 2010). Understanding how proteins interact
with other solution components, such as surfactants and excipients, could be infor-
mative and beneficial for the development of protein therapeutics. Unlike SPR, ITC
does not require immobilization or chemical modifications of the interacting com-
pounds (Pierce et al. 1999). This enables ITC to be applied to more types of interac-
tions that occur with proteins. In its standard format, this technique requires a
significant amount of protein and the experiment includes long-duration titrations.
The high-throughput development of ITC has been aimed at reducing the sample
consumption as well as improving the measurement speed. Recently ITC instru-
ments with reduced sample volumes have become commercially available, includ-
ing iTC200 and Auto-iTC200 calorimeters from MicroCal (GE Healthcare,
Waukesha, WI), as well as Nano ITC instrument (TA Instruments, New Castle, DE)
which offers higher sensitivity and titration speed and employs a significantly
smaller sample cell compared to the two-cell MicroCal VP-ITC instrument (Peters
et al. 2009). The Auto-iTC200 can run up to 384 titrations automatically through the
use of a temperature-controlled autosampler. A miniaturized ITC microcalorimeter
can reduce the protein requirement up to seven times (Verhaegen et al. 2000). Forty-
eight samples in each array can be measured with a reaction volume of 10–20 µL.
In addition, closed-chamber microfluidic calorimeters with thermopile heat sensors
have also been described in the literature (Lee et al. 2009). These calorimeters
achieve enhanced sensitivity by surrounding the measurement chamber with a vac-
uum. Flow microfluidic devices can use as little as 10–20 µL of sample volume
(Lerchner et al. 2008). In addition to the single-chamber settings, further advance-
ments utilizing array technologies might be possible.
2.2.7 Differential Scanning Calorimetry/Differential

Scanning Fluorometry
Stability of protein therapeutics under thermal stress is believed to be critical and

often assessed during development (Bruylants et al. 2005). A low temperature of
unfolding can lead not only to protein instability but also a decreased energy barrier
for unfolding events caused by protein interactions such as protein–surface (Chang
et al. 1996), protein–solvent, (Schiffer and Dotsch 1996) or other interactions.
Screening approaches focused on protein thermostability can reveal undesired
protein therapeutic candidates or manufacturing and storage conditions early during
the development. Differential scanning calorimetry (DSC) is a quantitative method
widely used for determination of protein thermal stability (Privalov and Privalov
2000). However, a low-throughput DSC takes a significant amount of sample and
time to complete a typical experiment. Automated autosampling instruments such
16 F. He et al.
as the VP-Capillary DSC Platform from GE Healthcare (Piscataway, NJ), the

PYRIS Diamond™ DSC Autosampler from PerkinElmer (Waltham, MA), or the
Nano DSC Autosampler System™ from TA Instruments (New Castle, DE) signifi-
cantly improve sample preparation, but analysis time is still lengthy and the signifi-
cant amounts of protein required make analysis difficult for the early screening of
hundreds of drug candidates and formulation excipients. Improvements in microflu-
idic technology, creation of array-based calorimetry microchips, and the use of
microplates will enable higher-throughput calorimetry (Vermeir et al. 2007;
Lerchner et al. 2006). Microelectromechanical system (MEMS)-based calorimetry
requires only 1.2 µL of solution, and the sensitivity is about 5 mg/mL of protein
concentration (Wang et al. 2008). Since the dimensions of this device are approxi-
mately 5 × 5 mm, there are opportunities for microarray fabrication.
DSC can detect the unfolding transitions of several domains in a multidomain
protein such as an immunoglobulin. It is, however, often sufficient to screen the
thermal stability of proteins based only on their lowest melting temperature. Single-
domain proteins can, of course, also be characterized by determination of their sin-
gle transition. In cases such as this, a high-throughput method, known as differential
scanning fluorometry (DSF), based on the extrinsic fluorescence of probes sensitive
to the polarity of their environment can be used (Pantoliano et al. 2001a). Unfolding
of proteins usually exposes hydrophobic regions to the solution resulting in a sig-
nificant increase in the fluorescence of these probes when bound. The method was
originally used for screening of small molecule interactions with proteins and can
therefore be used to identify some types of excipients (Pantoliano et al. 2001b). The
unfolding temperature of a protein shifts after binding of a ligand and binding
parameters can be determined by this shift in the melting temperature. In other stud-
ies this method has been applied to the evaluation of crystallization and general
stability (Malawski et al. 2006; Ericsson et al. 2006). Because of low background
fluorescence in the presence of native antibodies, DSF has been successfully used
for mAb formulation development (He et al. 2010c). Examples of DSF scans
obtained during formulation screening for a therapeutic protein are shown in
Fig. 2.1. A 96-well plate with different formulations was screened to determine the
most thermally stable formulation. The midpoint of the transition determined by the
increase in fluorescence intensity was found to correlate well with the unfolding
transition measured by DSC. One significant disadvantage of DSF in applications
for protein formulation development is the high-fluorescence background in the
presence of detergents commonly used in protein formulations (see high-
fluorescence background in Fig. 2.1). New probes, such as the thiol-specific fluoro-
chrome N-[4-(7-diethylamino-4-methyl-3-coumarinyl) phenyl] maleimide (CPM),
seem to overcome this detergent problem and have been used for stability profiling
of membrane proteins under different solution and ligand conditions (Alexandrov
et al. 2008). Fluorescent plate readers with a thermostat are necessary to make these
measurements. The maximal temperature of a typical scan is high and standard
plate-based fluorometers usually cannot provide the necessary temperature range.
Instead real-time polymerase chain reaction (RT-PCR) instruments equipped
with 96, 384 and even 1024-well plates are well suited for DSF measurements.
Fig. 2.1 Representative differential scanning fluorometry spectra obtained with a therapeutic
protein on 96-well plate filled with different formulations for each well. Fluorescence intensity
in relative fluorescence units (RFU) is shown as a function of temperature. Protein concentration
is 1 g/L. The sample volume is 30 µL. Data were obtained using the Bio-Rad CFX96 RT-PCR
plate reader
Many companies that manufacture these instruments have realized the popularity of
this application and have included software options to obtain and analyze DSF data.
Another advantage of DSF is the wide range of protein concentration that can be
used. Depending on protein properties, transitions can be detected at as low as
0.05 mg/mL or as high as at 100 mg/mL. Another high-throughput method, the
ProteoStat™ assay from Enzo Life Sciences (Farmingdale, NY), provides an
improved thermal shift approach based on extrinsic fluorescence for assessment of
protein stability through monitoring protein aggregation, rather than protein
unfolding. These methods, which employ high-throughput technologies, further
expand the application of thermal analysis to modern pharmaceutics. While it is not
guaranteed that thermal stability correlates with a protein’s physical stability during
storage, better thermal stability usually indicates a greater energy requirement to
unfold the protein. When executed under similar experimental conditions, high-
throughput thermal analysis offers useful information that can be used to rank pro-
tein or formulation candidates. It seems safe to state that when all other properties
are comparable, the protein constructs or formulations which lead to better thermal
stability will always be more desirable.
18 F. He et al.
2.2.8 Light Scattering
Aggregation is one of the major problems in protein pharmaceutical development.

The presence of aggregated protein can compromise the purity, safety, and efficacy
of a drug product. Several separation and detection techniques are used to monitor
protein aggregation. Light scattering methods have the advantage of high sensitivity
due to the size of scatterers, which makes it possible to detect small amounts of
large protein aggregates in pharmaceutically relevant samples. There are two basic
types of such methods used in protein therapeutics development: static light scatter-
ing (SLS) and dynamic light scattering (DLS). SLS can be applied to determine the
protein’s molecular mass and the mean square radius of gyration. SLS is often used
with separation methods such as SEC or field-flow fractionation (FFF) for the pur-
pose of obtaining a more accurate estimate of the size of the components separated
by these techniques (Tarazona and Saiz 2003; McEvoy et al. 2011). Such online
scattering method is commonly known as multi-angle light scattering (MALS). The
throughput of SLS analysis used in this fashion often depends on the throughput of
the separation technique. Though significant advancements have been achieved in
numerous types of separation technologies, a variety of physical parameters, such
as high pressure and short equilibration times, can be problematic for the coupling
of light scattering detectors. Light scattering can also be coupled to plate readers.
Simple monitoring of the scattered light at fixed wavelengths and angles can pro-
vide sensitive detection of aggregates formation. The Stargazer-384TM system
(Harbinger Biotechnology and Engineering Corporation, Toronto, Canada), a 384-
well microplate reader, was used in a study to monitor colloidal stability of mAbs at
elevated temperatures (Goldberg et al. 2010). As previously discussed, the high-
throughput Avacta Optim® 1000 (Pall Corporation, Port Washington, NY) is avail-
able for light scattering as well as fluorescence measurements with a volume as low
as 1 µL. In these instruments the light scattering signal is used to monitor aggrega-
tion by detecting the increased intensity, similar to turbidity measurements but with
higher sensitivity. It should be noted, however, that if the aggregates formed are less
dense than the monomeric protein, decreases in scattering can also be seen.
Dynamic light scattering is based on the measurement of the fluctuations in inten-
sity of scattered light. An autocorrelation function, derived from fluctuation analysis,
can reveal the distribution of the hydrodynamic radii of protein molecules present in
solution (Schmidt 2010). Separation is not necessary, but resolution of species
depends on their size difference and concentration. Because of the exponential form
of the autocorrelation function, no more than two to four components can be com-
fortably resolved in the same solution. DLS plate readers have become very popular
in the biopharmaceutical industry for protein and vaccine characterization (Vincentelli
et al. 2004). Multiwell plate formats, small volumes, and automated procedures for
measurement and data analysis make the DLS method high throughput and easy to
apply. Antibody self-association has been studied with the help of gold nanoparticles
and their characterization by dynamic and static light scattering. Nanoparticle–
antibody conjugates displayed complex aggregation behavior dependent on pH and
Fig. 2.2 High-throughput method for viscosity measurements based on dynamic light scattering
determination of the diffusion coefficient of polystyrene beads externally added into a protein
solution
ionic strength of the solution. Use of a DLS plate reader was a significant part of the
high-throughput analytical development (Sule et al. 2011).
In biopharmaceutical drug development DLS has been used not only for direct
detection and characterization of aggregation but also for the study of large colloidal
structures. Certain large colloid-like aggregates have been shown to inhibit enzymes
leading to false-positive HTS leads. These so-called promiscuous inhibitors were
detected and screened by DLS using a plate reader (Feng et al. 2005). Results from
such high-throughput assays for promiscuous inhibitory aggregates have been used
to develop new computational models of this phenomenon. A method for quantita-
tive characterization of macromolecular interactions using DLS has been introduced
in a temperature-controlled plate reader format (Hanlon et al. 2010). This technique
enabled determination of equilibrium dissociation constants and thermodynamic
parameters. The low volume of plate-based DLS reduced the sample amount to a
few microliters per experiment, with detection limits in the femtomolar range.
Biopharmaceutical products are often formulated at high concentrations to
maximize delivery dosage and efficiency, and solutions of some proteins become
very viscous at high concentrations (Yadav et al. 2010), creating significant prob-
lems for processes like purification, filtration, and injection through syringes.
Standard methods for viscosity measurements have low throughput and require
large quantities of protein. Thus, there is increasing demand for higher-throughput
viscosity screening. A DLS assay based on measurement of the diffusion coeffi-
cient of beads added directly to the protein solution is high throughput and run in
a multiwell plate format (He et al. 2010a). As shown in Fig. 2.2 the Stokes–Einstein
20 F. He et al.
equation can be used to calculate the viscosity of a protein solution using the
known radius of the added beads and the measured diffusion coefficient.
Furthermore, DLS measurements of diffusion coefficients as a function of protein
concentration can be used to derive the interaction parameter, kD, which has been
shown to correlate with protein properties such as viscosity and particulation pro-
pensity (Yadav et al. 2010; He et al. 2011). It is widely accepted that the second
virial coefficient, B22, obtained by SLS measurements contains information on pro-
tein–protein interaction (Printz et al. 2012). The kD parameter derived from DLS
measurements offers a simple way to compare samples under similar conditions
(discussed in Chap. 3).
2.2.9 Design of Experiment and Data Analysis
All high-throughput methodologies mentioned above share a common ability to

generate a large amount of biophysical information on therapeutic proteins. This
enables more complex experimental designs at various stages of pharmaceutical
development. The Quality by Design (QbD) concept has gained popularity in
recent years among the biopharmaceutical industry and regulatory agencies
(Rathore and Winkle 2009; Rathore and Devine 2008). Design of product quality
is built on comprehensive understanding of a well-defined process and product
space where the protein therapeutic is in its most desired form. The principle of
design of experiment (DOE) is often applied to systematically evaluate the protein
of interest under a variety of conditions, which are often selected based on the
types of stresses that a protein therapeutic is subjected to during manufacturing,
storage, and administration. The degree of these stresses often exceeds reality, and
the experimental results can be used to predict the protein behavior when failures
occur during the product life cycle. The combination of high-throughput measure-
ment and DOE can also help enhance the statistical power when interpreting the
results. More importantly, the implementation of high-throughput methodology
offers opportunities to simultaneously assess a large number of samples. This is
particularly valuable when employing methods that are only used to qualitatively
rank order protein samples.
Application of high-throughput techniques results in large data sets, often
requiring mathematical tools for rigorous analysis. Statistics helps to establish cor-
relations among measured properties of a molecule. Commonly used statistical
analyses that are frequently applied to the development of protein therapeutics
include Gaussian modeling, analyses of variance (ANOVA), and the t-test. In addi-
tion, biophysical characterization often involves spectroscopic methods such as
CD, FTIR, and fluorescence, and the results usually include measured values as a
function of wavelength of the optic source. Such data can be analyzed by methods
of chemometrics including the singular value decomposition technique which has
been used to determine the maximum changes in protein properties caused by
particular factors like pH or ionic strength. This methodology has been applied to
the interpretation of CD and FTIR spectra obtained during the production of
antibodies (Greenfield 2006; Sellick et al. 2010). Another useful mathematical tool
is polynomial-based data fitting. This approach involves fitting an arbitrary poly-
nomial model to a limited data set and then using the same model to predict protein
behavior outside of the tested range. For example, discrete data at pH 5, 5.5, and 6
can be used to generate a polynomial model that best fit the experimental results.
The continuous mathematical model can then be used to predict results at pH 5.8,
and even at pH values outside of 5–6, if the assumption holds. The polynomial
method is especially effective when the source data set is large. Even more predic-
tive information can be obtained while considering multiple variables simultane-
ously (Sall et al. 2007).
2.3 Empirical Phase Diagrams as Tools to Interpret Results

from High-Throughput Biophysical Approaches
2.3.1 Combination of Biophysical Techniques

and Data Analysis
As discussed above and in other chapters in this text, high-throughput screening

(HTS) is usually performed with only one or two low-resolution techniques of the
type previously considered. The selection of the particular technique is typically
based on what is known about degradation pathways of the target, convenience, and
availability of appropriate instrumentation as well as speed. It has recently become
possible to combine the results from multiple techniques with the goal of providing
a more comprehensive picture of a protein’s structure and its response to various
environmental perturbations. This can be used to select optimal methods for HTS as
well as for various forms of comparative analysis. A number of methods are avail-
able for this purpose. We will consider here only the one that has been most thor-
oughly described in the literature [reviewed in Maddux et al. (2011)], but other
approaches such as Chernoff faces and star charts (Yau 2011) in which information
is encoded in facial features or abstract geometric shapes are under consideration.
The former method is known as the EPD. The word empirical is inserted in front of
the phrase “phase diagram” to differentiate it from the well-known thermodynamic
or equilibrium phase diagram since equilibrium conditions are not implied in the
former.
The basic idea behind the EPD is to represent the protein (application has also
been made to peptides, nucleic acids, virus-like particles, viruses, and bacteria
cells) as a vector in which the components of the vector are experimental values
obtained from the various methods employed as a function of solution variables.
Preparation of EPDs generally involves buffer subtraction of the data, peak
22 F. He et al.
selection for data analysis (entire spectra can also be used), averaging of multiple
data acquisitions, normalization, input matrix synthesis, singular value decompo-
sition, and finally color mapping of the most significant data using an RGB color
scheme. A detailed description of the method including the mathematics involved
is presented in Maddux et al. (2011). In general, far-UV circular dichroism (CD)
is used to monitor secondary structure although FTIR and Raman spectroscopies
can also be employed. Tertiary structure is most commonly analyzed with intrin-
sic fluorescence, near-UV CD, or high-resolution derivative absorption spectros-
copy. Dye binding using compounds such as 8-anilino naphthalene sulfonic acid
(ANS) is frequently used to probe the exposure of apolar regions in the protein.
Dissociation and association (including aggregation) are typically probed with
static and/or DLS (although see below). Overall thermal stability is often studied with
differential scanning calorimetry. The most common independent variables (i.e.,
forms of stress that have previously been employed) are temperature and pH although
a wide variety of other variables have been used as described below. Perhaps the
major limitation of the EPD approach has been the time and instrumentation neces-
sary to prepare an EPD. This has recently changed with the advent of equipment
capable of performing multiple different types of measurements simultaneously.
Originally an EPD typically required a fluorometer, CD spectropolarimeter, light
scattering system, and perhaps a DSC or FTIR spectrometer. Several newly
developed instruments now at least partially overcome this limitation. For example,
recent improvements in CD instruments now permit both near- and far-UV
spectra, near-UV absorption spectra, fluorescence, and SLS (turbidity or scatter-
ing at the fluorescence emission wavelength) to all be acquired simultaneously in
a four-position sample chamber under variable temperature conditions (Hu et al.
2011). This permits EPDs to be generated in less than a day. A similar “protein
machine” with a six-position sample chamber has also been recently described
(Maddux et al. 2012). Perhaps the simplest version of a system with rapid EPD
generation capability is a UV absorption spectrometer, typically of the diode array
variety, to permit sufficient resolution of derivative peaks (Kueltzo et al. 2003). At
high resolution (usually second), the derivative spectrum of a protein will usually
manifest distinct peaks for phenylalanine, tyrosine, and tryptophan (if present).
Since the residues are usually buried, Tyr is often interfacial, and Trp present in
highly variable environments, temperature, and pH-induced peaks shifts can fre-
quently provide a fairly detailed picture of a protein’s structural response to vari-
ous perturbations. Such data can easily be represented in the form of an EPD
(Kueltzo et al. 2003). Similarly, fluorescence microtiter plate-based fluorometers
have been developed which employ multiple fluorescence and light scattering
measurements as a function of temperature, as also described in the fluorescence
section. The latter is of especially high throughput, permitting the generation of
many EPDs in a single day. EPDs for four model proteins obtained from multiple
instruments, two CD-based spectrometers, and a high-throughput fluorometer are
shown in Fig. 2.3, where it can be seen that all produce similar EPDs although
small differences are apparent due to the various types of measurements used to
construct each EPD.
Fig. 2.3 Empirical phase diagrams (EPDs) of four model proteins (1) aldolase, (2) BSA,
(3) chymotrypsin, and (4) lysozyme constructed using data collected from various instruments:
(a) intrinsic fluorescence (FL) and static light scattering (SLS) data from a Photon Technology
International (PTI) fluorometer, and circular dichroism (CD) from an Applied Photophysics
Chirascan, (b) FL and SLS by an Avacta Optim 1000 microtiter plate fluorometer, (c) FL and CD
Applied Photophysics Chirascan, and (d) FL, SLS, CD, and UV absorbance by an Olis Protein
Machine
2.3.2 High-Throughput Characterization

and Preformulation Development
The EPD method provides a comprehensive overview of how a protein responds to

environmental alteration in the form of a colored diagram in which regions of dif-
ferent color correspond to different structural states of the target molecules. By
reference to the original data, native partially folded and molten globule, extensively
unfolded, dissociated, oligomerized, and various aggregated states can all be identi-
fied. This provides the scientist with clues to trouble spots in a protein and a basis
with which to select assays with which to screen for potential stabilizers. For exam-
ple, if aggregation or a particular structural change occurs under moderate tempera-
ture and/or pH conditions, one can select a less stable condition and one or more
techniques sensitive to selected degradation events for screening purposes. Typically,
a supplemental GRAS (generally regarded as safe) library containing a selection of
24 F. He et al.
buffers, sugars, sugar alcohols, amino acids, polymers, detergents, and osmolytes is
used. In the initial screen relatively high concentrations of compounds are used with
their concentration dependence and use in combination later optimized. It is usually
wise to employ at least two methods: one sensitive to aggregation (e.g., light scat-
tering) and one to structural change (e.g., fluorescence, CD) for this purpose. DSC
is also commonly employed especially due to the recent availability of highly sensi-
tive high-throughput instruments. It is also possible to prepare EPDs in the presence
of selected stabilizers to permit a more detailed analysis/comparison of their effects
on a protein. The information thus obtained by a temperature/pH EPD thus provides
a basis for buffer and excipient selection at an early stage of pharmaceutical devel-
opment. Although not yet published, a new version of the EPD has been developed
in which the colors have actual physical measuring in contrast to the arbitrary
assignment of color in the original EPD.
2.3.3 Additional Application of High-Throughput Methods

and the EPD
High-throughput methods and EPDs can also be applied to a wide variety of differ-
ent situations, some of which will be briefly described here. Two commonly encoun-
tered forms of stress in the protein therapeutic area are freeze/thaw and shear. It is
often necessary to freeze and then thaw both during development and manufactur-
ing situations. An EPD can be created using the number of freeze/thaw cycles under
defined conditions as an independent variable accompanied by temperature and pH
stress. All three variables (temperature, pH, freeze/thaw cycles) can be combined
into a three-dimensional representation in which the EPD is presented as a colored
surface. Shear stress is also often encountered in the development, manufacture
(especially filling), and shipping of protein pharmaceuticals. To explore this poten-
tial degrading stress, the intensity of the shear can be varied by a mechanical pro-
cess such as stirring, shaking, or some other forms of agitation, and this is used as a
variable in EPD production.
Protein concentration is another important variable that has assumed increasing
importance with the use of high-concentration formulations. This variable can be
typically evaluated over the range of 0.05–300 g/L depending on the solubility of
the protein and the methods employed in the analysis. Proteins usually alter their
structure to little or no extent as a function of protein concentration, but aggregation
and surface adsorption are both highly concentration dependent. Thus, aggregation-
sensitive techniques such as light scattering are often of special importance in pro-
tein concentration-dependent studies. Another common variable of particular
importance is ionic strength. In the Debye–Huckel charge shielding regime (0–0.15
ionic strength), a number of intermediate concentrations should be evaluated to
probe electrostatic effects. At higher salt concentrations both preferential hydration
and binding effects usually dominate with salt concentration into the molar range
appropriately examined.
It is also possible to create EPDs based on phenomena such as aggregation.

A variety of different types of aggregates have been identified in protein solutions
based on their relative size and the nature of a protein’s conformation (altered or
native) within the aggregate. A number of methods are available that are sensitive to
these features (see Chap. 9), permitting an aggregation-based EPD to be created.
Again, using variables like temperature, pH, ionic strength, freeze/thaw, and shear,
soluble protein aggregates can be detected by methods such as size-exclusion
HPLC, sedimentation velocity analytical ultracentrifugation, FFF, and DLS.
Complimentary structural data can be obtained by the methods described above
with FTIR and Raman spectroscopy especially useful because of the particulate
nature of such samples. Larger (i.e., submicron) aggregates can be characterized by
DLS including single-particle microscope-based approaches (nanoparticle tracking
analysis) and classic microscopy-based techniques (atomic force microscopy, scan-
ning electron microscopy, and transmission electron microscopy) although they are
difficult to quantitate and new methods such as quartz crystal microbalances and
nanomechanical resonators are seeing increasing use. Subvisible and visible parti-
cles can be analyzed by methods such as coulter counting, light obscuration, micro-
flow digital imaging, and various visual procedures. Using parameters such as size,
composition, structure, and particle number as dependent variables, an EPD can be
generated that provides a comprehensive picture of the nature of protein aggregates
that form under a wide variety of stress conditions.
Although not yet described in the scientific literature, chemical degradation can
also be analyzed in various high-throughput modes and be summarized in EPDs.
The application of EPDs described above has all been to various physical processes
in which covalent bonds are not broken. Of equal importance to protein degradation,
however, are chemical changes such as oxidation and deamidation events. Chemical
changes are usually quantitatively determined by peptide mapping combined with
mass spectrometry (MS) in the form of HPLC-MS experiments which permit both
the amount and location of a residue modification within a protein’s amino acid
sequence to be determined. To increase throughput, once the nature of any changes
has been identified, HPLC-MS analysis may be replaced by methods such as
RP-HPLC and capillary electrophoresis or isoelectric focusing. A convenient way
to present and analyze such data is in the form of rate constants for the individual,
for example, deamidation and oxidation events. This of course requires time-
dependent measurements. The rate constants can be used as the dependent variables
in an EPD since they are typically determined as function-independent variables
such as pH and temperature. Of special interest is the comparative use of physical
and chemical EPDs which permits an exploration of the relationship between struc-
tural changes and chemical events (and vice versa).
As a final example, EPDs can be used in a strictly comparative mode. Structural
comparisons between both similar and assumed identical proteins are often a very
important element of pharmaceutical analysis. For example, when manufacturing
changes are made, during the development of biosimilars or when investigating
mutant proteins, a detailed comparison of the various species is a critical part of the
analysis. A direct comparison of EPDs of the target molecules provides a convenient
26 F. He et al.
and sensitive way to see if structural identity has been obtained. As an example,
second-generation functional mutants of fibroblast growth factor one (FGF-1) have
been compared using EPDs and used to select molecules that are not dependent on
heparin for their activity (Alsenaidy et al. 2012). Such EPD-dependent comparisons
have even been performed with different rotavirus serotypes despite their individual
complexity (Esfandiary et al. 2010). Various mathematical (difference) methods
exist to facilitate such comparisons.
2.4 Advantages and Challenges of Implementing

High-Throughput Technology in Therapeutic
Protein Development: An Industry Perspective
High-throughput technologies can be applied across all aspects of biopharmaceuti-

cal discovery and development from the early stages of candidate screening and
selection to the later stages of formulation development. The benefits of a successful
high-throughput screening strategy in this environment are many. The most obvious
one is speed. Faster assays mean that sufficient data to drive a decision can be col-
lected within a shorter timeframe leading to more rapid decisions and ultimately an
accelerated development timeline. In the pharmaceutical industry where product
development is a protracted process, any acceleration of the timeline can mean get-
ting a promising drug candidate into clinical testing sooner and ultimately to market
faster, potentially providing a competitive advantage and leading to an earlier reve-
nue stream. A second advantage is that high-throughput assays typically require
smaller volumes and fewer samples than standard assay formats. This sample spar-
ing feature is especially important early in development in which the purification
process is an early stage of development and consequently the quantities of the
candidates being tested may be in short supply. Another advantage is that high-
throughput procedures permit many more drug candidates to be tested than would
be possible using a standard approach. In the early stages of discovery research
where many candidates are being evaluated based on screening assays, a good high-
throughput screen for binding permits many more candidates to be tested and allows
this to be accomplished within a shorter period of time. Finally, high-throughput
assays permit additional molecular features to be evaluated. A good example of this
advantage is found in formulation development where the influence of a variety of
solution conditions and potential excipients will need to be tested. The solution pH,
buffer salts used, excipients, surfactants, and the presence or absence of salt all need
to be evaluated, and this can be accomplished much more easily using high-
throughput approaches.
High-throughput assays have been a standard approach to the discovery of novel
small molecule drugs for several years. These assays were typically designed using
an isolated drug target, often the soluble domain of a membrane bound receptor, and
a library of small molecules, often containing tens of thousands of compounds,
would be screened with a readout indicating a binding event or inhibition of binding.

With the birth of the biotechnology industry approximately three decades ago and
the beginning of the development of recombinant proteins as therapeutic entities,
the need for high-throughput approaches for biotherapeutic drug discovery and
development has grown. The greater chemical complexity of proteins and the need
to maintain native three-dimensional structure during both processing and storage
of the drug substance and drug product have contributed to the application of the
various biophysical methods discussed in this chapter. Another property of many
proteins, the tendency to self-associate to form dimers and higher oligomers, in
some cases an undesirable property, has also demanded the use of highly selective
biophysical techniques. The development of high-throughput approaches to the
application of these biophysical methods has lagged behind other methods, but
the burgeoning biotechnology industry with its unique analytical requirements has
helped to recently accelerate this development.
The implementation of a high-throughput screen, however, is not a simple matter
of just miniaturizing an assay and running it in a 96-well plate. High-throughput
assays must be carefully developed and validated before use in drug development.
Special equipment is often necessary to successfully perform a high-throughput
assay. Depending on the readout of the assay, a special instrument with the capabil-
ity to measure signals from a plate may be needed. Special devices for heating or
cooling a plate are often required to maintain the temperature required for the assay,
and, in some cases, liquid handling equipment is necessary for setting up the
multiple conditions to be examined. One of the biggest challenges to implement a
high-throughput strategy is the handling of the data produced. High-throughput
methods have the potential to generate vast quantities of data which not only must
be collected, stored, and archived but also must be analyzed and then presented in a
form that can be understood by others. The generation of these very large data sets
has perhaps had an unintended consequence. As described above, statistical meth-
ods are an absolute requirement for the analysis of these data. This has led to a much
greater understanding of the relationships among the variables in an experiment.
At the extremes of an experiment are experimental design and presentation of
results. DOE has been critical to the success of high-throughput experimental
design. When working with multiple molecules or a large number of variables, it is
a physical impossibility to design and execute experiments that examine all possi-
bilities in a meaningful way, even with the use of high-throughput methods. DOE
provides a way of designing experiments that provide the necessary data while sig-
nificantly decreasing the number of experiments. This approach is expected to grow
within the biopharmaceutical industry as more high-throughput assays are brought
on line. Equally demanding is the presentation of the processed data from high-
throughput experiments. One approach described above which draws on the statisti-
cal analysis of large multiple data sets is the EPD in which a protein is represented
as a vector in which the components of the vector are experimental values obtained
from the various methods employed as a function of solution variables. The concept
and preparation of EPDs have been discussed earlier in this chapter, but the
outcome is a presentation using color mapping to denote the most significant data.
28 F. He et al.
This approach as well as others allows a very complex data set to be presented in a
way that is readily understandable.
Although high-throughput biophysical methods become increasingly important
in the development of therapeutic proteins, there are still biophysical methods that
have not been adapted to a high-throughput format. Examples include circular
dichroism and mass spectrometry which are still in their infancy but which, no
doubt, will have high-throughput adaptations in the near future. It is clear that there
is a need for new and better high-throughput biophysical assays and that the needs
of the biopharmaceutical industry will play a role in their development.
References
Ahrer K, Jungbauer A (2006) Chromatographic and electrophoretic characterization of protein

variants. J Chromatogr B Analyt Technol Biomed Life Sci 841(1–2):110–122
Alexandrov AI, Mileni M, Chien EY, Hanson MA, Stevens RC (2008) Microscale fluorescent
thermal stability assay for membrane proteins. Structure 16(3):351–359
Alsenaidy MA, Wang T, Kim JH et al (2012) An empirical phase diagram approach to investigate
conformational stability of “second-generation” functional mutants of acidic fibroblast growth
factor-1. Protein Sci 21(3):418–432
Andrew SM, Titus JA (2001) Purification of immunoglobulin G. Curr Protoc Immunol. Chapter
2:Unit 2.7
Anquetil PA, Brenan CJ, Marcolli C, Hunter IW (2003) Laser Raman spectroscopic analysis of
polymorphic forms in microliter fluid volumes. J Pharm Sci 92(1):149–160
Bravman T, Bronner V, Lavie K, Notcovich A, Papalia GA, Myszka DG (2006) Exploring “one-
shot” kinetics and small molecule analysis using the ProteOn XPR36 array biosensor. Anal
Biochem 358(2):281–288
Bruylants G, Wouters J, Michaux C (2005) Differential scanning calorimetry in life sciences:
thermodynamics, stability, molecular recognition and applications in drug design. Curr Med
Chem 12:2011–2020
Capelle MA, Gurny R, Arvinte T (2009) A high throughput protein formulation platform: case
study of salmon calcitonin. Pharm Res 26(1):118–128
Chang BS, Kendrick BS, Carpenter JF (1996) Surface-induced denaturation of proteins during
freezing and its inhibition by surfactants. J Pharm Sci 85(12):1325–1330
Coffman JL, Kramarczyk JF, Kelley BD (2008) High-throughput screening of chromatographic
separations: I. Method development and column modeling. Biotechnol Bioeng
100(4):605–618
Ericsson UB, Hallberg BM, Detitta GT, Dekker N, Nordlund P (2006) Thermofluor-based high-
throughput stability optimization of proteins for structural studies. Anal Biochem
357(2):289–298
Esfandiary R, Yee L, Ohtake S et al (2010) Biophysical characterization of rotavirus serotypes G1,
G3 and G4. Hum Vaccin 6(5):390–398
Fahrner RL, Knudsen HL, Basey CD et al (2001) Industrial purification of pharmaceutical antibod-
ies: development, operation, and validation of chromatography processes. Biotechnol Genet
Eng Rev 18:301–327
Feng BY, Shelat A, Doman TN, Guy RK, Shoichet BK (2005) High-throughput assays for promis-
cuous inhibitors. Nat Chem Biol 1(3):146–148
Fowler A, Swift D, Longman E et al (2002) An evaluation of fluorescence polarization and lifetime
discriminated polarization for high throughput screening of serine/threonine kinases. Anal
Biochem 308(2):223–231
Goldberg DS, Bishop SM, Shah AU, Sathish HA (2010) Formulation development of therapeutic
monoclonal antibodies using high-throughput fluorescence and static light scattering tech-
niques: role of conformational and colloidal stability. J Pharm Sci 100(4):1306–1315
Greenfield NJ (2006) Using circular dichroism spectra to estimate protein secondary structure. Nat
Protoc 1(6):2876–2890
Haidekker MA, Brady TP, Lichlyter D, Theodorakis EA (2005) Effects of solvent polarity and
solvent viscosity on the fluorescent properties of molecular rotors and related probes. Bioorg
Chem 33(6):415–425
Hanlon AD, Larkin MI, Reddick RM (2010) Free-solution, label-free protein-protein interactions
characterized by dynamic light scattering. Biophys J 98(2):297–304
Harn N, Allan C, Oliver C, Middaugh CR (2007) Highly concentrated monoclonal antibody solu-
tions: direct analysis of physical structure and thermal stability. J Pharm Sci 96(3):532–546
Hawe A, Sutter M, Jiskoot W (2008a) Extrinsic fluorescent dyes as tools for protein characteriza-
tion. Pharm Res 25(7):1487–1499
Hawe A, Friess W, Sutter M, Jiskoot W (2008b) Online fluorescent dye detection method for the
characterization of immunoglobulin G aggregation by size exclusion chromatography and
asymmetrical flow field flow fractionation. Anal Biochem 378(2):115–122
Hawe A, Filipe V, Jiskoot W (2010a) Fluorescent molecular rotors as dyes to characterize
polysorbate-containing IgG formulations. Pharm Res 27(2):314–326
Hawe A, Rispens T, Herron JN, Jiskoot W (2010b) Probing bis-ANS binding sites of different
affinity on aggregated IgG by steady-state fluorescence, time-resolved fluorescence and iso-
thermal titration calorimetry. J Pharm Sci 100(4):1294–1305
He F, Becker GW, Litowski JR, Narhi LO, Brems DN, Razinkov VI (2010a) High-throughput
dynamic light scattering method for measuring viscosity of concentrated protein solutions.
Anal Biochem 399(1):141–143
He F, Hogan S, Latypov RF, Narhi LO, Razinkov VI (2010b) High throughput thermostability
screening of monoclonal antibody formulations. J Pharm Sci 99(4):1707–1720
He F, Phan DH, Hogan S et al (2010c) Detection of IgG aggregation by a high throughput method
based on extrinsic fluorescence. J Pharm Sci 99(6):2598–2608
He F, Woods CE, Becker GW, Narhi LO, Razinkov VI (2011) High-throughput assessment of
thermal and colloidal stability parameters for monoclonal antibody formulations. J Pharm Sci
100(12):5126–5141
Hu L, Olsen C, Maddux N, Joshi SB, Volkin DB, Middaugh CR (2011) Investigation of protein
conformational stability employing a multimodal spectrometer. Anal Chem
83(24):9399–9405
Jezek J, Rides M, Derham B et al (2011) Viscosity of concentrated therapeutic protein composi-
tions. Adv Drug Deliv Rev 63(13):1107–1117
Jiskoot W, Randolph TW, Volkin DB et al (2012) Protein instability and immunogenicity: road-
blocks to clinical application of injectable protein delivery systems for sustained release.
J Pharm Sci 101(3):946–954
Karlsson R (2004) SPR for molecular interaction analysis: a review of emerging application areas.
J Mol Recognit 17(3):151–161
Khossravi M, Kao YH, Mrsny RJ, Sweeney TD (2002) Analysis methods of polysorbate 20: a new
method to assess the stability of polysorbate 20 and established methods that may overlook
degraded polysorbate 20. Pharm Res 19(5):634–639
Kueltzo LA, Ersoy B, Ralston JP, Middaugh CR (2003) Derivative absorbance spectroscopy and
protein phase diagrams as tools for comprehensive protein characterization: a bGCSF case
study. J Pharm Sci 92(9):1805–1820
Kung CE, Reed JK (1989) Fluorescent molecular rotors: a new class of probes for tubulin structure
and assembly. Biochemistry 28(16):6678–6686
Lakowicz JR (2006) Principles of fluorescence spectroscopy, 3rd edn. Springer, New York
Lee W, Fon W, Axelrod BW, Roukes ML (2009) High-sensitivity microfluidic calorimeters for
biological and chemical applications. Proc Natl Acad Sci USA 106(36):15225–15230
30 F. He et al.
Lerchner J, Wolf A, Wolf G et al (2006) A new micro-fluid chip calorimeter for biochemical
applications. Thermochim Acta 445:144–150
Lerchner J, Maskow T, Wolf G (2008) A new micro-fluid chip calorimeter for biochemical appli-
cations. Thermochim Acta 47:991–999
Li CH, Nguyen X, Narhi L et al (2011) Applications of circular dichroism (CD) for structural
analysis of proteins: qualification of near- and far-UV CD for protein higher order structural
analysis. J Pharm Sci 100(11):4642–4654
Mach H, Arvinte T (2011) Addressing new analytical challenges in protein formulation develop-
ment. Eur J Pharm Biopharm 78(2):196–207
Mach H, Middaugh CR (2011) Ultraviolet spectroscopy as a tool in therapeutic protein develop-
ment. J Pharm Sci 100(4):1214–1227
Mach H, Bhambhani A, Meyer BK et al (2011) The use of flow cytometry for the detection of
subvisible particles in therapeutic protein formulations. J Pharm Sci 100(5):1671–1678
Maddux NR, Joshi SB, Volkin DB, Ralston JP, Middaugh CR (2011) Multidimensional methods
for the formulation of biopharmaceuticals and vaccines. J Pharm Sci 100(10):4171–4197
Maddux NR, Rosen IT, Hu L, Olsen CM, Volkin DB, Middaugh CR (2012) An improved method-
ology for multidimensional high-throughput preformulation characterization of protein confor-
mational stability. J Pharm Sci 101(6):2017–2024
Malawski GA, Hillig RC, Monteclaro F et al (2006) Identifying protein construct variants with
increased crystallization propensity – a case study. Protein Sci 15(12):2718–2728
McEvoy M, Razinkov V, Wei Z, Casas-Finet JR, Tous GI, Schenerman MA (2011) Improved par-
ticle counting and size distribution determination of aggregated virus populations by asym-
metric flow field-flow fractionation and multiangle light scattering techniques. Biotechnol Prog
27(2):547–554
Otsuki S, Ishikawa M (2010) Wavelength-scanning surface plasmon resonance imaging for label-
free multiplexed protein microarray assay. Biosens Bioelectron 26(1):202–206
Owicki JC (2000) Fluorescence polarization and anisotropy in high throughput screening: perspec-
tives and primer. J Biomol Screen 5(5):297–306
Pantoliano MW, Petrella EC, Kwasnoski JD et al (2001) High-density miniaturized thermal shift
assays as a general strategy for drug discovery. J Biomol Screen 6:429–440
Peters WB, Frasca V, Brown RK (2009) Recent developments in isothermal titration calorimetry
label free screening. Comb Chem High Throughput Screen 12(8):772–790
Pierce MM, Raman CS, Nall BT (1999) Isothermal titration calorimetry of protein–protein interac-
tions. Methods 19(2):213–221
Printz M, Kalonia DS, Friess W (2012) Individual second virial coefficient determination of mono-
mer and oligomers in heat-stressed protein samples using size-exclusion chromatography-light
scattering. J Pharm Sci 101:363–372
Privalov GP, Privalov PL (2000) Problems and perspectives in microcalorimetry of biological mac-
romolecules. Methods Enzymol 323:31–62
Rathore AS, Devine R (2008) PDA workshop on “Quality by Design for Biopharmaceuticals:
Concepts and Implementation”, May 21–22, 2007, Bethesda, Maryland. PDA J Pharm Sci
Technol 62(5):380–390
Rathore AS, Winkle H (2009) Quality by design for biopharmaceuticals. Nat Biotechnol
27(1):26–34
Rich RL, Myszka DG (2000) Advances in surface plasmon resonance biosensor analysis. Curr
Opin Biotechnol 11(1):54–61
Rosenberg AS (2006) Effects of protein aggregates: an immunologic perspective. AAPS
J 8(3):E501–E507
Sall J, Creighton L, Lehman A (2007) JMP start statistics: a guide to statistics and data analysis
using JMP, 4th edn. SAS Press, Cary
Schäfer G, Schmidt H (2006) High-throughput spectroscopic viscosity measurement of nanocom-
posite sols with ETC-effect. J Sol–Gel Sci Technol 38(3):241–244
Schiffer CA, Dotsch V (1996) The role of protein-solvent interactions in protein unfolding. Curr
Opin Biotechnol 7(4):428–432
Schmidt R (2010) Dynamic light scattering for protein characterization. Encyclopedia of Analytical
Chemistry, 20–24
Sellick CA, Hansen R, Jarvis RM et al (2010) Rapid monitoring of recombinant antibody produc-
tion by mammalian cell cultures using Fourier transform infrared spectroscopy and chemomet-
rics. Biotechnol Bioeng 106(3):432–442
Shire SJ, Shahrokh Z, Liu J (2004) Challenges in the development of high protein concentration
formulations. J Pharm Sci 93(6):1390–1402
Siebert F, Hildebrandt P (2008) Vibrational spectroscopy in life science. Wiley-VCH, Weinheim
Stackhouse N, Miller AK, Gadgil HS (2011) A high-throughput UPLC method for the character-
ization of chemical modifications in monoclonal antibody molecules. J Pharm Sci
100(12):5115–5125
Sule SV, Sukumar M, Weiss WF, Marcelino-Cruz AM, Sample T, Tessier PM (2011) High-
throughput analysis of concentration-dependent antibody self-association. Biophys J
101(7):1749–1757
Susanto A, Knieps-Grünhagen E, von Lieres E, Hubbuch J (2008) High throughput screening for
the design and optimization of chromatographic processes: assessment of model parameter
determination from high throughput compatible data. Chem Eng Technol 31(12):1846–1855
Swadesh J (2001) HPLC: practical and industrial applications, 2nd edn. CRC Press, Boca Raton
Szapacs ME, Urbanski JJ, Kehler JR et al (2010) Absolute quantification of a therapeutic domain
antibody using ultra-performance liquid chromatography-mass spectrometry and immunoas-
say. Bioanalysis 2(9):1597–1608
Takeda H, Fukumoto A, Miura A, Goshima N, Nomura N (2006) High-throughput kinase assay
based on surface plasmon resonance suitable for native protein substrates. Anal Biochem
357(2):262–271
Tarazona MP, Saiz E (2003) Combination of SEC/MALS experimental procedures and theoretical
analysis for studying the solution properties of macromolecules. J Biochem Biophys Methods
56(1–3):95–116
Torres FE, Recht MI, Coyle JE, Bruce RH, Williams G (2010) Higher throughput calorimetry:
opportunities, approaches and challenges. Curr Opin Struct Biol 20(5):598–605
Verhaegen K, Baert K, Simaels J, Van Driessche W (2000) A high-throughput silicon microphysi-
ometer. Sens Actuators A Phys 82(1):186–190
Vermeir S, Nicolai BM, Verboven P et al (2007) Microplate differential calorimetric biosensor for
ascorbic acid analysis in food and pharmaceuticals. Anal Chem 79(16):6119–6127
Vincentelli R, Canaan S, Campanacci V et al (2004) High-throughput automated refolding screen-
ing of inclusion bodies. Protein Sci 13(10):2782–2792
Wagner M, Reiche K, Blume A, Garidel P (2012) Viscosity measurements of antibody solutions
by photon correlation spectroscopy: an indirect approach – limitations and applicability for
high-concentration liquid protein solutions. Pharm Dev Technol. doi:10.3109/10837450.2011.
649851
Wang L, Wang B, Lin Q (2008) Demonstration of MEMS-based differential scanning calorimetry
for determining thermodynamic properties of biomolecules. Sens Actuators B Chem
134:953–958
Wassaf D, Kuang G, Kopacz K et al (2006) High-throughput affinity ranking of antibodies using
surface plasmon resonance microarrays. Anal Biochem 351(2):241–253
Wiendahl M, Schulze Wierling P, Nielsen J et al (2008) High throughput screening for the design
and optimization of chromatographic processes – miniaturization, automation and paralleliza-
tion of breakthrough and elution studies. Chem Eng Technol 31(6):893–903
Xiang Y, Liu Y, Lee ML (2006) Ultrahigh pressure liquid chromatography using elevated tempera-
ture. J Chromatogr A 1104(1–2):198–202
Yadav S, Liu J, Shire SJ, Kalonia DS (2010) Specific interactions in high concentration antibody
solutions resulting in high viscosity. J Pharm Sci 99(3):1152–1168
Yau N (2011) Visualize this: the FlowingData guide to design, visualization, and statistics. Wiley,
Indianapolis
Zhao H, Graf O, Milovic N et al (2010) Formulation development of antibodies using robotic
system and high-throughput laboratory (HTL). J Pharm Sci 99(5):2279–2294
Chapter 3
Techniques for Higher-Order Structure
Determination
James Kranz, Fatma AlAzzam, Atul Saluja, Juraj Svitel,

and Wasfi Al-Azzam
3.1 Introduction: Protein Structure
Proteins are the building blocks for major parts of living systems that are r esponsible
for all critical cellular functions. Their roles vary from catalyzing reactions
(enzymes) to facilitating movement (cytoskeletal and motor proteins) and acting as
messengers for signal transduction, to name a few. The ability of proteins to carry
out such broad structural and functional roles is due to their unique three-dimensional
structures, as well as thermodynamic characteristics that allow the functional form
of a protein to retain stability under many different environmental conditions, such
as changing pH or temperature. Understanding protein structure is a prerequisite for
the development of protein therapeutics, from the identification of potential targets
for therapeutic agents, being developed and administered as the therapeutic agent or
being used in more novel applications as carriers of therapeutic agents and sustained
released devices, etc. Control, optimization, and quantitation of protein structure are
critical to the success of protein pharmaceuticals.
J. Kranz • W. Al-Azzam (*)

Biopharmaceuticals Development, GlaxoSmithKline, 709 Swedeland Road,
UW2108, King of Prussia, PA 19406, USA
e-mail: [email protected]; [email protected]
F. AlAzzam
TechnoPharmaSphere LLC, 822 Geddy Lane, Downingtown, PA 19335, USA
A. Saluja
Drug Product Science & Technology, Bristol-Myers Squibb, 1 Squibb Drive,
New Brunswick, NJ 08903, USA
J. Svitel
Process and Product Development, Amgen Inc., One Amgen Center Drive,

34 J. Kranz et al.
Protein structure starts with the linear polymer composed of 20 different amino
acids covalently linked through peptide bonds; this is the primary structure or
sequence. The peptide bonds are also responsible for the formation of intraresidue
hydrogen bonding that gives rise to secondary structures. Each of the 20 different
amino acids has a specific side chain which determines the chemical and functional
properties of the residue; they are classified as nonpolar, polar, acidic, basic, or
aromatic.
The primary sequence can adopt stable configurations of local short-range sec-
ondary structures mediated by intraresidue peptide bond or “backbone” interac-
tions. Common examples of secondary structures are α-helices, β-sheets, turns, and
unstructured (random coil) segments. The structure of an α-helix is stabilized by
H-bonding between the nitrogen of the amide and the carbonyl carbons of peptide
bonds spaced four residues apart. β-sheets are composed of two or more different
linear sequences of at least 5–10 amino acids, stabilized by H-bonding between the
nitrogen of the amide of one amino acid and the carbonyl carbons of the amino acid
in the adjacent strand. β-sheets can contain strands which are far away from each
other in the linear amino acid sequence, but adjacent in the folded structure of the
protein. Parallel β-pleated sheets consist of strands that have the same (parallel)
orientation on the linear sequence or primary structure of the protein, while in the
more common antiparallel β-sheet, the amino acids at the strand–strand interface
are coming from opposite directions. Other secondary structures such as bends and
turns may exist in specific structures or conformations.
The final functional three-dimensional, or tertiary, structure of a protein is sta-
bilized by a large number of noncovalent interactions that are mediated almost
exclusively by side-chain interactions between amino acids that can be from very
different regions of the primary structure, but are brought into close proximity as
the protein folds. Tertiary structures result from further bonding between side
chains within the protein and with any water that may be present around the pro-
tein. The “hydrophobic effect,” which is a thermodynamic description of the driv-
ing forces stabilizing the tertiary or native structure of a protein, is due in large part
to the side-chain-specific interaction of H2O with a protein (Sharp et al. 1991;
Murphy and Freire 1992). The hydrophobic effect tends to facilitate the partition-
ing of polar amino acids on the outside of the folded protein, with nonpolar amino
acids clustering in the inside of the protein in its final stabilized structure. Examples
of tertiary contacts include hydrophobic (van der Waals) interactions, dipole- and
induced-dipole interactions, disulfide bonds, electrostatic and ionic interactions,
and hydrogen bonds.
Multidomain proteins are common, where different conserved functional
domains are present within a primary sequence and fold into separate domains often
connected only by the single strand of amino acids of the primary structure. These
domains can show cooperative behavior, with one domain influencing the folding of
other domains, or the folding of individual domains may be thermodynamically
independent of the rest of the protein. For example, different families of kinases
exist with a somewhat conserved catalytic domain but differ in terms of the auxiliary
3 Techniques for Higher-Order Structure Determination 35
domains that regulate kinase folding, function, and specificity. Likewise, many
proteins contain pro-domains that function only to facilitate proper folding, then
these are excised via proteolysis to produce the final folded active structure.
Quaternary structure refers to the association of folded domains from different
amino acid chains. It is in some ways an extension of tertiary interactions, arising
from ordered interactions between two or more polypeptides subunits. The bonds
in each monomer are the same as those found in the tertiary structure. The poly-
peptide subunits are attached either by peptide strand or noncovalent interactions.
Common quaternary structures include dimers (transcription factors), trimers
(some cytokines), and higher-ordered oligomers. Allostery is one additional
dynamic property that is common among multimeric proteins, where different
functional properties are observed for unique quaternary structures that are only
available to an assembly of monomers. One well-characterized example of allo-
stery is hemoglobin that gives rise to cooperative O2 binding as well as its differen-
tial affinity for O2 in the lungs and tissues based on the different conditions
(e.g., pH) of the surrounding tissues.
Structural analysis of protein therapeutics has been an important focus for the
biopharmaceutical industry, as described in the first chapter of this book. The goal
of successful protein therapeutic development is the creation of a therapeutic pro-
tein with native-like structure, good bioactivity, and stability that elicits a minimal
immune response. Moreover, protein conformational stability can correlate phe-
nomenologically with in vivo stability. Protein aggregation and chemical modifica-
tion are the two most common routes for protein degradation and loss, with
aggregation depending more on protein structure and stability (Putnam et al. 2010).
Aggregates are heterogeneous mixtures of native folded species, nonnative unfolded,
or partially unfolded species, self-associated into larger molecules or polymers.
Multi-monomeric aggregates may continue growing in regular shapes such as
fibrils, or as amorphous particles, and can range in size from oligomers ( nanometers)
to subvisible (microns) or visible (hundreds of microns) particulates (Stefani and
Dobson 2003).
Manufacturing processes are continually optimized during therapeutic devel-
opment to improve product yield and product quality, while changes in formula-
tion and delivery methods can also occur in response to information on the target
patient population. All of these changes could alter the physicochemical proper-
ties of the protein in use, including chemical modification, self-association, and
ultimately the tertiary structure and protein folding. Thus, there is a growing need
for the development of biophysical tools that are sensitive to subtle changes in
protein higher-order structure (HOS) during biopharmaceutical development.
This chapter provides a brief review of technologies that are commonly used and
available to characterize various higher-order structural properties, with an
intended application toward protein biotherapeutics. Applications of these tech-
niques throughout the product lifecycle are provided in many subsequent chapters
in this volume.
36 J. Kranz et al.
3.2 Spectroscopic Techniques: Structural Characterization
3.2.1 Absorption Spectroscopy
Spectroscopy and spectroscopic techniques predate our knowledge of the existence

of proteins and were in fact instrumental in the earliest characterization of protein
chemistry and structure. Some of the earliest examples of absorption spectroscopy
in proteins involve porphyrin-containing enzymes (Chance 1953; Chance and
Pappenheimer 1954; Stern 1937). Electromagnetic radiation at a certain wavelength
λ (or frequency ν = c/λ) interacts with a sample, followed by measurement of some
properties of the radiation that emerges. Absorbance is the fraction of incident radi-
ation that is absorbed or dissipated. The absorption spectrum of the material repre-
sents the fraction of photons that are absorbed by a material over the range of
frequencies monitored.
It is challenging to succinctly explain the interaction of light with matter, setting
aside a quantum mechanical description. Light is a rapidly oscillating electromag-
netic field; likewise, molecules are comprised of bonded atoms with their own dis-
tributions of spin states with distinct electromagnetic properties. When the frequency
(and energy) of an incident photon matches that of an allowed electronic transition
of a molecule, that electron absorbs the photonic energy jumping from its ground
state (highest occupied molecular orbital or HOMO) into a transitory excited state
(lowest unoccupied molecular orbital or LUMO). This metastable electronic state
relaxes back to the ground state through various mechanisms, including radiative
ones (fluorescence, phosphorescence), non-radiative pathways such as vibrational
modes (Raman spectroscopy, FTIR), or solvent quenching.
Both the amplitude and frequencies of emitted light provide sources of informa-
tion on the molecular environment and dynamics within a system. These are often
summarized in a Jablonski diagram (Fig. 3.1), which schematically depicts which
transitions are observed for absorption and relaxation pathways, among the numer-
ous possible transitions. The vertical axis depicts the energy (and frequency) of
transitions, showing that the highest energy transition (S0 → S2) will occur at the low-
est frequency, with all other transitions being of lower energy and corresponding
higher frequency. What is relevant to an understanding of proteins and their HOS is
how the local environment can influence which transitions are or are not observed
and how a change in the probability of observing different transitions relates to an
understanding of fluctuations within a protein sample. In proteins the amide back-
bone and aromatic amino acids and disulfide bonds are the intrinsic chromophores
involved in absorption spectroscopy. The backbone absorbance spectrum consists
of two major electron transitions, the π to π* transition at 195 nm and the weaker n
to π* transition occurring at about 220 nm, the intensity of which depends on the
secondary structure of the protein.
Among applications to macromolecules, there are unique chemistries that afford
distinct opportunities for selective interrogation. Most proteins contain aromatic
amino acids, whose pi-orbitals have lower energy (higher frequency) absorbance
Fig. 3.1 A Jabloński diagram depicting fluorophore absorption, radiative, and non-radiative

relaxation pathways. Singlet ground, first, and second electronic states are represented by S0, S1,
and S2. Within each of these energy levels are vibrational energy levels. Transitions between states
are shown as vertical lines, representing the instantaneous nature of light absorption, that is, the
Franck–Condon principle. Energy losses through non-radiative conversions give rise to the Stokes’
shift, a difference in frequency between absorptive and radiative transitions
bands compared to a typical carbon–carbon bond. The indole side chain of

tryptophan has the most intense absorbance with a maximum around 280 nm and a
weaker transition around 292 nm. Tyrosine has a weaker absorbance, with the stron-
gest transition occurring at 276 and weaker transitions as shoulders at 267 and
280 nm. Phenylalanine has the weakest absorbance which occurs at 250–270 nm.
Modifications of any of these residues including oxidation and ionization can affect
the absorbance spectra of these chromophores. The disulfide bonds have a very
weak absorbance signal from 250 to 300 nm which can also contribute to the inten-
sity of the protein absorbance in this part of the spectrum.
Absorption frequencies are relatively insensitive to molecular environment of
these aromatic residues, which provides a useful means of quantifying protein con-
centrations (Beer’s Law or the Beer–Lambert equation) or in any number of appli-
cations for detecting proteins (chromatography):
 I
A = log 10   = ecl, (3.1)
 I0 

where A is the measured absorbance, I0 is the original intensity of the light and I is
the intensity of the light transmitted through the sample, l is the length of the sample
path length the light was shown through, c is the concentration of the chromo-
phores, and ε is the absorbance of a one molar solution of the chromophores, often
38 J. Kranz et al.
referred to as the extinction coefficient of a chromophore. Taking the derivative of

the spectrum can allow the detection of changes in the wavelength of maximum
absorbance due to changes in solvent, and this is occasionally used to follow changes
in conformation or aggregation state.
As the subsequent sections will reveal, relaxation pathways can be greatly influ-
enced by the molecular environment of the chromophores within the folded struc-
ture of the protein and of the protein itself. These different relaxation pathways have
been exploited as sensitive tools to probe the secondary, tertiary, and quaternary
structures of proteins.
3.2.2 Fluorescence
3.2.2.1 Intrinsic Protein Fluorescence
Fluorescence is arguably one of the most widely used experimental techniques

employed in the study of macromolecules and macromolecular interactions, both
in vitro and in vivo. Many thorough reviews representing the array of applications
of this technique are available in the literature (Cantor and Schimmel 1980;
Lakowicz 2006; Royer 2006). Over the last few decades, continued improvements
in experimental techniques and in computational analyses have revolutionized the
use of spectroscopic techniques in understanding mechanisms of protein folding
and perturbations of protein structure (Bartlett and Radford 2009; Royer 1995;
Royer and Scarlata 2008).
The phenomenon of fluorescence was first observed by Sir George G. Stokes in
1852 (Stokes 1852), noting that the fluorescence emission typically occurs at lower
energy or longer wavelengths than the incident light. For Stokes, the source of UV
excitation was the sun and a blue glass filter (from a stained glass window) which
selectively transmitted light <400 nm. This incident light was absorbed by a glass of
quinine; fluorescence was filtered through a glass of white wine, which selectively
transmitted light >450 nm and subsequently detected by eye. Energy losses between
excitation and emission are universal for fluorescent molecules in solution, giving
rise to the Stokes’ shift (Fig. 3.1).
In proteins, tryptophan is the strongest intrinsic fluorophore, followed by tyro-
sine and then phenylalanine. Proteins tend to contain relatively fewer tryptophans
than tyrosines and phenylalanines, and due to its strong excited-state dipole, trypto-
phan fluorescence tends to be more sensitive to general environmental factors. Both
Trp and Tyr absorb light around 280 nm, and due to the overlap of their excitation
and emission spectra, if the two amino acids are located close enough in the folded
protein, this commonly results in energy transfer from Tyr to Trp. Thus, the Trp
emission is the one most often observed in proteins.
Due it its aromatic character, tryptophans are often buried or partially buried in
the hydrophobic core of protein interiors. Any disruption or destabilization of pro-
tein structure can lead to a change in solvent exposure and thus the fluorescent
properties of tryptophan residues (Bartlett and Radford 2009; Royer 1995; Royer
and Scarlata 2008). This is most clearly observed in systems with a single trypto-
phan (Epstein et al. 1971; Flanagan et al. 1992; Hynes and Fox 1991; Otto et al.
1994), but also in studies of proteins containing multiple tryptophan residues
(LeTilly and Royer 1993; Mann et al. 1993; Royer 1993). While the emission spec-
trum maximum of a buried tryptophan generally exhibits a red shift (lower energy,
higher frequency) upon protein unfolding, the peptide backbone and a number of
the amino acid side chains can alter tryptophan fluorescence spectra, generally
through excited-state quenching or electron transfer (Adams et al. 2002).
3.2.2.2 Fluorescence and Conformational Dynamics
All proteins are inherently flexible in solution, existing as a distribution of a popula-

tion of low- and high-energy conformational states that interconvert on different
time scales. There are a variety of advanced applications that utilize fluorescence as
a probe of conformational dynamics which are described in detail elsewhere (Cantor
and Schimmel 1980; Lakowicz 2006; Royer 1995, 2006; Bartlett and Radford 2009;
Royer and Scarlata 2008; Kamerzell et al. 2011). Time-resolved fluorescence has
been used to measure protein dynamics that occur in time scales ranging from sub-
nanoseconds to seconds, via multiple techniques. Some are steady-state measure-
ments, while others are kinetic. In terms of relevance to characterization of
higher-ordered protein structure, and fluctuations in structure, their use varies with
respect to the simplicity of the measurement and the type of information obtained.
Applications include studies of local and global motions, flexibility, heterogeneity,
and intermolecular interactions and generally employ a variety of techniques
(Kamerzell et al. 2011; Ramsey et al. 2009).
Intrinsic fluorescence is commonly used in time-resolved investigations of pro-
tein folding and unfolding. Unfolding is induced by exposure to temperature, chem-
ical denaturants, or other solution conditions. Returning the protein to the original
conditions will initiate refolding. The sophistication of these experiments continues
to improve, with advances in detection speeds and data analysis. Generally single
wavelength intensities are monitored as a function of the folding or unfolding time
course via stopped flow or temperature-jump techniques.
Anisotropy measurements provide information on the size and shape of proteins
and have been applied extensively in studies of protein–protein interactions.
Anisotropy, A, is based on the principle of photoselective excitation of fluorophores
by plane-polarized light:
(I − I ⊥ )
A= . (3.2)
(I + 2 I ⊥ )

Fluorophores that are physically oriented such that their electronic dipole
moment is aligned in the plane of incident light are selectively excited. After rotational
40 J. Kranz et al.
or vibrational diffusion occurs, some fraction of molecules have their dipoles

oriented out of the plane of incident light, giving a decrease in the signal detected
parallel to incident light and an increase in the signal detected perpendicular.
Steady-state anisotropy employs plane-polarized light, with a pair of polarizing fil-
ters before and after the sample; the result is a value of fluorescence anisotropy or
polarization that is specific to a sample composition and temperature.
Time-resolved measurements of fluorescence or anisotropy use rapid laser-based
excitation and rapid detection to follow the decay kinetics. Also called lifetime mea-
surements, time-resolved experiments are useful in estimating molecular rotational
correlation times and protein hydrodynamic radii and how these change in ways that
are sensitive to formulations (Ramsey et al. 2009). The relationship between the
anisotropy, A, and the lifetime and rotational correlation time is given by the Perrin
equation (Perrin 1926):
A0 t
−1 = , (3.3)
A tc
where A0 is the limiting anisotropy, τ is the fluorescence lifetime, and τc is the rota-
tional correlation time of the macromolecule. For globular (approximately spheri-
cal) proteins, the rotational correlation time is a function of the apparent molecular
weight, M, of a protein by
hV hM
tc = = (n + h), (3.4)
RT RT
where η is the viscosity, n- is the specific volume of the protein, h is the relative
hydration, T is temperature (K), and R is the universal gas constant. For a typical
- ~ 0.73 mL/g and hydration levels are h ~ 0.23 g of H O per gram of pro-
protein, n 2
tein. This expression predicts that the correlation time of a hydrated protein is on
average ~30 % larger than would be predicted for a fully anhydrous protein (spheri-
cal approximation).
The apparent volume of a protein can be measured experimentally, solely from
measurements of protein anisotropy as a function of varying viscosity; changes in
viscosity are typically achieved by variable temperature combined with viscous
cosolvents, like glycerol. The Perrin equation is substituted with terms from Eq. 3.4,
giving
1 1  tRT 
= + . (3.5)
A A0  t chV 

The use of Eq. 3.4 to measure the apparent volume of a protein is one of the earli-
est applications of fluorescence in a biochemical system (Weber 1952a, b). As dis-
cussed in subsequent sections, the apparent volume of a protein can vary significantly
in response to changes in formulations, aggregation onset, and protein unfolding.
Another laser-based technique is fluorescence correlation spectroscopy, or FCS,

which measures time-dependent fluctuations in fluorescent intensity arising from
Brownian motions. This technique is now commonplace in studies of in vivo protein
compartmentalization and changes in expression and can also be used to measure
protein dynamics in simple in vitro systems (Foldes-Papp et al. 2002). Applications
to in vitro systems involve observations of single molecules undergoing folding/
unfolding transitions, generally occurring on the microsecond to second time scale.
Design of a correlation spectroscopy experiment to observe fluctuations in pro-
tein structure requires that a conformational change gives rise to a change in fluo-
rescent dye intensity. The majority of FCS experiments involve molecules freely
diffusing in and out of the small volume being monitored, with fluctuations provid-
ing insights into diffusion, either passive or active, within cellular environments.
3.2.2.3 Fluorescence Quenching
Quenching refers generically to any molecular process that decreases the intensity
of fluorescence. These include molecular rearrangements, energy transfer, excited-
state reactions, and quenching through collisions. Fluorescence quenching has been
widely studied as a source of information on biological systems. During collisional
quenching, the quenching agent must diffuse to the fluorophore during the lifetime
of the excited state. Upon contact, there is energy transfer via interactions of the
excited-state dipole and the dipole (or induced dipole) of the quencher, facilitating
return of the fluorophore to the ground state without photon emission.
The requirement of intermolecular contact between the quencher and fluoro-
phore has obvious applications to biochemical systems. Quenching measurements
can be indicative of the relative accessibility of fluorophores to quenchers. For exam-
ple, in systems with a few fluorophore groups, either buried in the protein interior of
soluble proteins or in lipid environments of transmembrane proteins, quenching
rates can be used to determine the permeability of proteins or membranes to the
extrinsic quencher. Likewise, quenching rates and lifetimes of fluorophore and
quencher groups can be useful in measurements of molecular diffusion. Details on
collisional quenching theory and applications are prevalent in the literature (Cantor
and Schimmel 1980; Lakowicz 2006).
3.2.2.4 Fluorescence Resonance Energy Transfer
Fluorescence resonance energy transfer (FRET) is a technique that utilizes interac-

tions between fluorophores with overlapping energy states to determine distances
between them (Lakowicz 2006). In this technique the two molecules of interest are
labeled at specific sites with donor or acceptor fluorophores which have overlapping
emission/excitation spectra. As the distance between the molecules decreases, the
efficiency of energy transfer from the donor to the acceptor molecule increases.
Based on the characteristics of the labels, the actual distance between them can be
42 J. Kranz et al.
calculated (dos Remedios and Moens 1995). This technique can also be used to
follow changes in the distance between parts of the same polypeptide with changes
in solution condition, interactions with other proteins, etc.
Most applications of FRET utilize extrinsic fluorophores, ones that involve
cofactors as in the process of photosynthesis or more commonly are covalently
attached to a protein via chemical means. These have found widespread applica-
tions in labeling of cellular proteins used in studies of cellular trafficking or in vivo
protein–protein interactions (Pollok and Heim 1999; Miyawaki 2011; Jares-Erijman
and Jovin 2003) and in studies of single molecules (Ferreon et al. 2011; Kubelka
et al. 2004). Within proteins there is intrinsic fluorescence energy transfer between
the tyrosine, the donor amino acid, and tryptophan, the acceptor molecule. These
effects are perhaps most observable in simplified, highly purified protein solutions
such as protein therapeutics or in studies of protein folding kinetics, where limits of
low signal intensity among Tyr and Trp FRET pairings are not so problematic.
In general, however, endogenous fluorophores are not heavily used for FRET-based
studies of proteins.
3.2.2.5 Extrinsic Fluorescence
Almost 60 years ago, Weber demonstrated that the quantum yield of 1-anilino-
phthalene-8-sulfonic acid (1,8-ANS) increased significantly from a highly quenched
state in water (0.004) when bound to bovine serum albumin (0.75) (Daniel and
Weber 1966; Weber and Daniel 1966; Weber and Laurence 1954). The mechanism
of this effect involves a relief from collisional quenching from H2O molecules due
to direct binding to a hydrophobic drug binding site on albumin. More commonly,
the nonspecific binding of ANS to hydrophobic surfaces of proteins has been used
as a general indicator of partial or complete protein unfolding.
In contrast to intrinsic protein fluorescence, which arises from the naturally fluo-
rescent amino acids present in a given protein, extrinsic fluorescent dyes offer addi-
tional possibilities in the characterization of proteins. These can be covalently
attached to proteins (via amines as with lysine or the amino group of the N-terminus
or via free thiol groups on cysteine) or may interact with a protein via noncovalent-
binding interactions. While ANS and other extrinsic fluorescent dye probes are a
useful general probe of a change in the conformation of a protein (Hawe et al. 2008,
2011), they provide only a qualitative, nonlocalized means of investigating struc-
tural perturbations. As described in Chap. 2, these properties have been exploited
for high throughput assays, especially as screening assays, where qualitative rank-
ing of different molecules or conditions is the desired outcome. When more quanti-
tative assessment is needed, extrinsic fluorescence is generally complimented by
other techniques.
Extrinsic dyes work via excited-state reactions. These are molecular processes
which change the electronic or chemical structure of the excited state prior to relax-
ation via fluorescence. The best known example of an excited-state reaction is that
of phenol deprotonation, which occurs much more readily from its excited state due
Fig. 3.2 A Jablonski diagram depicting fluorescence with solvent relaxation
to a shift of electrons from the phenolic hydroxyl onto the phenol ring, which
acidifies the hydroxyl group. Collisional quenching and solvent relaxation are other
well-characterized examples of excited-state reactions, and the ones most relevant
to extrinsic dyes.
Solvent relaxation is one aspect of the local environment of a polar fluorophore
that can significantly influence its emission spectrum. This is the origin of the
Stokes’ shift. Spectral shifts result from the interaction of the excited-state fluoro-
phore dipole with solvent dipoles, leading to another energy transfer pathway,
depicted in Fig. 3.2. This pathway for energy transfer is due in large part to the
kinetics of different phenomena and can result in large Stokes’ shifts of ANS and
similar dyes. Absorption lifetimes are affectively instantaneous (10−15 s) with respect
to atomic motions, the Franck–Condon principle. The excited-state lifetime of fluo-
rophores are typical 10−9 to 10−8 s. Solvent motions are generally on the order of
10−10 s, which allows sufficient time for solvent dipoles to align with the fluorophore
excited dipole and for non-radiative transfer to occur. Fluorescence still occurs, but
at a lower, relaxed energy level for the excited state. With the loss of the fluorophore
dipole upon fluorescence emission, the aligned solvent molecules are at a higher
energy state, which relaxes to the original ground state through non-radiative pro-
cesses. The end result is a fluorescence that is red shifted (lower energy) from the
original absorption frequency.
Steady-state fluorescence spectroscopy is the most common experimental tech-
nique used to follow the interactions of extrinsic dyes with proteins. Protein binding
44 J. Kranz et al.
affects the solvent–dye interactions, resulting in changes in emission wavelength or

intensity. Other applications discussed above can be used in conjunction with the
addition of extrinsic dye for the detection of protein structural changes as well,
including time-resolved fluorescence (lifetime measurements), anisotropy (rota-
tional motions), and fluorescence correlation spectroscopy (translation/diffusion).
3.2.3 Fourier Transform Infrared Spectroscopy
The use of Fourier transform infrared (FTIR) spectroscopy to measure the second-
ary structure of protein molecules has dramatically expanded in the last few decades,
encouraged by the need to measure protein secondary structure in samples with
various physical forms without sample manipulation. This method can be used to
analyze molecules ranging from a few amino acids to large protein complexes, in
different physical forms including solutions, lyophilized or crystallized solids, gas,
and chemically modified or embedded in polymers (Perez et al. 2002; Carpenter
et al. 1998; Haris and Chapman 1995; Chalmers 2002; Griebenow and Klibanov
1995). Additionally, modern FTIR spectrophotometers require only small volumes
of solution for analysis, 10–100 μL, though at a relatively high concentration of
10 mg/mL and higher (Perez et al. 2002; Zuber et al. 1992). Compared to other
spectroscopic techniques, FTIR can produce high-quality spectra relatively easily
without the complications of background fluorescence, light scattering or problems
related to the size of the protein. However, the presence of water absorption, buffer
components, or other biological molecules produces problems of spectral overlap
that can only be successfully subtracted and resolved by mathematical approaches
(Susi et al. 1967; Venyaminov and Kalnin 1990a, b; Dong et al. 1990; Dousseau and
Pezolet 1990). Thus, FTIR spectroscopy can be used to study various biological
systems including proteins in pharmaceutical and non-pharmaceutical formulations
(Carpenter et al. 1998; Haris and Chapman 1995; Griebenow and Klibanov 1995;
Krimm and Bandekar 1986).
FTIR spectroscopy typically refers to the absorption of infrared light in the mid-
IR region (400−4,000 cm−1). The Michelson interferometer (Fig. 3.3) is the central
part of an FTIR spectrometer and is unique among other spectroscopic instrumen-
tation. Infrared radiation from the source is collected and collimated before it
strikes the beam splitter. The beam splitter ideally transmits one-half of the radia-
tion and reflects the other half. Both transmitted and reflected beams strike mirrors,
which reflect the two beams back to the beam splitter. Thus, one-half of the infra-
red radiation has first been reflected from the beam splitter to the moving mirror
(Mirror 2) and then back to the beam splitter prior to being shone through the
sample. The other half of the infrared radiation going to the sample first goes
through the beam splitter and then is reflected from the fixed mirror (Mirror 1)
back to the beam splitter. When these two optical paths are reunited, interference
occurs at the beam splitter because of the difference in optical paths caused by the
scanning of the moving mirror.
Fig. 3.3 Schematic of the principle of Michelson interferometer in FTIR spectrometer
The interference signal measured by the detector as a function of the optical path
length difference is called the interferogram (Chalmers 2002; Griffiths and de
Haseth 1986; Smith 1996). In order to extract and present information as an absor-
bance or transmittance spectrum, the interferogram has to be Fourier transformed,
generating a single beam spectrum of IR absorption as a function of wavelength.
Instrumental and atmospheric contributions are superimposed in the primary IR
spectrum. To eliminate these contributions, a background spectrum obtained with-
out a sample must be recorded and subtracted. Finally, the sample spectrum must be
normalized, by dividing the sample spectrum signal, I, and by the background sig-
nal, I0. The result is a %-transmittance spectrum (as T = I/I0). For liquids, transmit-
tance is related to absorbance A as
I   1
A = log10  0  = log10   . (3.6)
 I T

Physically, FTIR is a type of vibrational spectroscopy measuring the fluctuating
covalent bonds within a protein; these give rise to specific wavelengths for different
modes of vibration, the intensity relating to the number of bonded atoms in a com-
mon chemical environment. There are many modes of vibration a chemical bond can
exhibit in a given chemical structure which result in IR absorbance, even for simple
molecules such as CO2 (Fig. 3.4). Given the vast number of normal modes, the vibra-
tional FTIR spectrum is complex with many vibrational bands overlapping. However,
it is often possible to select a spectral region that provides specific structural informa-
tion. Table 3.1 shows spectral regions of common chemical structure vibrations. For
analysis the spectra are usually deconvoluted. Better visual resolution of the different
bands can be achieved by taking the second derivative, and the data is often presented
as the deconvoluted or second derivative of the raw spectrum.
46 J. Kranz et al.
Fig. 3.4 Vibrational modes of CO2 molecule
Table 3.1 Common vibrations with chemical structure assignments

and spectral region
X–H vibrations Bond Wave numbers (cm−1)
Hydroxyl O–H 3,610–3,640
Amines N–H 3,300–3,500
Aromatic rings C–H 3,000–3,100
Alkenes C–H 3,020–3,080
Alkanes C–H 2,850–2,960
Triple bonds 2,500–1,900
Double bonds 1,900–1,500
Heavy atoms 1,500–
Protein secondary structural information is measured by FTIR via monitoring

the vibrations exhibited within the peptide backbone. Each vibration mode absorbs
IR electromagnetic radiation at specific frequencies depending on the peptide bond
orientation and the protein secondary structure associated with a given residue. The
vibrational modes of C=O stretching and N–H bending around the peptide bond,
called amide I and amide II vibrational modes, are shown in Fig. 3.5. Up to nine
different characteristic IR bands, the amide A, B, I, II, III, IV, V, VI, and VII have
been reported for protein molecules measured by FTIR (Miyazawa et al. 1956;
Bandekar 1992). Amide III and IV are very complex bands resulting from a mixture
Fig. 3.5 Schematic
representation of examples of
peptide bond vibrations from
amide I and amide II
of several coordinate displacements (Fu et al. 1999; Schweitzer-Stenner et al. 2002).

The out-of-plane motions are found in amide V, IV, and VII (Krimm and Bandekar
1986; Bandekar 1992). The amide A band around 3,500 cm−1 and amide B around
3,100 cm−1 originate from a Fermi resonance between the first overtone of amide II
and the N–H stretching vibration (Haris et al. 1986, 1990). The use of these amide
vibrations to extract protein secondary structure has been limited due to the com-
plexity of these bands, the possible interference arising from side chains such as the
ionized side chains of aspartic acid and arginine and the overlap of different vibra-
tions giving rise to these bands (Dousseau and Pezolet 1990). Therefore, the major-
ity of the work in developing FTIR has been focused on analyzing amide I vibrations
to monitor protein secondary structure (D'Antonio et al. 2012).
Absorption associated with the amide I band leads to stretching vibrations of the
C=O bond of the amide. This band is sensitive to protein conformation because of
its involvement in hydrogen bonding in different elements of secondary structure.
Studies with proteins of known structure have been used to systematically correlate
features of the amide I band to secondary structure content (Byler and Susi 1986;
Surewicz and Mantsch 1988). The second derivative amide I IR spectrum of lyso-
zyme is shown in Fig. 3.6. Assignments of each peak under the amide I band to
specific secondary structure elements have been a matter of interest for many
researches and have been assigned using model peptides, X-ray crystallography,
and mathematical modeling. Table 3.2 shows typical amide I second derivative band
assignments to various secondary structures.
In order to quantify the percentage content of each secondary structure element
in a protein, a curve fitting procedure can be applied to estimate quantitatively the
area of each component representing a type of secondary structure (Dousseau and
Pezolet 1990; Byler and Susi 1986; Susi and Byler 1986; Jackson and Mantsch
1995). The percentages of the different secondary structure elements obtained by
curve fitting of amide I second derivative IR spectra for various proteins have been
in good agreement with the secondary structure information obtained from their
X-ray crystallographic structures. An example of the curve fitting of lysozyme
amide I FTIR second derivative spectrum is also shown in Fig. 3.6.
48 J. Kranz et al.
Fig. 3.6 Amide I FTIR spectrum second derivative of lysozyme in 50 mM phosphate buffer at pH
7 and 25°C. Red line is showing the derivative curve fit. Dotted line is showing the curve fit of
individual peaks
Table 3.2 Common band assignments of protein secondary structure

in amide I region
Secondary structure assignment Spectral region (cm−1)
Alpha-helix 1,645−1,662
Beta sheet 1,613−1,637, 1,695−1,682
Unordered 1,645−1,637
Turns 1,682−1,662
FTIR spectroscopy is widely used in the development of protein therapeutics.

It has been used to measure the secondary structure of the targeted protein as part of
HOS characterization. Quantifying the secondary structure components such as
α-helix, β-sheet, bends, and turns has been used to better understand the HOS prod-
uct quality attributes of the protein under study and used to support process develop-
ment, formulation development, stability studies, and comparability studies; some
of these applications are discussed in the last half of this book. Qualification of this
technique for use in comparability and filing documents is described in Chap. 5.
3.2.4 Circular Dichroism
Nearly all organic molecules and macromolecules are “optically active,” arising
from a lack of chemical symmetry. For proteins the symmetry is provided by the
three-dimensional structure of the folded protein. The CD spectrum arises from an
electronically symmetrical chromophore in an asymmetric environment. The chro-
mophores that absorb light in the UV range are the peptide backbone and the aro-
matic amino acids and disulfide bonds. The result is electromagnetic interactions
between neighboring chromophores that can be detected spectroscopically. There
are a number of ways in which optically active samples can alter properties of trans-
mitted light, including linearly polarized light, circularly polarized light, circular
birefringence, and circular dichroism (CD) spectroscopy. With CD spectroscopy,
one measures the differential absorption of either left-hand or right-hand compo-
nents of circularly polarized light, the output being a measure of ellipticity, θ, of a
sample as a function of different wavelengths. Because this is a difference spectrum,
the signal can be either positive or negative, with the blank at zero.
The contribution of different secondary structural elements to the total CD signal
is additive, but also wavelength-dependent (Table 3.3). Thus, CD can be used to
estimate the relative fraction of different secondary structural elements of a protein,
α-helix, β-sheet, and random coil, directly from the CD spectrum without any other
experimental input, as described extensively by Greenfield (Greenfield and Fasman
1969; Greenfield 1996, 2006a, b, c, 2007) as well as many others. However, the
structural environment around the aromatic residues can result in signals in the far-
UV CD spectrum that can complicate this analysis.
In the near-UV CD region the combination of aromatic and disulfide environ-
ments present in the protein structure results in a complex spectrum which is like a
Table 3.3 Features of protein CD spectra

Far-UV CD
Range (190–240 nm)
Source of signal Peptide bond in asymmetric environment; protein secondary structure
α-Helix Negative at 208 nm
Negative at 222 nm
Positive at 192 nm
β-Sheet Negative at 218 nm
Positive at 196 nm
Random coil Positive at 212 nm
Negative at 195 nm
Near-UV CD (260–320 nm): aromatics in asymmetric environment of folded
protein; 3° structure
Range (260–320 nm)
Source of signal Aromatics in asymmetric environment; protein tertiary structure
50 J. Kranz et al.
finger print for that particular protein. Each of the aromatic amino acids has a char-
acteristic wavelength profile that corresponds to their absorbance spectra. Trp shows
a peak close to 290 nm with fine structure between 285 and 305 nm; Tyr has a peak
between 270 and 285 nm, with a shoulder at longer wavelengths often obscured by
the Trp band; Phe shows bands with fine structure between 250 and 265 nm.
Disulfide bonds also absorb in the near-UV region (weak broad absorption bands
from 250 to 280 nm), the changes in the dihedral angle of the disulfide bond will
result in a change in the signal in this region of the spectrum. The actual shape and
magnitude of the near-UV CD spectrum of a protein will depend on the protein
primary sequence, the number of each type of aromatic amino acids present, their
mobility, and the nature of their environment (H-bonding, polar groups, and polariz-
ability). While specific residue assignments cannot be made, changes in the protein
fold result in changes in the environment of the aromatic amino acids and disulfide
bonds and thus in the near-UV CD spectrum. As with fluorescence, the complexity
of the aromatic CD spectrum increases with increasing numbers of aromatic groups.
CD spectroscopy has been used extensively to characterize protein folding and
unfolding, both as a function of denaturants and temperature. It is a truly general
technique that is not limited to constraints of molecular size or chemical composition
and is a “label-free” approach that does not require secondary conjugation or chemi-
cal modification to obtain data on protein structural composition. Moreover, modern
improvements in CD instrumentation and data fitting have made it straightforward
to collect full spectra as a function of temperature, providing a means to quantitate
the presence of subtle fluctuations in protein structure that occurs at temperatures
below the primary protein unfolding transition.
3.2.5 Vibrational Optical Activity
Vibrational optical activity (VOA) is the field of spectroscopy associated with com-
bining infrared absorption or Raman scattering with the optical activity, manifested
as either optical rotation or CD. These include infrared vibrational circular dichro-
ism (VCD) and vibrational Raman optical activity (ROA). VCD is an extension of
electronic circular dichroism (ECD) described above, using an infrared spectrome-
ter (FTIR) to measure the difference in the IR intensity for left minus right circularly
polarized radiation, while ROA measures the difference in intensity of right minus
left circularly polarized Raman scattered radiation in several different instrument
configurations (Nafie 2011). VOA measurements are complementary to typical
spectroscopic methods such as CD and FTIR spectroscopy, but with additional sen-
sitivity in monitoring subtle changes in tertiary structure (Nafie 2011; Nafie and
Dukor 2007; Lakhani et al. 2009; Cao et al. 2008; Keiderling et al. 2006; Zhu et al.
2006; Barron 2006).
In a typical optical activity instrument, light emerging from a Michelson interfer-
ometer is polarization modulated using a photoelastic modulator (PEM), with light
oscillating between left and right circularly polarized states at the PEM frequency
Fig. 3.7 Diagram of a typical FTIR-VCD optical path
(Fig. 3.7). The infrared light is modulated by the Michelson interferometer and then
enters the VCD portion of the setup. By passing through a linear polarizer and a
PEM, the Fourier-modulated IR beam is further modulated at the PEM frequency.
To obtain a VCD spectrum, the doubly modulated signal is first demodulated at the
PEM frequency by a lock-in amplifier, then Fourier transformed by a computerized
system (Keiderling et al. 1999).
A chiral molecule interacts differently with left and right circularly polarized light, in
contrast to classical IR spectroscopy in which vibrational excitation occurs with nonpo-
larized IR radiation and thus does not vary due to differences in chirality. In certain
applications of VCD, it is possible to infer relative absolute configuration between pairs
of structurally related molecules. Such correlations are strengthened due to the large
number of distinct vibrational bands in a typical mid-infrared and Raman scattering
spectrum. This high level of sensitivity to minor structural change gives optical activity
technologies, VCD and Raman optical activity, a high level of analytical power to eluci-
date both absolute molecular structure and conformation (Nafie and Dukor 2007).
Aromatic residues and polysaccharides are known to have overlap in the far-UV
CD region, but show good separation in a VCD spectrum (Shi et al. 2006; Dukor
and Keiderling 1991). The sensitivity of VCD to interactions of dipoles and the
combination of sign, pattern, and band shapes with frequency allow VCD to provide
local structural information and to distinguish between different secondary struc-
tures (Shi et al. 2006; Dukor and Keiderling 1991). Moreover, VCD, like the other
types of vibrational spectroscopy, can be used across different types of samples
including high concentration solutions (>50 mg/mL), solids, protein aggregates,
and foreign organic particles. VCD is useful for monitoring formation of amyloid-
like fibrils, with an order of magnitude increase in the VCD intensity upon fibril
formation, reporting on the long-range supramolecular chirality of protein fibrils
(Shi et al. 2006; Dukor and Keiderling 1991; Meyer et al. 2004; Baumruk and
Keiderling 1993; Yoder et al. 1997). Moreover, VCD spectroscopy should be trans-
parent to many commonly used formulation components that can interfere with CD
and FTIR spectroscopy measurements. Glycine is commonly used in protein
therapeutic drug product formulations, but the carboxyl group of free glycine in the
52 J. Kranz et al.
Fig. 3.8 Vibrational circular

dichroism (a) and FTIR
(b) spectra of lysozyme;
native, centrifuged, and
supernatant. Samples were
treated at pH 2 and heated at
60°C for 2 days
formulation is particularly problematic for IR spectroscopic characterization of

proteins from amide I overlap (Carpenter et al. 1998; Meyer et al. 2004). However,
in the case of VCD, glycine does not interfere with the amide I region since its C=O
stretch is lower in frequency.
VCD has been used to examine fibril formation in lysozyme and insulin. Under
conditions of acidic pH and incubation at 60°C, the fibril formation produced a
VCD spectral signature in the amide I centered at 1,627 cm−1 and amide II centered
at 1,515 cm−1 in IR region for both proteins, which was not observed in the
corresponding IR spectrum, shown in Fig. 3.8 (Ma et al. 2007; Kurouski et al.
2010). More recent studies have revealed spontaneous interconversion between
opposing supramolecular chiral forms of fibril structure and correlations between
the signs and magnitudes of the VCD spectra and the morphologies of fibrils as
observed by SEM and AFM microscopies (Kurouski et al. 2012).
Raman optical activity (ROA) scattering spectra can also provide very useful
information about changes in protein secondary structure. Additional ROA studies
have focused on the amide III region where particular sensitivity to partially or
Fig. 3.9 ROA spectra of human IgG2 in solutions at pH 3 and 7
inherently unfolded protein structure can be studied by monitoring protein tertiary

structure through signals from the aromatic amino acids. ROA is not impeded by
background water scattering, and hence the entire spectral region from 2,000 to
100 cm−1 can be used as a diagnostic of overall secondary and tertiary protein struc-
tures (Zhu et al. 2006). Several studies have been published illustrating the ability
of ROA to monitor aggregate formation in various protein therapeutics in various
formulations of monoclonal antibodies. Two classes of human monoclonal antibod-
ies, namely, IgG1 and IgG2, were studied using Raman and ROA at pH 7 and pH 3
(Li and Li 2009). The ROA spectra for both of these two antibodies exhibit large but
very similar differences at these two pH values (Fig. 3.9). The data demonstrate the
sensitivity of ROA to structural differences in HOS of these two antibodies, with
structural differences arising from a net unstructuring or from changes in the stereo-
conformational structure of the glycosidic tails (Li and Li 2009). These features are
readily measured using ROA, but are not observed by traditional Raman spectros-
copy (Li and Li 2009).
Both VCD and ROA are relatively new techniques which are used primarily to
provide in-depth characterization of specific proteins, but are not yet widely employed.
3.2.6 Raman Spectroscopy
Like infrared (IR) spectroscopy, Raman spectroscopy is a method for probing struc-
tures and interactions of molecules by measuring the energies (or frequencies) of
molecular vibrations. The main advantage of Raman spectroscopy for applications
54 J. Kranz et al.
Table 3.4 Raman bands features in Raman effect spectra

and their vibrational origin
Frequency range (cm−1) Band assignment
2,700–3,100 C–H alkyl free vibration
2,230 C≡N bond stretch
2,190–2,300 C≡C bond stretch
2,100–2,140 C≡C bond stretch
1,650–1,750 C=O bond stretch
1,600–1,675 C=C bond stretch
1,580–1,620 C=C bond stretch
990–1,010 Aromatic ring breathing
650–860 C–Cl stretch
to proteins and their complexes is the fact that liquid water (both H2O and D2O) does
not confound the Raman effect. Also like FTIR samples in different physical forms
can be analyzed including solutions, suspensions, gels, precipitates, fibers, single
crystals, and amorphous solids, with no complicated sample preparation necessary.
Raman spectroscopy is nondestructive, with high specificity for certain structures
and applicability to large supermolecular assemblies (Nafie 1996).
IR and Raman differ fundamentally in the mechanism of interaction between
radiation and matter (Benevides et al. 2004). Raman scattering is inelastic light
scattering which occurs at wavelengths that are shifted from the incident light by
the energies of the molecular vibrations. Typical applications include structure
determination, multicomponent qualitative analysis, and quantitative analysis
(Nafie 1996; Dong et al. 1998). The vibrational spectrum of a molecule is com-
posed of bands representing active normal vibrations. The spectrum depends on
the masses of the atoms in the molecule, the strength of their chemical bonds, and
the atomic arrangement. Consequently, groups of atoms connected by certain types
of bonds have certain characteristic vibrations in the Raman spectra. Table 3.4
shows a set of common chemical structures with their frequency regions in Raman
spectra (Socrates 2004).
A Raman system typically consists of four major components (Fig. 3.10a): an
excitation source such as a laser, a sample illumination system, light collection
optics, a wavelength selector which can be either a filter or spectrophotometer, and
the detector which can be a photodiode array, a charge coupled device, or a photo-
multiplier tube. A sample is normally illuminated with a laser beam in the ultravio-
let (UV) visible (Vis) or near infrared (NIR) range (Dong et al. 1998). Other
state-of-the-art Raman spectrometer systems have been described for various bio-
logical applications such as UV resonance Raman [UVRR (Brennan et al. 1997;
Hashimoto et al. 1993; Asher et al. 1993)], Raman optical activity [ROA (Nafie
1996; Grauw et al. 1997; Barron et al. 1996)], and confocal Raman microscopy
(Goldstein et al. 1996). In all cases scattered light is collected with a lens and sent
through an interference filter or spectrophotometer to obtain the Raman spectrum of
the sample being analyzed. In the case of Raman microscopy, the microscope is
Fig. 3.10 Schematic of (a) basic Raman spectroscopy components and (b) energy level diagram
showing the states involved in Raman signal. The line thickness is roughly proportional to the
signal strength from the different transitions
used to visualize the sample and ensure that the spectrum is from a particular p article
or region of the sample.
Raman spectra are obtained when the light interacts with the molecule and dis-
torts (polarizes) the cloud of electrons round the nuclei to form a short-lived meta-
stable state from which photons are quickly reradiated (Fig. 3.10b). If only electron
cloud distortion is involved in scattering, the photons will be scattered with very
small frequency changes. This elastic scattering reaction is called Rayleigh scatter-
ing and is the dominant process following electron cloud distortion. However, if
light scattering induces relaxation through nuclear motions, then energy will be
transferred either from the incident photon to the molecule or from the molecule to
the scattered photon. In these cases the process is inelastic and the energy of the
scattered photon is different from that of the incident photon by one vibrational unit.
It is an inherently weak process in that only one in every 106–108 photons scattered
results in Raman scattering. However, it is very sensitive to subtle structural changes
and with the development of high-power modern lasers and microscopes this tech-
nique is now much more readily available and in use.
Changes in the molecular geometry or conformation that characterize many
molecular biological phenomena can produce relatively large changes in Raman
band positions, which are usually referred to as frequency shifts. Such shifts in the
Raman band frequencies are the basis for applications of the technique to analyze
protein secondary structures, monitor tertiary structure changes, determine side-
chain conformations, and detect intramolecular interactions. Because the molecular
geometry and force field are also sensitive to interactions between molecules,
Raman can be effectively applied to probe protein intermolecular interactions,
including the formation on specific complexes and large assemblies.
A typical protein spectrum contains virtually all of the fundamental vibrational
information on a protein or nucleic acid molecule, except for hydrogen stretch
modes which generate bands in the 2,400–3,600 cm−1 Raman interval (Barron et al.
2000). Raman intensities associated with protein secondary structure are typically
produced by the vibrations of the C=O and C–N chemical groups stretching in the
56 J. Kranz et al.
Table 3.5 Amide I and amide III Raman bands of representative polypeptides and proteins with
secondary structure assignments
Frequency (cm−1) Frequency (cm−1) Secondary structure
Molecule amide I amide III assignment
α-Poly-l-alanine 1,655 1,265–1,348 α-Helix
α-Poly-l-glutamate 1,652 1,290 α-Helix
α-Poly-l-lysine 1,645 1,295–1,311 α-Helix
β-Poly-l-alanine 1,669 1,226–1,243 β-Strand
β-Poly-l-glutamate 1,672 1,236 β-Strand
β-Poly-l-lysine 1,670 1,240 β-Strand
Poly-l-lysine, pH 4 1,665 1,243–1,248 Irregular
Poly-l-glutamate, pH 11 1,656 1,249 Irregular
cI (Lamda) repressor 1,675 1,245 Turns/irregular
(1–102)
Bacteriophage P22 subunit 1,655 1,235 β-Strand/turns
peptide bonds. Stretching vibrations of C–C, C–N, and C=O bonds also produce
intense Raman bands, especially if they involve the concerted symmetrical displace-
ments of side-chain skeletons. The Raman bands associated with the bending and
stretching modes of individual hydrogens in the protein substituent (e.g., C–H,
N–H, and O–H) are generally weak, but their collective spectral intensities, which
result from the large numbers of such groups in a protein, can be significant.
Generally, application of Raman spectroscopy to protein secondary structure
analysis focuses on changes in the amide I spectral region (1,640–1,680 cm−1),
which is primarily a carbonyl stretching mode, as well as the amide III (1,230–
1,310 cm−1), which combines both in-plane N–H bending and C–N stretching
motions. Table 3.5 lists the Raman frequencies assigned to amide I and amide III
bands of representative polypeptides and proteins of differing secondary structures.
Other conformation-sensitive bands have been identified in Raman and ultraviolet
resonance Raman (UVRR) spectra of peptide model compounds and proteins. For
example, amide II (1,500 and 1,600 cm−1), which involves substantial C–N stretch-
ing, is useful in studies of protein α-helical content (Wang et al. 1991). Polarized
Raman spectra of α-helical proteins also generate a useful secondary structure
marker near 1,340–1,345 cm−1, which can be exploited for α-helix orientation anal-
yses (Tsuboi et al. 2000). Additional amide-related modes near 1,390 cm−1 have
been proposed in UVRR spectra which have been assigned to the UVRR the over-
tone of amide V and appear to be sensitive to the local conformation of the peptide
linkage (Wang et al. 1991).
Certain Raman signatures are unique to protein tertiary structure, notably vibra-
tional frequencies of side-chain conformations. For example, the Raman band
between 2,500 and 2,600 cm−1 resulting from the cysteine sulfhydryl bond (S–H)
stretching vibration is a unique probe of local SH structure and dynamics (Raso
et al. 2001). In-plane vibrations of the rings of aromatic side chains such as trypto-
phan, tyrosine, and phenylalanine are also expected to produce Raman bands of
high intensity. In addition, vibrations that involve the displacement of heavy atoms
such as sulfur in C–S stretching modes of methionine and cysteine, S–S stretching
of cysteine, S–H stretching of cysteine, and Zn–S stretching in zinc metalloproteins
are expected to be relatively intense.
The indolyl moiety of tryptophan generates many prominent Raman bands, sev-
eral of which have been correlated with the local environment and geometry of the
tryptophan side chain in proteins (Miura et al. 1991; Siamwiza et al. 1975). For
example, normal mode W3 generates an intense and sharp Raman band in the
1,540–1,560 cm−1 interval. The tryptophan residue also generates a Fermi doublet
with components at 1,360 and 1,340 cm−1, with an intensity ratio (I1360/I1340) that
increases with increasing hydrophobicity of the indolyl ring environment and thus
serves as an indicator of local hydropathy. The normal mode W17 Raman band near
880 cm−1 is sensitive to indolyl N–H hydrogen-bond donation which exhibits a rela-
tively high value when the indolyl moiety is located in a highly hydrophobic envi-
ronment, such as the hydrophobic core of a globular protein. Normal mode W18,
which is an indole ring-breathing vibration, generates bands near 755 cm−1 in Raman
spectra of proteins with increased intensity when the hydrophobicity of the indolyl
ring environment decreases.
Raman spectroscopy becomes an effective technology for various applications in
protein biotechnology due to its capability to simultaneously measure the secondary
and tertiary structure of a protein sample and to be applied to various physical
forms. In the field of protein therapeutics, Raman spectroscopy has been used to
measure protein HOSs in solution, in lyophilized form, or embedded in biocompat-
ible polymer, throughout process and formulation development. Additionally,
Raman spectroscopy can be very useful investigating incidents during stability stud-
ies and during manufacturing.
3.2.7 Other Structural Characterization Methods
X-ray Crystallography. High-resolution atomic structures of proteins have been

integral to our understanding of protein function and fundamental principles of pro-
tein structure. As much as molecular biology or genomics, X-ray crystallography
has been part of the exponential growth and success in biomedical research dating
back to our earliest understanding of macromolecular structures. The advent of
molecular replacement (non-crystallographic symmetry) has greatly simplified and
accelerated crystallographic structure determination and allowed homology model-
ing to expand structural analysis and comparisons to proteins that have not been (or
cannot be) crystallized (Rossmann 2001).
Crystallography has also had an impact in immunology, both in understanding
the various antibody–protein interactions and in how to modulate and optimize
these interactions from the theoretical to the practical (Sondermann and Oosthuizen
2002; Herr et al. 2003; Davies and Cohen 1996; Chothia et al. 1989; Al-Lazikani
et al. 1997). The rapidity with which a structure can be refined once a crystal is
58 J. Kranz et al.
available makes crystallography an important tool for understanding the HOS of

proteins. The challenges with X-ray crystallography of proteins are the following:
(1) many proteins do not crystallize readily, if at all, as is especially the case for the
majority of integral-membrane proteins and proteins with flexible regions; (2) the
crystallographic state is not always indicative of the functionally relevant state of a
protein, in a crowded solution in a cellular environment; and (3) the flexible regions
of a protein usually do not show up in the structure derived from X-ray crystallog-
raphy, even if the protein crystallizes.
Nuclear magnetic resonance (NMR) and Isotope Exchange. NMR spectroscopy is
one of the most information-rich experimental techniques, affording atomic-level
information on protein structure and dynamics. Like X-ray crystallography, NMR is
useful in determining tertiary structures of macromolecules, generally producing
ensembles of structures that reflect the solution state of a protein and glycoproteins
(Billeter et al. 1992, 2008; Wormald et al. 2002; Markley et al. 2003; Ziarek et al.
2011; Shen et al. 2008). It is often the method of choice for investigating protein
motion, as it enables measurement of different dynamic processes from very rapid
(ns) fluctuations through slow conformational transitions (μs) and global unfolding
(ms-s) (Sapienza and Lee 2010; Tamm et al. 2007; Nashine et al. 2010; Manley and
Loria 2012; Chill and Naider 2011; Baldwin and Kay 2009; Boehr et al. 2006;
Wand 2001), integral-membrane and membrane-associated proteins (Kielec et al.
2009), and in different enzymatically active or functional states (Boehr et al. 2006;
Mittermaier and Kay 2009; Eisenmesser et al. 2005; Loria et al. 2008). Historically,
NMR has been more suitable to studies of small (<30 kDa) proteins and protein
fragments, but continuing evolution of new pulse sequences and relaxation tech-
niques have led to significant advancements in studies of larger proteins and pro-
tein–ligand complexes. For biopharmaceutical development, NMR can provide
insights into protein fluctuations, as well as interactions between but also how pro-
teins and excipient and how this affects protein dynamics (Skinner and Laurence
2008, 2010).
Hydrogen–deuterium isotope exchange (HX) is an effective tool for investigat-
ing protein fluctuations that relate to unfolding and aggregation. Historically, HX
studies employed NMR to monitor the water-catalyzed exchange of peptide back-
bone amides, providing information on the hydrogen-bonding stability of each and
every residue in a protein. Structural fluctuations of local and sub-global unfolding/
folding events will expose/protect amides from exchange, with the exchange con-
stant related to the free energy for that residue. These stability values relate to the
contribution of each residue towards the global stability of the protein and provide
insights into solvent accessibility, regional flexibility, folding pathways, and how
these are modified by ligand and excipient binding events (Skinner et al. 2012;
Englander et al. 1997; Bai et al. 1995; Englander 2000; Li et al. 2008).
Other methods, including neutron diffraction, mass spectrometry, and FTIR,
have been adapted to detect HX in proteins, generally involving trapping of
exchangeable amides by pH jumps, followed by detection of peptide fragments gen-
erated by low pH proteolysis (Englander 2006; Tsutsui and Wintrode 2007; Hamuro
et al. 2005). These techniques are more rapid, since they do not require backbone
resonance assignments necessary for NMR-based HX studies, but have diminished
resolution with respect to changes in stability on a per peptide, rather than a per resi-
due basis, providing a more heterogeneous view of protein fluctuations.
3.3 T
hermodynamic and Hydrodynamic Techniques:
Shape and Size
3.3.1 Calorimetric Methods
There are two general calorimetric approaches for investigating the thermodynam-
ics of proteins and protein–ligand (or excipient) interactions: titration calorimetry
(Velazquez-Campoy et al. 2004; Ladbury and Chowdhry 1996; Jelesarov and
Bosshard 1999; Weber and Salemme 2003) and scanning calorimetry (Jelesarov
and Bosshard 1999; Weber and Salemme 2003; Sanchez-Ruiz 2011; Spink 2008;
Krishinan and Brandts 1978; Brandts and Lin 1990). All calorimetric techniques are
“label-free,” that is, they do not require the presence of a particular fluorophore or
special labeling, but rather rely on direct measurement of thermodynamic changes
that arise during an experiment.
3.3.1.1 Isothermal Titration Calorimetry
With titration calorimetry, a ligand solution is injected into a solution of protein

(or vice versa) while the change in heat, either evolved or absorbed, upon binding
is observed. These are primarily isothermal titration calorimetry (ITC) experi-
ments, which allow the determination of the binding constant (binding free energy),
the binding enthalpy, and the stoichiometry for the interaction. This technique has
its origins in understanding ligand-binding energetics and developing an under-
standing of molecular recognition. It is equally well suited for the study of excipi-
ent interactions, to differentiate an effect on protein stability via either a direct
binding mechanism or a change in the preferential hydration/dehydration effect.
Careful analysis of the heat generated upon mixing of solutions with different
compositions must be taken into consideration in order to avoid artifacts during
ITC experiments.
3.3.1.2 Differential Scanning Calorimetry
One of the most routine methods for candidate and formulation screening during
biopharmaceutical development is differential scanning calorimetry (DSC). In DSC
a reference cell (H2O or buffer) and a sample cell (containing the protein solution
60 J. Kranz et al.
being analyzed) are heated at a constant rate of change in temperature, and the
difference in the energy provided to the two cells, or heat capacity, in order to
achieve this rate is recorded. The output of DSC experiments is an evolution of heat
(enthalpy) corresponding to the net thermodynamic loss of tertiary structure and
rehydration of unfolded protein and a midpoint temperature, Tm, that represents the
point at which the sample is equally populated by folded and unfolded forms. The
thermodynamics of a reversible reaction can be obtained from these experiments,
while for irreversible reactions, relative differences in conformational stability or
stability to aggregation can be obtained from differences in Tm or in the temperature
at which unfolding begins. For multidomain proteins such as antibodies, it is often
possible to resolve the transition from each domain, providing information about
individual parts of the protein that is difficult to obtain by other techniques. The
application of DSC is somewhat limited due to the fact that it is a low-throughput
method that requires about 1 mg of protein.
DSC has applications in measuring ligand binding as well (Brandts and Lin
1990). Usually binding of a ligand increases the stability of the protein, increas-
ing the protein Tm in a manner that depends both on ligand affinity and on con-
centration. As described in Chap. 2, a higher-throughput approach that involves
heating a solution in the presence of an extrinsic fluorescent probe, which fluo-
resces only in the presence of unfolded protein, has been developed and is
broadly applied when material and time are limited. This technique is variously
referred to as differential scanning fluorimetry (Goldberg et al. 2010; He et al.
2009a, b; Li et al. 2011), ThermoFluor™ (Mezzasalma et al. 2007; Matulis et al.
2005), or a thermal shift assay (Kranz and Schalk-Hihi 2011; Cimmperman et al.
2008; Zubriene et al. 2009; Norvaisas et al. 2012; Zhang and Monsma 2010).
Each of these has been adapted to plate-based unfolding studies, which reduce
sample requirements by several orders of magnitude and allow for parallel
unfolding or ligand-binding studies. Thermal shift assays are quite useful in
excipient screens, allowing for rapid identification of stabilizing conditions for a
given protein. DSC can then be used as a sort of gold standard, providing support
for the results obtained with the higher-throughput methods and increasing
understanding of the underlying reactions.
Light scattering studies have been utilized extensively when structural informa-
tion is needed for protein molecules dissolved in solution (Van Holde et al. 2006).
For this purpose, both static light scattering (SLS) and dynamic light scattering
(DLS) studies (most commonly using light in the visible range) have been
employed; the former is primarily used for molecular weight and dimensional
analysis, and the latter is most useful for effective hydrodynamic size measure-
ments. Both these techniques and some key areas of application in protein devel-
opment are presented below.
3.3.2.1 Static Light Scattering
SLS studies refer to the analysis of the intensity of the light scattered from a
molecule in solution. In simple terms, scattering can be defined as deflection of an
incident beam of light by a molecule. However, a number of phenomena take place
from the instant that light impacts the molecule to the moment it is dispersed. When
focused on a molecule, linearly polarized light, an electromagnetic wave moving in
a particular direction (e.g., x-axis) with its electric vector oscillating perpendicular
to this direction (e.g., z-axis) induces corresponding oscillation in the electrons of
the molecule. Such oscillations of the electron in the molecule result in generation
of oscillating dipoles and subsequent emission of some of the incident energy in the
form of light in a direction different from that of the incident light. When the size of
the scattering molecule is significantly smaller than the wavelength of the incident
light beam (about 50-fold), a phenomenon described as Rayleigh scattering occurs,
the ratio of the scattered light intensity to that of the incident light can be expressed
as either of the following two expressions depending on the angle of observation
with respect to either the z- or x-axis:
i 16p 4a 2 sin 2 j
= , (3.7)
I0 r2l 4

i 8p 4a 2
= 2 4 (1 + cos2 q ), (3.8)
I0 r l

where i is the intensity of the scattered light, i0 is intensity of the incident light, a is the
molecular polarizability, π is the angle between the direction of scattered light and the
direction of the oscillation of the electric vector (z-axis), r is the distance between the
point of observation and the position of the scattering particle, l is the wavelength of the
incident light, and q is the angle between the direction of the scattered light and that of
the incident beam (x-axis). It is apparent from this expression that (a) there will be no
scattering along the direction of the electric vector oscillation of the incident light, that
is, when π = 0, (b) the scattering intensity decreases with increasing distance from the
molecule and more significantly with increasing wavelength and (c) the scattering inten-
sity will be symmetrical in the forward and backward direction, that is, there is no angu-
lar dependence of the scattering light intensity. Equation 3.8, expressing the scattering
from a single molecule, can be expanded further to represent the scattering from a col-
lection of such molecules in a dilute solution to
2
 dn 
2p 2 n02  
iq  dc 
= cM (1 + cos2 q ), (3.9)
I0 r l N
2 4

where iq is the cumulative scattered intensity from the collection of molecules
observed at an angle q from the direction of the incident light, n0 is the refractive
62 J. Kranz et al.
index of the solvent, dn/dc is the refractive index increment of the solution per unit
concentration of the added solute, N is Avogadro’s number, c is the molecule or
solute concentration, and M is the molecular weight of the solute. Equation 3.9 can
be further simplified to the more commonly used expression:
iq r2
Rq = = KcM , (3.10)
I 0 (1 + cos2 q )

where Rq is defined as the Rayleigh ratio. In Eq. 3.10, K is an instrumental con-
stant, since it is fixed for a given instrument and particular solute–solvent system,
defined as
2
 dn 
2p n  
2 2
 dc 
0
K= . (3.11)
l N
4
For a dilute solution consisting of an ensemble of associated states of the solute

of interest (for instance aggregated states of different sizes of the protein monomer),
Eq. 3.10 is represented as
Rq = KcM w , (3.12)
with Mw being the weight-averaged molecular weight of the solute since each asso-
ciated state contributes to the average molecular weight based on its total weight
fraction instead of its number fraction. Equation 3.7 is commonly employed in pro-
tein development studies to determine the molecular weight of protein in solutions
as a function of any given solution variable or over the shelf life of the product. It is
assumed here that the solution being studied is sufficiently dilute so that the inter-
molecular interaction between multiple protein molecules does not affect the overall
scattering behavior. In other words, the expression applies to ideal solutions.
However, this assumption is often invalid, and intermolecular interactions signifi-
cantly impact the scattering behavior of the molecules and solution.
In protein development, the most common application of static light scattering
has been its use for the determination of Mw of the protein in relatively dilute solu-
tions (Wyatt 1993). Since the formation of aggregates and other higher-order spe-
cies is considered a potential patient safety risk, early detection of propensity to
aggregate is critical to successful product development (Goldberg et al. 2011).
In addition to molecular weight measurements, static light scattering is also
employed to obtain dimensional information. For this to be possible, the size of the
molecule of interest has to be large enough to be treated as a point scatterer. In this
case, the scattering behavior for even a single molecule varies depending on the
location of the molecule where the incident light strikes. The scattered light thus
varies with the angle of measurement. With multiple light waves now capable of
striking the molecule simultaneously at different points, interference between the
light scattered from these multiple scattering locations needs to be accounted for.
The key attribute derived from the scattering data from larger molecules is the radius
of gyration (RG) defined as the root mean square average of the distance of the scat-
tering element from the center of mass of the molecule. For Rayleigh scattering to
be valid, l /RG should be in excess of about 50. Thus, for instruments equipped with
a 600 nm or higher laser, molecules with a radius of gyration of less than 12 nm can
be treated as point scatterers and will thus not exhibit any angular dependence of the
scattering intensity. Obviously in such cases, it will not be possible to determine the
radius of gyration of the particle and the primary application of light scattering will
be for the determination of molecular weight and the association state (e.g., mono-
mer, dimer). For globular proteins, the average hydrodynamic diameter (discussed
in the next section) has been found to be ~3 times the radius of gyration, and thus
when using a 600 nm laser, Rayleigh scattering will be valid for molecules with
radii as high as 36 nm (Mandel 1993). This is usually the case for most proteins in
development for therapeutic application. For example, for immunoglobulins, one of
the larger proteins frequently encountered in development with a mass of ~150 kD,
the hydrodynamic diameter is ~12 nm. In cases where information regarding the
dimensions of the molecular radius of gyration is required, using instruments with
incident light at less than 600 nm (ultraviolet light or X-ray) can help. But even this
approach has limitations since below 300 nm, protein molecules start absorbing,
and below about 160 nm, water itself starts absorbing. X-rays with wavelengths
below 10 nm, to which most materials are transparent, have been used at small
angles to overcome these limitations for analyzing dimensions of small and globular
protein molecules.
Static light scattering has been an excellent tool for characterizing the molecular
weight and the higher-order species in protein solutions. However, it has found lim-
ited application for effective size measurement in solution given the lack of angular
dependence for most of the molecules currently being developed. In instances where
hydrodynamic properties need to be measured, biophysical chemists have often
relied on the technique of DLS which is the focus of the next section.
3.3.2.2 Dynamic Light Scattering
Unlike static light scattering experiments, which capture the time-averaged scat-
tered light intensity of a solution sample, DLS is concerned with the fluctuations in
the intensity of the scattered light over relatively small time periods for a small
volume fraction of the total solution. On a molecular level, whereas static measure-
ments track the total scattered intensity from a collection of molecules, dynamic
studies track the fluctuations in the intensity of light scattered from a collection of
molecules as a function of time. This concept can be expanded to analysis of a small
volume element (or fraction) of the total solution wherein a large number of such
molecules are in constant motion. This motion is primarily due to Brownian motion,
but can also be due to intermolecular interactions. As a result of this constant motion
and the constantly changing location and orientation of these molecules in the finite
64 J. Kranz et al.
volume element, the light scattered from that volume element exhibits fluctuations
in intensity. This fluctuation in scattered intensity is used to calculate the average
size of the molecule in solution. Since water (or solvent) molecules closely associ-
ate with the molecule of interest and diffuse with it as a single entity, DLS measures
the size of the hydrated (or solvated) molecule.
A solution of larger molecules will undergo intensity fluctuations at a slower
rate, due to slower movements in and out of the portion of the volume being moni-
tored, compared to a solution of smaller-sized molecules. A DLS instrument tracks
these intensity fluctuations over nano- and microseconds and in one (and the most
commonly employed) variations of the technique correlates the intensity at time, t
with that at time t + Δt, t + 2Δt, and so on through an autocorrelation function repre-
sented as
〈i(t )•i(t + ∆t )〉
g ( 2 ) ( ∆t ) = , (3.13)
〈i 〉 2
where g(2) is the second-order autocorrelation function, i(t) is the scattered intensity
at time, t i(t + Δt) is the scattered intensity following an interval of Δt, and the
brackets (〈〉) indicate the averaged product of scattered intensity over time. The
symbol 〈i〉 represents the average values of scattered intensity. At short time inter-
vals (lower multiples of Δt), the correlation between scattered intensity values will
be high since the molecules would not have moved significantly from their initial
position and there is consistency in scattering states. Thus, higher and more stable
values of the autocorrelation function are obtained at shorter time intervals. As the
time delays become longer (at higher multiples of Δt), correlation between scattered
intensities decreases, since there is decreasingly little correlation between scattering
states, and more movement of the molecules in and out of the volume being moni-
tored. When a single component of the solution dominates the scattering from the
solution (monodisperse solutions), the autocorrelation function decays exponen-
tially such that
g ( 2 ) ( ∆t ) = 1 + a •exp( −2q 2 D∆t ), (3.14)
{ }
l n g ( 2 ) ( ∆t ) − 1 = ln a − 2q 2 D∆t , (3.15)
4pn q
q= sin   , (3.16)
l  2
where q is the scattering vector, n represents the solution refractive index, q is the
scattering angle, l is the wavelength of incident light, a is an instrumental constant,
and D is the translational diffusion coefficient (referred to simply as the diffusion
coefficient) since it characterizes translational movement of a molecule from one
point to another rather than rotational movements about a given axis. Equation 3.15
can be fitted to the system-generated autocorrelation function data over time to cal-
culate the diffusion coefficient of a given component of the solution. It must be
noted that, similar to Rayleigh scattering, Eq. 3.15 can be used for the determination
of the diffusion coefficient only as long as the dimensions of the scattering mole-
cules are small compared to the wavelength of light. If this condition is not satisfied
and the molecular size approaches that of the wavelength of incident light, rota-
tional as well as internal movements of the molecules also contribute to the change
in the scattering pattern and intensity with time. In this case, the calculated autocor-
relation function does not accurately reflect the translational diffusion coefficient of
the molecule.
From the above discussion, it is apparent that DLS directly measures the diffu-
sion coefficient, not the size, of a given solute in solution. The effective or the
hydrodynamic size is subsequently derived through the Stokes–Einstein relation
assuming a spherical geometry for the molecule:
kT kT
Dm = = . (3.17)
f 3phdH

In Eq. 3.17, f is the frictional coefficient or a measure of drag on the diffusing
molecule, k is the Boltzmann constant, T is the absolute temperature, and η is the
solution viscosity. Therefore, the diffusion coefficient, sometimes called the self-
diffusion coefficient (Ds), of the solute, is inversely related to the hydrodynamic
diameter (dH) or the effective size of the hydrated solute molecule (Eberstein et al.
1994; Schümmer 1978). It is important to note that the Stokes–Einstein equation is
valid for spheres, and thus DLS measurements provide only an apparent hydrody-
namic size of the molecule assuming the particle under examination to diffuse in a
manner similar to a sphere of the same size. Consequently, globular proteins in DLS
measurements exhibit a smaller measured diameter (or larger Ds) compared to ran-
dom coils of the same molecular weight since random coils, due to more complex
and interfered motions and interactions with the solvent, diffuse more slowly in the
solution. Thus, DLS measurements incorporate the effect of both hydration and
shape of the molecule in measuring the size of the molecule. Any change in the
shape or hydration of the molecule that results in a change in the diffusion behavior
of molecules can be detected by DLS studies. Conformational changes in the pro-
tein molecules in solution, occurring as a function of solution environment, can thus
be readily detected via this technique.
The use of DLS is usually limited for particles in the submicron size (<1–2 μm).
Most common instruments have a practical upper limit of ~1 μm. Fortunately, most
protein monomers are significantly smaller, in low nanometer range, and thus can
be studied via DLS. Additionally, the formation of higher molecular weight or
aggregated species can also be detected since a considerable population of aggre-
gated species can exist at sizes below 1 μm.
The form of the Stokes–Einstein relation expressed in Eq. 3.17 assumes an ideal
solution behavior where intermolecular interactions, or in the case of protein solu-
tions protein–protein interactions, do not affect the movement of protein molecules
and thus do not impact the measured diffusion coefficient. However, this may not be
true in all cases. For ideal solutions, the diffusion coefficient depends on the
66 J. Kranz et al.
h ydrodynamic size of the protein molecule (in addition to viscosity and t emperature).
In other words, for systems in which protein–protein interactions are too weak to
influence the diffusion of the protein molecule, the diffusion coefficient is indepen-
dent of the protein concentration, and the measured dH is an absolute measure of the
effective hydrodynamic diameter of the protein. However, for solutions in which
protein–protein interactions affect translational movement of the protein molecules,
the diffusion coefficient is a function of the protein concentration as well. Thus,
DLS measurements should be carried out over a series of protein concentrations
instead of a single concentration in order to tease out the effect of molecular interac-
tions on molecular size measurements.
As discussed in several other chapters in this volume, DLS has become a potent
tool for use during candidate and formulation screening, where the ability to obtain
high throughput, relative qualitative results is perfectly suited for the requirements
of these stages of therapeutic development.
3.3.3 Analytical Ultracentrifugation
Analytical ultracentrifugation (AUC) is one of the oldest biophysical techniques

used for characterizing the association and conformation of proteins and macromol-
ecules. The basic principles of AUC were pioneered during the 1930s by Svedberg,
with few changes or innovations occurring in the technique over the next several
decades. With the growth of the biotechnology industry, there has been a concurrent
increase in the need for a technique to monitor protein self-association across a
large size range. There are two types of AUC analysis currently in use, sedimenta-
tion velocity analytical ultracentrifugation (SV-AUC) and sedimentation equilib-
rium analytical ultracentrifugation (SE-AUC).The development of SV-AUC with
software analysis using the c(s) distribution model has resulted in the use of AUC in
therapeutic protein development for HOS characterization and as an orthogonal
method to size exclusion chromatography for protein aggregation analysis.
3.3.3.1 Sedimentation Velocity
During an SV-AUC experiment, there are three forces acting on a single protein
molecule: (1) the gravitational force mω2r where m is the protein mass, ω2 is the
rotor angular velocity, and r is the distance from the center of rotation; (2) the buoy-
ancy force –mν(bar)ρω2 where ν (bar) is the protein partial specific volume and ρ is
the solvent density; and (3) the frictional force s(kT/D)ω2 where k is the Boltzmann
constant, T is the absolute temperature, and D is the diffusion constant. The sedi-
mentation coefficient s, measured in Svedberg units, is defined as s = v/ω2 where v is
the absolute velocity. The fundamental equation in AUC, the Svedberg equation, is
derived from the balance of the three forces acting on a single molecule:
s M (1 − ns )
= , (3.18)
D RT
where M is the molar mass, and R is the gas constant. The sedimentation and
diffusion of a protein in solution are measured in SV-AUC experiments as a series
of sedimentation profiles and can be described by the Lamm equation:
dc  1  d   dc  
=  rD   − sw 2 r 2 c  , (3.19)
dt  r  dr   dr  

where χ is the concentration of the protein at a distance r from the center of rotation
at time t for a given angular velocity ω. Protein samples are usually not analyzed
under standard conditions, that is, aqueous solutions at 20°C; therefore, for the pur-
pose of comparing various samples, the s values obtained under different nonstan-
dard conditions can be converted to the standard conditions using the equation:
 h   h   1 − vr20,w 
s20,w = s  T,w   s   , (3.20)
 h20,w   hw   1 − vrT,w 
where η20,w is the viscosity of water at 20°C, ηT,w is the viscosity of water at the
experimental temperature T, ηw is the viscosity of water at the experimental tem-
perature and ηs is the viscosity of the solvent at the same experimental temperature,
ρ20,w is the density of water, and ρT,s is the is the density of the solvent at the experi-
mental temperature. The theory behind protein sedimentation and AUC has been
described in detail in multiple books (van Holde et al. 2006; Machtle and Borger
2006; Scott and Schuck 2005; Philo 2005). Advanced AUC techniques and strate-
gies for experimentation and data analysis were described in several critical reviews
and tutorial articles (Philo 2006; Liu et al. 2006; Schuck et al. 2002).
In an SV-AUC experiment a protein sample is centrifuged, usually over several
hours, while the progress of centrifugation is monitored by a detection system, most
commonly UV absorbance. A typical result of SV-AUC analysis is presented in
Fig. 3.11. The primary information obtained from an AUC experiment is the sedi-
mentation coefficient. This can be converted (with certain presumptions about
shape) into other information better understood by general protein chemists work-
ing in formulation development, that is, molecular weight, MW distribution
(Fig. 3.11b), or hydrodynamic radius.
There are multiple approaches and also choices that are made during the perfor-
mance and subsequent data analysis of an AUC experiment that are important for
the interpretation of the data. The sample needs to be diluted to be in the linear
range of the detector (below optical density of 1.5 for UV absorbance) in buffer. As
many protein therapeutics are formulated at high concentration above 100 mg/mL,
the dilution can be significant. The first crucial decision is what buffer to use for
dilution; two approaches currently exist, to either use buffer without excipients that
68 J. Kranz et al.
a 5 b
5
4 4
3 3
c(s)
c(M)
2 2
1 1
magnified 20x magnified 20x
0 0
5 10 15 20 0 200 400 600 800 1000
Sedimentation coefficient (Svedberg) Molecular weight (kDa)
Fig. 3.11 SV-AUC analysis of an IgG analyzed in formulation buffer diluted to 0.5 mg/mL. Panel
a shows size distribution obtained using continuous c(s) distribution model, and panel b presents
c(M) distribution model
might interfere with the analysis or to use the original formulation buffer. Many
buffers used in protein formulation contain substantial concentrations of excipients,
mostly sugars, to enhance long-term stability. Dynamic gradients, caused by the
presence of the excipients, can have detrimental effects on the reliability of the data
obtained.
The sedimentation coefficient is the crucial parameter for HOS determination
and for detection of changes in HOS and/or changes in protein conformation and
shape. The value obtained from AUC experiments is in general dependent on pro-
tein concentration, as demonstrated in Fig. 3.12, and also solution conditions such
as density and viscosity. In order to eliminate the effect of protein concentration on
the sedimentation coefficient, it is necessary to extrapolate to zero protein concen-
tration. This is labor intensive because it requires at least several data points and
considerable experimental time. A more generally accepted approach is to compare
s values measured at the same protein concentration, for example, 0.5 mg/mL.
Similarly, the effect of density and viscosity on the sedimentation coefficients can
be compensated for by extrapolating the sedimentation coefficient s to standard
conditions at zero excipient concentration and 20°C (s20,w). The sedimentation coef-
ficient of an antibody monomer can be measured by AUC with great precision and
therefore changes in sedimentation coefficient values can be used to detect small
changes in protein conformation.
There are some situations where measurement of changes in s values has valu-
able application in formulation development: The conformational changes detected
by AUC can be also detected by other techniques, such as the spectrophotometric
techniques and ITC, described earlier in this chapter. However in some situations,
especially when the conformational changes are small, they may not be detected by
Fig. 3.12 The concentration

dependence of the
sedimentation coefficient of
an IgG antibody. IgG was
dialyzed and diluted into very
low salt containing buffer at
pH 5.2 and in the same buffer
with addition of 150 mM of
sodium chloride. Data were
fit to the equation
s = s0(1 + ksc)−1 and values of
ks = 37.8 cm3/g for low salt
buffer, and ks = 6.48 cm3/g for
150 mM sodium chloride
were obtained
spectrophotometric techniques. The most common uses of sedimentation velocity

AUC in protein therapeutic development are:
1. To follow subtle changes in conformation due to improper folding under pH or
thermal stress, by monitoring subtle changes in the sedimentation coefficient
2. Demonstration of comparability in the conformation of therapeutic proteins
from different lots or processes and as an orthogonal method to verify the results
of size exclusion chromatography
For both of these applications the molar mass must remain constant; changes in
MW, for example, due to protein deglycosylation or clipping, will result in a change
of sedimentation coefficient as well.
The most common application of SV-AUC in HOS determination during drug
development is demonstrating comparability of conformation and especially self-
association for proteins from different lots, or after changes to the manufacturing
and purification processes, or the same protein produced at different locations.
These comparability studies are a routine part of filing documents. Figure 3.13 dem-
onstrates comparability of an IgG protein formulated in the same formulation buffer
but produced at two different locations (same sequence); for the purpose of this
comparability, the processes are labeled A and B.
These two processes showed very good comparability in both conformation, ana-
lyzed as sedimentation coefficient, and the number and quantity of high molecular
species (not presented here). The sedimentation coefficient distributions presented in
Fig. 3.13 are summarized in Table 3.6. As can be seen in the Table 3.6, variations of
sedimentation coefficients for both processes and also overall are below 0.5 %, which
70 J. Kranz et al.
5 5
Process A Process B
4 4
3 3
c(s)
c(s)
2 2
Lot 3 Lot 3
1 1
Lot 2 Lot 2
Lot 1 Lot 1
0 0
2 4 6 8 10 2 4 6 8 10
Sedimentation coefficient (Svedberg) Sedimentation coefficient (Svedberg)
Fig. 3.13 Comparability of size distributions of therapeutic monoclonal antibody samples

manufactured by two different processes. Three lots of each process, each lot in triplicate, were
analyzed in formulation buffer at identical conditions including protein concentration 0.5 mg/mL
Table 3.6 Comparability of sedimentation coefficients of monoclonal antibody samples

manufactured by two different processes. Sedimentation coefficients were obtained as weighted
average of sedimentation coefficients of monomers in the range 4S–6S
Process A Process B
Lot Sedimentation coefficient Sedimentation coefficient
(#) (average ± s.d.) Lot (#) (average ± s.d.)
1 5.000 ± 0.015 1 4.998 ± 0.006
2 5.020 ± 0.024 2 5.021 ± 0.016
3 5.018 ± 0.023 3 5.013 ± 0.009
Process 1 average ± s.d. Process 2 average ± s.d.
5.013 ± 0.021 5.011 ± 0.014
Overall average ± s.d.
5.012 ± 0.018
is remarkable considering that these were produced at two different locations at dif-
ferent points of time and analyzed by AUC in the original formulation buffer contain-
ing a considerable amount of excipients.
3.3.3.2 Sedimentation Equilibrium
The SV-AUC method described in the previous section is currently extensively used
in protein development for aggregate characterization and for characterization and
for studies on the comparability of conformation. SE-AUC is an equilibrium method
and therefore does not provide any information on the molecular shape, unfolding,
Fig. 3.14 Sedimentation
1.5
equilibrium experiment of an
IgG antibody. The sample
was analyzed at three
Absorbance (280 nm)

different concentrations 0.1,
0.2, and 0.4 mg/mL at speeds 1.0
6,000, 9,000, and 12,000 rpm.
The data were analyzed using
single ideal species model,
and molecular mass within
±3 kDa of the theoretical 0.5
value was obtained
0.0
6.9 7.0 7.1

r (cm)
conformation, or changes in protein structure. SE-AUC is a valuable complemen-

tary method used to determine the molecular weight of proteins and their oligo-
meric state.
When equilibrium is reached in an SE-AUC experiment, the equilibrium of a
single ideal species is
2
/ 2 RT )( r 2 − r02 )
c(r ) = c(r0 )e ( M (1−nr )w , (3.21)

where c(r) is the protein concentration at the radial position r, r0 is the reference
radial position, M is molecular weight, ν (bar) is the protein partial specific volume,
ρ is the solvent density, ω2 is the rotor angular velocity, R is the gas constant, and T
is the absolute temperature. If more species are present, dimer or oligomers, the
right side of the equation becomes the sum of all species. For proteins that revers-
ibly self-associate, the actual equilibrium association constant can also be obtained
from SE-AUC experiments. The strength of this method is that it can detect all types
of self-association including those that are very weak and not readily detected by
other methods. For situations where non-ideality is present, the molar mass in the
equation can be replaced by the apparent molecular mass, dependent on concentra-
tion, to compensate for non-ideality.
An SE-AUC experiment is done at much lower rotor speeds compared to sedi-
mentation velocity experiments and a series of protein samples at various concen-
trations is spun at various rotor speeds (Fig. 3.14). It can take days or weeks, in
some situations, for a protein solution to reach sedimentation and diffusion equilib-
rium during an SE-AUC experiment. In comparison, SV-AUC experiments take
only a few hours to complete.
72 J. Kranz et al.
3.3.4 Rheology
Rheology is the study of flow behavior of matter either in the liquid or the solid
state. However, for most practical purposes, it applies to liquids and semisolids and
other matter which undergo some degree of deformation or flow (viscous) and loss
of some or all of the applied energy as opposed to a material in which all of the
applied energy can be recovered (elastic, e.g., a perfect spring) (Schümmer 1978).
Viscosity (η) of a fluid is its resistance to flow, that is, its “internal friction” and is
the simplest expression of the rheological properties of a material. It can be defined
as the pressure [shear stress, τ] required to deform or to make a fluid flow at a unit
rate [shear rate, γ] in a direction perpendicular to the plane of applied pressure:
t
h= . (3.22)
g
Most of the fluids that we encounter in daily life exhibit viscosity which is a
function of the shear rate, that is, the viscosity will change as one varies the rate at
which the fluid is deformed. Such liquids are referred to as non-Newtonian. On the
other hand, for Newtonian fluids like water viscosity is not a function of the shear
rate but rather remains constant over a broad and practical range of shear rates
(Macosko 1994; Shaw 1992).
Rheological behavior of protein dispersions is intimately linked to the charac-
teristics of the protein itself as well as environmental factors. The characteristic
properties of the protein that are important in this aspect are its conformation,
shape, size, solubility, swellability, and charge which are influenced by environ-
mental conditions like temperature, concentration, pH, and ionic strength
(Hermansson 1972, 1979).
The viscosity of protein solutions is largely dependent on the (1) hydrodynamic
behavior of the protein molecules in solution and (2) protein–protein interactions
between adjacent molecules and their spatial distribution. Consequently, the study
of the behavior of protein solutions at infinite dilution of the protein when protein–
protein interactions are negligible can provide meaningful insight into the hydrody-
namic properties of the protein molecule which in turn can be related to the shape
and size or conformational changes in the protein. The problem of relating viscosity
of the colloidal dispersions with the fundamental nature of the dispersed particles
has been the subject of much experimental investigations and theoretical consider-
ations (Harding 1980, 1981; Lundqvist 1999). Such an effect is defined in terms of
viscosity functions such as relative (ηrel), specific (ηsp), and reduced or reduced spe-
cific viscosity (ηred):
h
hrel = , (3.23)
h0

h − h0
hsp = hrel − 1 = , (3.24)
h0

hred = hsp / c = (hrel − 1) / c, (3.25)

where η is the viscosity of the solution of concentration c and η0 is the viscosity of
the solvent. Intrinsic viscosity [η] is an intrinsic function of the dissolved/dispersed
macromolecule and is defined as the limiting value, at zero concentration, of relative
viscosity increment.
Two factors that contribute to this characteristic property of the dispersed particle
are particle shape and size/volume as summarized by the following relationship:
[h] = v •Vs , (3.26)
where, ν is the molecular shape factor known as the viscosity increment or universal
shape function (Kauzmann 1959; Yang 1961) and Vs is known as the swollen spe-
cific volume and is a measure of the solvent associated with the macromolecule. It
is defined as the volume of the macromolecule in solution per unit of anhydrous
mass of the macromolecule. Intrinsic viscosity is thus indirectly dependent on the
axial ratio of the molecule that governs its shape. It increases with increased axial
ratio and is therefore minimum for a sphere (Tanford and Buzzell 1956; Harding
1997). As a result of this dependence, intrinsic viscosity has been used as a com-
bined measure of the hydrodynamic volume and molecular shape.
Changes in intrinsic viscosity or in reduced viscosity at low concentration have
also been used as a measure of the extent of denaturation, which also reflects the
relationship with changing axial ratio (Suryaprakash and Prakash 2000; Ahmad and
Salahuddin 1976). Interestingly, for globular proteins unlike random coils, intrinsic
viscosity has been found to be independent of the molecular weight and depends
solely on the shape and size of the protein. In an excellent review, Harding (Harding
1997) has listed the values of intrinsic viscosity of numerous proteins and peptides.
A relatively simple method has been proposed in the literature for calculating the
axial ratio of proteins based on viscosity measurements and data fitting (Monkos
and Turczynski 1991). Rheological measurements have also been used to determine
the unfolding onset for protein in solution in the presence of an increasing concen-
tration of a denaturant like urea (Tu and Breedveld 2005).
A general application of intrinsic viscosity measurements has been to assess the
denaturation or conformational change of a protein molecule in solution. For a poly-
peptide poly-(his-ala-glu) studied in the pH range of 2.97–9.70, the intrinsic viscos-
ity varied from 8 mL/g at pH 2.97 to 33 mL/g at pH 4.99 to 55 mL/g at pH 9.70
indicating an increasing unfolding or conformational change with increase in solu-
tion pH. This is consistent with a general increase in the net negative charge on the
molecule due to the ionization of the glutamic acid residues and the deprotonation
of histidine residues with increasing solution pH. From the example it is apparent
that solution pH and ionic strength have a marked effect on the hydrodynamic
74 J. Kranz et al.
p roperties of the protein and therefore on the flow behavior of the system on the
whole. This effect is mediated through the effect on charge associated with the pro-
tein molecule, its folding and unfolding kinetics, and the solubility of the protein
that changes with the pH. In addition to the absolute shape and size effects, electro-
static charge can significantly influence the intrinsic viscosity of a protein molecule
in solution, especially if the molecule carries multiple charges in solution
(polyelectrolyte) like proteins. In theory three such contributions have been sum-
marized for a compact globular protein molecule: a “primary effect” due to the dif-
fuse double layer surrounding the protein molecule, a “secondary effect” due to
repulsion between the double layers of different molecules, and a “tertiary effect”
arising from the intermolecular repulsions that affect the shape of the molecule.
These three have been collectively referred to as “electroviscous” effects. Under
conditions of high net charge on the protein molecule, the electroviscous effect
contributes significantly to solution flow behavior. However, under conditions of
low molecular charge and/or high charge screening, as is the case with high ionic
strength solutions, electroviscous effects are not very prominent.
During the processing and delivery of protein therapeutics, there are multiple
steps that can be limited by viscosity, including filtration, chromatography, and
delivery through syringes. Thus, viscosity measurements can be important during
formulation and candidate screening and during process development.
3.4 Conclusions
A very important characteristic of proteins is the unique, folded structure that is

important for function and stability. For a protein, deviations from the preferred
structure or ensemble of structures are just as critical to establishing functionality as
is the primary amino acid sequence. Understanding these facts and the ability to
characterize protein structures provide key tools to develop various protein thera-
peutics and other biotechnological applications. We have provided an overview of
the most common technologies that monitor proteins higher-ordered structural qual-
ity attributes, specifically techniques aimed at quantifying changes in secondary
structure, tertiary structures, and quaternary structures. The chapter further provided
an overview of technologies to characterize protein thermal characteristics such as
calorimetric techniques to monitor protein thermodynamic properties, with more
attention paid to those approaches that require significant biophysical expertise.
Continuous efforts to develop new technologies or improve existing technologies
are needed to better understand protein higher-ordered structural distributions in
protein biotherapies. In Chaps. 6–10 the applications of these techniques during all
stages of protein therapeutic development will be illustrated.
Acknowledgements The authors would like to thank Dr. Lee Olszewski, Dr. Doug Nesta,
Dr. Aston Liu, and Dr. Joseph Rinella at GlaxoSmithKline Biopharm Development for supporting
this publication. We thank Dr. Rina K. Dukor from BioTools Inc. and Prof. Laurence A. Nafie from
Syracuse University for their insights and helpful discussion.
References
Adams PD, Chen Y, Ma K, Zagorski MG, Sonnichsen FD, McLaughlin ML, Barkley MD (2002)
Intramolecular quenching of tryptophan fluorescence by the peptide bond in cyclic hexapep-
tides. J Am Chem Soc 124:9278–9286
Ahmad F, Salahuddin A (1976) Reversible unfolding of the major fraction of ovalbumin by guani-
dine hydrochloride. Biochemistry 15(23):5168–5175
Al-Lazikani B, Lesk AM, Chothia C (1997) Standard conformations for the canonical structures
of immunoglobulins. J Mol Biol 273:927–948
Asher SA, Bormett RW, Chen XG, Lemmon DH, Cho N (1993) UV resonance Raman spectros-
copy using a new cw laser source: convenience and experimental simplicity. Appl Spectrosc
47:628–633
Bai Y, Sosnick TR, Mayne L, Englander SW (1995) Protein folding intermediates: native-state
hydrogen exchange. Science 269:192–197
Baldwin AJ, Kay LE (2009) NMR spectroscopy brings invisible protein states into focus. Nat
Chem Biol 5:808–814
Bandekar J (1992) Amide modes and protein conformation. Biochim Biophys Acta 1120:123–143
Barron LD (2006) Structure and behaviour of biomolecules from Raman optical activity. Curr
Opin Struct Biol 16:638–643
Barron LD, Hecht L, Bell AF, Wilson G (1996) Recent developments in Raman optical activity of
biopolymers. Appl Spectrosc 50:619–629
Barron LD, Hecht L, Blanch EW, Bell AF (2000) Solution structure and dynamics of biomolecules
from Raman optical activity. Prog Biophys Mol Biol 73:1–49
Bartlett AI, Radford SE (2009) An expanding arsenal of experimental methods yields an explosion
of insights into protein folding mechanisms. Nat Struct Mol Biol 16:582–588
Baumruk V, Keiderling TA (1993) Vibrational circular dichroism of proteins in H2O solution. J Am
Chem Soc 115:6939–6942
Benevides JM, Overman SA, Thomas GJ Jr (2004) Raman spectroscopy of proteins. Curr Protoc
Protein Sci, Chapter 17, Unit 17.18
Billeter M, Vendrell J, Wider G, Aviles FX, Coll M, Guasch A, Huber R, Wuthrich K (1992)
Comparison of the NMR solution structure with the X-ray crystal structure of the activation
domain from procarboxypeptidase B. J Biomol NMR 2:1–10
Billeter M, Wagner G, Wuthrich K (2008) Solution NMR structure determination of proteins revis-
ited. J Biomol NMR 42:155–158
Boehr DD, Dyson HJ, Wright PE (2006) An NMR perspective on enzyme dynamics. Chem Rev
106:3055–3079
Brandts JF, Lin LN (1990) Study of strong to ultratight protein interactions using differential scan-
ning calorimetry. Biochemistry 29:6927–6940
Brennan JF, Wang Y, Ramachandra RD, Feld MS (1997) Near-infrared Raman spectrometer sys-
tems for human studies. Appl Spectrosc 51:201–208
Byler DM, Susi H (1986) Examination of the secondary structure of proteins by deconvolved FTIR
spectra. Biopolymers 25:469–487
Cantor RC, Schimmel PR (1980) Biophysical Chemistry: techniques for the study of biological
structure and function, vol 2. W. H. Freeman, San Francisco
Cao XL, Dukor RK, Nafie LA (2008) Reduction of linear birefringence in vibrational circular
dichroism measurement: use of a rotating half-wave plate. Theor Chem Acc 119:69–79
Carpenter JF, Prestrelski SJ, Dong A (1998) Application of infrared spectroscopy to development
of stable lyophilized protein formulations. Eur J Pharm Biopharm 45:231–238
Chalmers J, Griffiths P (eds) (2002) Handbook of vibrational spectroscopy. Wiley, Chichester
Chance B (1953) The carbon monoxide compounds of the cytochrome oxidases. I. Difference
spectra. J Biol Chem 202:383–396
Chance B, Pappenheimer AM Jr (1954) Kinetic and spectrophotometric studies of cytochrome b5
in midgut homogenates of cecropia. J Biol Chem 209:931–943
76 J. Kranz et al.
Chill JH, Naider F (2011) A solution NMR view of protein dynamics in the biological membrane.
Curr Opin Struct Biol 21:627–633
Chothia C, Lesk AM, Tramontano A, Levitt M, Smith-Gill SJ, Air G, Sheriff S, Padlan EA, Davies
D, Tulip WR et al (1989) Conformations of immunoglobulin hypervariable regions. Nature
342:877–883
Cimmperman P, Baranauskiene L, Jachimoviciute S, Jachno J, Torresan J, Michailoviene V,
Matuliene J, Sereikaite J, Bumelis V, Matulis D (2008) A quantitative model of thermal stabi-
lization and destabilization of proteins by ligands. Biophys J 95:3222–3231
Daniel E, Weber G (1966) Cooperative effects in binding by bovine serum albumin. I. The binding
of 1-anilino-8-naphthalenesulfonate. Fluorimetric titrations. Biochemistry 5:1893–1900
D'Antonio J, Murphy BM, Manning MC, Al-Azzam WA (2012) Comparability of protein thera-
peutics: quantitative comparison of second-derivative amide I infrared spectra. J Pharm Sci
101:2025–2033
Davies DR, Cohen GH (1996) Interactions of protein antigens with antibodies. Proc Natl Acad Sci
USA 93:7–12
Dong A, Huang P, Caughey WS (1990) Protein secondary structures in water from second-
derivative amide I infrared spectra. Biochemistry 29:3303–3308
Dong J, Dinakarpandian D, Carey PR (1998) Extending the Raman analysis of biological samples
to the 100 micromolar concentration range. Appl Spectrosc 52:1117–1122
dos Remedios CG, Moens PD (1995) Fluorescence resonance energy transfer spectroscopy is a
reliable “ruler” for measuring structural changes in proteins. Dispelling the problem of the
unknown orientation factor. J Struct Biol 115:175–185
Dousseau F, Pezolet M (1990) Determination of the secondary structure content of proteins in
aqueous solutions from their amide I and amide II infrared bands. Comparison between classi-
cal and partial least-squares methods. Biochemistry 29:8771–8779
Dukor RK, Keiderling TA (1991) Reassessment of the random coil conformation: vibrational CD
study of proline oligopeptides and related polypeptides. Biopolymers 31:1747–1761
Eberstein W, Georgalis Y, Saenger W (1994) Molecular interactions in crystallizing lysozyme
solutions studied by photon correlation spectroscopy. J Cryst Growth 143:71–78
Eisenmesser EZ, Millet O, Labeikovsky W, Korzhnev DM, Wolf-Watz M, Bosco DA, Skalicky JJ,
Kay LE, Kern D (2005) Intrinsic dynamics of an enzyme underlies catalysis. Nature
438:117–121
Englander SW (2000) Protein folding intermediates and pathways studied by hydrogen exchange.
Annu Rev Biophys Biomol Struct 29:213–238
Englander SW (2006) Hydrogen exchange and mass spectrometry: a historical perspective. J Am
Soc Mass Spectrom 17:1481–1489
Englander SW, Mayne L, Bai Y, Sosnick TR (1997) Hydrogen exchange: the modern legacy of
Linderstrom-Lang. Protein Sci 6:1101–1109
Epstein HF, Schechter AN, Chen RF, Anfinsen CB (1971) Folding of staphylococcal nuclease:
kinetic studies of two processes in acid renaturation. J Mol Biol 60:499–508
Ferreon AC, Moran CR, Gambin Y, Deniz AA (2011) Single-molecule fluorescence studies of
intrinsically disordered proteins. Methods Enzymol 472:179–204
Flanagan JM, Kataoka M, Shortle D, Engelman DM (1992) Truncated staphylococcal nuclease is
compact but disordered. Proc Natl Acad Sci USA 89:748–752
Foldes-Papp Z, Demel U, Domej W, Tilz GP (2002) A new dimension for the development of
fluorescence-based assays in solution: from physical principles of FCS detection to biological
applications. Exp Biol Med (Maywood) 227:291–300
Fu K, Griebenow K, Hsieh L, Klibanov AM, Langer R (1999) FTIR characterization of the sec-
ondary structure of proteins encapsulated within PLGA microspheres. J Control Release
58:357–366
niques: role of conformational and colloidal stability. J Pharm Sci
niques: role of conformational and colloidal stability. J Pharm Sci 100(4):1306–1315
Goldstein SR, Kidder LH, Herne TM, Levin IW, Lewis EN (1996) The design and implementation
of a high-fidelity Raman imaging microscope. J Microsc 184:35–45
Grauw CJ, Otto C, Greve J (1997) Linescan Raman microspectrometry for biological applications.
Appl Spectrosc 51:1607–1612
Greenfield NJ (1996) Methods to estimate the conformation of proteins and polypeptides from
circular dichroism data. Anal Biochem 235:1–10
Greenfield NJ (2006a) Using circular dichroism spectra to estimate protein secondary structure.
Nat Protoc 1(6):2876–2890
Greenfield NJ (2006b) Determination of the folding of proteins as a function of denaturants, osmo-
lytes or ligands using circular dichroism. Nat Protoc 1:2733–2741
Greenfield NJ (2006c) Analysis of the kinetics of folding of proteins and peptides using circular
dichroism. Nat Protoc 1:2891–2899
Greenfield NJ (2007) Using circular dichroism collected as a function of temperature to determine
the thermodynamics of protein unfolding and binding interactions. Nat Protoc 1:2527–2535
Greenfield N, Fasman GD (1969) Computed circular dichroism spectra for the evaluation of pro-
tein conformation. Biochemistry 8:4108–4116
Griebenow K, Klibanov AM (1995) Lyophilization-induced reversible changes in the secondary
structure of proteins. Proc Natl Acad Sci USA 92:10969–10976
Griffiths PR, de Haseth JA (1986) Fourier transform infrared spectroscopy. Wiley Interscience,
New York
Hamuro Y, Weber PC, Griffin PR (2005) High-throughput analysis of protein structure by hydro-
gen/deuterium exchange mass spectrometry. Methods Biochem Anal 45:131–157
Harding SE (1980) The combination of the viscosity increment with the harmonic mean rotational
relaxation time for determining the conformation of biological macromolecules in solution.
Biochem J 189(2):359–361
Harding SE (1981) A compound hydrodynamic shape function derived from viscosity and molecu-
lar covolume measurements. Int J Biol Macromol 3(5):340–341
Harding SE (1997) The intrinsic viscosity of biological macromolecules. Progress in measure-
ment, interpretation and application to structure in dilute solution. Prog Biophys Mol Biol
68(2–3):207–262
Haris PI, Chapman D (1995) The conformational analysis of peptides using Fourier transform IR
spectroscopy. Biopolymers 37:251–263
Haris PI, Lee DC, Chapman D (1986) A Fourier transform infrared investigation of the structural
differences between ribonuclease A and ribonuclease S. Biochim Biophys Acta 874:255–265
Haris PI, Chapman D, Harrison RA, Smith KF, Perkins SJ (1990) Conformational transition
between native and reactive center cleaved forms of alpha 1-antitrypsin by Fourier transform
infrared spectroscopy and small-angle neutron scattering. Biochemistry 29:1377–1380
Hashimoto S, Ikeda T, Takeuuchi H, Harada I (1993) Utilization of a prism monochromator as a
sharp-cut bandpass filter in ultraviolet Raman spectroscopy. Appl Spectrosc 47:1283–1285
Hawe A, Sutter M, Jiskoot W (2008) Extrinsic fluorescent dyes as tools for protein characteriza-
tion. Pharm Res 25:1487–1499
Hawe A, Rispens T, Herron JN, Jiskoot W (2011) Probing bis-ANS binding sites of different affin-
ity on aggregated IgG by steady-state fluorescence, time-resolved fluorescence and isothermal
titration calorimetry. J Pharm Sci 100:1294–1305
He F, Hogan S, Latypov RF, Narhi LO, Razinkov VI (2009a) High throughput thermostability
screening of monoclonal antibody formulations. J Pharm Sci 99:1707–1720
He F, Phan DH, Hogan S, Bailey R, Becker GW, Narhi LO, Razinkov VI (2009b) Detection of IgG
aggregation by a high throughput method based on extrinsic fluorescence. J Pharm Sci
99:2598–2608
Hermansson AM (1972) Functional properties of proteins for foods. Lebensin-Wiss Technol 5:24
78 J. Kranz et al.
Hermansson AM (1979) Aspects of protein structure, rheology and texturization. Food Texture
Rheol [Proc Symp] 265–282
Herr AB, Ballister ER, Bjorkman PJ (2003) Insights into IgA-mediated immune responses from
the crystal structures of human FcalphaRI and its complex with IgA1-Fc. Nature
423:614–620
van Holde KE, Johnson WC, Ho (2006) Methods for the Separation and Characterization of
Macromolecules. In: van Holde KE, Johnson WC, Ho PS. Principles of Physical Biochemistry.
2nd ed. Upper Saddle River, NJ: Pearson Prentice Hall:213–275
Hynes TR, Fox RO (1991) The crystal structure of staphylococcal nuclease refined at 1.7 A resolu-
tion. Proteins 10:92–105
Jackson M, Mantsch HH (1995) The use and misuse of FTIR spectroscopy in the determination of
protein structure. Crit Rev Biochem Mol Biol 30:95–120
Jares-Erijman EA, Jovin TM (2003) FRET imaging. Nat Biotechnol 21:1387–1395
Jelesarov I, Bosshard HR (1999) Isothermal titration calorimetry and differential scanning calo-
rimetry as complementary tools to investigate the energetics of biomolecular recognition.
J Mol Recognit 12:3–18
Kamerzell TJ, Esfandiary R, Joshi SB, Middaugh CR, Volkin DB (2011) Protein-excipient interac-
tions: mechanisms and biophysical characterization applied to protein formulation develop-
ment. Adv Drug Deliv Rev 63:1118–1159
Kauzmann W (1959) Some factors in the interpretation of protein denaturation. Adv Protein Chem
14:1–63
Keiderling TA, Silva RA, Yoder G, Dukor RK (1999) Vibrational circular dichroism spectroscopy
of selected oligopeptide conformations. Bioorg Med Chem 7:133–141
Keiderling TA, Kubelka J, Hilario J (eds) (2006) Vibrational circular dichroism of biopolymers:
summary of methods and applications. CRC, Boca Raton
Kielec JM, Valentine KG, Wand AJ (2009) A method for solution NMR structural studies of large
integral membrane proteins: reverse micelle encapsulation. Biochim Biophys Acta
1798:150–160
Kranz JK, Schalk-Hihi C (2011) Protein thermal shifts to identify low molecular weight fragments.
Methods Enzymol 493:277–298
Krimm S, Bandekar J (1986) Vibrational spectroscopy and conformation of peptides, polypep-
tides, and proteins. Adv Protein Chem 38:181–364
Krishinan KS, Brandts JF (1978) Scanning calorimetry. Methods Enzymol 49:3–14
Kubelka J, Hofrichter J, Eaton WA (2004) The protein folding “speed limit”. Curr Opin Struct Biol
14:76–88
Kurouski D, Lombardi RA, Dukor RK, Lednev IK, Nafie LA (2010) Direct observation and pH
control of reversed supramolecular chirality in insulin fibrils by vibrational circular dichroism.
Chem Commun (Camb) 46:7154–7156
Kurouski D, Dukor RK, Lu X, Nafie LA, Lednev IK (2012) Spontaneous inter-conversion of insu-
lin fibril chirality. Chem Commun (Camb) 48:2837–2839
Ladbury JE, Chowdhry BZ (1996) Sensing the heat: the application of isothermal titration calorim-
etry to thermodynamic studies of biomolecular interactions. Chem Biol 3:791–801
Lakhani A, Malon P, Keiderling TA (2009) Comparison of vibrational circular dichroism instru-
ments: development of a new dispersive VCD. Appl Spectrosc 63:775–785
Lakowicz JR (2006) Principles of fluorescence spectroscopy, 3rd edn. Springer, New York
LeTilly V, Royer CA (1993) Fluorescence anisotropy assays implicate protein–protein interactions
in regulating trp repressor DNA binding. Biochemistry 32:7753–7758
Li CH, Li T (2009) Application of vibrational spectroscopy to the structural characterization of
monoclonal antibody and its aggregate. Curr Pharm Biotechnol 10:391–399
Li Y, Williams TD, Topp EM (2008) Effects of excipients on protein conformation in lyophilized
solids by hydrogen/deuterium exchange mass spectrometry. Pharm Res 25:259–267
Li Y, Mach H, Blue JT (2011) High throughput formulation screening for global aggregation
behaviors of three monoclonal antibodies. J Pharm Sci 100:2120–2135
Liu J, Andya JD, Shire SJ (2006) A critical review of analytical ultracentrifugation and field flow
fractionation methods for measuring protein aggregation. AAPS J 8:E580–E589
Loria JP, Berlow RB, Watt ED (2008) Characterization of enzyme motions by solution NMR
relaxation dispersion. Acc Chem Res 41:214–221
Lundqvist R (1999) Molecular weight studies on hydroxypropyl methyl cellulose. Part 2. Intrinsic
viscosity. Int J Polym Anal Charact 5(1):61–84
Ma S, Cao X, Mak M, Sadik A, Walkner C, Freedman TB, Lednev IK, Dukor RK, Nafie LA (2007)
Vibrational circular dichroism shows unusual sensitivity to protein fibril formation and devel-
opment in solution. J Am Chem Soc 129:12364–12365
Machtle W, Borger L (2006) Sedimentation velocity. In: Machtle W, Borger L (eds) Analytical
ultracentrifugation of polymers and nanoparticles. Heidelberg Springer:47–96
Macosko CW (1994) In: Macosko CW (ed) Rheology: principles, measurements and applications.
Wiley-VCH, New York, p 550
Mandel M (1993) Applications of dynamic light scattering to polyelectrolytes in solution. In:
Brown W (ed) Dynamic light scattering: the method and some applications. Oxford University
Press, New York, pp 319–371
Manley G, Loria JP (2012) NMR insights into protein allostery. Arch Biochem Biophys
519:223–231
Mann CJ, Royer CA, Matthews CR (1993) Tryptophan replacements in the trp aporepressor from
Escherichia coli: probing the equilibrium and kinetic folding models. Protein Sci 2:1853–1861
Markley JL, Ulrich EL, Westler WM, Volkman BF (2003) Macromolecular structure determina-
tion by NMR spectroscopy. Methods Biochem Anal 44:89–113
Matulis D, Kranz JK, Salemme FR, Todd MJ (2005) Thermodynamic stability of carbonic anhy-
drase: measurements of binding affinity and stoichiometry using ThermoFluor. Biochemistry
44:5258–5266
Meyer JD, Bai SJ, Rani M, Suryanarayanan R, Nayar R, Carpenter JF, Manning MC (2004) Infrared
spectroscopic studies of protein formulations containing glycine. J Pharm Sci 93:1359–1366
Mezzasalma TM, Kranz JK, Chan W, Struble GT, Schalk-Hihi C, Deckman IC, Springer BA, Todd
MJ (2007) Enhancing recombinant protein quality and yield by protein stability profiling.
J Biomol Screen 12:418–428
Mittermaier AK, Kay LE (2009) Observing biological dynamics at atomic resolution using NMR.
Trends Biochem Sci 34:601–611
Miura T, Takeuchi H, Harada I (1991) Raman spectroscopic characterization of tryptophan side
chains in lysozyme bound to inhibitors: role of the hydrophobic box in the enzymatic function.
Biochemistry 30:6074–6080
Miyawaki A (2011) Proteins on the move: insights gained from fluorescent protein technologies.
Nat Rev Mol Cell Biol 12:656–668
Miyazawa T, Shimanouchi T, Mizushima S (1956) Characteristic infrared bands of monosubsti-
tuted amides. J Chem Phys 24:408
Monkos K, Turczynski B (1991) Determination of the axial ratio of globular proteins in aqueous
solution using viscometric measurements. Int J Biol Macromol 13(6):341–344
Murphy KP, Freire E (1992) Thermodynamics of structural stability and cooperative folding
behavior in proteins. Adv Protein Chem 43:313–361
Nafie LA (1996) Vibrational optical activity. Appl Spectrosc 50:14A–26A
Nafie LA (2011) Vibrational optical activity: principles and applications. Wiley, Chichester
Nafie LA, Dukor RK (eds) (2007) Applications of vibrational optical activity in the pharmaceuti-
cal industry. Wiley, Chichester
Nashine VC, Hammes-Schiffer S, Benkovic SJ (2010) Coupled motions in enzyme catalysis. Curr
Opin Chem Biol 14:644–651
Norvaisas P, Petrauskas V, Matulis D (2012) Thermodynamics of cationic and anionic surfactant
interaction. J Phys Chem B 116:2138–2144
Otto MR, Lillo MP, Beechem JM (1994) Resolution of multiphasic reactions by the combination
of fluorescence total-intensity and anisotropy stopped-flow kinetic experiments. Biophys J
67:2511–2521
80 J. Kranz et al.
Perez C, Castellanos IJ, Costantino HR, Al-Azzam W, Griebenow K (2002) Recent trends in sta-
bilizing protein structure upon encapsulation and release from bioerodible polymers. J Pharm
Pharmacol 54:301–313
Perrin F (1926) J Phys Radium 1:390–401
Philo JS (2006) Is any measurement method optimal for all aggregate sizes and types? AAPS J
8:E564–E571
Philo JS (2005) Analytical ultracentrifugation. In: Borchardt RT, Middaugh CR, series eds.
Biotechnology: pharmaceutical aspects, Jiskoot W, Crommelin D, vol eds. Volume III: Methods
for structural analysis of protein pharmaceuticals. AAPS Press, Arlington, pp 379–412
Pollok BA, Heim R (1999) Using GFP in FRET-based applications. Trends Cell Biol 9:57–60
Putnam WS, Prabhu S, Zheng Y, Subramanyam M, Wang YM (2010) Pharmacokinetic, pharma-
codynamic and immunogenicity comparability assessment strategies for monoclonal antibod-
ies. Trends Biotechnol 28:509–516
Ramsey JD, Gill ML, Kamerzell TJ, Price ES, Joshi SB, Bishop SM, Oliver CN, Middaugh CR
(2009) Using empirical phase diagrams to understand the role of intramolecular dynamics in
immunoglobulin G stability. J Pharm Sci 98:2432–2447
Raso SW, Clark PL, Haase-Pettingell C, King J, Thomas GJ Jr (2001) Distinct cysteine sulfhydryl
environments detected by analysis of Raman S-hh markers of Cys–>Ser mutant proteins. J Mol
Biol 307:899–911
Rossmann MG (2001) Molecular replacement–historical background. Acta Crystallogr D Biol
Crystallogr 57:1360–1366
Royer CA (1993) Understanding fluorescence decay in proteins. Biophys J 65:9–10
Royer CA (1995) Fluorescence spectroscopy. Methods Mol Biol 40:65–89
Royer CA (2006) Probing protein folding and conformational transitions with fluorescence. Chem
Rev 106:1769–1784
Royer CA, Scarlata SF (2008) Fluorescence approaches to quantifying biomolecular interactions.
Methods Enzymol 450:79–106
Sanchez-Ruiz JM (2011) Probing free-energy surfaces with differential scanning calorimetry.
Annu Rev Phys Chem 62:231–255
Sapienza PJ, Lee AL (2010) Using NMR to study fast dynamics in proteins: methods and applica-
tions. Curr Opin Pharmacol 10:723–730
Schuck P, Perugini MA, Gonzales NR, Howlett GJ, Schubert D (2002) Size-distribution analysis
of proteins by analytical ultracentrifugation: strategies and application to model systems.
Biophys J 82:1096–1111
Schümmer P. Mechanics of Non-Newtonian Fluids. Von W. R. Schowalter. Pergamon Press,
Oxford–Frankfurt 1978. 1. Aufl., IX, 300 S., zahlr. Abb. u. Tab., geb., $ 35.00. Chemie
Ingenieur Technik. 1979;51(7):766–766
Schweitzer-Stenner R, Eker F, Huang Q, Griebenow K, Mosz P, Kozlowski P (2002) Structure
analysis of dipeptides in water by exploring and utilizing the structural sensitivity of amide III
by polarized visible Raman, FTIR spectroscopy and DFT based normal coordinate analysis.
J Phys Chem B 106:4294–4304
Scott DJ, Schuck P (2005) A brief introduction to the analytical ultracentrifugation of proteins for
beginners. In: Scott DJ, Harding SE, Rowe AJ (eds) Analytical ultracentrifugation techniques
and methods. RSC Publishing, Cambridge, pp 1–25
Sharp KA, Nicholls A, Friedman R, Honig B (1991) Extracting hydrophobic free energies from
experimental data: relationship to protein folding and theoretical models. Biochemistry
30:9686–9697
Shaw DJ (1992) Rheology. Colloidal and surface chemistry, 4th edn. Butterworth-Heinemann,
Boston
Shen Y, Lange O, Delaglio F, Rossi P, Aramini JM, Liu G, Eletsky A, Wu Y, Singarapu KK, Lemak
A, Ignatchenko A, Arrowsmith CH, Szyperski T, Montelione GT, Baker D, Bax A (2008)
Consistent blind protein structure generation from NMR chemical shift data. Proc Natl Acad
Sci USA 105:4685–4690
Shi Z, Chen K, Liu Z, Kallenbach NR (2006) Conformation of the backbone in unfolded proteins.
Chem Rev 106:1877–1897
Siamwiza MN, Lord RC, Chen MC, Takamatsu T, Harada I, Matsuura H, Shimanouchi T (1975)
Interpretation of the doublet at 850 and 830 cm−1 in the Raman spectra of tyrosyl residues in
proteins and certain model compounds. Biochemistry 14:4870–4876
Skinner AL, Laurence JS (2008) High-field solution NMR spectroscopy as a tool for assessing
protein interactions with small molecule ligands. J Pharm Sci 97:4670–4695
Skinner AL, Laurence JS (2010) Probing residue–specific interactions in the stabilization of pro-
teins using high-resolution NMR: a study of disulfide bond compensation. J Pharm Sci
99:2643–2654
Skinner JJ, Lim WK, Bedard S, Black BE, Englander SW (2012) Protein dynamics viewed by
hydrogen exchange. Protein Sci 21:996–1005
Smith BC (1996) Fundamentals of fourier transform infrared spectroscopy. CRC, Boca Raton
Socrates G (2004) Infrared and Raman characteristic group frequencies: tables and charts, 3rd edn.
Wiley, Chichester
Sondermann P, Oosthuizen V (2002) X-ray crystallographic studies of IgG-Fc gamma receptor
interactions. Biochem Soc Trans 30:481–486
Spink CH (2008) Differential scanning calorimetry. Methods Cell Biol 84:115–141
Stefani M, Dobson CM (2003) Protein aggregation and aggregate toxicity: new insights into pro-
tein folding, misfolding diseases and biological evolution. J Mol Med (Berl) 81:678–699
Stern KG (1937) Spectroscopy of catalase. J Gen Physiol 20:631–648
Stokes GG (1852) On the change of refrangibility of light. Philos Trans R Soc Lond
142:463–562
Surewicz WK, Mantsch HH (1988) New insight into protein secondary structure from resolution-
enhanced infrared spectra. Biochim Biophys Acta 952:115–130
Suryaprakash P, Prakash V (2000) Unfolding of multimeric proteins in presence of denaturants.
A case study of helianthinin from Helianthus annuus L. Nahrung 44(3):178–183
Susi H, Byler DM (1986) Resolution-enhanced Fourier transform infrared spectroscopy of
enzymes. Methods Enzymol 130:290–311
Susi H, Timasheff SN, Stevens L (1967) Infrared spectra and protein conformations in aqueous
solutions. I. The amide I band in H2O and D2O solutions. J Biol Chem 242:5460–5466
Tamm LK, Lai AL, Li Y (2007) Combined NMR and EPR spectroscopy to determine structures of
viral fusion domains in membranes. Biochim Biophys Acta 1768:3052–3060
Tanford C, Buzzell JG (1956) The viscosity of aqueous solutions of bovine serum albumin between
pH 4.3 and 10.5. J Phys Chem 60:225–231
Tsuboi M, Suzuki M, Overman SA, Thomas GJ Jr (2000) Intensity of the polarized Raman band
at 1340–1345 cm−1 as an indicator of protein alpha-helix orientation: application to Pf1 fila-
mentous virus. Biochemistry 39:2677–2684
Tsutsui Y, Wintrode PL (2007) Hydrogen/deuterium exchange-mass spectrometry: a powerful tool
for probing protein structure, dynamics and interactions. Curr Med Chem 14:2344–2358
Tu RS, Breedveld V (2005) Microrheological detection of protein unfolding. Phys Rev E Stat
Nonlin Soft Matter Phys 72(4 Pt 1):041914
Van Holde KE, Johnson WC, Ho PS (2006) Principles of physical biochemistry. Pearson/Prentice
Hall, Upper Saddle River
Velazquez-Campoy A, Ohtaka H, Nezami A, Muzammil S, Freire E (2004) Isothermal titration
calorimetry. Curr Protoc Cell Biol, Chapter 17, Unit 17.18
Venyaminov S, Kalnin NN (1990a) Quantitative IR spectrophotometry of peptide compounds in
water (H2O) solutions. I. Spectral parameters of amino acid residue absorption bands.
Biopolymers 30:1243–1257
Venyaminov S, Kalnin NN (1990b) Quantitative IR spectrophotometry of peptide compounds in
water (H2O) solutions. II. Amide absorption bands of polypeptides and fibrous proteins in
alpha-, beta-, and random coil conformations. Biopolymers 30:1259–1271
Wand AJ (2001) Dynamic activation of protein function: a view emerging from NMR spectros-
copy. Nat Struct Biol 8:926–931
82 J. Kranz et al.
Wang Y, Purrello R, Jordan T, Spiro TG (1991) UVRR spectroscopy of the peptide bond. J Am
Chem Soc 113:6359–6368
Weber G (1952a) Polarization of the fluorescence of macromolecules. II. Fluorescent conjugates
of ovalbumin and bovine serum albumin. Biochem J 51:155–167
Weber G (1952b) Polarization of the fluorescence of macromolecules. I. Theory and experimental
method. Biochem J 51:145–155
Weber G, Daniel E (1966) Cooperative effects in binding by bovine serum albumin. II. The binding
of 1-anilino-8-naphthalenesulfonate. Polarization of the ligand fluorescence and quenching of
the protein fluorescence. Biochemistry 5:1900–1907
Weber G, Laurence DJ (1954) Fluorescent indicators of adsorption in aqueous solution and on the
solid phase. Biochem J 56:xxxi
Weber PC, Salemme FR (2003) Applications of calorimetric methods to drug discovery and the
study of protein interactions. Curr Opin Struct Biol 13:115–121
Wormald MR, Petrescu AJ, Pao YL, Glithero A, Elliott T, Dwek RA (2002) Conformational stud-
ies of oligosaccharides and glycopeptides: complementarity of NMR, X-ray crystallography,
and molecular modelling. Chem Rev 102:371–386
Wyatt PJ (1993) Light scattering and the absolute characterization of macromolecules. Anal Chim
Acta 272(1):1–40
Yang JT (1961) The viscosity of macromolecules in relation to molecular conformations. Adv
Protein Chem 16:323–401
Yoder G, Polese A, Silva RAGD, Formaggio F, Crisma M, Broxterman QB, Kamphuis J, Toniolo
C, Keiderling TA (1997) Conformational characterization of terminally blocked L-(alpha Me)
Val homopeptides using vibrational and electronic circular dichroism. 3(10)-helical stabiliza-
tion by peptide–peptide interaction. J Am Chem Soc 119:10278–10285
Zhang R, Monsma F (2010) Fluorescence-based thermal shift assays. Curr Opin Drug Discov
Devel 13:389–402
Zhu F, Tranter GE, Isaacs NW, Hecht L, Barron LD (2006) Delineation of protein structure classes
from multivariate analysis of protein Raman optical activity data. J Mol Biol 363:19–26
Ziarek JJ, Peterson FC, Lytle BL, Volkman BF (2011) Binding site identification and structure
determination of protein-ligand complexes by NMR a semiautomated approach. Methods
Enzymol 493:241–275
Zuber G, Prestrelski SJ, Benedek K (1992) Application of Fourier transform infrared spectroscopy
to studies of aqueous protein solutions. Anal Biochem 207:150–156
Zubriene A, Matuliene J, Baranauskiene L, Jachno J, Torresan J, Michailoviene V, Cimmperman
P, Matulis D (2009) Measurement of nanomolar dissociation constants by titration calorimetry
and thermal shift assay – radicicol binding to Hsp90 and ethoxzolamide binding to CAII. Int J
Mol Sci 10:2662–2680
Chapter 4
Biophysical Techniques for Protein Size
Distribution Analysis
Ziping Wei and Alla Polozova
4.1 Introduction
Size distribution analysis of biopharmaceutical protein products is commonly

performed to ensure product quality. Protein size heterogeneity usually originates
from a combination of the following factors: (1) aggregation, (2) reversible self-
association, (3) fragmentation, (4) conformation/shape heterogeneity, and (5)
glycosylation. These contributing factors are briefly described below.
Aggregation: Protein aggregation is an association of proteins to form multimeric spe-
cies that can be soluble aggregates or insoluble particles ranging from several nano
meters to millimeters in size (Cromwell et al. 2006a; Mahler et al. 2009). Protein
aggregates can be of multiple different types, including covalently or non-covalently
linked, reversible or irreversible, and with native or denatured structures.
Self-association: Some protein molecules are prone to transient concentration-
dependent reversible self-association, which can lead to significant size heterogene-
ity (Cromwell et al. 2006b, 2006a).
Fragmentation: Protein fragmentation causes the formation of smaller species in a
range of sizes depending on the cleavage site.
Conformation/shape heterogeneity: Proteins often have a broad range of conforma-
tional heterogeneity leading to multiple different molecular shapes and a wider dis-
tribution of hydrodynamic sizes (Vendruscolo 2007; Manta et al. 2011).
Z. Wei
Analytical Development, Novavax, Rockville, MD 20850, USA
A. Polozova (*)
Biopharmaceutical Development, Medimmune, 1 Medimmune Way,
Gaithersburg, MD 20878, USA

84 Z. Wei and A. Polozova
Glycosylation: Glycosylation of glycoproteins influences their hydrodynamic sizes,

and glycosylation heterogeneity can contribute to size heterogeneity (Wen et al.
1996).
Given the diversity of potential size variants, multiple analytical methods are
needed to assess and characterize protein products (Mahler et al. 2009; Philo 2006).
Biophysical methods commonly used for size distribution analysis can be classified
based on their principles: size exclusion, field-flow fractionation, sedimentation,
light scattering, microscopy, light obscuration, electrical zone sensing, and mass
spectrometry. This chapter describes these common size distribution analysis
methods, including their advantages and limitations (Table 4.1).
4.2 Analytical Methods for Size Distribution Analysis
4.2.1 Size-Exclusion Chromatography
High-performance size-exclusion chromatography (HPSEC) is one of the most

frequently used analytical methods for detection and quantitation of soluble protein
aggregates and fragments. High-throughput capability and a good resolution
between monomers and soluble aggregates are important advantages of HPSEC.
HPSEC separates molecules based on partitioning between porous stationary
phase and aqueous mobile phase. The protein molecules are eluted according to
their hydrodynamic size rather than actual molecular weight. Monomeric
conformational variants of monomer may have different hydrodynamic sizes and,
therefore, elute as separate peaks. For example, partially denatured monomers with
an elongated shape have been shown to elute before the monomer peak (Philo
2006). Protein glycosylation with large oligosaccharides often affects hydrodynamic
size and can have a noticeable effect on the peak position in the HPSEC profile
(Wen et al. 1996).
Reversible self-association can be detected by HPSEC as either broadening of
monomer peaks or through the presence of shoulder peaks. Although HPSEC can
be useful for analysis of protein self-association (Moore et al. 1999), the results
should always be interpreted with caution. In some cases nonspecific interactions
with the column matrix can cause distorted peak shape (Wang et al. 1996).
Other limitations of HPSEC include the potential loss of some aggregate species
due to selective absorption on the stationary phase (Liu et al. 2006; Gabrielson et al.
2007a) or retention by the column of very large aggregate species exceeding the
upper HPSEC limit of about 0.1 μm. In addition, higher ionic strength, choice of
salts, or different pH of HPSEC mobile phase may change protein conformation,
induce new aggregates, or cause dissociation of weakly associated aggregates and
thus bias the results (Washabaugh and Collins 1986; Philo 2006).
4
Table 4.1 Analytical methods for protein size distribution analysis

Method category Method Size Pros and cons Methods references
Size-exclusion HPSEC <0.1 μm Pros: fast, high throughput Liu et al. (2006), Gabrielson
chromatography Cons: limited size range, limited choice of et al. (2007a), Philo
buffers, possible matrix interactions (2006)
Field-flow fractionation AF4 <1 μm (normal Pros: high throughput, broad dynamic range, Gabrielson et al. (2007a),
mode) and works with wide range of buffers Litzen et al. (1993), Liu
1–50 μm (steric Cons: requires highly trained analysts. Membrane et al. (2006), Cao et al.
mode) interactions are possible. Separation mode (2009a, b)
crossover at ~1 μm
Sedimentation AUC <0.1 μm Pros: works with wide range of buffers and Liu et al. (2006), Gabrielson
conditions, no matrix interactions et al. (2007b), Rivas et al.
Cons: low throughput, requires highly trained (1999)
analysts
Light scattering Static light <1 μm Pros: absolute measurement Follmer et al. (2004), Oliva
scattering Cons: requires highly trained analysts, does not et al. (2004)
resolve different sizes
Dynamic light <1 μm Pros: fast, easy, can be high throughput Jossang et al. (1988),
scattering Cons: sensitive to dust, does not resolve different Follmer et al. (2004),
sizes in heterogeneous samples Philo (2006, 2009)
Turbidity NA Pros: compendial test, fast, easy Tatford et al. (2004)
Cons: no quantitative information on size, counts
Biophysical Techniques for Protein Size Distribution Analysis
Nanoparticle 0.02–1 μm Pros: unbiased size measurements in heteroge- Filipe et al. (2010)
tracking neous samples
analysis Cons: limited dynamic range (~107–109 particles/
mL)
Light obscuration Light 1–150 μm Pros: compendial test. Counts one particle at a Fujishita et al. (1995), Narhi
obscuration time, does not rely on distribution models et al. (2009)
Cons: sizing of protein particles is not reliable.
Does not work well with high protein
concentrations, turbid and viscous samples
85
(continued)
86
Table 4.1 (continued)

Electrical sensing zone Electrical 0.4–1,200 μm Pros: wide dynamic range, counts single particle Narhi et al. (2009)
sensing zone at a time
Cons: limited concentration range, requires
electrolyte solution and multiple measure-
ments with different apertures to cover the
entire range
Microscopy EM <100 μm Pros: high resolution Roux (1999)
Cons: sample preparation artifacts possible
AFM <1 μm Pros: high resolution Thomson (2005), Fechner
Cons: sample preparation and imaging artifacts et al. (2009)
possible
Light >0.5 μm Pros: fast, easy Barber et al. (1989)
microscopy Cons: low precision, difficult to detect translucent
particles
Flow microscopy Flow >0.5 μm Pros: fast, provides information on counts, size, Narhi et al. (2009), Sharma
microscopy and morphology et al. (2007), Huang et al.
Cons: difficult to detect translucent particles (2009), Wuchner et al.
(2010)
Mass spectrometry Electrospray or >0.1 μm Pros: detailed structural information, high Krutchinsky et al. (2000),
MALDI MS sensitivity and resolution Kukrer et al. (2010), Liu
Cons: requires highly trained analysts. Potential et al. (2008), North et al.
dissociation of non-covalently linked (2009), Wang et al. (2011)
aggregates Limitation of m/z range.
Qualitative or semiquantitative method
Z. Wei and A. Polozova
4
New technology Gas-phase 0.003–0.25 μm Pros: fast, direct measurement, high resolution Pease et al. (2008)
electropho- Cons: limited concentration range, electrospray
retic buffer different from formulation buffer
mobility
Resonant mass >0.1 μm Pros: measures single particle at a time, direct Burg et al. (2007)
measure- mass measurement, can distinguish different
ment types based on density
Cons: low sensitivity to particles with density
close to that of surrounding solution
Biophysical Techniques for Protein Size Distribution Analysis
87
4.2.2 Field-Flow Fractionation
Asymmetric flow field-flow fractionation (AF4) is an important orthogonal

separation-based matrix-free technique with a wide dynamic range, capable of anal-
ysis of fragments, monomers, and aggregates up to 50 μm in size (Gabrielson et al.
2007a; Litzen et al. 1993; Liu et al. 2006; Cao et al. 2009a). AF4 separation is car-
ried out in a flat trapezoidal channel equipped with a semipermeable membrane.
Analytes are separated based on their hydrodynamic size by a cross flow through
the semipermeable membrane, perpendicular to the elution flow. The pore size of
the semipermeable membrane is selected to retain the analytes, but it allows a free
flow of the mobile phase. In the typical AF4 separation, the elution flow is the fast-
est in the middle of the channel and slower near the top and the bottom. Cross flow
pushes larger species closer to the semipermeable membrane at the bottom of the
channel; therefore, larger species elute later, after smaller species. Recent publica-
tions demonstrated that AF4 can be successfully applied for analysis of large pro-
tein aggregates, and collected data are consistent with other aggregate analysis
methods such as HPSEC and flow microscopy (Cao et al. 2009b).
One of the important advantages of AF4 is a wider selection of mobile phases
compared to HPSEC, because the former is compatible with a broader range of pH
and salt concentrations. Other advantages include a wide dynamic range for
separation of different species and the absence of a stationary phase.
Limitations of AF4 include potential nonspecific interactions of certain analytes
with the semipermeable membrane, lack of information on particle counts, and a
crossover of the separation mode in the 1 μm range described below. Nonspecific
interactions of analytes with the semipermeable membranes may lead to distorted
elution profiles. The possibility of such interactions should be carefully assessed for
each sample type. If needed, mobile phase should be optimized to minimize the
nonspecific interactions. Separation mode crossover may pose a significant
challenge for AF4 analysis of highly heterogeneous samples. For species larger than
1 μm, the steric elution mode becomes dominant. It results in a faster elution of
larger analytes, which is the opposite of the normal elution order in the typical AF4
profile (Yohannes et al. 2011). Therefore, the analysis of samples containing both
sub- and above-micron species can be compromised, which can limit AF4 practical
applicability to submicron range.
4.2.3 Analytical Ultracentrifugation
Analytical ultracentrifugation (AUC) is widely used for measurements of aggregate

and fragment levels, determination of aggregate stoichiometry, and analysis of
reversible self-association (Liu et al. 2006; Gabrielson et al. 2007b; Rivas et al.
1999). AUC can also provide information on distribution of protein conformations,
if they are related to different molecular shapes. Two major AUC methods are
4 Biophysical Techniques for Protein Size Distribution Analysis 89
sedimentation velocity and sedimentation equilibrium. In a typical sedimentation

velocity experiment, AUC cell channels loaded with a sample and a reference buffer
are scanned by either absorbance or interference optics as analytes sediment to the
bottom of the cells at relatively high speed, about 45,000 rpm. Collected scans
contain information on the shape of the sedimentation boundary as a function of
time. Sedimentation boundaries depend on both the size and shape of the sedimenting
molecules. Molecular weight distribution can be derived from the analysis of
multiple sedimentation boundaries if certain assumptions about the shape of the
molecule can be made.
Sedimentation equilibrium experiments involve spinning at lower speeds until
sedimentation-diffusion equilibrium is established. Usually, several different speeds
and sample concentrations are evaluated in one sedimentation equilibrium
experiment. Molecular weight distribution information is derived from the analysis
of scans by fitting to theoretical distribution models (e.g., monomers, monomers
and non-dissociable aggregates, reversible dimers, and trimers).
Similarly to AF4, important advantages of AUC are the minimization of matrix
interactions, which may result in loss of some aggregate species, and compatibility
with many buffers in a broad pH range. In addition, AUC allows sample analysis at
predetermined concentrations, and it is the method of choice for studying self-
association behavior. Limitations of AUC include the following: a relatively high
(about 1 %) limit of quantitation for small aggregates and fragments, a relatively
narrow (up to approximately 0.1 μm) dynamic range, low throughput, and the
requirement for highly trained analysts (Narhi et al. 2009).
Multiple static and dynamic light scattering techniques are widely used for protein
size heterogeneity analysis, including sizing and quantitation of aggregates (Follmer
et al. 2004; Philo 2006, 2009), distribution of glycosylation (Wen et al. 1996), and
reversible self-association (Wu et al. 1997; Attri and Minton 2005). In methods based
on static or classical light scattering, the molecular weight of molecules in solution is
derived from the light scattering theory equations. These equations link scattered
light intensity and concentration to the molecular weight. Multiangle light scattering
(MALS) applications are most commonly used for proteins in solution (Oliva et al.
2004). In a typical MALS measurement, light scattering intensity measured at three
or more angles is determined and used to calculate the molecular weight. For mole-
cules larger than 10 nm, the radius of gyration can be also calculated based on the
dependance of the light scattering intensity on the scattering angle.
Dynamic light scattering (DLS) relies on the analysis of an autocorrelation func-
tion derived from measurements of scattered light intensity fluctuations due to
Brownian motion of molecules in solution (Jossang et al. 1988). The translational
diffusion coefficient Dτ of protein molecules in the samples is derived from the
autocorrelation function. The hydrodynamic radius Rh is calculated using the

Stokes–Einstein equation linking Dτ, temperature, and viscosity of the solution.
Both static and DLS methods can be used to determine the presence of protein
aggregates in solution and to measure aggregates up to 1 μm in size, including high-
throughput screening applications. A composition gradient technique based on
static light scattering measurements of increasing protein concentrations can be
applied for analysis of self-associating systems (Attri and Minton 2005).
The light scattering methods work particularly well for sizing of monodisperse
samples. However, they have significant limitations in the analysis and quantitation
of size distributions in polydisperse samples. Results can be skewed to larger sizes,
because scattering intensity depends on the sixth power of the size, and thus, larger
species scatter significantly more light. If separation of different species (e.g., by
HPSEC or AF4) is carried out prior to analysis, it significantly improves the sizing
capabilities of the light scattering methods. MALS or DLS coupled with HPSEC or
AF4 as on-line detectors are powerful methods for the characterization of aggre-
gates. In particular, MALS and DLS detectors in tandem can provide information
on the molecular weight, size, and shape of aggregates (Rambaldi et al. 2011).
Proteins with high levels of glycosylation can be analyzed with a triple-detector
HPSEC method with consecutively attached MALS, UV, and refractive index detec-
tors. This method allows on-line determination of proportions and distributions of
oligosaccharide across different fractions (Wen et al. 1996).
Solution turbidity measurement is a low-resolution light scattering-based
technique that can provide a rough assessment of size heterogeneity levels. Solution
turbidity can be determined quantitatively by intensity loss of the transmitted beam
or by the intensity of the light scattered at a given angle. The presence of protein
aggregates or particles often results in an increase in turbidity. Therefore, the degree
of turbidity in a protein solution may be used to qualitatively assess the extent of
aggregation (Tatford et al. 2004).
Nanoparticle tracking analysis (NTA) is a new light scattering-based technique
that is rapidly gaining use for the analysis of submicron particles (Filipe et al. 2010).
Particle size measurements are based on tracking the distance traveled by individual
particles over a given amount of time and do not rely on the scattering intensity. Due
to this, the NTA technique can provide relatively unbiased measurements of particle
sizes even in samples with heterogeneous populations. A rather narrow dynamic
range of particle counts suitable for the measurements (~107–109 particle/mL) is a
limitation to its applicability.
4.2.5 Light Obscuration
Light obscuration techniques (Fujishita et al. 1995; Narhi et al. 2009) can be applied
to the enumeration and size determination of large aggregates and particles (microm-
eter size). Measurements are based on blockage of laser light by individual particles
passing the detector in the flow cell. The change in light intensity reaching the
detector is converted to the particle size based on calibration with known particle
standards. Light obscuration is often used as a routine monitoring test for injectable
drugs to detect levels of subvisible particles (≥10 μm). Counting of a single particle
at a given time is an important advantage of this method, because it offers a direct
measurement of particle counts that does not require distribution models. However,
high-concentration protein solutions are not always compatible with this method.
For example, the presence of large amounts of background protein can reduce
contrast between the particles and the solution and interfere with measurements. In
addition, sizing of protein particles by light obscuration methods is often not reliable
because they have different optical properties and block different amounts of light
compared to commonly used polystyrene bead calibration standards. Introduction
of calibration standards similar in optical properties to large protein aggregates
would improve the sizing accuracy; however, such standards are still under
development and not yet commercially available.
4.2.6 Electrical Sensing Zone Analysis
Electrical sensing zone (ELS) analysis can be applied to large aggregates ranging in
size from 0.4 μm up to 1,200 μm (Narhi et al. 2009). Particle volume is a primary
parameter measured by this method. The sample is suspended in a weak electrolyte
solution. When large aggregates and particles pass through a small orifice with
applied voltage, they displace a volume of the electrolyte solution equivalent to their
volume. This results in a decrease in electric current, registered as a pulse. The size
information is derived from the magnitude of the pulse, which is directly proportional
to the particle volume. In addition to the size information, the ELS technique also
provides information on the particle counts for a given sample volume. The
requirements for an electrolyte solution and the use of multiple apertures to analyze
the entire aggregate and particle distribution may limit the applicability of this
technique.
4.2.7 Microscopy
Electron microscopy (EM) and atomic force microscopy (AFM) offer a high-
resolution capability for the characterization of conformational heterogeneity and
aggregates at the molecular level (Roux 1999; Thomson 2005). In electron
microscopy methods, the image is formed by a beam of electrons either passing
through (transmission EM) or reflecting off the specimen (scanning EM), while sets
of magnets serve as condenser and objective lenses. Because an electron beam is
used for imaging, EM offers very good spatial resolution. The theoretical resolution
limit is less than 10 Å; however, in practical applications with biological samples,
the resolution limit is ~1 nm, which still allows a detailed analysis of structure
(Roux 1999). The most promising high-resolution tool for detailed analysis of con-
formational heterogeneity and aggregate structure is cryo-transmission EM tomog-
raphy (Sandin et al. 2004; Banyay et al. 2004). A small droplet of aqueous protein
solution is flash-frozen to form a vitreous ice and directly analyzed in the TEM
without further sample manipulation or preparation. The three-dimensional images
reconstructed by computer software from a set of tilted-plane images provide
detailed information about aggregate order and structure. However, this technique is
still evolving, and, at this time, it is not routinely applied to the analysis of hetero-
geneity in protein solutions.
In AFM, the image is obtained by scanning the surface, usually in a raster pattern,
with very thin tip attached to a cantilever. The spatial resolution and sensitivity of
AFM depend on the mechanical properties of the cantilever and the sharpness of the
tip. Usually, 1 nm or better resolution can be achieved (Engel and Muller 2000).
AFM can provide valuable information on the structure of individual protein
molecules (Thomson 2005), as well as aggregates (Lee et al. 2011). However,
results should be interpreted with caution because this method might be prone to
sample preparation artifacts arising from adsorption on the support surfaces and
sample-tip interactions (Fechner et al. 2009).
Optical microscopy is one of the oldest techniques available for analysis of very
large aggregates, and it can be a useful screening tool for an initial assessment.
Resolution of optical microscopy is defined by the diffraction limit, which is
~0.5 μm for visual light. Large aggregates can be analyzed directly in solution;
however, some translucent particles might be difficult to see due to their low con-
trast with the surrounding solution. Alternatively, large aggregates can be captured
on a filter and detected under oblique light illumination (Barber et al. 1989),
although some thin and fragile particles might be lost during the sample preparation
process.
4.2.8 Flow Microscopy
Over the past several years, flow microscopy has become an increasingly popular
tool for the analysis of large micron-size aggregates in protein solutions (Narhi et al.
2009; Sharma et al. 2007; Huang et al. 2009). This technique offers the combined
advantages of a particle counter and an optical microscope integrated into one
instrument (Sharma et al. 2007; Wuchner et al. 2010). The sample solution flows
through a narrow flow cell, typically 100–300 μm thick. Images magnified by a
10–20× microscope objective are captured by a high-speed camera operating at
8–10 frames per second and automatically analyzed by the image processing
software for presence of particles. Images of all detected particles in the sample are
collected in a database, and information on counts, particle size, transparency, and
morphology can be generated by the software. Large aggregates (particles) in the
2–300 μm range can be detected and analyzed. The ability to provide information
on morphology, shape, and optical properties of large aggregates are important

advantages of flow microscopy. However, results depend on the contrast between
the protein particles and the solution, which may limit the sensitivity of this method
to translucent particles, especially for smaller particle sizes.
4.2.9 Mass Spectrometry
Mass spectrometry (MS) separates analytes based on their mass and charge (m/z
ratio). Electrospray ionization (ESI) and matrix-assisted laser desorption ionization
(MALDI) techniques are commonly used for the analysis of proteins. MS offers
high sensitivity and good mass resolution and has been a powerful tool used in a
large number of studies for determining protein molecular weights and investigating
protein complexes (Bich and Zenobi 2009). Although MS has been recently reported
to be a useful online detection tool for HPSEC analysis of protein aggregates
(Kukrer et al. 2010), there have been very limited applications of this technique to
study protein aggregation for several reasons (Wang et al. 2011). Firstly, the poten-
tial dissociation of non-covalently linked aggregates during ionization or transfer
into the gas phase could limit the applicability of MS-based techniques for size
distribution analysis. Secondly, most MS mass analyzers have upper-range limita-
tions in analysis of high molecular weight species. A time-of-flight (TOF) mass
analyzer provides a relatively wide m/z range and is often used for protein aggregate
analysis (Krutchinsky et al. 2000). Specialized detectors can be used to make the
m/z range of MALDI MS suitable for analysis of larger proteins and complexes
(Bich and Zenobi 2009; Bahr et al. 1996; Wenzel et al. 2005). Lastly, MS is usually
limited to a qualitative analysis of small oligomeric protein aggregates. However, a
recent study showed that MS was capable of probing the early stage of protein
aggregation in a semiquantitative fashion (Wang et al. 2011).
MS is a powerful tool to determine the structures, sequences, and sites of
N-linked and O-linked glycosylation (North et al. 2009). MS is widely used to
determine the molecular weights of protein fragments and is also capable of
identification of cleavage sites of fragments based on the accurate molecular mass
and known posttranslational modifications of fragments (Liu et al. 2008). Therefore,
MS provides more detailed structural information than other techniques for
glycoprotein size analysis and protein fragment analysis.
4.2.10 New Technologies
4.2.10.1 Gas-Phase Electrophoretic Mobility
Electrospray differential mobility analysis, also known as gas-phase electrophoretic

mobility molecular analysis, is a new method for measuring low-order soluble
protein aggregates based on the separation of charge-to-aerodynamic size ratio
(Pease et al. 2008). Size distribution of protein species in the range of 3–250 nm can
be obtained, including distinct peaks for IgG monomers to pentamers. It provides a
fast and quantitative readout of the aggregate distribution. The limitations of the
method are that it works within a limited protein concentration range and the
electrospray step cannot be performed in buffer containing nonvolatile salts.
Additionally, the electrospray buffer may affect the aggregation state of proteins.
4.2.10.2 Resonant Mass Measurement
Resonant buoyant mass measurement of individual aggregates and particles is a novel

technique that is capable of covering a broad range of size distributions for aggregates
and particles larger than 0.1 μm (Burg et al. 2007). Particles flowing through a thin
microfluidic channel are detected via a change in the resonance frequency. This tech-
nique measures one particle at a time and it can accommodate viscous solutions (up to
50 cP) without dilution. However, this technique relies on the density difference
between the particles and the surrounding solution. Because the density of some pro-
tein aggregates and particles can be close to that of the aqueous solutions containing
them, the sensitivity of this method can be impaired in this situation.
4.2.11 Discussion
Therapeutic protein products can have a very broad heterogeneous size distribution,
making size distribution analysis a challenging task. Because no single analytical
method is capable of covering the entire size range of all types of size variants, it is
critical to apply complementary tools to monitor and characterize the entire size dis-
tribution. Depending on the specific requirements for a product, different information,
such as molecular weight, distribution of different species, size, counts, or morphol-
ogy may be needed. Multiple biophysical methods presented in this chapter are capa-
ble of providing this information. Each method may provide a different type of
information depending on a measurement principle, and correlation between orthogo-
nal methods should be thoroughly assessed and understood.
As each analytical method has its own strengths and weaknesses, the choice of
analytical method(s) should be tailored to the sample type and to the intended
purpose. To ensure robust and meaningful size distribution measurements and to
avoid false conclusions, care must be taken in method selection and result
interpretation. Regardless of the method(s) chosen, the use of appropriate controls
is critical because sample preparation and handling may have a significant impact
on the results. Overall, a good knowledge of factors which may influence size
heterogeneity during sample handling and a thorough understanding of physical
principles of the chosen analytical techniques are keys for reliable measurements of
size distribution in protein solutions.
References
Attri AK, Minton AP (2005) Composition gradient static light scattering: a new technique for
rapid detection and quantitative characterization of reversible macromolecular hetero-
associations in solution. Anal Biochem 346:132–138
Bahr U, Rohling U, Lautz C, Strupat K, Schurenberg M, Hillenkamp F (1996) A charge detector
for time-of-flight mass analysis of high mass ions produced by matrix-assisted laser desorption/
ionization (MALDI). J Mass Spectrom Ion Process 153:9–21
Banyay M, Gilstring F, Hauzenberger E, Ofverstedt LG, Eriksson AB, Krupp JJ, Larsson O (2004)
Three-dimensional imaging of in situ specimens with low-dose electron tomography to analyze
protein conformation. Assay Drug Dev Technol 2:561–567
Barber TA, Lannis MD, Williams JG (1989) Method evaluation: automated microscopy as a com-
pendial test for particulates in parenteral solutions. J Parenter Sci Technol 43:27–47
Bich C, Zenobi R (2009) Mass spectrometry of large complexes. Curr Opin Struct Biol
19:632–639
Burg TP, Godin M, Knudsen SM, Shen W, Carlson G, Foster JS, Babcock K, Manalis SR (2007)
Weighing of biomolecules, single cells and single nanoparticles in fluid. Nature
446:1066–1069
Cao S, Pollastrini J, Jiang Y (2009a) Separation and characterization of protein aggregates and
particles by field flow fractionation. Curr Pharm Biotechnol 10:382–390
Cao S, Jiao N, Jiang Y, Mire-Sluis A, Narhi LO (2009b) Sub-visible particle quantitation in protein
therapeutics. Pharmeur Bio Sci Notes 2009:73–79
Cromwell ME, Hilario E, Jacobson F (2006a) Protein aggregation and bioprocessing. AAPS
J 8:E572–E579
Cromwell MEM, Felten C, Flores H, Liu J, Shire SJ (2006b) Self-association of therapeutic
proteins: implications for product development. In: Murphy RM, Tsai AM (eds) Misbehaving
proteins: protein (Mis)folding, aggregation, and stability. Springer, New York
Engel A, Muller DJ (2000) Observing single biomolecules at work with the atomic force micro-
scope. Nat Struct Biol 7:715–718
Fechner P, Boudier T, Mangenot S, Jaroslawski S, Sturgis JN, Scheuring S (2009) Structural infor-
mation, resolution, and noise in high-resolution atomic force microscopy topographs. Biophys
J 96:3822–3831
Filipe V, Hawe A, Jiskoot W (2010) Critical evaluation of Nanoparticle Tracking Analysis (NTA)
by NanoSight for the measurement of nanoparticles and protein aggregates. Pharm Res
27:796–810
Follmer C, Pereira FV, Da Silveira NP, Carlini CR (2004) Jack bean urease (EC 3.5.1.5) aggrega-
tion monitored by dynamic and static light scattering. Biophys Chem 111:79–87
Fujishita O, Sendo T, Hisazumi A, Otsubo K, Aoyama T, Oishi R (1995) The evaluation of sizing
accuracy of particle counters for parenteral drugs. PDA J Pharm Sci Technol 49:267–271
Gabrielson JP, Brader ML, Pekar AH, Mathis KB, Winter G, Carpenter JF, Randolph TW (2007a)
Quantitation of aggregate levels in a recombinant humanized monoclonal antibody formulation
by size-exclusion chromatography, asymmetrical flow field flow fractionation, and
sedimentation velocity. J Pharm Sci 96:268–279
Gabrielson JP, Randolph TW, Kendrick BS, Stoner MR (2007b) Sedimentation velocity analytical
ultracentrifugation and SEDFIT/c(s): limits of quantitation for a monoclonal antibody system.
Anal Biochem 361:24–30
Huang CT, Sharma D, Oma P, Krishnamurthy R (2009) Quantitation of protein particles in paren-
teral solutions using micro-flow imaging. J Pharm Sci 98:3058–3071
Jossang T, Feder J, Rosenqvist E (1988) Photon correlation spectroscopy of human IgG. J Protein
Chem 7:165–171
Krutchinsky AN, Ayed A, Donald LJ, Ens W, Duckworth HW, Standing KG (2000) Studies of
noncovalent complexes in an electrospray ionization/time-of-flight mass spectrometer. In:
Chapman J (ed) Mass spectrometry of proteins and peptides. Humana Press, Totowa
Kukrer B, Filipe V, Duijn EV, Kasper PT, Vreeken RJ, Heck AJR, Jiskoot W (2010) Mass spectro-
metric analysis of intact human monoclonal antibody aggregates fractionated by size-exclusion
chromatography. Pharm Res 27:2197–2204
Lee H, Kirchmeier M, Mach H (2011) Monoclonal antibody aggregation intermediates visualized
by atomic force microscopy. J Pharm Sci 100:416–423
Litzen A, Walter JK, Krischollek H, Wahlund KG (1993) Separation and quantitation of monoclo-
nal antibody aggregates by asymmetrical flow field-flow fractionation and comparison to gel
permeation chromatography. Anal Biochem 212:469–480
fractionation methods for measuring protein aggregation. AAPS J 8:E580–E589
Liu H, Gaza-Bulseco G, Lundell E (2008) Assessment of antibody fragmentation by reversed-
phase liquid chromatography and mass spectrometry. J Chromatogr B Analyt Technol Biomed
Life Sci 876:13–23
Mahler HC, Friess W, Grauschopf U, Kiese S (2009) Protein aggregation: pathways, induction
factors and analysis. J Pharm Sci 98:2909–2934
Manta B, Obal G, Ricciardi A, Pritsch O, Denicola A (2011) Tools to evaluate the conformation of
protein products. Biotechnol J 6:731–741
Moore JM, Patapoff TW, Cromwell ME (1999) Kinetics and thermodynamics of dimer formation
and dissociation for a recombinant humanized monoclonal antibody to vascular endothelial
growth factor. Biochemistry 38:13960–13967
Narhi LO, Jiang Y, Cao S, Benedek K, Shnek D (2009) A critical review of analytical methods for
subvisible and visible particles. Curr Pharm Biotechnol 10:373–381
North SJ, Hitchen PG, Haslam SM, Dell A (2009) Mass spectrometry in the analysis of N-linked
and O-linked glycans. Curr Opin Struct Biol 19:498–506
Oliva A, Llabres M, Farina JB (2004) Applications of multi-angle laser light-scattering detection
in the analysis of peptides and proteins. Curr Drug Discov Technol 1:229–242
Pease LF 3rd, Elliott JT, Tsai DH, Zachariah MR, Tarlov MJ (2008) Determination of protein
aggregation with differential mobility analysis: application to IgG antibody. Biotechnol Bioeng
101:1214–1222
Philo JS (2006) Is any measurement method optimal for all aggregate sizes and types? AAPS
J 8:E564–E571
Philo JS (2009) A critical review of methods for size characterization of non-particulate protein
aggregates. Curr Pharm Biotechnol 10:359–372
Rambaldi DC, Reschiglian P, Zattoni A (2011) Flow field-flow fractionation: recent trends in
protein analysis. Anal Bioanal Chem 399:1439–1447
Rivas G, Stafford W, Minton AP (1999) Characterization of heterologous protein-protein interac-
tions using analytical ultracentrifugation. Methods 19:194–212
Roux KH (1999) Immunoglobulin structure and function as revealed by electron microscopy. Int
Arch Allergy Immunol 120:85–99
Sandin S, Öfverstedt L-G, Wikström A-C, Wrange Ö, Skoglund U (2004) Structure and flexibility
of individual immunoglobulin G molecules in solution. Structure 12:409–415
Sharma DK, King D, Moore PD, Oma P, Thomas D (2007) Flow microscopy for particulate analy-
sis in parenteral and pharmaceutical fluids. Eur J Parenter Pharm Sci 12:97–101
Tatford OC, Gomme PT, Bertolini J (2004) Analytical techniques for the evaluation of liquid pro-
tein therapeutics. Biotechnol Appl Biochem 40:67–81
Thomson NH (2005) Imaging the substructure of antibodies with tapping-mode AFM in air: the
importance of a water layer on mica. J Microsc 217:193–199
Vendruscolo M (2007) Determination of conformationally heterogeneous states of proteins. Curr
Opin Struct Biol 17:15–20
Wang YJ, Shahrokh Z, Vemuri S, Eberlein G, Beylin I, Busch M (1996) Characterization, stability,
and formulations of basic fibroblast growth factor. Pharm Biotechnol 9:141–180
Wang G, Johnson AJ, Kaltashov IA (2011) Evaluation of electrospray ionization mass spectrometry
as a tool for characterization of small soluble protein aggregates. Anal Chem 84:1718–1724
Washabaugh MW, Collins KD (1986) The systematic characterization by aqueous column

chromatography of solutes which affect protein stability. J Biol Chem 261:12477–12485
Wen J, Arakawa T, Philo JS (1996) Size-exclusion chromatography with on-line light-scattering,
absorbance, and refractive index detectors for studying proteins and their interactions. Anal
Biochem 240:155–166
Wenzel RJ, Matter U, Schutheis L, Zenobi R (2005) Analysis of megadalton ions using cryodetec-
tion MALDI time-of-flight mass spectrometry. Anal Chem 77:4329–4337
Wu G, Hummel D, Herman A (1997) Analysis of the solution behavior of protein pharmaceuticals
by laser light scattering photometry. ACS Symp Ser 675:168–185
Wuchner K, Buchler J, Spycher R, Dalmonte P, Volkin DB (2010) Development of a microflow
digital imaging assay to characterize protein particulates during storage of a high concentration
IgG1 monoclonal antibody formulation. J Pharm Sci 99:3343–3361
Yohannes G, Jussila M, Hartonen K, Riekkola ML (2011) Asymmetrical flow field-flow fraction-
ation technique for separation and characterization of biopolymers and bioparticles.
J Chromatogr A 1218:4104–4116
Chapter 5
Qualification of Biophysical Methods
for the Analysis of Protein Therapeutics
Yijia Jiang, Cynthia Li, and John Gabrielson
5.1 Introduction
As described elsewhere (Jiang and Narhi 2006) and in other chapters of this book,
biophysical methods are routinely used by the biopharmaceutical industry to gain
information on protein higher-order structure and stability throughout the therapeu-
tic protein development life cycle. These methods include SV-AUC and SE-HPLC-LS
for size distribution analysis and molecular weight determination of species sepa-
rated by SE-HPLC; CD, FTIR, and DSC for secondary and tertiary structure and
overall conformation and thermal stability analysis; and light obscuration (LO)- and
micro-flow imaging (MFI)-based methods for subvisible particle analysis. The
results of these biophysical analyses have been included in regulatory filings as
required by ICH guidelines M4Q (R1), Q5E, and Q6B (ICH M4Q 2001; ICH Q5E
2005; ICH Q6B 1999). These results are included in the filing documents to demon-
strate that the therapeutic protein shows the predicted and expected properly folded
higher-order structure and retains the higher-order structure after manufacturing
process, site, and/or formulation changes. ICH guidelines M4Q (R1) (ICH M4Q
2001) on common technical document (CTD) for the registration of pharmaceuti-
cals for human use and Q6B (ICH Q5E 2005) on specifications state that for desired
Y. Jiang (*)
Process and Product Development, Amgen Inc., One Amgen Center Drive, 30E-1-C,
C. Li
Process and Product Development, Amgen Inc., One Amgen Center Drive,
J. Gabrielson
Analytical Sciences, Amgen Inc., 4000 Nelson Road, Longmont CO 80503, USA

100 Y. Jiang et al.
product and product-related substances, details should be provided on primary,

secondary, and higher-order structure when relevant. ICH guideline Q5E (ICH Q6B
1999) specifies: “following a manufacturing process change, manufacturers should
attempt to determine that higher order structure (secondary, tertiary, and quaternary
structure) is maintained in the product.”
Unlike the methods used for product release testing, for which the qualification
and validation are regularly performed as governed by ICH guideline Q2 (R1) (ICH
Q2 (R1) 2005) to show precision, accuracy, linearity, LOD, and LOQ of the meth-
ods, little has been done and published to standardize an approach to qualify char-
acterization methods in general, and the biophysical in particular, until recently. The
lack of qualification of biophysical methods used is mainly due to the qualitative
nature of some of these methods. In recent years, in response to increasing regula-
tory requirements and scrutiny, more effort has been invested in understanding the
precision and sensitivity of the biophysical techniques, which has resulted in several
publications and presentations (Jiang et al. 2011; Li et al. 2011; Wen et al. 2011;
Gabrielson et al. 2010). In this chapter we review the qualification of the biophysi-
cal methods including AUC, CD, FTIR, DSC, SE-HPLC-LS, MFI, and LO-based
methods and demonstrate how qualification of these methods enables better knowl-
edge of the methods and objective interpretation of the results. The general consid-
erations described here can be applied to other biophysical methods as appropriate
as well.
5.2 Qualification of CD
Circular dichroism (CD) spectroscopy is a technique that has been widely used in
structural biology for examining secondary structure, tertiary structure, conforma-
tional changes, and folding and binding interactions involving protein molecules. It
is also used by the biopharmaceutical industry as a characterization tool to study the
effect of manufacturing processes, formulation compositions and conditions, stor-
age conditions, and delivery systems on protein conformation (Johnson 1988,
1999a; Woody 1994, 1996; Manning and Woody 1989; Greenfield and Fasman
1969; Yang et al. 1986). This information is often included in regulatory filings fol-
lowing changes in any of these manufacturing conditions (ICH Q5E 2005; ICH
Q6B 1999). In recent years, a new extension of the conventional CD method, syn-
chrotron radiation circular dichroism (SRCD), has been shown to have a better sen-
sitivity for the higher-order structure of a protein by producing data to lower
wavelengths with higher information content and improved signal-to-noise levels
relative to those obtained using conventional benchtop instruments (Wallace 2005;
Lees and Wallace 2002; Miles et al. 2007; Wallace and Janes 2010). SRCD
spectroscopy can provide important static and dynamic structural information on
proteins in solution, including secondary structures of intact proteins and their
domains, protein stability, the differences between wild-type and mutant proteins, the
identification of natively disordered regions in proteins, and the dynamic p rocesses of
5 Qualification of Biophysical Methods for the Analysis of Protein Therapeutics 101
protein folding and membrane insertion and the kinetics of enzyme reactions (Miles
and Wallace 2006; Meersman et al. 2010; Drechsler et al. 2009). A new web-based
bioinformatics resource, the Protein Circular Dichroism Data Bank (PCDDB), has
also been created recently by Dr. Wallace’s laboratory which enables archiving,
access, and analyses of CD and SRCD spectra and supporting data (Powl et al.
2010; Whitmore and Wallace 2004; Wallace et al. 2006; Lees et al. 2006).
Unfortunately many of the formulation buffers used for therapeutic proteins interfere
with the SRCD, especially at the lower wavelengths required, and this technique can
usually not be applied to comparability evaluations or similar studies.
Even though CD spectroscopy has received considerable attention and been
widely used for biological molecules over the last 20 years by both industry and
academic scientists, the technique is mostly used in research and development labo-
ratories rather than quality control laboratories. This is mainly due to the complex-
ity of the analysis, the need for an experienced spectroscopist to run the studies, and
the qualitative nature of the spectra. The lack of defined best practices in making the
CD measurement and the variability in the data obtained from different laboratories
also make this an inappropriate technique for routine use in quality control labora-
tories. There have been a few publications describing standardization of best prac-
tices in making CD measurements (Van Stokkum et al. 1990; Kelly et al. 2005).
To understand the extent of comparability of CD data, the National Physical
Laboratory (NPL) has conducted an interlaboratory CD spectroscopy study in
recent years. Schiffmann et al. published a report titled “CD spectroscopy: an inter-
laboratory study” (Jones et al. 2004) and Ravi et al. (2010) published two papers
about the international comparability in spectroscopic measurements of protein
structure by CD (Schiffmann et al. 2004; Ravi et al. 2010). Both studies involved
academic and industry laboratories worldwide. The studies have highlighted some
of the current difficulties in obtaining comparable CD spectra from different instru-
ments and possible solutions to some of these problems, for instance, the iterative
spectral alignment method (ISAM). Both studies show that the uncertainties
involved in a CD measurement are rather complex, involving components relating
to calibration, path length, sample concentration, and instrumentation (Ravi et al.
2010; Whitmore and Wallace 2008; Miles et al. 2003, 2004, 2005). Therefore, a
careful calibration of the CD instrument and exact path lengths of the cuvettes
must be carried out to ensure equivalence of CD measurements when using dif-
ferent instruments and cuvettes. The calibration and standardization of CD
instruments has been described by Dr. Wallace et al. (Whitmore and Wallace 2008;
Miles et al. 2003). The authors suggested the procedure for proper calibration of a
CD spectrophotometer and proposed a method for standardization of protein spectra
obtained on any instrument. For example, calibrations should be done at more than
one wavelength in order to cover the wavelength range measured in a protein spec-
trum, and the linearity of the response should be demonstrated; the ratios of the
measured and expected ellipticity values at each wavelength should be used to
develop a model function for corrections.
For application of the method in the biopharmaceutical industry, one of the
important questions is the degree of comparability of the CD data between different
102 Y. Jiang et al.
lots of protein generated from different processes or formulation conditions. In Ravi

and Schiffmann’s studies, they used multivariate statistical methods, such as princi-
pal component analysis (PCA) and soft independent modeling of class analogies
(SIMCA), to investigate the comparability of CD spectroscopic data (Jones et al.
2004; Schiffmann et al. 2004; Ravi et al. 2010). These methods can also be used in
an industrial environment to assess the comparability of different lots of products.
In the last few years, more numerical methods to compare the entire spectra have
been developed and used by researchers in the biopharmaceutical industry, for
example, using 3× standard deviation of reference spectra (3× SD) to assess spectral
similarity (Li and Coworkers, personal observation, unpublished data). The method
uses historical data to assess the precision of the instrument. These methods are
suitable for analytical product comparability studies, and the data obtained are com-
pared to a reference spectrum. However, none of these methods have been used to
systematically assess the sensitivity of the CD method.
Another method commonly used by researchers in the field is the deconvolution
of the far-UV CD spectrum of a protein to contributions of individual secondary
structural elements (alpha-helix, beta-sheet, beta-turn, and unordered conformation,
as well as the contribution in the 195–240 nm region from aromatic amino acids)
followed by determination of the percentage of each component in the protein sec-
ondary structure. Comparability is then determined by comparing changes in the
percentage of each component (Greenfield 1996; Johnson 1999b; King and Johnson
1999; Perczel et al. 1991; Perczel and Fasman 1993; Venyaminov et al. 1991).
However, the method is not optimized for assessing overall similarity of protein
structure between samples from different processes or formulations. Also, the quan-
tification of the secondary structure composition can vary depending on the meth-
ods and parameters used and involves considerable mathematical manipulations
including spectra manipulation, curve fitting, area integration, and normalization.
There is no analogous deconvolution of tertiary structure components that give rise
to spectra in the near-UV CD region due to the lack of defined and pure side-chain
conformational CD features in these spectra. Therefore, deconvolution of CD spec-
tra is not routinely carried out for protein comparability assessments in the biophar-
maceutical industry.
There are very limited published data on the systematic qualification of CD
methods to evaluate the precision and sensitivity of the CD measurement. In a recent
publication by Li and colleagues, the authors used Thermo Electron OMNIC soft-
ware QC compare function (Cover and Hart 1967; TQ Analyst Algorithm 2007–2010)
to compare the entire CD spectra in both the near- and far-UV CD regions to sys-
tematically qualify a CD method based on its precision and sensitivity (Li et al.
2011).
The Thermo Electron OMNIC software QC compare function was used by Li
et al. to quantitatively assess the spectral similarity of different CD spectra, in order
to determine the precision and sensitivity of near- and far-UV CD spectroscopy
(Fig. 5.1). Proteins from different structural families in their stable storage buffer or
under stress conditions were analyzed. The results from this precision study showed
that (1) for the near-UV CD, the analysis has good precision in terms of spectral
Fig. 5.1 (a) Far-UV CD a
spectra of protein 1–4 6000
(spectral similarity compared
to that of protein 1 is 100% 4000
protein 1; 5.0% protein 2;
44.9% protein 3; 28.7% 2000
CD Ellipticity
protein 4). (b) Near-UV CD
spectra of protein 1–4 0
(spectral similarity compared
to that of protein 1 is 100% -2000 Protein-1
Protein-2
protein 1; 41.0% protein 2;
Protein-3
3.1% protein 3; 12.9% -4000
Protein-4
protein 4)
-6000
200 210 220 230 240
Wavelength (nm)
Tyr
b Phe Trp
80 S-S
40
CD Ellipticity
Protein-1
-40
Protein-2
Protein-3
-80 Protein-4
-120
250 300
Wavelength (nm)
similarity when comparing the spectra obtained with the same instrument and
cuvette independent of the analyst, location, and date; and (2) for far-UV CD, the
precision varies depending on the structural nature of the protein and the buffer
composition. For proteins with helical secondary structure in buffers with little
interference in the spectra, far-UV CD analysis shows low variability in spectral
similarity. For proteins with predominantly beta-sheet secondary structure and in
buffers where the excipients have absorbance or CD signals in this region, far-UV
CD analysis shows higher variability in spectral similarity. The sensitivity assess-
ment of CD spectroscopy is challenging because the unfolding of a protein by some
downstream processing conditions, for example, low pH or Gdn, is often reversible
or partially reversible when the stresses are removed. Thus a standard curve that
correlates spectral similarity to the percentage of denatured form of the protein can-
not be generated empirically. To address this issue, Li et al. used a computer-
simulated blending study of the spectra of the native protein and the unfolded form
104 Y. Jiang et al.
of the same protein to generate a standard curve that correlates the spectral s imilarity
of the measured spectrum to the native spectrum with the amount of unfolded form
of the protein present. To determine the sensitivity of CD spectroscopy, one would
have to combine the precision assessment of the technique under the required test
conditions (e.g., protein concentration and buffer) with the spectral similarity cor-
relation of the simulated blending study of denatured and native proteins. The
sensitivity of the method can be expressed as the percentage of denatured form of
the protein when the spectral similarity of the sample spectrum compared to the
native/control protein exceeds the precision of the method. Under different condi-
tions and with a different protein, the precision and sensitivity of the CD analysis
may also change.
The results from Li et al.’s qualification study indicate that CD spectroscopy can
be qualified for characterizing protein secondary and tertiary structures and for
comparability studies of biopharmaceuticals. With proper calibration of the instru-
ments and cuvettes, and standardized good operating practices, CD spectroscopy is
quite reproducible and able to detect changes induced in the secondary and tertiary
structure of a protein by lower pH and other possible denaturing reagents that are
present during therapeutic protein manufacturing processes. The sensitivity of the
technique to detect the changes in conformation that might be induced by the manu-
facturing process is dependent on the experimental variability and the nature and
extent of the structural changes in the particular protein being analyzed.
The results shown above provide the ground work on how the CD method should
be calibrated and qualified and demonstrate that the method is sufficiently precise to
monitor protein structure changes beyond the variability of the method.
5.3 Qualification of FTIR
Fourier transform infrared (FTIR) spectroscopy can be used to obtain information

about the higher-order structure of a protein both in solution and in the solid state
(Carpenter et al. 1998; Fu et al. 1999; Haris and Chapman 1995; Kong and Yu
2007). It is routinely employed as a characterization method in the biopharmaceuti-
cal industry to determine the higher-order structure of protein therapeutics (Gross
and Zeppezauer 2010). It is often included as part of the biophysical characteriza-
tion methods to assess the secondary structure of protein therapeutics before and
after changes to process, formulation, manufacturing, and storage conditions, and
the data are included in regulatory filings (Jiang and Narhi 2006; Jiang et al. 2008).
However, the qualification of the FTIR method for protein secondary structure anal-
ysis has mostly remained elusive, mainly due to the lack of a consistent method
applied to quantitatively determine the similarity of the FTIR spectra and hence the
comparability of the protein samples analyzed. Visual comparisons of a sample
spectrum to a reference have typically been used as a way to verify proper folding
of a protein sample. To objectively qualify the FTIR method, a mathematical algo-
rithm is needed to compare the FTIR spectra quantitatively and demonstrate the
precision and sensitivity of the method for protein secondary structure analysis.
Work has been done to quantify the percentages of different secondary structural
components for proteins in solution by FTIR (Vonhoffa et al. 2010; Susi and Byler
1987). There are two primary ways of doing this. The first uses curve fitting of the
second-derivative or self-deconvoluted spectra to obtain the relative amounts of dif-
ferent types of secondary structure based on the band areas (Susi and Byler 1986;
Byler and Susi 1986). The second involves peak fitting of the non-deconvolved and
baseline-corrected amide I bands and then obtaining the percentage of secondary
structures by correlating with the shape and intensity using an interval partial least
squares algorithm (Vedantham et al. 2000; Surewicz and Mantsch 1988). These
methods are helpful in determining the major secondary structure components and
estimating their percentages in a protein. However, these methods are not optimized
for assessing overall similarity of protein structure between samples from different
processes or formulations. The quantification of the secondary structure composi-
tion can vary depending on the methods and parameters used, and it involves con-
siderable mathematical manipulations including spectra manipulation, curve fitting,
area integration, and normalization. Therefore, they are not routinely carried out for
protein secondary structure assessment and qualification of FTIR by analysts in the
biopharmaceutical industry.
One other mathematical approach explored by Prestrelski et al. (1993) is to com-
pare the FTIR spectra quantitatively by using the correlation coefficient function.
Occasionally, the correlation coefficient value of the uncorrected spectra did not
agree with a visual assessment of the spectral similarity due to an offset in baselines
which led to an artificially low value. Conversely, if the spectra were baseline cor-
rected and peak positions were similar, but there were differences in relative peak
heights, the correlation value would be unreasonably high. To avoid this inconsis-
tency, Kendrick has developed a method to quantify the overlap of second-derivative
FTIR spectra to determine structural similarity between proteins (Kendrick et al.
1996). In this approach area-normalized second-derivative spectra were used and
compared. The authors found that quantifying the area of overlap between area-
normalized spectra provides a reliable, objective method to compare overall spectral
similarity and avoids the problems associated with calculation of the correlation
coefficient. However, due to normalization of the spectra, the area of overlap
approach cannot distinguish between true differences in the magnitude of FTIR
signals. Recently, D’antonio et al. 2012 showed a modified area of overlap method
that improved the differentiation power for quantitative comparison of FTIR spec-
tra. The authors also compared four different algorithms including the correlation
coefficient and area of overlap and summarized their strength and weakness for
studying comparability of protein therapeutics.
A new approach using Thermo Electron OMNIC software QC compare function
(Cover and Hart 1967; TQ Analyst Algorithm 2007–2010) has been applied to
directly compare FTIR spectra with minimal mathematical manipulation (Fig. 5.2),
which led to the identification of important performance characteristics for the qual-
ification of the FTIR method (Jiang et al. 2011). The QC compare function was
originally intended and used as a library searching tool for small-molecule identifi-
cation by FTIR (TQ Analyst Algorithm 2007–2010). It correlates the spectral
106 Y. Jiang et al.
2.0
1.0
0.0
Arbitrary unit
-1.0
-2.0
-3.0
-4.0
10-4
1720 1700 1680 1660 1640 1620 1600
Wavenumbers (cm-1)
Fig. 5.2 Second-derivative FTIR spectra of protein 1–5 (spectral similarity compared to that of
protein 1 is 100% protein 1 (pink); 46.3% protein 2 (purple); 36.3% protein 3 (green); 9.7% pro-
tein 4 (blue); 12.1% protein 5 (black))
information in the specified region of the sample and the reference spectra to
determine the similarity between the two. The results of the method are reported as
a value between 0% and 100%, which indicates how well the sample spectrum
matches the reference spectrum. A spectral similarity value of 100% indicates the
two spectra are identical.
The selection of an appropriate approach to numerically compare the spectra
should depend on the capability of the method, the ease of access and use of the
method, and the purpose of the study.
With the OMNIC QC compare algorithm, Jiang and colleagues evaluated the pre-
cision and sensitivity of FTIR for the analysis of protein secondary structure through
a multisite/instrument and multi-analyst study in an effort to qualify the method
(Jiang et al. 2011). Proteins containing different types of secondary structures such
as alpha-helical and beta-sheet were included to ensure that the precision assessment
would apply to secondary structure analysis of all proteins regardless of the specific
structural type. In addition inter-day repeatability and the effect of protein concentra-
tion differences on the precision of the method were also evaluated. The overall FTIR
spectral similarities of these analyses on the same proteins were compared. The sen-
sitivity of the method was evaluated by comparing the reference spectrum of a pro-
tein to that of the partially or fully unfolded protein after exposure to the denaturing
condition and also by blending studies where the spectrum of a reference was mixed
with that of a denatured protein. Standard curves were generated where the spectral
similarity was plotted as a function of the percentage of the unfolded protein.
Results by Jiang et al. (2011) demonstrate that the FTIR method for the analysis
and characterization of protein secondary structure is precise with the standard
deviations of the characteristic FTIR band frequencies <1.1 cm−1 and the spectral
similarity >93% for replicate measurements. The FTIR spectra remain the same
whether collected on the same day or during a three-day period and when the pro-
tein concentration deviates from the target concentration (≥30 mg/mL) by ≤10%
regardless of the maker of the instrument used or the structural family of the protein.
The authors also demonstrate that the method is sensitive and appropriate for assess-
ing protein secondary structure and changes in conformation resulting from stresses
that can be encountered during common biopharmaceutical manufacturing pro-
cesses. The sensitivity of the FTIR method is dependent on the extent of the struc-
tural changes induced and the magnitude of the resulting changes in the spectra.
The precision of the FTIR method is closely related to the signal-to-noise ratio
of the instrument, the sample concentration, the buffer components, and the spectral
region of interest. The sensitivity of the method to detect structural change depends
on the nature of the sample and relative spectral changes corresponding to the struc-
ture changes. Therefore, protein-specific qualification/verification of the method
may need to be performed to ensure the understanding of precision and sensitivity
of the method.
5.4 Qualification of DSC
DSC has been used widely to study thermal dynamics of protein folding. The melt-
ing/thermal transition temperature(s) of proteins can be used to understand the energy
barrier for protein unfolding and thermal stability of proteins (Pyrpassopoulos et al.
2006; Johnson et al. 1995). DSC has also been used to study the binding interactions
of proteins with other molecules and the effect of mutations and the presence of
carbohydrates on protein thermal stability (Celej et al. 2006; Bruylants et al. 2005;
Johnston et al. 2011; Protasevich et al. 2010; Wen et al. 2008). The application of
DSC has been reviewed extensively in the area of characterization of macromole-
cules and their interactions during pharmaceutical product development (Chiu and
Prenner 2011; Bruylants et al. 2005). However, even though the method has been
used extensively to monitor and show changes in protein thermodynamic parame-
ters such as enthalpy and melting temperature (Tm), there are very limited published
data on the systematic qualification of the method to show that the DSC measure-
ments are precise, accurate, and sensitive for the analysis of protein conformation
and thermal stability. A protocol on measuring protein thermostability by DSC
was published by Makhatadze (Coligan et al. 2001). It provides a detailed proce-
dure for conducting DSC analysis including sample preparation and interpretation
of the results, as well as calibration of the instrument and maintenance of DSC
cells. Calibration of DSC was described by Gmelin and Sarge (Gmelin and Sarge
1995). They provided recommendations for instrument-independent calibration of
temperature, heat, heat flow rate of the scanning calorimeter, operation procedures,
108 Y. Jiang et al.
calibration substance, and data treatment algorithms which can be used for all
instruments.
In the early days of DSC analysis, various data analytical algorithms were devel-
oped and used to assess the effect of instrument configuration and scan rate on DSC
measurement results (Lopez and Freire 1987). During the use of DSC for purity
analysis of drug product, instrumental differences, sample size, heating rate, and
details of the calculation of such analysis were examined (Yoshii 1997). It was found
that changing the sample size or heating rate resulted in a difference in the effect of
purity value between two instruments. Collectively controlling the heating rate dur-
ing DSC measurement appears to be critical in obtaining reproducible results.
The precision of DSC measurements across different instruments and labs using
the same sample (polymeric material) and parameters has been assessed. The study
was organized by Empa (Swiss Federal Institute for materials testing and research)
(Schmid 2012; Affolter et al. 2001). It collected measured data on glass transition
(Tg) and melting point (Tm) and evaluated those using statistical methods to assess
the precision of DSC. The results show that variability of DSC measurements is
mainly caused by differences in the analyst, instrument, and calibration of the
instrument. For one-point temperature measurements such as Tg and Tm, good agree-
ments were observed with a standard deviation of repeatability at 0.3–1.0°C and
that of reproducibility at 1.0–2.1°C. The standard deviation increases significantly
with user-defined data evaluation.
A recent paper by Wen et al. (2011) explored the qualification of DSC for its
applications in thermal stability analysis of proteins. The authors assessed the preci-
sion and sensitivity of the DSC method through a multisite, instrument, and analyst
study using several proteins from different structural families including monoclonal
antibodies and cytokines and parameters including Tm and profile similarity. The
same experimental parameters such as heating/scan rate and similar protein concen-
trations were used for measurement on all instruments with the exception of one.
The results show that the Tm values obtained for the same protein by the same or
different instruments and/or analysts are quite reproducible, varying generally
<1.1°C with the same instrument and <1.5°C across different instruments.
The profile similarity values obtained using the OMNIC QC compare function
(discussed above in the CD and FTIR sections) for the same protein from the same
instrument are also high (> ~95%). The variability of inter-day DSC measurements
is very similar to that of the intra-day measurements (only slightly higher by
~0.1°C). The sensitivity of the DSC method for assessing protein thermal stability
and conformational changes was evaluated by several experiments including ana-
lyzing samples perturbed by pH 3 or 6 M guanidine HCl (Gdn). The results show
that DSC is able to detect changes in protein conformation caused by low pH and
denaturant to levels as low as 10% denatured protein present in the sample. The
authors then applied DSC to the analysis of pH stability and buffer screening of a
protein and candidate screening. They demonstrated that DSC is an appropriate
method for assessing protein thermal stability and conformational changes that may
result from manufacturing processes, formulation, and storage conditions and can
also be used to compare relative stability of candidate molecules.
Validation of DSC for pharmaceutical analysis under cGMP was described by

Weissburg et al. (2002). The validation included development of a validation plan
(i.e., scope of validation, criterion for completion of validation), design documents,
user requirements, calibration and maintenance procedures, IQ/OQ, standard opera-
tion procedures, and test scripts. The authors provided details about validation of
the computer-controlled system to be 21CFR Part 11 compliant and of the thermal
accuracy and precision using a NIST-certified indium standard at 0.5 and 50°C per
minute scan rate employing the DSC testing script. Validation of DSC was achieved
when the test results obtained met the specifications. The validation showed that the
DSC instrument and the computer program and system used met 21 CFR Part 11,
cGMP, and the user predefined requirements.
The cases shown above demonstrate that DSC can be qualified and validated for
the purpose of assessing protein higher-order structure changes for regulatory fil-
ings. The qualification and/or validation should focus on demonstrating the preci-
sion of the method and that the method is suitable for its applications. When the
instrument is properly calibrated and maintained, DSC is accurate, precise, and suf-
ficiently sensitive to detect changes in proteins resulting from process and formula-
tion changes and point mutations of the protein primary structure.
5.5 Qualification of SV-AUC
Sedimentation velocity analytical ultracentrifugation (SV-AUC) is used as a charac-

terization tool in biotherapeutic development primarily to assess product purity; that
is, it measures the relative abundance of size variants in a protein solution. Although
nonroutine, it is commonly applied in product comparability studies, protein refer-
ence standard qualifications, product quality investigations, and other product char-
acterization activities (Berkowitz 2006; Shire 1994).
The SV-AUC analysis is an orthogonal method to size-exclusion (SE) HPLC,
and it is in this context that it is perhaps most useful, because it does not suffer from
many of the potential limitations of SE-HPLC: physical disruption of the sample,
extensive dilution into what is often a different solution environment, and lack of
separation of product size variants close to the column exclusion limit (Gabrielson
et al. 2007a; Philo 2006; Carpenter et al. 2010). Because it employs a fundamentally
different mode of separation, SV-AUC can be used to verify the accuracy of
SE-HPLC methods when they are developed and remediated and to confirm the
suitability of SE-HPLC for routine use.
In addition to measuring the relative concentration of size variants, SV-AUC can
be used to determine their mass and shape. The sedimentation boundary profiles
provide information about the sedimentation and diffusion rates of size variants
present in the sample (Dam et al. 2004; Laue et al. 1992; Lebowitz et al. 2002;
Schuck et al. 2002). The sedimentation coefficient is directly measured from the
time-dependent displacement of the boundaries. For a size variant of known mass,
consistency of the measured sedimentation coefficient provides confirmation that
110 Y. Jiang et al.
the overall shape and charge of the size variant has been maintained when multiple
samples are measured under the same solution conditions. However, the accuracy
and precision of the measured sedimentation coefficient decreases as the concentra-
tion of the size variant decreases (Gabrielson et al. 2007b). In very pure samples
with low concentrations of fragmented and aggregated size variants (<1% of the
total protein mass), only the sedimentation coefficient of the most abundant species
(typically monomer) can be reliably determined.
In the biopharmaceutical industry, SV-AUC is primarily used as an orthogonal
method to SE-HPLC to quantitatively measure protein purity and secondarily used
to measure protein monomer sedimentation coefficients (Gabrielson et al. 2010).
Qualification of an SV-AUC method should reflect the fact that it will be used to
quantitatively measure these two attributes. Furthermore, two performance charac-
teristics of an SV-AUC method should be considered during qualification: (1) evalu-
ation of the method’s precision and (2) confirmation that the method is sufficiently
sensitive to provide quantitative results, thereby allowing meaningful conclusions to
be drawn.
Assessing the precision of SV-AUC, and determining which type(s) of precision
to evaluate, should be governed by how the method will be used. For example, if
SV-AUC analysis can typically be completed by one analyst using one instrument
for product characterization studies required during biotherapeutic development,
then it may not be necessary to study the reproducibility of the method. However, if
the number of samples requires multiple runs on different days, then understanding
the intermediate precision of the method is critical.
Experimental factors that impact the precision of SV-AUC measurements have
been reviewed elsewhere (Gabrielson et al. 2007b, 2010; Schuck et al. 2002;
Gabrielson and Arthur 2011; Arthur et al. 2009; Pekar and Sukumar 2007) and
include the instrument, rotor, centerpiece and other cell components, overall align-
ment of channels, temperature equilibration, analysis software and settings, and
sample characteristics. Even under the tightest experimental control currently
achievable, normal variation of these factors leads to variability in results of approx-
imately 0.4% for size variant quantification (Gabrielson and Arthur 2011; Arthur
et al. 2009) and about 0.02 S for sedimentation coefficient measurements (Gabrielson
and coworkers, personal observation, unpublished data). Much of the imprecision
arises from the specific centerpieces used (Arthur et al. 2009; Pekar and Sukumar
2007) and from the alignment of cells (Pekar and Sukumar 2007), factors that can
vary as much within a run as they do across runs or even across laboratories.
A robust qualification design for an SV-AUC method should therefore account for
normal variations from all relevant experimental factors, that is, those factors that
reflect how the method will be applied during development of a biotherapeutic.
Studies to evaluate SV-AUC method performance indicate that performance
parameters including linear range, limit of detection (LOD), limit of quantitation
(LOQ), and measurement precision are, in general, independent of the specific
protein product analyzed (Gabrielson et al. 2007b; Gabrielson and Arthur 2011;
Arthur et al. 2009). Variability in measured results is primarily due to experimental
factors, such as centerpiece differences, and depends to a lesser extent on sample
characteristics. Therefore, after sufficient data have been acquired to account for
relevant experimental factors, it is possible to qualify SV-AUC analysis of a new
protein product by simply verifying that the precision of sedimentation coefficient
measurements and the precision of size variant measurements are within a range
expected from previous experience with the method. Additional qualification exper-
iments may be warranted if the expected method performance is not achieved for a
particular product or sample type, for example, protein samples with reversible self-
association. In those cases, protein-specific intermediate precision should be more
rigorously defined.
The ability of SV-AUC to detect changes to the product, both to its size distribu-
tion profile and to the sedimentation coefficient of the monomer, is essential for
successful application of the method in biotherapeutic development. Therefore,
method qualification should include experiments demonstrating the suitability of
SV-AUC for detecting such changes. This may be accomplished by manipulating a
sample to perturb the protein structure and to produce higher levels of size variants.
A suitable amount of degradation can usually be obtained from thermal, UV, or pH
stresses. For example, subjecting a monoclonal antibody sample to low pH condi-
tions will often cause the sedimentation coefficient of the monomer to shift and
induce formation of high molecular weight size variants (aggregates), which can be
used to confirm the sensitivity of the method.
The sensitivity of the method can best be determined when changes to the mea-
sured output are correlated with changes deliberately made to the product. Changes
to the sedimentation coefficient of a protein and the amount of aggregation mea-
sured by SV-AUC for a different protein, both induced by exposure to low pH, are
shown in Fig. 5.3.
The data presented in Fig. 5.3 are instructive for several reasons. First, the
response of the method to a given stress condition may or may not be linear. In panel
A, the sedimentation coefficient responds nearly linearly to pH changes from 5.5 to
3.5 and then declines nonlinearly at more acidic pH, whereas in panel B the maxi-
mum aggregation levels are measured in a narrow pH range near 3.5; aggregation
levels decline markedly at higher and lower pH. Second, it is often convenient to
express the method sensitivity by normalizing the response, either by dividing by
the range of the independent variable (i.e., by calculating the slope if the response is
linear) or by dividing by an estimate of the variability of the method to express the
change in response as a number of standard deviations. For example, in panel A of
Fig. 5.3, a pH change from 5.5 to 3.5 results in a 0.2 Svedberg shift in the sedimen-
tation coefficient. The sensitivity of the method to pH could then be expressed as 0.1
Svedberg per pH unit (or five standard deviations per pH unit when normalized by
the method variability of 0.02 S). Finally, method sensitivity studies are most useful
when they evaluate a range wider than what the product will likely encounter during
manufacture, shipping, and long-term storage. This allows interpolation of the
response, rather than extrapolation to conditions which have not been studied.
Any number of stress conditions can be used to demonstrate the sensitivity of
SV-AUC. These may include thermal degradation, light/UV exposure, pH modifica-
tion, chemical denaturation, or primary structural modifications (e.g., oxidation).
112 Y. Jiang et al.
Fig. 5.3 Sedimentation
coefficients (a) and
aggregation levels (b)
measured by SV-AUC for
protein product samples
across the pH range indicated
Many considerations should inform the selection of product stress conditions with
which to determine the sensitivity of the method, including the likelihood that the
product will encounter a certain stress during manufacturing and the intended
change to the product. That is, a general stress like elevated temperature may be
desired to produce aggregation, whereas a more targeted stress like oxidation may
be useful to correlate the sedimentation rate of the protein with changes to its
primary structure. And finally, the intended purpose of the method should also be
considered when selecting product stress conditions. A different stress condition
may be desirable if the method will be used exclusively to characterize a drug sub-
stance manufacturing process compared to if the method will be used to evaluate
drug product comparability.
5.6 Qualification of SE-HPLC-LS
During biotherapeutic development, it is often valuable to measure the molecular

weights of size variants present in product samples. Molecular weight determina-
tion with sufficient accuracy facilitates identification of these size variants (such as
dimer and larger aggregate species), which is required for comprehensive product
characterization.
Addition of light scattering (LS) detection to a size separation method like
SE-HPLC makes it possible to determine the molecular weights of size variants
after elution from the column (Kendrick et al. 2001; Wen et al. 1996). Under some
conditions, the LS detection is also more sensitive for large size variants than con-
ventional detection (such as UV, RI) alone. Both of these LS capabilities arise from
the fact that the intensity of the light scattered by an eluting species is proportional
not only to its concentration but also to its molecular weight (Wen et al. 1996; Wyatt
1993). However, because the LS detector response is a function of two variables,
molecular weight and concentration, neither can be calculated directly from the LS
signal alone. The molecular weight of each eluting species can only be determined
by using a separate measurement of protein concentration (i.e., a simultaneously
collected UV or RI chromatogram) to deconvolute the molecular weight and con-
centration information for each peak.
LS analysis is commonly applied as a characterization tool in product compara-
bility studies, characterization studies to support product licensure, and product
quality investigations. For molecular weight measurements by LS, the accuracy of
the measured result is the key method performance parameter that should be evalu-
ated during method qualification. The precision of the method may be determined
in a straightforward manner and is useful to include in the design of qualification
experiments. Accounting for different columns and mobile phases in these experi-
ments is critical if characterization studies might be performed over multiple days
and analysis sequences. Nonetheless, it is the accuracy of the method, that is, how
much the measured mass differs from the true mass, which dictates the usefulness
of the method for product characterization. Therefore, the method should be quali-
fied to demonstrate that it is sufficiently accurate for the intended purpose of deter-
mining the molecular weights of size variants separated by SE-HPLC (or another
flow separation method such as field flow fractionation).
Comparison of the measured mass of a protein standard to its known mass is an
important way to assess the accuracy of SE-HPLC-LS. This confirms that the sys-
tem is capable of providing results with sufficient accuracy. Nevertheless, the
accuracy of results for unknown species cannot be inferred from the accuracy of a
protein standard because the measurement accuracy depends on factors which may
be different for each protein and size variant, such as the extinction coefficient and
concentration. Therefore, propagation of error analysis provides a convenient
approach by which to assess the accuracy of the measured result for any protein
size variant.
First, the signal and noise associated with each detector (LS, UV, and/or RI) can
be determined directly from a chromatogram. If a multiangle LS detector is used,
114 Y. Jiang et al.
signal and noise estimates should be determined from the 90° detector. (For isotropic
scatters, such as proteins, the 90° detector has the highest signal-to-noise ratio.)
Each detector’s signal is taken to be the maximum signal (in volts) of the peak of
interest. The noise for each detector is estimated from a region without protein sig-
nal (i.e., a baseline region) using any one of several noise calculation approaches
such as the square root of the mean squared error.
The governing equations that relate the signals of each detector to the physical
attributes of the protein are provided in Eqs. 5.1, 5.2, and 5.3 (Wen et al. 1996):
2
 dn 
SLS = kLS Mc   , (5.1)
 dc 

SUV = kUV elc, (5.2)

 dn 
SRI = kRI   c, (5.3)
 dc 
where kLS (V mol L mL−2), kUV (V), and kRI (V g mg−1) are detector proportionality
constants and l (cm) is the path length of the UV cell, also a constant. The concen-
tration, c (mg mL−1), of the size variant can be determined from either Eq. 5.2 or
Eq. 5.3, depending on the choice of detector, when the extinction coefficient ε
(mL mg−1 cm−1) and the refractive index increment dn/dc (mL g−1) are provided. The
molecular weight of the size variant, M (g mol−1), is then determined from Eq. 5.1.
The error of the molecular weight measurement can be estimated by applying a
propagation of error analysis to Eq. 5.1. The generalized form of the error propaga-
tion formula (Taylor 1997) is shown in Eq. 5.4 for a response, f, that is a function of
independent variables, x, y, and z:
2 2 2
 ∂f   ∂f   ∂f 
e f =  ex  +  ey  +  ez  ,
 ∂x   ∂y   ∂z  (5.4)

where ef is the expected error in the calculated value of f and ex, ey, and ez are error
estimates for variables x, y, and z, respectively.
5.6.1 Case 1: LS-UV Detectors
When LS and UV detectors are connected in series, the size variant concentration is
calculated from Eq. 5.2, which is then used to calculate the size variant molecular
weight from Eq. 5.1. It is assumed that some degree of uncertainty exists in the
protein extinction coefficient and refractive index increment, both of which are pro-
vided by the analyst. Propagation of error using Eq. 5.4 leads to the following
expression for the expected error in the molecular weight calculation (Eq. 5.5):
2 2 2 2
eM e  e  e   e 
=  SLS  +  SUV  +  e  + 4  dn / dc  . (5.5)
M  SLS   SUV   e   dn / dc 

As expected, the molecular weight accuracy depends on the error in the LS and
UV detector signals as well as uncertainty in the extinction coefficient and refrac-
tive index increment, which may not be explicitly known for peaks in the chromato-
gram. The error in each detector’s signal can be approximated by its noise (i.e.,
eSLS ≈ NLS) by assuming that inter-detector band broadening is either negligible or
accounted for by the data analysis software.
5.6.2 Case 2: LS-RI Detectors
It is possible to determine the size variant concentration from Eq. 5.3 when RI

detection is used instead of UV detection. In this case, the extinction coefficient is
not required for the calculation, but the refractive index increment is required.
Uncertainty in this parameter, along with errors in the LS and RI detector signals, is
propagated using Eq. 5.4 to determine an expression for the molecular weight error
(Eq. 5.6):
2 2 2
eM e  e   e 
=  SLS  +  SRI  +  dn / dc  . (5.6)
M  SLS   SRI   dn / dc 

In this case, the error expression has only three terms: error in each detector’s
signal and uncertainty in the protein refractive index increment.
5.6.3 Case 3: LS-UV-RI Detectors
In the special case where all three detectors are connected in series, the refractive
index increment can be measured using the RI detection. This can be shown by
reexpressing Eq. 5.1 after solving for M and using Eqs. 5.2 and 5.3 to eliminate the
dn/dc term, as shown in Eq. 5.7:
SLS SUV
M=k , (5.7)
2
SRI e

where k is a collection of constants. Using all three detectors, an a priori estimate of
the refractive index increment is not required, but the extinction coefficient is required
to determine the concentration. Following similar error propagation analysis, an
expression for the error of the molecular weight calculation can be determined:
116 Y. Jiang et al.
2 2 2 2
eM e  e  e  e 
=  SLS  +  SUV  + 4  SRI  +  e  . (5.8)
M  SLS   SUV   SRI   e 

Because all three detectors are used, errors in each of their signals appear in the
overall error expression. Uncertainty in the protein extinction coefficient will also
impact the accuracy of the molecular weight determination.
Using the appropriate error expression based on the choice of detectors used, and
accounting for uncertainty in the protein extinction coefficient and/or refractive
index increment, allows an assessment of molecular weight accuracy to be per-
formed for any peak in any chromatogram. Thus, an evaluation of accuracy is not
made at a single point in time but is a dynamic assessment that varies with protein
load, size variant mass, and system noise. The accuracy also improves as analysts
apply more accurate estimates of extinction coefficients and refractive index
increments.
A graphical representation of the molecular weight error estimates from the three
detector arrangements is provided in Fig. 5.4. Fractional error in the molecular
weight (eM/M) is plotted as a function of the size variant concentration for different
levels of uncertainty in the extinction coefficient and/or refractive index increment.
At high protein concentrations (where S » N), the fractional errors in the detector
signal terms in Eqs. 5.5, 5.6, and 5.8 are negligible. Therefore, for cases 2 and 3 in
which LS and RI detection are used, the fractional error in molecular weight
approaches the fractional uncertainty in either dn/dn (LS-RI, Eq. 5.6, Fig. 5.4b) or ε
(LS-UV-RI, Eq. 5.8, Fig. 5.4c) at high protein concentrations. When UV detection
is used without RI detection, the limiting error in the molecular weight is higher
than when RI detection is used because uncertainty in both dn/dn and ε contribute
to the error estimate (LS-UV, Eq. 5.5, Fig. 5.4a). At low concentrations (where
S ~ N), the LS signal error term dominates the molecular weight error expression
(all panels of Fig. 5.4).
To show that the instrumentation is performing properly and the calculated molar
masses are meaningful, the measured masses by LS should be compared to a
standard of known mass such as bovine serum albumin. At least one such standard
should be analyzed within each sequence to confirm the system’s suitability. Then,
the molecular weights of size variants in the sample elution profile can be reported
with associated error estimates derived from Eqs. 5.5, 5.6, or 5.8.
An understanding of both the precision and accuracy of measured molecular
weights using LS detection should be considered when determining whether the
method will be appropriate for specific applications in the biopharmaceutical indus-
try. Addition of LS detection to an SE-HPLC method may be useful for product
profile characterization, contaminant identification, and product comparability
assessments. Nevertheless, the performance required of the method may be differ-
ent for each of these applications, which is why qualification of the method is criti-
cal to enable selection of appropriate uses of LS detection.
Fig. 5.4 Molecular weight fractional error estimates derived from error propagation analysis
(Eqs. 5.5, 5.6, and 5.8) and plotted across a range of protein concentrations. In each panel, different
curves represent different levels of uncertainty for ε and/or dn/dc. The curves were generated with
0, 2.5%, 5%, and 10% error for ε and dn/dc to encompass the expected range of uncertainty in
these parameters. (a) LS-UV detectors (Eq. 5.5). Fractional errors in ε/dn/dc from lowest curve to
highest: 0/0 (lowest curve); 0.025/0 (curve 2); 0.05/0 and 0/0.025 (indistinguishable, curve 3);
0.025/0.025 (curve 4); 0.05/0.025 (curve 5); 0/0.05, 0.1/0, and 0.025/0.05 (indistinguishable, curve
6); 0.05/0.05 and 0.1/0.025 (indistinguishable, curve 7); 0.1/0.05 (curve 8); 0/0.1, 0.025/0.1, and
0.05/0.1 (indistinguishable, curve 9); and ε = 0.1, dn/dc = 0.1 (highest curve). (b) LS-RI detectors
(Eq. 5.6). Fractional errors in dn/dc from lowest curve to highest: 0, 0.025, 0.05, and 0.1.
(c) LS-UV-RI detectors (Eq. 5.8). Fractional errors in ε from lowest curve to highest: 0, 0.025,
0.05, and 0.1. Data shown in all panels were generated using the following values: NUV = 0.00035
V; NLS = 0.0055 V; NRI = 0.00025 V; kUV = 10 V; kLS = 0.0015 V mol L mL−2; kRI = 20 V g mg−1;
M = 34,000 g mol−1; ε = 0.48 mL mg−1 cm−1; l = 1 cm; and dn/dn = 0.165 mL g−1
118 Y. Jiang et al.
5.7 Qualification of LO- and MFI-Based Methods
The importance of subvisible particle (SbVP) analysis of protein therapeutics has

increased as it has become an area of intense regulatory scrutiny due to concerns
around the potential safety impact of protein aggregates in the micron size range,
especially potential immunogenicity (Carpenter et al. 2009). The available methods
have been reviewed (Narhi et al. 2009; Cao et al. 2012; Mahler et al. 2009; Zölls
et al. 2011), with the authors showing the utilities of LO-, MFI-based, and conduc-
tivity and light scattering-based methods and microscopy for the analysis of subvis-
ible particles.
The amount of subvisible particles allowed in therapeutics is currently governed
by the pharmacopeia of the individual countries in which a drug is marketed, such
as the USP <788> (United States Pharmacopeial Convention 2009) for the USA and
EP 2.9.19 for the EU (Council of Europe 2007). These compendial methods for the
quantification of subvisible particles were written for application to small-molecule
therapeutics delivered intravenously using the LO technique, with the primary con-
cern being the occlusion of capillaries by foreign particles due to contamination
(Cao et al. 2009; Haines-Nutt and Munton 1984; Taylor and Spence 1983; Alexander
and Veltman 1988). Applying USP <788> or EP 2.9.19 to protein therapeutics has
met some significant issues including false-positive and negative results caused by
trapped air bubbles and breaking down of fragile protein particles (Cao et al. 2009,
2010). The amount of sample needed for testing SbVP throughout the therapeutic
protein product development life cycle by the compendial method is also prohibitive
since the therapeutic protein is generally at high concentration and the availability
of the material at early development stages is very limited.
A modified LO method for SbVP quantification in protein therapeutics was
developed and qualified by Cao et al. (2010) recently. The authors described a modi-
fied LO method that can be used in procedures to analyze and control subvisible
particles in protein therapeutics. They introduced appropriate sample handling such
as vacuum degassing to minimize the false-positive and negative results. The
method includes the ability to analyze sample volumes as small as 1 mL and
quantification of subvisible particles down to 2 μm instead of 10 μm as specified by
USP <788> or EP 2.9.19. The method was qualified for accuracy, precision (includ-
ing repeatability and intermediate precision), and linearity using polystyrene parti-
cle counting standards (generally <10% RSD with sufficient amount of particles).
The USP is currently in the process of establishing a biologics-specific chapter
<787> that will closely resemble this modified LO method.
Protein samples can only be used to demonstrate precision and linearity, due to
the lack of protein particle counting standards. Much of the qualification and valida-
tion of methods for SbVP determination must therefore be done with particles of
known size and quantity made of polymers, The differences in morphology/shape,
refractive index, and other optical properties of protein particles and polystyrene
particle counting standards have made it difficult to quantify accurately the amount
of protein particles without an appropriate standard, especially at the lower size
range (i.e., <5 μm). The effect of the difference in refractive index between the
polystyrene calibration spheres and test particles on the apparent size of particles
determined by LO-based methods has been observed for other types of particles
such as lipids as well (Driscoll 2004). However, the demonstrated precision and
linearity, as well as the broadened size range of the modified LO method for protein
samples, enable the use of the method for monitoring effectiveness of filtration and
clearance of SbVPs during downstream processing steps, assessing the effect of
various process stresses on SbVP formation, and the contribution of primary con-
tainer components to the SbVP present (Cao et al. 2009, 2010).
The validation of the particle size method and instrument qualification have been
described by Rawle in the book on “Practical Approaches to Method Validation and
Essential Instrument Qualification”(Chan et al. 2011) and in the pharmacopeia
(United States Pharmacopeial Convention 2009; Council of Europe 2007) from dif-
ferent country/regions. The procedures for instrument calibration, performance
verification, and the use of particle counting standards are specified. Different
instrument vendors essentially follow the same calibration procedure for LO-based
instruments. This ensures the precision and accuracy of the measurements when
polystyrene particle counting standards are measured. The performance of the LO
method for protein therapeutics needs to be verified as shown by Cao et al. before it
is used for testing.
Even though LO-based methods have been demonstrated to be precise and are
widely used by the pharmaceutical industry to monitor SbVPs in their injectable
products, the information obtained from these methods is limited to the size and
concentration of the particles. In order to gain more understanding of the nature of
SbVP, MFI-based techniques were developed recently (Mahler et al. 2009; Huang
et al. 2009; Sharma et al. 2010a; Wuchner et al. 2010). Flow imaging techniques
allow the characterization of particle shape and morphology in addition to size and
concentration. It has been shown to be more sensitive than LO methods for measur-
ing subvisible proteinaceous particles in protein formulations.
Due to the potential challenges in detecting particles in opalescent mAb formula-
tions, the accuracy of MFI in determining size and particle counts in opalescent
solutions was investigated and compared to LO and membrane microscopy methods
by Sharma et al. (2010b). They used proteinaceous monoclonal antibody (mAb)
particles, generated either by chemical denaturation or agitation stress, polystyrene,
and glass particles as model systems for measurements in opalescent mAb solu-
tions. The authors demonstrate that the sizing and counting accuracies of MFI were
unaffected by the opalescence of the medium. Using glass particles as a model
system for proteinaceous particles, MFI was able to detect relatively low particle
concentrations (approximately 10 particle/mL) in opalescent solutions. MFI showed
excellent linearity (R(2) = 0.9969) for quantifying proteinaceous particles in opales-
cent solutions over a wide range of particle concentrations (approximately
20–160,000 particle/mL). The authors also show that the intensities of MFI particle
image are significantly different with respect to the transparency of proteinaceous
particles as a function of their size and mode of generation. The LO method signifi-
cantly underestimated proteinaceous particles, particularly those in the 2–10 μm
120 Y. Jiang et al.
size range. The less opaque proteinaceous particles, the more underestimated by the
LO method. Similar phenomenon was observed by Driscoll for lipid particles
(Driscoll 2004).
Polystyrene bead standards are generally used as a reference for calibrating and
validating particle analyzers. However, these standards, unlike protein particles, are
easily detected and do not challenge the sensitivity of optical instruments.
Instruments calibrated and verified only with beads can still exhibit significant dif-
ferences in measuring concentrations of protein particles. To ensure the accuracy of
the analysis of the size and concentration of suspended protein particles in protein
therapeutics, standards should resemble protein particles with respect to their trans-
parency and refractive index. Sharma et al. (2011) described the work on evaluating
a potential standard and the use of it to harmonize the performance of MFI instru-
ments. They have shown that the use of a suitable standard can significantly increase
measurement consistency when multiple instruments are used to characterize the
same protein particle suspension.
Similar to MFI, a micro-flow digital imaging (MDI) method was developed,
optimized, and qualified to detect and monitor protein particles in a high-
concentration IgG1 monoclonal antibody formulation by Wuchner et al. (2010).
The MDI method was qualified as a characterization assay with respect to accuracy
and precision of particle size and number using polystyrene standards (5–300 μm)
and protein particles (2 to >100 μm). The authors then used it to monitor the stabil-
ity profile of a 90 mg/mL IgG1 formulation stored at 2–8°C and −70°C for up to 18
months. The MDI assay results showed improved sensitivity in detecting subvisible
particles (≥5 μm) compared to the conventional LO-based method and a good over-
all correlation with the amount of visible protein particles (>70 μm). The MDI
results showed an accumulation of protein particles in certain size ranges and an
increase in the overall particle size distribution over time. The combined MDI
results and protein characterization studies have provided an enhanced understand-
ing of protein particulate formation during long-term storage of a high-concentration
IgG1 monoclonal antibody formulation.
In summary, the LO- and MFI-based methods can and should be qualified for
SbVP analysis in specific protein solutions. The methods offer sensitive and
reproducible measurement of particles in protein therapeutics solutions which

enables proper monitoring and control of SbVP levels during protein therapeutics
development, manufacturing, transportation, storage, and delivery.
5.8 Conclusion
The qualification of the biophysical methods including CD, FTIR, DSC, SV-AUC,
SE-HPLC-LS, and LO- and MFI-based methods has been reviewed. Consistent
instrument calibration and maintenance and proper sample handling are required to
ensure successful qualification of these methods for protein higher-order structure
and size distribution analysis. A numerical method was applied to compare the
overall spectra or profile similarity for the qualification of some of the methods such
as CD and FTIR. AUC and SE-HPLC-LS are first principle methods; therefore, the
qualification of these methods does not require the use of a standard. However, for
the proper qualification of the LO- and MFI-based methods, suitable particle
standards that closely resemble the optical properties and morphologies of protein
particles are still needed. For all of the biophysical methods, qualification is focused
on the precision of the methods and demonstration that the methods are suitable for their
intended applications. Successful qualification enables the understanding of the
method capability and reliable determination and control of product attributes.
Therefore, qualification of biophysical methods should be undertaken for protein
therapeutics to ensure the best possible knowledge of the product and its
manufacturing process.
References
Affolter S, Ritter A, Schmid M (2001) Interlaboratory tests on polymers by differential scanning

calorimetry (DSC): determination of glass transition temperature (Tg). Macromol Mater Eng
286(10):605–610
Alexander DM, Veltman AM (1988) Particulate contamination in solutions of antibiotics packed
as dry powders in vials. J Pharm Pharmacol 40(5):358–359
Arthur KK et al (2009) Detection of protein aggregates by sedimentation velocity analytical ultra-
centrifugation (SV-AUC): sources of variability and their relative importance. J Pharm Sci
98(10):3522–3539
Berkowitz SA (2006) Role of analytical ultracentrifugation in assessing the aggregation of protein
biopharmaceuticals. AAPS J 8(3):590–605
Bruylants G, Wouters J, Michaux C (2005) Differential scanning calorimetry in life science: ther-
modynamics, stability, molecular recognition and application in drug design. Curr Med Chem
12(17):2011–2020
Byler DM, Susi H (1986) Examination of the secondary structure of proteins by deconvoluted
FTIR spectra. Biopolymers 25:469–487
Cao S, Jiao N, Jiang Y, Mire-Sluis A, Narhi LO (2009) Sub-visible particle quantitation in protein
therapeutics. Pharmeuropa Bio & Scientific Notes 73–79
Cao S, Jiang Y, Narhi LO (2010) A light obscuration method specific for quantifying subvisible
particles in protein therapeutics. US Pharmacoepial Forum 36(3):824–834
Cao S, Narhi LO, Jiang Y, Rajan RS (2012) Analytical methods to measure sub-visible particu-
lates. In: Mahler HC, Jiskoot W (eds) Analysis of aggregates and particles in protein pharma-
ceuticals. Wiley, New York
Carpenter JF, Prestrelski SJ, Dong A (1998) Application of infrared spectroscopy to development
of stable lyophilized protein formulations. Eur J Pharm Biopharm 45(3):231–238
Carpenter JF et al (2009) Overlooking subvisible particles in therapeutic protein products: gaps
that may compromise product quality. J Pharm Sci 98(4):1201–1205
Carpenter JF et al (2010) Potential inaccurate quantitation and sizing of protein aggregates by size
exclusion chromatography: essential need to use orthogonal methods to assure the quality of
therapeutic protein products. J Pharm Sci 99(5):2200–2208
Celej MS, Dassie SA, González M, Bianconi ML, Fidelio GD (2006) Differential scanning calo-
rimetry as a tool to estimate binding parameters in multiligand binding proteins. Anal Biochem
350(2):277–284
122 Y. Jiang et al.
Chan CC, Lam H, Zhang XM (2011) Practical approaches to method validation and essential
instrument qualification. Wiley, New York
Chiu MH, Prenner EJ (2011) Differential scanning calorimetry: an invaluable tool for a detailed
thermodynamic characterization of macromolecules and their interactions. J Pharm Bioallied
Sci 3(1):39–59
Coligan JE et al (2001) Measuring protein thermostability by differential scanning calorimetry.
In: Makhatadze GI (ed) Current protocols in protein science/editorial board. Wiley, New York
Council of Europe (2007) Particulate contamination: sub-visible particles, general chapter 2.9.19.
Council of Europe, Strasbourg
Cover TM, Hart PE (1967) Nearest neighbor pattern classification. IEEE Trans Inf Theory
IT-13(1):21–27
D’antonio J, Murphy BM, Manning MC, Al-Azzam WA (2012) Comparability of protein thera-
peutics: quantitative comparison of second-derivative amide I infrared spectra. J Pharm Sci.
doi:10.1002/jps.23133
Dam J et al (2004) Calculating sedimentation coefficient distributions by direct modeling of sedi-
mentation velocity concentration profiles. In: Methods in enzymology. Academic.
pp 185–212
Drechsler A, Miles AJ, Norton RC, Wallace BA, Separovic F (2009) Effect of lipid on the confor-
mation of the n-terminal region of equinatoxin II: a synchrotron radiation CD study. Eur
Biophys J 39:121–127
Driscoll DF (2004) Examination of selection of light-scattering and light-obscuration acceptance
criteria for lipid injectable emulsions. Pharm Forum 30:2244–2253
Fu K, Griebenow K, Hsieh L, Klibanov AM, Langer R (1999) FTIR characterization of the sec-
ondary structure of proteins encapsulated within PLGA microspheres. J Controlled Release
58:357–366
Gabrielson JP, Arthur KK (2011) Measuring low levels of protein aggregation by sedimentation
velocity. Methods 54(1):83–91
Gabrielson JP et al (2007a) Quantitation of aggregate levels in a recombinant humanized monoclo-
nal antibody formulation by size-exclusion chromatography, asymmetrical flow field flow frac-
tionation, and sedimentation velocity. J Pharm Sci 96(2):268–279
Gabrielson JP et al (2007b) Sedimentation velocity analytical ultracentrifugation and SEDFIT/
c(s): limits of quantitation for a monoclonal antibody system. Anal Biochem 361:24–30
Gabrielson JP, Arthur KK, Stoner MR et al (2010) Precision of protein aggregation measurements
by sedimentation velocity analytical ultracentrifugation in biopharmaceutical applications.
Gmelin E, Sarge St.M (1995) Calibration of differential scanning calorimeters. IUPAC, Pure Appl
Chem 67:1789–1800
Greenfield NJ (1996) Methods to estimate conformation of proteins and polypeptides from CD
data. Anal Biochem 235(1):1–10
Greenfield NJ, Fasman GD (1969) Computed circular dichroism spectra for the evaluation of pro-
tein conformation. Biochemistry 8:4108–4116
Gross PC, Zeppezauer M (2010) Infrared spectroscopy for biopharmaceutical protein analysis.
J Pharm Biomed Anal 53:29–36
Haines-Nutt RF, Munton TJ (1984) Particle size measurement in intravenous fluids. J Pharm
Pharmacol 36(8):534–536
Haris PI, Chapman D (1995) The conformational analysis of peptides using Fourier transform IR
spectroscopy. Biopolym (Pept Sci) 37:251–263
teral solutions using micro-flow imaging. J Pharm Sci 98(9):3058–3071
ICH M4Q (2001) The common technical document for the registration of pharmaceuticals for
human use: quality. US Department of Health and Services, FDA, CDER
ICH Q2(R1) (2005) Validation of analytical procedures: test and methodology. In: ICH Harmonised
Tripartite guideline in international conference on harmonisation of technical requirements for
registration of pharmaceuticals for human use
ICH Q5E (2005) Comparability of biotechnological/biological products subject to changes in their

manufacturing process. US Department of Health and Services, FDA, CDER & CBER
ICH Q6B (1999) Test procedures and acceptance criteria for biotechnological and biological
products
Jiang Y, Narhi LO (2006) Applying selective biophysical techniques in assessing the comparabil-
ity of protein therapeutics – case studies. J Am Pharm Rev 9(5):34–43
Jiang Y, Li C, Ramachander R, Wen J, Narhi LO (2008) Meeting new regulatory expectations of
characterization methods-Qualification of biophysical analyses. J Am Pharm Rev February,

online issue, Russell Publishing. http://pharmoutsourcing.com/ViewArticle.aspx?ContentID=2221.
Accessed 7 Feb 2012
Jiang Y, Li C, Nguyen X, Muzammil S, Towers E, Gabrielson JP, Narhi LO (2011) Qualification
of FTIR spectroscopic method for protein secondary structural analysis. J Pharm Sci
100(11):4631–4641
Johnson WC Jr (1988) Secondary structure of proteins through circular dichroism spectroscopy.
Annu Rev Biophys Biophys Chem 17:145–166
Johnson WC Jr (1999a) Analyzing protein circular dichroism spectra for accurate secondary struc-
tures. Proteins 35(3):307–312
Johnson WC (1999b) Analysing protein circular dichroism spectra for accurate secondary struc-
tures. Proteins 35(3):307–312
Johnson CR, Morin PE, Arrowsmith CH, Freire E (1995) Thermodynamic analysis of the struc-
tural stability of the tetrameric oligomerization domain of p53 tumor suppressor. Biochemistry
34(16):5309–5316
Johnston MJW, Frahm G, Li X, Durocher Y, Hefford MA (2011) O-linked glycosylation leads to
decreased thermal stability of interferon alpha 2b as measured by two orthogonal techniques.
Pharm Res 28(7):1661–1667
Jones C et al (2004) Val-CiD Best Practice Guide: CD spectroscopy for the quality control of bio-
pharmaceuticals. National Physical Laboratory. DQL-AS 008 ISSN: 1744–0602. British
Library Document Supply Centre, Boston Spa, Wetherby, West Yorkshire, LS23 7BQ
Kelly SM, Jess TJ, Price NC (2005) How to study proteins by circular dichroism. Biochim Biophys
Acta 1751:119–139
Kendrick BS, Dong A, Allison SD, Manning MC, Carpenter JF (1996) Quantitation of overlap of
infrared second derivative spectra to determine structural similarity between proteins. J Pharm
Sci 85:155–158
Kendrick BS et al (2001) Online size-exclusion high-performance liquid chromatography light
scattering and differential refractometry methods to determine degree of polymer conjugation
to proteins and protein–protein or protein–ligand association states. Anal Biochem
299(2):136–146
King SM, Johnson WC (1999) Assigning secondary structure from protein coordinate data.
Proteins 35:313–320
Kong J, Yu S (2007) Fourier transform infrared spectroscopic analysis of protein secondary struc-
tures. Acta Biochim Biophys Sin 39(8):549–559
Laue TM et al (1992) Computer-aided interpretation of analytical sedimentation data for proteins.
In: Harding SE, Rowe AJ, Horton JC (eds) Analytical ultracentrifugation in biochemistry and
polymer science. Royal Society of Chemistry, Cambridge, pp 90–125
Lebowitz J, Lewis MS, Schuck P (2002) Modern analytical ultracentrifugation in protein science:
a tutorial review. Protein Sci 11(9):2067–2079
Lees JG, Wallace BA (2002) Synchrotron radiation circular dichroism and conventional circular
dichroism spectroscopy: a Comparison. Spectroscopy 16:121–125
Lees JG, Miles AJ, Wien F, Wallace BA (2006) A reference database for circular dichroism spec-
troscopy covering fold and secondary structure space. Bioinformatics 22:1955–1962
Li C, Nguyen X, Narhi LO, Chemmalil L, Towers E, Muzammil S, Gabrielson JP, Jiang Y (2011)
Applications of circular dichroism for structural analysis of proteins: qualification of near- and
far-UV CD for protein higher order structural analysis. J Pharm Sci 100(11):4642–4654
124 Y. Jiang et al.
Lopez MO, Freire E (1987) Dynamic analysis of differential scanning calorimetry data. Biophys
Chem 27(1):87–96
Mahler HC, Friess W, Grauschopf U, Kies S (2009) Protein aggregation: pathways, induction fac-
tors and analysis. J Pharm Sci 98:2909–2934
Manning MC, Woody RW (1989) Theoretical study of the contribution of aromatic side chains to
the circular dichroism of basic bovine pancreatic trypsin inhibitor. Biochemistry
28:8609–8613
Meersman F, Atilgan C, Miles AJ, Bader R, Shang W, Matagne A, Wallace BA, Koch MHJ (2010)
Consistent picture of the reversible thermal unfolding of hen egg-white lysozyme from experi-
ment and molecular dynamics. Biophysical J 99:2255–2263
Miles AJ, Wallace BA (2006) Synchrotron radiation circular dichroism spectroscopy of proteins
and applications in structural and functional genomics. Chem Soc Reviews 35:39–51
Miles AJ, Wien F, Lees JG, Rodger A, Janes RW, Wallace BA (2003) Calibration and standardisa-
tion of synchrotron radiation circular dichroism (SRCD) amplitudes and conventional circular
dichroism (CD) spectrophotometers. Spectroscopy 17:653–661
Miles AJ, Wien F, Wallace BA (2004) Redetermination of the extinction coefficient of camphor-b-
sulfonic acid, A calibration standard for circular dichroism spectroscopy. Anal Biochem
335:338–339
Miles AJ, Wien F, Lees JG, Wallace BA (2005) Calibration and standardisation of synchrotron
radiation and conventional circular dichroism spectrometers. Part 2: factors affecting magni-
tude and wavelength. Spectroscopy 19:43–51
Miles AJ, Hoffman SV, Tao Y, Janes RW, Wallace BA (2007) Synchrotron radiation circular
dichroism (SRCD) spectroscopy: new beamlines and new applications in biology. Spectroscopy
21:245–255
Narhi LO, Jiang Y, Cao S, Benedek K, Shnek D (2009) A critical review of analytical methods for
subvisible and visible particles. Curr Pharm Biotechnol 10(4):373–381
Pekar A, Sukumar M (2007) Quantitation of aggregates in therapeutic proteins using sedimenta-
tion velocity analytical ultracentrifugation: practical considerations that affect precision and
accuracy. Anal Biochem 367:225–237
Perczel A, Fasman GD (1993) Effect of spectral window size on circular dichroism spectra decon-
volution of proteins. Biophys Chem 48:19–29
Perczel A, Hollosi M, Tusnady G, Fasman GD (1991) Convex constraint analysis: a natural decon-
volution of circular dichroism curves of proteins. Protein Eng 4:669–679
8(3):564–571
Powl AM, O'Reilly AO, Miles AM, Wallace BA (2010) Synchrotron radiation circular dichroism
spectroscopy-defined structure of the C-terminal domain of NaChBac and its role in channel
assembly. Proc Natl Acad Sci 107:14064–14069
Prestrelski S, Tedeschi N, Arakawa T, Carpenter J (1993) Biophys J 65(2):661–671
Protasevich I, Yang Z, Wang C, Atwell S, Zhao X, Emtage S, Wetmore D, Hunt JF, Brouillette CG
(2010) Thermal unfolding studies show the disease causing F508del mutation in CFTR ther-
modynamically destabilizes nucleotide-binding domain 1. Protein Sci 19(10):1917–1931
Pyrpassopoulos S, Vlassi M, Tsortos A, Papanikolau Y, Petratos K, Vorgias CE, Nounesis G
(2006) Equilibrium heat-induced denaturation of chitinase 40 from Streptomyces thermoviola-
ceus. Proteins 64(2):513–523
Ravi J, Rakowska PD, Garfagnini T et al (2010) International comparability in spectroscopic
measurements of protein structure by circular dichroism: CCQM-P59.1, Metrologia 47(6):
631–641
Ravi J, Schiffmann D, Tantra R et al (2010) International comparability in spectroscopic measure-
ments of protein structure by circular dichroism: CCQM P59, Metrologia 47:1A
Schiffmann D, Butterfield DM, Yardley RE, Knight A, Windsor SA, Jones C (2004) Val-CiD
Appendix A: CD spectroscopy: an inter-laboratory study. DQL-AS 009 ISSN: 1744–0602
Schmid M (2012) Precision of DSC measurements: results of interlaboratory tests. http://www.
eurostar-science.org/MAP1/Schmid.pdf. Accessed 7 Feb 2012
Schuck P et al (2002) Size-distribution analysis of proteins by analytical ultracentrifugation: strat-

egies and application to model systems. Biophys J 82(2):1096–1111
Sharma DK, King D, Oma P, Merchant C (2010a) Micro-flow imaging: flow microscopy applied
to sub-visible particulate analysis in protein formulations. AAPS J 12(3):455–464
Sharma DK, Peter O, Pollo MJ, Sukumar M (2010b) Quantification and characterization of subvis-
ible proteinaceous particles in opalescent mab formulations using micro-flow imaging. J Pharm
Sci 99(6):2628–2642
Sharma DK, King D, Merchant C (2011) Reference material development for calibration and veri-
fication of image-based particle analyzers. Int J Pharm 416(1):293–295
Shire SJ (1994) Analytical ultracentrifugation and its use in biotechnology. In: Schuster TM, Laue
TM (eds) Modern analytical ultracentrifugation. Birkhauser, Boston
Surewicz WK, Mantsch HH (1988) New insight into protein conformation from infrared spectros-
copy. Biochim Biophys Acta 952:115–130
Susi H, Byler M (1986) Resolution enhanced fourier transform infrared spectroscopy of enzymes.
Meth Enzymol 130:290–311
Susi H, Byler M (1987) Fourier transform infrared study of proteins with parallel b-chains. Arch
Biochem Biophys 258(2):465–469
Taylor JR (1997) An introduction to error analysis: the study of uncertainties in physical measure-
ments, 2nd edn. University Science Books, Sausalito
Taylor SA, Spence J (1983) Particles in small volume injections. J Pharm Pharmacol
35(12):769–773
TQ Analyst Algorithm (2007–2010) Thermo OMNIC software manual, Thermo Electron Scientific
Instruments, LLC., Madison, WI, USA (part of Thermo Fisher Scientific Inc.)
United States Pharmacopeial Convention, Inc (2009) Particular matter in injections, general chap-
ter <788>, USP 32, NF 27. United States Pharmacopeial Convention, Inc., Rockville
Van Stokkum IHM, Spoelder HJW, Bloemendal M, Van Grondelle R, Groen FCA (1990)
Estimation of protein secondary structure and error analysis from CD spectra. Anal Biochem
191:110–118
Vedantham G, Sparks G, Sane SU, Tzannis S, Przybycien TM (2000) A holistic approach for
protein secondary structure estimation from infrared spectra in H2O solutions. Anal Biochem
285:33–49
Venyaminov SW, Baikalov I, Chuen-Shang C, Yang JT (1991) Some problems of CD analysis of
protein conformation. Anal Biochem 198:250–255
Vonhoffa S, Condliffeb J, Schiffterb H (2010) Implementation of an FTIR calibration curve for fast
and objective determination of changes in protein secondary structure during formulation
development. J Pharm Biomed Anal 51(1):39–45
Wallace BA (2005) Shining new light on protein structure and function thru synchrotron radiation
circular dichroism (SRCD) spectroscopy. Australian Biochemist 36:47–50
Wallace BA, Janes RW (2010) Synchrotron radiation circular dichroism (SRCD) spectroscopy: an
enhanced method for examining protein conformations and protein interactions. Biochem Soc
Trans 38:861–873
Wallace BA, Whitmore L, Janes RW (2006) The protein circular dichroism data bank (PCDDB):
a bioinformatics and spectroscopic resource. Proteins 62:1–3
Weissburg RP, Burris FE, Ferrell TK, Montgomery ER (2002) Validation of DSC and TGA for
pharmaceutical analysis under current good Manufacturing Practices (cGMP). Am Pharm Rev
5(2):76–81
absorbance, and refractive index detectors for studying proteins and their interactions. Anal
Biochem 240(2):155–166
Wen J, Jiang Y, Narhi LO (2008) Effect of carbohydrate on thermal stability of antibodies. Am
Pharm Rev 11:98–104
126 Y. Jiang et al.
Wen J, Arthur K, Chemmalil L, Muzammil S, Gabrielson JP, Jiang Y (2011) Applications of dif-
ferential scanning calorimetry for thermal stability analysis of proteins: qualification of DSC.
J Pharm Sci. doi:10.1002/jps.22820
Whitmore L, Wallace BA (2004) DICHROWEB, An online server for protein secondary structure
analyses from circular dichroism spectroscopic data. Nucleic Acids Res 32:668–673
Whitmore L, Wallace BA (2008) Protein secondary structure analyses from circular dichroism
spectroscopy: methods and reference databases. Biopolymers 89:392–400
Woody RW (1994) Contributions of tryptophan side chains to the far-ultraviolet circular dichroism
of proteins. Eur Biophys J 23:253–262
Woody RW (1996) Theory of circular dichroism of proteins, circular dichroism and the conforma-
tional analysis of biomolecules. In: Fasman GD (ed) Circular dichroism and the conforma-
tional analysis of biomolecules. Plenum, New York
Wuchner K, Büchler J, Spycher R, Dalmonte P, Volkin DB (2010) Development of a microflow
IgG1 monoclonal antibody formulation. J Pharm Sci 99(8):3343–3361
Wyatt PJ (1993) Light scattering and the absolute characterization of macromolecules. Anal Chim
Acta 272:1–40
Yang JT, Wu CSC, Martinez HM (1986) Calculation of protein conformation from circular dichro-
ism. Methods Enzymol 130:208–269
Yoshii K (1997) Application of differential scanning calorimetry to the estimation of drug purity:
various problems and their solutions in purity analysis. Chem Pharm Bull 45(2):338–343
Zölls S, Tantipolphan R, Wiggenhorn M, Winter G, Jiskoot W, Friess W, Hawe A (2011) Particles
in therapeutic protein formulations, Part 1: overview of analytical methods. J Pharm Sci.
doi:10.1002/jps.23001
Chapter 6
Application of Biophysics to the Early
Developability Assessment of Therapeutic
Candidates and Its Application to Enhance
Developability Properties
Hasige Sathish, Nicolas Angell, David Lowe, Ambarish Shah,

and Steven Bishop
6.1 Introduction and Scope
For the past few decades, both the cost and the complexity of therapeutic protein
development have soared to levels that are becoming unsustainable (Adams and
Brantner 2010). Additionally, there is a significant increase in the attrition rate of
programs that can be attributed to, for example, complex disease pathways, an
increased safety bar, and a greater number of highly complex molecules in develop-
ment (Kola and Landis 2004). According to a recent analysis, the success rate for IND
to market is approximately 25–30% for biotech companies (DiMasi and Grabowski
2007). Both the spiraling cost associated with drug development and the demands
from end users for low-cost prescription drugs are forcing the industry to look for
ways to increase efficiencies. One approach is cycle time reduction, a strategy that
translates to a reduced and more efficient development timeline so that a drug can
move faster to phase I and faster to market. This chapter focuses on current and future
challenges in developing therapeutic biological drugs and how the use of novel bio-
physical tools to screen for potential developability and manufacturability risks can
improve product quality, streamline development, increase manufacturing efficiency,
H. Sathish (*) • A. Shah • S. Bishop

Formulation Sciences, Biopharmaceutical Development, MedImmune, One Medimmune Way,
Gaithersburg, MD 20878, USA
e-mail: [email protected]; [email protected]; [email protected]
N. Angell
Process and Product Development, Amgen Inc., One Amgen Center Drive, MS 30E-1-C,
D. Lowe
Antibody Discovery and Protein Engineering, Medimmune Ltd, The Aaron Klug Building,
Granta Park, Cambridge, Cambridgeshire CB21 6GH, UK
128 H. Sathish et al.
and increase end-user convenience. Apart from selecting low-risk molecules, early
developability screening can also aid in identification of platform compatible leads,
which can significantly reduce both development cost and time.
6.2 Current Challenges and Future Trends

in Therapeutic Drug Development
Monoclonal antibodies (mAbs) represent the fastest growing class of therapeutics in

the pharmaceutical marketplace, with 147 human and 167 humanized mAbs having
entered clinical development between 1985 and 2008 (Nelson et al. 2010). Because
mAbs cannot readily enter cells or cross the blood–brain barrier, mAbs are typically
developed against particular target classes, such as soluble proteins (e.g., cytokines
and growth factors) or extracellular domains of receptors (Carter 2011). These bio-
logical limitations have resulted in a large number of similar mAbs targeting the
same therapeutic indications. Hence, much of the discovery work is focused upon
identifying novel therapeutic molecular targets and choosing or engineering proper-
ties that will potentially confer a clinical or commercial advantage. A third industry
challenge is that of optimizing development/manufacturing design spaces to achieve
a competitive advantage in terms of efficiency and timeline.
6.2.1 Novel Targets
More novel target classes, such as G-protein-coupled receptors (GPCRs), ion chan-
nels, and intracellular targets may not be evolutionarily ideally suited for antibody
therapy. Therefore, these targets may require the development of therapeutic mole-
cules that are inherently more novel than the prototypical IgG, such as alternative
protein fold scaffolds or engineered receptor ligands. Recent advances in protein
expression and production techniques, as well as improved in vitro libraries and
selection strategies, have started to expand the coverage of these targets that are
implicated in a range of different diseases (Lundstrom 2005; Ackerman and
Clapham 1997; Naylor and Beech 2009). However, the types of therapeutic mole-
cules (discussed in Sect. 6.2.2) that need to be engineered to bind these novel targets
are far less well characterized in terms of large-scale development and manufacture
than prototypical mAbs and are trending towards being highly complex, resulting in
a greater attrition rate during discovery and development (Ma and Zemmel 2002).
6.2.2 Engineering Desired Properties
The field of antibody engineering has matured rapidly over the past decade so that
properties such as affinity, potency, specificity, pharmacokinetics, and effector
6 Application of Biophysics to the Early Developability Assessment of Therapeutic… 129
function are now readily tunable (Igawa et al. 2011). Also, the pharmaceutical
industry is increasingly focusing on engineering molecules with novel mechanisms
of action (Ma and Zemmel 2002).
For molecules with an established mechanism of action, the focus is on enhanced
safety, higher affinity to target, and increased half-life. The first generation of mAbs
to reach the market typically exhibited dissociation constants (Kds) in the micromo-
lar range, in line with the in vivo ceiling of the B-cell response (Foote and Eisen
1995; Batista and Neuberger 1998). In vitro affinity maturation methods and tech-
nologies [reviewed in Igawa et al. (2011)] have allowed researchers to go far beyond
these limits, reaching femtomolar Kds in many cases (Boder et al. 2000; Finch et al.
2011). Increasing the affinity of an antibody to a given antigen is achieved by target-
ing the variable domains, as these are regions of the molecule that have evolved to
interact directly with the antigen, particularly the complementarity determining
regions (CDRs). Affinity maturation typically results in several mutations in the
CDR loops. As the CDR loops are surface exposed, engineering changes may result
in changes to biophysical behavior, such as an increase in aggregation propensity
and loss of stability (Chennamsetty et al. 2009a; Pepinsky et al. 2010).
Over the last decade, there has been an increased effort in glycoengineering and
the engineering of Fc variants in order to modulate effector functions [reviewed in
Kubota et al. (2009)] and improve pharmacokinetic properties through modified
interaction with the neonatal Fc receptor (Presta 2008; Stewart et al. 2011;
Dall’Acqua et al. 2006a, b). Glycosylation in IgG molecules is essential for struc-
tural integrity, solubility, and stability (Krapp et al. 2003; Mimura et al. 2000;
Kayser et al. 2011). Any modulation of glycosylation can, in turn, have significant
impact on solubility, manufacturability, and long-term storage. Likewise, engi-
neering the Fc region of the IgG molecule to modulate effector function or phar-
macokinetics may have unintended effects on the biophysical properties of the
molecule, such as stability and aggregation propensity. The unintended effects of
glycoengineering and the engineering of Fc variants may not become apparent
until larger-scale manufacture and formulation to support clinical studies.
The search for molecules with novel mechanisms of action has resulted in an
increased development of antibody-based alternative formats, such as multivalent
antibodies (Carter 2011). Recently, a new set of protein therapeutics based on alter-
native protein scaffolds to immunoglobulin domains has entered clinical develop-
ment (Gebauer and Skerra 2009). Each of these novel therapeutics is likely to
present their own challenges for development, which will doubtless become appar-
ent over the coming decade.
6.2.3 CMC Challenges
Chemistry, manufacturing, and controls (CMC) design spaces are another area of
challenge in the commercialization of protein therapeutics. For molecules with a
precedented mechanism of action, enhanced product performance can be achieved
with superior drug product dosing and presentation. While IV administration

remains an option to achieve the necessary high dose, the goal of providing greater
patience convenience, lower dosing frequency, and self-administration limits the
administration route to subcutaneous (SC) injection. SC injection typically
requires the development of a high-concentration liquid formulation, which
imposes an additional burden on the formulation and device design space and
often limits the types of approaches that can be used to provide an acceptable sta-
bility and delivery profile.
Medical devices designed to enhance patient convenience are often a key aspect
of product development. Typical devices include prefilled syringes (PFS) and auto
injectors (AI). While the range of available devices is growing to include reconstitu-
tion and delivery of lyophilized protein formulations, the most common delivery
system for self-administration is a high-concentration drug product in a PFS con-
tained within an AI. These liquid-delivery devices pose limitations that must be
considered during product development. Often certain parameters can be adjusted
within the scope of the device, such as spring force and needle gauge to allow a
formulation to “fit” a device, but many more limitations arise with high-concentration
formulations. Injection time, force, and associated pain are all patient-related fac-
tors, but additional aspects related to stability and material compatibility must also
be addressed.
Finally, from a commercial standpoint, the logistics of therapeutic protein manu-
facturing, fill/finish operation and storage must be considered a global activity.
Exposure of a therapeutic protein with inherent instabilities to manufacturing
stresses during these processes can trigger instabilities. Maintaining consistency
and the stability of the therapeutic protein throughout bioprocess, fill/finish opera-
tion, transportation, shelf life, and compatibility with the administration units will
be extremely challenging for unstable molecules. Typically, a detailed understand-
ing of the sensitivity of a protein to various unit operations is determined, and a
formulation control is designed or developed accordingly to maintain stability of the
protein during various process steps. In general, a stable molecule will have a larger
optimization window, whereas a protein with several inherent instabilities will have
a very narrow optimization window, thus restricting the development of a cost-
effective and commercial-friendly process and product that can be implemented on
a global basis.
6.3 Rationale for Early Developability Screening
As explained in Sect. 6.2.2, there is greater focus at the discovery stage on obtaining
attributes such as higher affinity, enhanced safety, and increased half-life, which
translates to highly complex, often novel molecules. From the development per-
spective, there is an increased demand to enhance product performance and patient
comfort by use of a high-concentration liquid formulation in specialty devices, with
an additional burden of decreasing the overall development timeline. Taken together,
these demands translate into attempts to develop novel molecules in a challenging

configuration and presentation under a reduced timeline. One approach that is gain-
ing widespread acceptance across the industry to tackle the complex challenge of
increasing development efficiency and reducing the attrition rate is the early identi-
fication of liabilities and other molecular properties important for smooth and effi-
cient CMC development. This approach is termed developability. Conceptually, the
ideal therapeutic protein is easily manufactured, expressed at a high titer, easily
purified with acceptable purity at a given yield, amenable to formulation at high
concentration, and capable of presentation in a format or medical device that offers
the best for patient and clinical convenience. A judicious candidate selection pro-
cess that incorporates both biological activity, affinity, and safety along with devel-
opability properties such as stability, high-concentration solution properties like
viscosity and opalescence, and manufacturability is critical in selecting such an
ideal protein. This approach can use either a screening strategy where high-risk
molecules are flagged or incorporate some of the developability properties at the
earlier design phase. In the small-molecule world, assessment of “drug-like” prop-
erties is an integral part of drug discovery. In fact, the pharmaceutical industry as
such is evolving from a “gatekeeper or screening type” strategy to a “drug-like
properties by design” approach (Hageman 2006).
Protein therapeutics are typically highly complex large molecules that can have
several instabilities. The reduction or elimination of these instabilities requires
simultaneous optimization of multiple parameters. While early screening of mole-
cule properties may lead to early identification of developability risks and aid in
streamlining manufacturing and resource allotment, it is not practical or efficient to
address all of these liabilities at the early discovery stage, when both resources and
material availability are often limited. Hence identification of the critical instabili-
ties and evaluation and development of novel high-throughput biophysical methods
or tools to screen for and reduce or eliminate such instabilities is a prerequisite for
successful developability studies.
The frontloading of developability properties assessment requires combining
computational and experimental screening approaches. Application of in silico
analysis is an emerging area and, over the past years, significant advancements have
been made, especially in the area of in silico protein aggregation prediction
(Chennamsetty et al. 2009a, b; Lauer et al. 2012; Cellmer et al. 2007). Such a pre-
dictive tool can be used to identify and eliminate, or reengineer structural and
sequence liabilities to generate proteins with improved stability and manufacturability
(Fowler et al. 2005). Unfortunately most of the available in silico tools can only be
applied to predict native aggregation behavior. For prediction of the most common
nonnative aggregation, as well as additional chemical instabilities and solution
properties at high concentration (e.g., phase separation and viscosity), the field still
relies on experimental approaches. Apart from the issue with material and resource
availability, the strategies and methods that can be used during an early developabil-
ity assessment should target the screening and optimization of multiple parameters
in an extremely dynamic environment. In short, early developability assessment
methods should allow low cost and rapid exploration of multiple parameters with
Table 6.1 Summary of common protein instabilities

Instabilities
1. Physical • Conformational stability
Thermal denaturation
Cold denaturation
Pressure-induced denaturation
Chemical denaturation
• Colloidal stability
• Surface/interfacial-induced denaturation
2. Chemical • Deamidation
• Asp isomerization
• Fragmentation
• Glycation
• Oxidation
• Disulfide scrambling
the objective of quick and cheap screening. Also key is better predictive strategies,
which should have the ability to predict the behavior of the protein during manufac-
turing and storage.
6.4 Typical Protein Liabilities
Macromolecules in general and proteins in particular are highly complex in their

structure and have various instabilities. These protein instabilities can be broadly
classified as physical or chemical instabilities (Table 6.1). Typical physical lia-
bilities that should be screened for in an early assessment strategy are those
intrinsic stability parameters that can be used to predict overall stability and
aggregation propensity and undesirable solution properties, such as solubility,
high viscosity, phase separation, and opalescence. Chemical instabilities that
should be screened for include deamidation, fragmentation, and oxidation pro-
pensity (Chan and Carter 2010).
6.5 Use and Description of Biophysical Tools in Molecule

Selection and Development
A typical early screen for a mAb might initially focus on a property, such as speci-
ficity or strength (affinity) of binding. Subsequent secondary screens and assays
typically include specificity against related molecules to the target and binding to
species homologues (to support preclinical safety studies). In this way a “screening
cascade” of different assays is performed, reducing the number of potential

antibody variants for subsequent development. As the number of individual vari-
ants is reduced, the amount and purity of sample is typically concomitantly
increased. Ideally, screens for developability characteristics, such as solubility,
expression, chemical stability, and reduced propensity for aggregation would be
placed as early in the screening cascade as possible in order to maximize the
potential sequence diversity available in a large population of variants. For exam-
ple, laboratories have reported the use of modified phage-display selections to
preferentially enrich for those variants that have increased thermal stability
(Jespers et al. 2004; Famm et al. 2008). Taking this approach even further, antibody-
fragment libraries have been designed to maximize stability, expression levels,
and aggregation resistance (Prassler et al. 2011). High-throughput (HT) assays
are ideally suited for performance early in the screening cascade, allowing bio-
physical properties to be preferentially selected for, alongside other properties,
such as affinity and specificity.
There are a wide array of biophysical techniques and strategies available to
determine protein instabilities (Table 6.2). Each of these techniques has inherent
advantages and disadvantages with respect to its ability to detect a meaningful
change in the structure of a protein and to translate this change into information
that can be used to rank order, or exclude or include, a candidate in a preclinical
development program. The sensitivity of the technique combined with the predic-
tive nature of the analysis can help determine the suitability of a method for
screening preclinical candidates. Often the techniques are employed in specific
studies designed to elicit a behavior that can then be directly measured. The crite-
ria used for identifying the ideal analytical methods and testing strategies are typi-
cally those associated with the phase of project development. For the screening of
large-molecule libraries, methods that offer advantages in speed and throughput
while using little material are preferred. At this point, the types of methodologies
that should be employed become limited to those that are robust and sensitive
towards specific behaviors that enable selection of a lead candidate for CMC
development.
For the biophysical assessment of a large number of candidate molecules, HT
approaches are preferable. In general, available techniques should support minia-
turization, simple data visualization, and interpretation to help obtain more infor-
mation in well-focused studies. The techniques that are most attractive are those
that can robustly predict or measure specific molecular attributes that indicate if a
candidate meets a defined developability criteria that are specific to its target indi-
cation or is within an acceptable developability design space either though mech-
anistic rules or phenomological correlation. Such a technique should allow for
exploration of a larger number of instabilities or parameters more quickly with a
lower cost. In general the methods should be flexible, miniature (consume mini-
mal amounts of material), and employable in a HT mode such that a battery of
different screens can be performed on the same sample simultaneously or in a
highly parallel array.
Table 6.2 List of most commonly used biophysical tools

High-throughput
Instabilities Method amenable
Conformational stability Differential scanning calorimetry Yes
Fluorescence spectroscopy Yes
• Extrinsic
• Intrinsic
Absorption spectroscopy Yes
Circular dichroism Yes
Fourier transform infrared spectroscopy No
Raman spectroscopy No
NMR No
X-ray No
Dynamic light scattering Yes
Colloidal stability Light scattering Yes
• SLS
• DLS
Membrane osmometer No
Self-interaction chromatography Yes
Analytical ultracentrifugation No
• Sedimentation velocity
• Sedimentation equilibrium
Interfacial stability Tensiometry No
Extensional rheology No
Chemical stability RPHPLC No
Bioanalyzer Yes
LC/MS No
Solution properties
Viscosity Rheometry No
DLS Yes
Phase separation Nephleometry Yes
Right angle light scattering No
Absorption spectroscopy (transmission) No
6.6 Molecular Attributes and Instabilities

that Should Be Screened Early
6.6.1 Aggregation
The propensity for a molecule to aggregate is one of the single most influential and
challenging factors for formulation development. Candidates that are inherently
resistant towards aggregation are considered superior. Nonnative aggregation is a
multistep process and can be explained by the Lumry–Eyring framework. Each
state has a different molecular configuration and energy, and each reaction step
proceeds through an energy barrier (ΔG*, difference in energy or activation free
energy). The step that has highest ΔG* is the rate-limiting step (Chi et al. 2003a).
The first step in this process, which is a reversible process, gives rise to an
aggregation-competent partially unfolded intermediate (I) state. In the second step,

the partially unfolded intermediate self-associates to from irreversible higher-order
aggregates (A). The final undesired A state can be controlled either by increasing
conformational stability of the native (N) state (ΔG) which results in a decreased
population of aggregation-competent partially unfolded intermediates (I state) and/
or by increasing the colloidal stability of the partially unfolded I state such that
self-association of this state is minimized (Chi et al. 2003b). During development,
controlling aggregation is achieved by screening for stabilizing excipients and
solution conditions to develop appropriate formulation controls. Alternatively, such
stabilization can be achieved by either selecting or designing conformationally and/
or colloidally stable candidates. Mutations that increase conformational stability
increase the free energies of the unfolded and transition states relative to the native
state. Increasing the free energy of transition state relative to native results in a
larger ΔG*, which shifts the equilibrium towards native state. The net result is a
decreased population of aggregation-competent partially unfolded state.
Both conformational and colloidal stability have an indirect but significant effect
on surface- and interface-mediated properties. For example, adsorption of protein to
various surfaces and interfaces during process, fill/finish, and storage can not only
lead to significant reduction in yield but quite often can also induce structural
perturbation in a protein resulting in reduced binding and activity as well as aggre-
gation and particle formation (Sethuraman and Belfort 2005). Similarly shear stress
on a protein during process and fill/finish operation along with mixing of the hydro-
phobic air–liquid interface with the bulk solution is known to induce structural per-
turbations and subsequent aggregations (Bekard et al. 2011; Thomas and Geer
2011; Biddlecombe et al. 2007; Bee et al. 2009). Typically, both surface- and
interface-mediated aggregation and particle formation is controlled during develop-
ment either by avoiding the surfaces that can have deleterious effect on a particular
protein or by designing and developing appropriate formulation controls such that
the adsorption process is avoided or approximately controlled. Alternatively, such
surface and interface effects can be reduced by increasing the conformational
stability of a protein (Karlsson et al. 2005; Norde 1986; Wendorf et al. 2004). In a
study with human carbonic anhydrase mutants, Karlsson et al. (2005) showed that
adsorption and desorption characteristics can be controlled by a protein-engineering
approach. An early investment to either design or select molecules with better con-
formational stability can potentially reduce some of the process and fill/finish
related surface- and interface-mediated effects. Such an approach (stability by
design) will not only result in a wider process and formulation development window
but can also result in development of a platform process and fill/finish conditions.
Thermal stability using differential scanning calorimetry (DSC) is a widely
acceptable measure and predictor of a protein’s conformational stability and can be
used to form a basis for screening candidate molecules under varying buffer condi-
tions (Remmele et al. 1998). New techniques, strategies, and instrumentation are
now available that provide thermal stability in a fraction of the time typically
associated with DSC and use very low amounts of material (He et al. 2010; Goldberg
et al. 2011). These methodologies include spectroscopy-based techniques coupled
with thermal ramping (reviewed in Chap. 2). These techniques can provide through-
put on the order of 96–384 samples per hour and typically only require 10s of
micrograms. Aggregation and particle formation can also be directly monitored by
techniques such as dynamic light scattering (DLS), HIAC light obscuration, micro-
flow imaging (MFI), and electrical charge-based methods (e.g., Coulter counter). Of
these methods, DLS offers the greatest flexibility for early-stage use based on
throughput (plate-based DLS) and material required. However, recent advances
have resulted in improvements in particle detection including scanning ion occlu-
sion spectroscopy (qNano), chip-based mass measurements (Archimedes), and
nanoparticle tracking analysis (NanoSight). Each of these methods offers the pos-
sibility of increased sensitivity and throughput with a decrease in sample consump-
tion, resulting in more screening possibilities.
6.6.2 Chemical Stability
As shown in Table 6.1, proteins have multiple liabilities that are prone to chemical
degradation during manufacture and storage. The rate of degradation depends on
both the solution condition and the structure of the molecule (Sinha et al. 2009;
Vlasak and Ionescu 2011). Typically some of the most common liabilities like
deamidation, nonenzymatic hydrolysis, and oxidation are highlighted very early and
easily by in silico primary sequence liability analysis. Theoretically, molecules with
the fewest chemical liabilities will allow one to develop a wider process and formu-
lation design space. For molecules with identified liabilities, there will be an addi-
tional burden of determining the effect of the liabilities on product quality and
designing appropriate process and formulation controls. Unlike physical stability
studies, the slow rates of reaction associated with accessing chemical stability results
in time-consuming forced degradation studies. For example, evaluation of the effect
of deamidation on function and stability requires forced degradation studies at ele-
vated temperature and pH, followed by liquid chromatography–mass spectrometry
(LCMS) and subsequent selected ion monitoring (SIM). Similarly fragmentation
propensity can be determined either using size-based separation techniques or side-
chain chemistry-based analytical methods like reversed-phase high-performance
liquid chromatography (RPHPLC). Even though HT analytics are available to fol-
low fragmentation, the slow rate of reaction of forced degradation studies that can
accurately predict nonenzymatic fragmentation under storage conditions typically
requires time-consuming incubation studies in the range of 4–6 weeks.
6.6.3 Solution Properties at High Protein Concentrations
Other important developability criteria that are often neglected during early screen-
ing are the solution properties of a molecule at high concentration. Solution properties
such as solubility, phase separation, and viscosity can often become the rate-limiting
step during processing, filtration, drug delivery, etc., as solution properties of a mol-
ecule are governed by various attractive and repulsive nonspecific intermolecular
interactions [reviewed in Saluja and Kalonia (2008) and Tadros (2011)]. Under con-
centrated solution conditions, the intermolecular distance becomes less than the
molecule’s diameter. This reduction in intermolecular distance will not only increase
the collision frequency but will also increase the duration of such an encounter
(Scherer et al. 2010; Minton 2005). Thus at high concentrations, the nature of, and
relative contributions from, various nonspecific interactions is dictated by both solu-
tion conditions and the surface properties of a molecule, including net charge,
charge distribution, and surface hydrophobicity. From a development perspective,
the consequences of such unfavorable interactions are a change in the protein’s
solution behavior potentially resulting in issues such as poor solubility, phase sepa-
ration, and/or increased viscosity (Saluja and Kalonia 2008; Chari et al. 2009).
Typically, such unfavorable interactions can be controlled during development by
modulating solution conditions like ionic strength and/or pH to develop appropriate
formulation controls (Salinas et al. 2010; Yadav et al. 2012; Liu et al. 2005), but in
some cases, depending on strength of intermolecular interactions, it may not be
possible to achieve this control. Quite often the solution conditions might have an
unintended opposite effect on various solution properties and overall stability
(Salinas et al. 2010).
The rheological properties of a molecule under a given buffer system have typi-
cally been ascertained using a cone- and plate-type instrument to determine viscos-
ity. The dimensions of the cone and plate instruments have been dramatically
reduced over the years, resulting in a correlating decrease in the volume of material
required. However, the time associated with these measurements has not fallen to
the same degree. New techniques such as viscosity–DLS (see He et al., Chap. 2, this
volume) and chip-based measurements provide an attractive alternative to the tradi-
tional cone and plate approach. Alternatively, solution properties can also be
predicted by accurate measurements of net protein–protein interactions through col-
loidal stability (Yadav et al. 2012; George and Wilson 1994; Mehta et al. 2012). The
osmotic second virial coefficient (B22), which is a thermodynamic parameter, is often
used to directly quantify overall protein–protein interactions (Chi et al. 2003a, b;
Valente et al. 2005; Payne et al. 2006). A positive B22 value indicates an overall
dominance of repulsive interactions, where as a negative B22 value indicates an over-
all attractive intermolecular interactions. The B22 of a colloidally active system can
be accurately measured by various techniques like DLS (Yadav et al. 2011), static
light scattering (SLS), analytical ultracentrifugation (AUC), osmometry, and self-
interaction chromatography (SIC) (Le Brun et al. 2010).
One of the main reasons for limited screening of these undesirable solution prop-
erties during early development is the limited availability of HT techniques. Most of
the available techniques mentioned above are low throughput in nature and also
require large amounts of material. But recently significant progress has been made
in this area, and several HT methods are available to either directly measure some
of the solution properties (reviewed in Chap. 2) or accurately predict solution
properties (Johnson et al. 2009; Bajaj et al. 2007; Sule et al. 2011; Tessier et al.
2008). Availability of such HT techniques allows one to screen and select candidate
molecules with desired surface properties or interaction potential, which is a mole-
cule that has a B22 value close to 0 under the desired solution conditions. Applying
an approach of selecting or engineering a molecule with desired surface properties,
or interaction potential, will enable the formulation scientist to focus mostly on
aggregation or other chemical degradations, thus widening the formulation develop-
ment design space.
One consequence of such an approach is that there is a potential for conflict
between the different parameters that are being simultaneously screened. For exam-
ple, if antibody affinity is found to correlate with an unfavorable biophysical char-
acteristic, such as aggregation propensity, careful consideration would need to be
given as to which characteristic would be most economic to engineer/mitigate
against the aggregation propensity. Examples of candidate therapeutic antibodies
that have been extensively reengineered to improve biophysical properties and as
such improve developability have begun to be described (Pepinsky et al. 2010).
6.7 Case Studies
6.7.1 Case Study 1
As stated in Sect. 6.2, for molecules with a precedented mechanism of action, the
focus is on enhanced safety, higher affinity to target, and increased half-life. The
efficacy and potency of a molecule can be fine-tuned by efficient mediation of effec-
tor functions (Anderson et al. 1997; Weng and Levy 2003). This case study demon-
strates how engineering of Fc variants to modulate these activities can have an
unintentional effect on a molecule’s developability and how this unintentional effect
can be screened for and detected early by applying appropriate biophysical tools. In
this case a triple mutant S239D/A330L/I332E (resulting in mAbA-3M) was intro-
duced into the CH2 domain of a human mAb (mAbA). Compared to its wild type,
these mutations resulted in an approximately 10–100 fold increase in mAbA-3M
binding to its receptor, but at a price of reduced stability (Fig. 6.1) (Oganesyan et al.
2008; Anandakumar et al. 2008). Subsequent developability-accelerated stability
studies indicated a significant increase in the aggregation rate at 40 °C (Fig. 6.2).
Based on various biophysical studies like DSC and denaturant unfolding, it was
speculated that the reduced conformational stability was due to minor structural
perturbations in the region around the CH2 domain. By solving the structure of the
mutant, Oganesyan et al. determined that the mutations resulted in opening of the
cleft in the CH2 domain of the Fc fragment (Oganesyan et al. 2008). They concluded
that the enhanced interactions with human Fcγ(gamma)RIIIA could be due to
enhanced Fc openness as well as the introduction of hydrophobic and electrostatic
interactions at the corresponding interface (Oganesyan et al. 2008). Even though
Fig. 6.1 Comparison

of conformational stability
of mAbA and its 3M mutant
(a) Overlay of the DSC
thermogram; figure adapted
from Oganesyan et al. (2008)
(b) Overlay of urea-induced
unfolding obtained
by monitoring changes
in intrinsic tryptophan
fluorescence of the protein
at a concentration of 1 mg/mL
developability studies flagged the mAbA-3M as having a problematic aggregation

potential, it was selected as the lead candidate because of its increased potency.
Investigators were able to move mAbA-3M forward because appropriate formula-
tion controls and resources were allotted based on the information obtained during
developability studies.
6.7.2 Case Study 2
Case Study 2 examines three protein candidates to ascertain if any have potential
liabilities with respect to manufacturing or formulation. Three IgG2 monoclonal
Fig. 6.2 Comparison of stability of mAbA and its 3M mutant obtained under accelerated condi-
tions. Stability of both proteins was monitored by incubating the protein at 50 mg/mL at 40 °C and
monitoring the changes in purity and aggregation by HPSEC
antibody (mAb) candidates against the same target, molecule H, molecule J, and
molecule K, were provided for assessment. Using biophysical and accelerated incu-
bation studies, these molecules were screened to see if they could adhere to a for-
mulation and processing platform.
6.7.2.1 Formulation Stability
DSC was used to monitor the thermal stability of the candidates in the formulation
at 1 mg/mL. At pH 5.2, all the candidates exhibited an apparent Fab Tm between
77 °C and 80 °C. The observation of relatively higher Fab Tm value indicated lower
stability/developability risk. This was further confirmed by an incubation study per-
formed at 70 mg/mL mAb concentration in an acetate pH 5.2 buffer. Samples were
incubated in the buffer and stored at 4 °C, 25 °C, or 37 °C for 2–4 weeks. For sam-
ples stored at 37 °C, all candidates exhibited similar increases in HMW (high
molecular weight) species (between 0.3% and 0.4% increase from t = 0) over a
period of 4 weeks. As expected, no significant change in the SE-HPLC % main peak
(MP) was observed for any of the three candidates upon storage at 4 °C for 4 weeks
or at 25 °C for 2 weeks. To simulate the pH jump upon physiological administra-
tion, candidates were subjected to pH change from pH 5.2 to ~7.0 and held for 24 h
at 37 °C. An increase in % HMW species was noted for all the candidates after 24 h
incubation at pH 7 at 37 °C as detected by SE-HPLC: for molecule H (0.8%)

followed by molecule J (0.68%) and molecule K (0.67%)
6.7.2.2 Process Stability
To test the compatibility with the possible solution conditions encountered during
purification, the candidates were exposed to representative buffers at pH values
between 3.0 and 7.4. Potential changes in thermal stability, higher-order structure,
and resulting degradation were monitored by applying various biophysical tools and
SE-HPLC analysis. At pH 7.4 as expected the DSC thermogram showed three endo-
therms corresponding to the unfolding of the CH2, Fab, and CH3 domains. However,
at pH 3.0, only a single endotherm was observed for all candidates indicating either
increased inter-domain interactions (cooperative unfolding) or unfolding of some of
the domains (even at temperatures below the starting point of the DSC analysis) at
low pH. The conformation of molecule K appeared to be the most stable, as indi-
cated by the higher Tm of the Fab domain at pH 3.5 and 5.0, followed by molecule
J and molecule H. When reversed into PBS by dialysis, no significant irreversible
change was observed for any of the three candidates, and the DSC thermograms
were superimposable.
The tertiary structure of the candidates was studied by near-UV CD. The spectra
of the three candidates exhibited similar features, with varying intensities of mean
residue ellipticity (MRE). The spectra for all candidates at pH 3.5 and 5.0 were simi-
lar suggesting that no conformational changes occurred at pH 3.5. At pH 3.0, the
near-UV CD spectrum of all candidates showed a marked loss in MRE that indicated
the protein molecules had lost a significant amount of tertiary structure at this pH.
Most of the loss in the near-UV CD signal occurred in the region between 250 nm
and 280 nm suggesting a marked change in the disulfide environment at pH 3.0. The
near-UV CD spectra of the reversibility samples in PBS overlay very well with the
protein in pH 5 reversed into PBS at pH 7.2 indicating that there were no irreversible
changes in the tertiary structure of any of the candidates induced by exposure to
lower pH. The effect of low pH on the protein conformation and its reversibility
were further evaluated by ANS-binding studies, which revealed an increase in pro-
tein surface hydrophobicity with a decrease in pH. Upon reversing the solution con-
dition to PBS, all candidates exhibited full reversibility in surface hydrophobicity.
Implications of low pH-induced structural perturbation on candidate stability
were evaluated by incubating the candidates at different pH. None of the candidates
exhibited a decrease in % main peak by SE-HPLC after 6 h of exposure to actual
solution pH values (pH 3.7, 5.0 and 7.4). Whereas, when exposed to pH 3.0, all
candidates exhibited aggregation and a loss in % main peak determined by
SE-HPLC. Molecule K exhibited the most aggregation (84%) followed by molecule
H (78%) and molecule J (77%). The results obtained by SEC were further supported
by DLS studies, which showed the presence of HMW species at pH 3.0 for all three
candidates. At pH 3.5 and 5.0, the size distribution was superimposable and the
hydrodynamic diameter (dH) was close to that of a monomer (approximately
10–11 nm), suggesting a more monomer-like species at higher solution pH. The
polydispersity index (PdI, an indicator of size heterogeneity in the sample with
lower values <0.1 suggesting a homogeneous population) also showed a decreasing
trend from pH 3.0 to 5.0. For samples reversed into PBS, a decrease in size distribu-
tion was not observed in the pH 3.0 samples indicating that some irreversible
changes were produced. However, full reversibility was observed for samples
exposed to pH values of 3.5 and 5 and then dialyzed into PBS, consistent with
observations made from the other biophysical techniques tested during this study.
This result was expected since the samples at pH 3.5 and 5.0 exhibited a homoge-
neous size distribution.
Biophysical and stability data obtained during early developability screening
indicated all three candidates were low-risk molecules from a developability/
manufacturability perspective and adhered to the formulation and processing
platforms.
6.8 Conclusion
Recent advances in the understanding of protein structure, stability, and function

properties combined with advances in equipment technology are enabling more
comprehensive and accurate developability and manufacturability screening during
early research stages. This early screening is starting to result in the selection of
molecules more closely fitting the desired target product profile (i.e., CMC proper-
ties that maintain desired biological properties), with the promise of reduced
resources and faster progression through the development phase. This process does
not entirely eliminate the risk associated with taking a selected candidate forward
(as biological properties are typically of higher priority than CMC properties); apart
from minimizing the risk of taking forward a candidate that is difficult to develop,
such an approach of selecting low-risk molecule can also result in enhanced product
quality and increase end-user convenience.
Depending upon the target product profile, screening tools for CMC properties
should be selected and used based on the value added from a cost-benefit and risk-
benefit analysis. In an increasingly competitive global landscape, a wiser and more
focused application of novel, cutting-edge biophysical technologies can provide
sustainable benefits, such as early identification of developability fingerprints to
allot appropriate resources and streamlining of development and manufacturing
activities. Realizing these benefits can result in a significant increase in overall
R&D output and productivity. Similar to the strategy employed in the small-
molecule world, the biotech industry, in general, is starting a gatekeeper approach
by making the screening of developability properties an integral part of preclinical
drug discovery. With a better understanding of structure, stability, and activity rela-
tionships and successful integration of in silico and high-throughput experimental
tools, the industry should evolve from a “gatekeeper” (or screening) strategy to “a
stability by design” approach.
Acknowledgements Authors wish to thank Nancy Craighead at MedImmune for critical review
of this chapter.
References
Ackerman MJ, Clapham DE (1997) Ion channels–basic science and clinical disease. N Engl J Med
336(22):1575–1586
Adams CP, Brantner VV (2010) Spending on new drug development1. Health Econ
19(2):130–141
Anandakumar R, Harn NR, Sathish HA, LeachWL, Oliver CN, Bishop SM (2008) American
Association of Pharmaceutical Scientist, National Biotech Conference, 2008
Anderson DR, Grillo-Lopez A, Varns C, Chambers KS, Hanna N (1997) Targeted anti-cancer
therapy using rituximab, a chimaeric anti-CD20 antibody (IDEC-C2B8) in the treatment of
non-Hodgkin’s B-cell lymphoma. Biochem Soc Trans 25(2):705–708
Bajaj H, Sharma VK, Kalonia DS (2007) A high-throughput method for detection of protein self-
association and second virial coefficient using size-exclusion chromatography through simul-
taneous measurement of concentration and scattered light intensity. Pharm Res
24(11):2071–2083
Batista FD, Neuberger MS (1998) Affinity dependence of the B cell response to antigen: a thresh-
old, a ceiling, and the importance of off-rate. Immunity 8(6):751–759
Bee JS, Stevenson JL, Mehta B, Svitel J, Pollastrini J, Platz R, Freund E, Carpenter JF, Randolph
TW (2009) Response of a concentrated monoclonal antibody formulation to high shear.
Biotechnol Bioeng 103(5):936–943
Bekard IB, Asimakis P, Bertolini J, Dunstan DE (2011) The effects of shear flow on protein struc-
ture and function. Biopolymers 95(11):733–745
Biddlecombe JG, Craig AV, Zhang H, Uddin S, Mulot S, Fish BC, Bracewell DG (2007)
Determining antibody stability: creation of solid–liquid interfacial effects within a high shear
environment. Biotechnol Prog 23(5):1218–1222
Boder ET, Midelfort KS, Wittrup KD (2000) Directed evolution of antibody fragments with mon-
ovalent femtomolar antigen-binding affinity. Proc Natl Acad Sci USA 97(20):10701–10705
Carter PJ (2011) Introduction to current and future protein therapeutics: a protein engineering
perspective. Exp Cell Res 317(9):1261–1269
Cellmer T, Bratko D, Prausnitz JM, Blanch HW (2007) Protein aggregation in silico. Trends
Biotechnol 25(6):254–261
Chan AC, Carter PJ (2010) Therapeutic antibodies for autoimmunity and inflammation. Nat Rev
Immunol 10(5):301–316
Chari R, Jerath K, Badkar AV, Kalonia DS (2009) Long- and short-range electrostatic interactions
affect the rheology of highly concentrated antibody solutions. Pharm Res 26(12):2607–2618
Chennamsetty N, Voynov V, Kayser V, Helk B, Trout BL (2009a) Design of therapeutic proteins
with enhanced stability. Proc Natl Acad Sci USA 106(29):11937–11942
Chennamsetty N, Helk B, Voynov V, Kayser V, Trout BL (2009b) Aggregation-prone motifs in
human immunoglobulin G. J Mol Biol 391(2):404–413
Chi EY, Krishnan S, Randolph TW, Carpenter JF (2003a) Physical stability of proteins in aqueous
solution: mechanism and driving forces in nonnative protein aggregation. Pharm Res
20(9):1325–1336
Chi EY, Krishnan S, Kendrick BS, Chang BS, Carpenter JF, Randolph TW (2003b) Roles of
conformational stability and colloidal stability in the aggregation of recombinant human granu-
locyte colony-stimulating factor. Protein Sci 12(5):903–913
Dall’Acqua WF, Cook KE, Damschroder MM, Woods RM, Wu H (2006a) Modulation of the
effector functions of a human IgG1 through engineering of its hinge region. J Immunol
177(2):1129–1138
Dall’Acqua WF, Kiener PA, Wu H (2006b) Properties of human IgG1s engineered for enhanced
binding to the neonatal Fc receptor (FcRn). J Biol Chem 281(33):23514–23524
DiMasi JA, Grabowski HG (2007) The cost of biopharmaceutical R&D: is biotech different?
Manag Decis Econ 28(4–5):469–479
Famm K, Hansen L, Christ D, Winter G (2008) Thermodynamically stable aggregation-resistant
antibody domains through directed evolution. J Mol Biol 376(4):926–931
Finch DK, Sleeman MA, Moisan J, Ferraro F, Botterell S, Campbell J, Cochrane D, Cruwys S,
England E, Lane S, Rendall E, Sinha M, Walker C, Rees G, Bowen MA, Schneider A, Liang
M, Faggioni R, Fung M, Mallinder PR, Wilkinson T, Kolbeck R, Vaughan T, Lowe DC (2011)
Whole-molecule antibody engineering: generation of a high-affinity anti-IL-6 antibody with
extended pharmacokinetics. J Mol Biol 411(4):791–807
Foote J, Eisen HN (1995) Kinetic and affinity limits on antibodies produced during immune
responses. Proc Natl Acad Sci USA 92(5):1254–1256
Fowler SB, Poon S, Muff R, Chiti F, Dobson CM, Zurdo J (2005) Rational design of aggregation-
resistant bioactive peptides: reengineering human calcitonin. Proc Natl Acad Sci USA
102(29):10105–10110
Gebauer M, Skerra A (2009) Engineered protein scaffolds as next-generation antibody therapeu-
tics. Curr Opin Chem Biol 13(3):245–255
George A, Wilson WW (1994) Predicting protein crystallization from a dilute solution property.
Acta Crystallogr D Biol Crystallogr 50(Pt 4):361–365
niques: role of conformational and colloidal stability. J Pharm Sci 100:1306–1315
Hageman MJ (2006) Solubility, solubilization and dissolution in drug delivery during lead optimi-
zation. Springer, New York
He F, Becker GW, Litowski JR, Narhi LO, Brems DN, Razinkov VI (2010) High-throughput
dynamic light scattering method for measuring viscosity of concentrated protein solutions.
Igawa T, Tsunoda H, Kuramochi T, Sampei Z, Ishii S, Hattori K (2011) Engineering the variable
region of therapeutic IgG antibodies. MAbs 3(3):243–252
Jespers L, Schon O, Famm K, Winter G (2004) Aggregation-resistant domain antibodies selected
on phage by heat denaturation. Nat Biotechnol 22(9):1161–1165
Johnson DH, Parupudi A, Wilson WW, DeLucas LJ (2009) High-throughput self-interaction chro-
matography: applications in protein formulation prediction. Pharm Res 26(2):296–305
Karlsson M, Ekeroth J, Elwing H, Carlsson U (2005) Reduction of irreversible protein adsorption
on solid surfaces by protein engineering for increased stability. J Biol Chem
280(27):25558–25564
Kayser V, Chennamsetty N, Voynov V, Forrer K, Helk B, Trout BL (2011) Glycosylation influ-
ences on the aggregation propensity of therapeutic monoclonal antibodies. Biotechnol
J 6(1):38–44
Kola I, Landis J (2004) Can the pharmaceutical industry reduce attrition rates? Nat Rev Drug
Discov 3(8):711–715
Krapp S, Mimura Y, Jefferis R, Huber R, Sondermann P (2003) Structural analysis of human
IgG-Fc glycoforms reveals a correlation between glycosylation and structural integrity. J Mol
Biol 325(5):979–989
Kubota T, Niwa R, Satoh M, Akinaga S, Shitara K, Hanai N (2009) Engineered therapeutic anti-
bodies with improved effector functions. Cancer Sci 100(9):1566–1572
Lauer TM, Agrawal NJ, Chennamsetty N, Egodage K, Helk B, Trout BL (2012) Developability
index: a rapid in silico tool for the screening of antibody aggregation propensity. J Pharm Sci
101(1):102–115
Le Brun V, Friess W, Bassarab S, Muhlau S, Garidel P (2010) A critical evaluation of self-interaction
chromatography as a predictive tool for the assessment of protein-protein interactions in pro-
tein formulation development: a case study of a therapeutic monoclonal antibody. Eur J Pharm
Biopharm 75(1):16–25
Liu J, Nguyen MD, Andya JD, Shire SJ (2005) Reversible self-association increases the viscosity
of a concentrated monoclonal antibody in aqueous solution. J Pharm Sci 94(9):1928–1940
Lundstrom K (2005) Structural genomics of GPCRs. Trends Biotechnol 23(2):103–108
Ma P, Zemmel R (2002) Value of novelty? Nat Rev Drug Discov 1(8):571–572
Mehta CM, White ET, Litster JD (2012) Correlation of second virial coefficient with solubility for
proteins in salt solutions. Biotechnol Prog 28(1):163–170
Mimura Y, Church S, Ghirlando R, Ashton PR, Dong S, Goodall M, Lund J, Jefferis R (2000) The
influence of glycosylation on the thermal stability and effector function expression of human
IgG1-Fc: properties of a series of truncated glycoforms. Mol Immunol 37(12–13):697–706
Minton AP (2005) Influence of macromolecular crowding upon the stability and state of associa-
tion of proteins: predictions and observations. J Pharm Sci 94(8):1668–1675
Naylor J, Beech DJ (2009) Extracellular ion channel inhibitor antibodies. Open Drug Discov
J 1:36–42
Nelson AL, Dhimolea E, Reichert JM (2010) Development trends for human monoclonal antibody
therapeutics. Nat Rev Drug Discov 9(10):767–774
Norde W (1986) Adsorption of proteins from solution at the solid–liquid interface. Adv Colloid
Interface Sci 25(4):267–340
Oganesyan V, Damschroder MM, Leach W, Wu H, Dall’Acqua WF (2008) Structural characteriza-
tion of a mutated, ADCC-enhanced human Fc fragment. Mol Immunol 45(7):1872–1882
Payne RW, Nayar R, Tarantino R, Del Terzo S, Moschera J, Di J, Heilman D, Bray B, Manning
MC, Henry CS (2006) Second virial coefficient determination of a therapeutic peptide by self-
interaction chromatography. Biopolymers 84(5):527–533
Pepinsky RB, Silvian L, Berkowitz SA, Farrington G, Lugovskoy A, Walus L, Eldredge J, Capili
A, Mi S, Graff C, Garber E (2010) Improving the solubility of anti-LINGO-1 monoclonal
antibody Li33 by isotype switching and targeted mutagenesis. Protein Sci 19(5):954–966
Prassler J, Thiel S, Pracht C, Polzer A, Peters S, Bauer M, Norenberg S, Stark Y, Kolln J, Popp A,
Urlinger S, Enzelberger M (2011) HuCAL PLATINUM, a synthetic Fab library optimized for
sequence diversity and superior performance in mammalian expression systems. J Mol Biol
413(1):261–278
Presta LG (2008) Molecular engineering and design of therapeutic antibodies. Curr Opin Immunol
20(4):460–470
Remmele RL Jr, Nightlinger NS, Srinivasan S, Gombotz WR (1998) Interleukin-1 receptor (IL-
1R) liquid formulation development using differential scanning calorimetry. Pharm Res
15(2):200–208
Salinas BA, Sathish HA, Bishop SM, Harn N, Carpenter JF, Randolph TW (2010) Understanding
and modulating opalescence and viscosity in a monoclonal antibody formulation. J Pharm Sci
99(1):82–93
Saluja A, Kalonia DS (2008) Nature and consequences of protein-protein interactions in high
protein concentration solutions. Int J Pharm 358(1–2):1–15
Scherer TM, Liu J, Shire SJ, Minton AP (2010) Intermolecular interactions of IgG1 monoclonal
antibodies at high concentrations characterized by light scattering. J Phys Chem B
114(40):12948–12957
Sethuraman A, Belfort G (2005) Protein structural perturbation and aggregation on homogeneous
surfaces. Biophys J 88(2):1322–1333
Sinha S, Zhang L, Duan S, Williams TD, Vlasak J, Ionescu R, Topp EM (2009) Effect of protein
structure on deamidation rate in the Fc fragment of an IgG1 monoclonal antibody. Protein Sci
18(8):1573–1584
Stewart R, Thom G, Levens M, Guler-Gane G, Holgate R, Rudd PM, Webster C, Jermutus L, Lund
J (2011) A variant human IgG1-Fc mediates improved ADCC. Protein Eng Des Sel
24(9):671–678
Sule SV, Sukumar M, Weiss WF, Marcelino-Cruz AM, Sample T, Tessier PM (2011) High-
throughput analysis of concentration-dependent antibody self-association. Biophys
J 101(7):1749–1757
Tadros T (2011) Interparticle interactions in concentrated suspensions and their bulk (rheological)
properties. Adv Colloid Interface Sci 168(1–2):263–277
Tessier PM, Jinkoji J, Cheng YC, Prentice JL, Lenhoff AM (2008) Self-interaction nanoparticle
spectroscopy: a nanoparticle-based protein interaction assay. J Am Chem Soc 130(10):
3106–3112
Thomas CR, Geer D (2011) Effects of shear on proteins in solution. Biotechnol Lett
33(3):443–456
Valente JJ, Payne RW, Manning MC, Wilson WW, Henry CS (2005) Colloidal behavior of proteins:
effects of the second virial coefficient on solubility, crystallization and aggregation of proteins
in aqueous solution. Curr Pharm Biotechnol 6(6):427–436
Vlasak J, Ionescu R (2011) Fragmentation of monoclonal antibodies. MAbs 3(3):253–263
Wendorf JR, Radke CJ, Blanch HW (2004) Reduced protein adsorption at solid interfaces by sugar
excipients. Biotechnol Bioeng 87(5):565–573
Weng WK, Levy R (2003) Two immunoglobulin G fragment C receptor polymorphisms indepen-
dently predict response to rituximab in patients with follicular lymphoma. J Clin Oncol
21(21):3940–3947
Yadav S, Scherer TM, Shire SJ, Kalonia DS (2011) Use of dynamic light scattering to determine
second virial coefficient in a semidilute concentration regime. Anal Biochem 411(2):292–296
Yadav S, Shire SJ, Kalonia DS (2012) Viscosity behavior of high-concentration monoclonal anti-
body solutions: correlation with interaction parameter and electroviscous effects. J Pharm Sci
101(3):998–1011
Chapter 7
Application of Biophysics in Formulation,
Process, and Product Characterization:
Selected Case Studies
Satish K. Singh, Qin Zou, Min Huang, and Muralidhara Bilikallahalli
7.1 Introduction
Biophysical techniques are essential tools in the development and characterization

of biotherapeutics. The information gleaned from these techniques about the
(higher-order) structure of the molecule provide insight into its behavior as process
and formulation conditions are changed, thus helping to identify the optimum con-
ditions for the molecule. Earlier chapters in this book have described the most
commonly used techniques such as circular dichroism (CD), fluorescence, FTIR
spectroscopy, analytical ultracentrifugation (AUC), and asymmetric flow field-flow
fractionation (AFFFF). In this chapter, we illustrate the utilization of these and
some other novel (variations of) techniques in various aspects of biotherapeutics
development.
7.2 Case Study 1: Biophysical Analysis

for Comparability Exercise
The program for development of a biotherapeutic from discovery to commerciali-

zation, more often than not, involves changes to various parts of the manufactur-
ing process and/or formulation. Changes to the process and product are also made
post approval. In order to bridge quality, safety, and efficacy between product
S.K. Singh () • Q. Zou • M. Bilikallahalli

Pfizer Inc., 700 Chesterfield Parkway West, Chesterfield, MO 63017, USA
e-mail: [email protected]; [email protected]
M. Huang
Pfizer Inc., One Burtt Road, Andover, MA 01810, USA
148 S.K. Singh et al.
manufactured before and after such changes, comparability studies are performed
(ICH Q5E 2005). In order to show that a product is comparable before and after
the changes, they are required to have a high degree of similarity in their physico-
chemical and biological properties. Biophysical characterization is an important
component of comparability studies to demonstrate similarity in higher-order
structure of the product. For this purpose, low-resolution structural information is
obtained through spectroscopic techniques such as far- and near-UV circular
dichroism (CD) spectroscopy, fluorescence spectroscopy, and Fourier-transform
infrared (FTIR) spectroscopy. These methods provide a signature for the second-
ary and tertiary structure of the molecule. Hydrodynamic analyses such as light-
scattering spectroscopy, AUC, and AFFFF can provide information about
quaternary structure. The case study below on an IgG2 antibody is presented
wherein product characterization by various biophysical methods was used to
explain the observed differences before and after a process-scale and formulation
change and to establish criteria for comparability.
7.2.1 Methods
Far- and near-UV CD spectra were collected on two drug substance lots with pilot-
scale formulation (Lots 1 and 2) and eight other lots with full-scale formulation
(Lots 3–10). For heat denaturation experiments, CD spectra were taken every 5°C
from 25 to 75°C. Three separate denaturation experiments were done for the assess-
ment of experimental uncertainty. All spectra were corrected with matched buffer
blanks. The spectral similarity was assessed by Similarity Match analysis in TQ
Analyst (Thermo Scientific). The DSC thermogram was generated at a scan rate of
1°C/min, and the molar excess heat capacity was calculated and deconvoluted to a
non-two-state model. Tryptophan fluorescence spectra were collected using an exci-
tation wavelength of 290 nm. All spectra were corrected with matched buffer blank.
For the Analytical ultracentrifugation sedimentation velocity (AUC-SV) experi-
ments, the samples were diluted to 0.5 mg/mL with the matched buffer and spun at
45,000 rpm. The absorbance data at 280 nm was collected and analyzed using Sedfit
(Schuck 2000). For details on the methods used, including the underlying theory,
please see Chap. 3 in this volume.
7.2.2 Results
Simple visual examination of both the far- and near-UV CD spectra in Fig. 7.1
shows differences between the pilot-scale and full-scale materials. Differences in
far-UV CD spectra (Fig. 7.1a) are mainly at 215–218 nm, suggesting alteration in
β (beta)-sheet content. The near-UV CD spectra (Fig. 7.1a) differ largely at around
290 nm, mainly attributable to the change in chiral environment around tryptophan
7 Application of Biophysics in Formulation, Process, and Product Characterization… 149
Fig. 7.1 Far-UV (a) and near-UV (b) CD spectra of pilot-scale and full-scale drug substance lots
in clinical and commercial formulations, respectively. Lots 1 and 2 are pilot-scale lots; Lots 3–10
are full-scale lots. Lots 10 and 10* are the reference material in full-scale/commercial and pilot-
scale/clinical formulation, respectively (see Table 7.1)
residues (Strickland 1974; Kelly and Price 2000). Intrinsic tryptophan fluorescence
spectroscopy analysis also shows difference between pilot-scale and full-scale
materials (Fig. 7.2), which is consistent with the difference seen in the near-UV
CD spectra. However, the change in fluorescence is not easily interpreted. The
Fig. 7.2 Tryptophan emission fluorescence spectra of all different lots excited at 295 nm.
See Fig. 7.1 or Table 7.1 for lot information
bandwidth of the emission spectrum for the pilot-scale formulation is relatively

wider than that for the full-scale formulation, which suggests that some trypto-
phan residues might have greater exposure to the solvent in the former. Additional
information about the changes in higher-order structure comes from the AUC-SV
analysis where a sharp main peak at a reasonable frictional coefficient of f/f0
(=1.6) is seen for the commercial formulation, whereas the main species in the
pilot-scale formulation shows a wider peak at larger sedimentation coefficient and
a lower frictional coefficient f/f0 (=1.3) (Fig. 7.3). This suggests that the pilot-
scale product main peak could contain a molecule with a different conformation,
presumably less elongated, or that the sample has a heterogeneous population of
different molecular species due to rapid self-association that could decrease the
weight-averaged frictional coefficient (Dam and Schuck 2000). This can be fur-
ther confirmed by a concentration-dependent AUC-SV analysis (not discussed
here; see Case Study 2).
To quantitatively evaluate the similarity between the CD spectra, the Similarity
Match analysis implemented in TQ Analyst™ software (Thermo Scientific) was
applied to both far-UV and near-UV CD spectra using the spectra of the reference
material as the standards. The algorithm uses Gram–Schmidt orthogonalism to
extract the spectral information from the standard spectra and subtracts that from
the test spectra. The residual spectra are then evaluated for any spectral information
Fig. 7.3 AUC-SV profiles of a pilot-scale/clinical formulation lot (Lot 2) and a full-scale/
commercial formulation lot (Lot 10)
for comparison. This application provides a single number for spectral comparison
and supplements the qualitative visual assessment. The use of this commercial
algorithm also eliminates tedious and error-prone spreadsheet calculations pre-
sented previously (Zou and Luo 2010). The calculated match values for the ten
different lots versus the reference standard (Lot 10) are listed in Table 7.1; the dif-
ference between pilot-scale and full-scale lots is in agreement with the visual
inspection. To establish a possible cut-off value for the spectral difference between
“similar” and “dissimilar,” comparison among a series of spectra generated by heat
denaturation was conducted (Fig. 7.4a, b). The spectra at the temperatures lower
than the onset denaturation temperature (60°C) as determined by DSC (Fig. 7.5)
are compared pair-wise using the similarity analysis. The resulting similarity
match values are subject to the t-test at 95% confidence level to deduce the lower
limit of the similarity match value used as the spectral similarity cut-off values
(87.27% and 96.50% for far- and near-UV CD spectra, respectively). As shown in
Table 7.1, only one lot (Lot 6) in the full-scale material is not similar to the refer-
ence standard in the far-UV CD analysis. However, there is no difference in its
near-UV CD spectrum. Careful examination of the far-UV CD spectrum of this lot
reveals a high level of noise in this data, and therefore, the spectrum does not have
as good a signal to noise ratio, confounding the comparison to the other samples.
However, the similarity analysis is sensitive enough to pick up small differences in
Fig. 7.4 Far- (a) and near-UV (b) CD spectra of the IgG2 antibody at different temperatures
the spectra and should serve as a good tool for comparability studies. Other type of
data analyses such as data deconvolution for far-UV CD spectra can provide the
estimate of relative abundance of different secondary structures (Sreerama and
Woody 2000). However, it is judged to be not sufficiently sensitive to be used in
comparability studies (Zou and Luo 2010).
Fig. 7.5 The DSC thermogram of the reference standard (Lot 10). The traces are the fits from the
non-two-state model are shown
Table 7.1 Quantitative similarity analysis for pilot-scale and full-scale drug substance lots in
clinical and commercial formulations, respectively
Similarity match Similarity match
Drug substance value for far UV value for near UV
Lot # /formulation CD (%) CD (%)
10 (reference Full-scale/commercial 100 100
standard)
1 Pilot-scale/clinical 64.16 86.12
2 Pilot-scale/clinical 61.51 86.31
3 Full-scale/commercial 93.06 99.47
6 Full-scale/commercial 85.98a 99.60
10* (reference Pilot-scale/clinical 58.88 87.30
standard in clinical
formulation)
The values in bold are below the limits obtained from the heat denaturation study (87.27% and
96.50% for far- and near-UV CD, respectively)
a
Difference attributable to poor quality of spectrum collected
A change in the higher-order structure of a biotherapeutic has potential to impact

activity and stability. However, the structural perturbation could be reversible, in
which case the risk to potency and safety would be greatly reduced. The reversibil-
ity of higher-order structural changes was demonstrated by transferring the full-
scale reference material to pilot-scale formulation (Lot 10*) and comparing the
far- and near-UV CD spectra to the original full-scale formulation (Fig. 7.1a, b). It
is evident that changes in both secondary and tertiary structures are reversible, both
by spectral overlay and similarity match analysis. The similarity values of Lot 10*
are very different from Lot 10. However, a direct comparison between Lot 1 and Lot
10* showed a higher similarity value for far-UV CD spectra (92.22%) in contrast to
the much lower similarity value between Lot 10 and Lot 10* (58.88%) (Table 7.1).
For near-UV CD spectra, the similarity between Lot 1 and Lot 10* is 99.00% while
that between Lot 10 and Lot 10* is 87.30%, much lower than the similarity limit
(96.50%) obtained from the heat denaturation study. This formulation-related
structural reversibility was consistent with the observation that the activity is not
affected by the formulation change (data not shown).
7.2.3 Conclusions
An in-depth biophysical characterization with quantitative criteria enhances the

utility of the comparability exercise. This not only provides information needed for
product comparability but also helps to assess the impact of differences should they
be observed during the comparability exercise.
7.3 Case Study 2: Self-Association and Aggregation

of an Fc-Fusion Protein
Comprehensive characterization of a biotherapeutic before it goes into clinical trials

and later for commercialization is an expectation from both the industry and regula-
tory bodies (Kozlowski and Swann 2006). The characterization of biotherapeutics
can be generally divided into three areas: (1) biochemical characterization for pri-
mary sequence, chemical composition, and post-translational modification; (2) bio-
physical characterization for higher-order structure, association state in solution,
and ligand-binding mechanism; and (3) biological characterization for in vitro
activity, ex vivo potency, and mode of action. A thorough characterization effort
provides essential knowledge needed for the development of analytical methods
that are used to evaluate the critical quality attributes during manufacture. One of
the critical quality attributes for biotherapeutics is the level and type of aggregates,
which arise from intermolecular interactions and are potentially linked to immuno-
genicity. Size-exclusion chromatography (SEC) is the most common method to
Fig. 7.6 SEC chromatogram of representative stability samples of an Fc-fusion protein
quantify aggregates, but the upper limit of size range is limited to approximately
50 nm. Thus, this technique can assess oligomers but misses the larger aggregates
in the submicron and micron range (den Engelsman et al. 2011). Limitations of SEC
in terms of accuracy and specificity, when evaluating aggregates, have been dis-
cussed elsewhere (Philo 2009; Carpenter et al. 2010). Orthogonal methods such as
AUC and AFFFF are therefore recommended to supplement the SEC results.
The second case study presented in this chapter involves an Fc-fusion protein
which showed significant aggregation during storage at accelerated and stressed
conditions as measured by SEC. Application of AUC in two modes (sedimentation
velocity and gravitational sweep) provided some interesting insights into the self-
association behavior and impact of salt in the mobile phase for SEC.
7.3.1 Methods
The SEC analysis was performed using a TSKgel G3000SW column (Tosoh
Bioscience) with a mobile phase buffer of 200 mM sodium phosphate, 450 mM
NaCl, and pH 6.5. The elution profile was monitored at 214 nm. A sample volume
of 20 μL at 0.5 mg/mL was injected. AUC sedimentation velocity was run at
20,000 rpm and absorbance data at 280 nm was collected. The data were analyzed
using Sedfit (Schuck 2000). For the concentration-dependent study, samples at three
different concentrations were analyzed at 50,000 rpm, and the data were fitted using
Sedphat for reversible association (Schuck 2003). Gravitational sweep AUC
Fig. 7.7 AUC-SV sedimentation coefficient distribution for the stability samples of an Fc-fusion
protein
(AUC-gs) was performed between 3,000 rpm and 50,000 rpm, and the absorbance
data at 280 nm was analyzed using SedAnal (Stafford and Braswell 2004). For more
detailed discussion of the techniques, please see Chap. 3.
7.3.2 Results
Representative SEC chromatograms from the stability samples are shown in

Fig. 7.6. AUC-SV was used to verify the SEC results. AUC-SV was run at relativly
low speeds (20,000 rpm) on the same stability samples in an effort to capture the
larger aggregates (Fig. 7.7) that would otherwise sediment too quickly. AUC-SV
was performed in a matching formulation buffer that has a lower salt content
(compared to the SEC mobile phase). Comparison between AUC-SV and SEC
results showed significant differences in relative abundance and size profiles of the
aggregates (Tables 7.2 and 7.3). The SEC method underestimated the level of aggre-
gates, especially for the samples stressed at 35°C. The peak area recovery in SEC
was assessed to be reasonable and would not account for the loss of aggregates (data
not shown). The possibility that some of the aggregates are salt dependent and could
Table 7.2 Relative abundance of aggregates by SEC analysis for the

Fc-fusion protein stability samples
Sample treatment
(1 month) Total aggregates Aggregate 1 Aggregate 2
5 °C 1.6% 1.4% 0.2%
25 °C 3.7% 1.8% 1.8%
35 °C 20.9% 2.4% 18.5%
dissociate at high salt concentrations in the mobile phase was considered. The dis-
crepancy prompted further development of a more suitable SEC method. To further
characterize the aggregates profile, gravitational sweep AUC (AUC-gs) with speeds
between 3,000 and 45,000 rpm was employed. This technique expands the dynamic
range and allows detection of even larger aggregates if present, although quantita-
tion is difficult (Stafford and Braswell 2004). The resultant profile shown in Fig. 7.8
demonstrates that some very large aggregates are present in the stressed sample.
Comparison of the size distribution from the single speed at 20,000 rpm to that from
the variable speeds showed some interesting similarities as well as differences. Both
methods captured the large aggregate at the apparent sedimentation coefficient of
55 S (78 S in standard condition, Fig. 7.8), but the gravitational sweep method
detected even larger aggregates at 1,800 S. Due to the changing speed during the
sedimentation, it is not possible to determine the true size and mass of aggregates in
Fig. 7.8. However, if treated as compact spheres, the minimal size and molar masses
of the aggregates in an AUC-gs run can be estimated, and this data is provided in
Table 7.4. The results show that a broad size range of aggregates are formed in this
biotherapeutic after being subject to stress conditions.
In addition to the broad distribution of aggregates, the AUC profile at 20,000 rpm
shows two overlapping main peaks (Fig. 7.7). This suggests the existence of a
dynamic equilibrium during the sedimentation process. To further understand the
potential for reversible self-association, a concentration-dependent AUC-SV analy-
sis was applied to the sample at 5°C. Figure 7.9 shows the size distributions at three
different concentrations. The concentration-dependent shift in sedimentation coef-
ficient of the second peak suggests that there is rapid association/dissociation equi-
librium at the timescale of the sedimentation experiment (Schuck 2003). Global
data fitting to the Gilbert–Jenkins theory implemented in Sedphat gives a dimer
dissociation constant of 40 μM and the free energy of the association as −5.9
kcal/mol, suggesting a moderate strength protein–protein interaction (Horton and
Lewis 1992). It is likely that this association is dependent on salt concentration so
that this equilibrium does not occur in the SEC mobile phase, and the SEC analysis
therefore does not detect the presence of protein self-association. It may be possible
that the non-Fc part of the fusion protein has pH- and salt-dependent self-association
behavior too.
158
Table 7.3 AUC-SV results of the stability samples for the Fc-fusion protein
Other higher
Monomer Dimer Large aggregates order aggregates
Sample
treatment Mol. wt. Sed. coeff. Mol. wt. Sed. Coeff. Mol. wt. Sed. coeff.
(1 month) (kDa) (S20oC,w) Fraction % (kDa) (S20oC,w) Fraction % (kDa) (S20oC,w) Fraction % Fraction %
5 °C 85 4.1 21.5 222 7.7 78.0 1.2a
25 °C 86 4.7 28.5 202 8.0 68.7 3.0a
35 °C 7 78.2 33.9 27.8b
a
species that are larger than the “dimer” peak
b
species that are larger than the 7 MDa peak
S.K. Singh et al.
Fig. 7.8 Gravitational sweep AUC analysis of the 35°C 1-month stability sample of an Fc-fusion
protein showing the wide range of aggregate sizes present
Table 7.4 Size and molar mass of the hypothetical compact spheres for the
species in the gravitational sweep AUC (also see Fig. 7.8)
Minimum molecular weight Minimum hydrodynamic radius
Species (MDa) for compact spheres (nm) for compact spheres
1800 S 565.6 54.7
245 S 28.4 20.2
55 S 3.0 9.6
15 S 0.4 5
4.5 S 0.1 2.7
7.3.3 Conclusions
Orthogonal characterization based upon AUC in different modes, allowed the nature
of aggregation behavior of the Fc-fusion protein to be comprehensively elucidated
in contrast to the SEC analysis. The observed discrepancy prompted an improve-
ment in the SEC method, while detailed analysis of the AUC runs showed the exis-
tence of a reversible self-association behavior in the system at the timescale of the
AUC experiment.
Fig. 7.9 Concentration-dependent AUC-SV analysis of the 5°C 1-month sample showing the
concentration-dependent shift in position of the second peak
7.4 Case Study 3: Formation of Subvisible

and Visible Particles in a Monoclonal Antibody
Drug Substance Solution
Aggregation and more specifically proteinaceous (subvisible and visible) particles

have come to be regarded as critical quality attributes of biotherapeutics since they
have the potential to impact the safety and efficacy of these products (Carpenter
et al. 2009). The mechanism for generation of (non-native) proteinaceous particles
is generally considered in terms of the nucleation and growth model where small
aggregates eventually grow large enough to become subvisible (greater than 1 μm
in apparent diameter) or larger (Weiss et al. 2009; Roberts 2006). The model envi-
sions a continuum of aggregate sizes, ranging from dimer to multimers larger than
100 μm, present in solution. However, the kinetics of the step-wise growth model
allows for certain size ranges to not be populated to any significant extent.
Visible particles are defined as those that can be seen using a regular, consistent,
visual inspection method, usually larger than 100 μm with the actual size dependent
on the method developed by the individual manufacturer. The mechanism of visible
particle formation for an IgG4 antibody was explored using dynamic light scatter-
ing and is described below.
7.4.1 Methods
Dynamic light-scattering (DLS) experiments were performed on the drug substance

lot without further dilution. The data were collected and analyzed without any correc-
tion for solvent viscosity, so only apparent hydrodynamic size was calculated. The
correlation of scattering intensity was deconvoluted using the CONTIN algorithm to
obtain the size distribution. For more details on the technique, please see Chap. 3.
7.4.2 Results
Visible particles were observed in an IgG4 monoclonal antibody drug substance

after storage under the preferred conditions. However, SEC analysis did not show
any changes in soluble aggregate levels before the emergence of the visible parti-
cles. Concern was raised as to the suitability of the SEC method being stability
indicating since it seemed to have missed any early signs of physical instability.
AUC analysis on the drug substance agreed with the SEC results showing similar
abundance and distribution of soluble aggregates (data not shown). To understand
the mechanism of formation of the visible particles and potential subvisible parti-
cles, dynamic light scattering was employed. A DLS size distribution profile of the
drug substance stability sample that contained visible particles is shown in
Fig. 7.10. There is a peak around 8 μm, suggesting the presence of very large
Fig. 7.10 Dynamic light-scattering size distribution of the particles in drug substance stability
samples. Both backscattering (173°) and forward-scattering (13°) results from unfiltered and fil-
tered samples are shown
Fig. 7.11 Real-time monitoring of particle formation in the filtered samples by DLS at 37°C
aggregates/subvisible particles. However, the actual size information is not reliable

since there is the possibility of sedimentation of the larger particles during the
measurement itself, so that they are not detected. The amount of these aggregates
could also be low enough that they are below the level of detection. Even though
data on the exact sizes is not available, the results still provide valuable informa-
tion about the presence of very large aggregates. This was demonstrated by the
absence of the peak at around 8 μm in the same stability sample after filtration
through a 0.2 μm filter (Fig. 7.10). Visual inspection of the filtered sample also
showed no visible particles. Data in Fig. 7.10 also demonstrated that the forward
scattering at a scattering angle of about 13° is more sensitive to large particles than
the backscattering at 173°. A time-course experiment with DLS in the forward-
scattering mode was then carried out at 37°C, monitoring the formation of
submicron and subvisible particles in real-time at that temperature. Filtered
stability samples were used for this study. Results show the generation of submi-
cron particles after 24 h of incubation and then the rapid appearance of larger par-
ticles (~8 μm) within another half hour (Fig. 7.11).
The simplest scenario for the inability of SEC to detect any aggregation trend
before the appearance of visible particles is that the kinetics of formation of subvis-
ible and visible particles is relatively fast. Thus, there is no accumulation of small
aggregates that would be detectable by SEC analysis (Fig. 7.12).
Fig. 7.12 A schematic diagram illustrating the mechanism for visible particle formation in the
IgG4 antibody
7.4.3 Conclusions
A real-time DLS study was able to help elucidate the mechanism of visible particle
formation and explain the lack of detection of aggregation by SEC. Once aggregates
reach a certain size range, they rapidly coalesce to form larger visible particles, thus
effectively eliminating all but a small amount of aggregates, detectable by SEC,
from the solution. No aggregates in intermediate size ranges are detectable, and
therefore, no changes are seen in the SEC results, despite the appearance of visible
particles.
7.5 Case Study 4: Probing Aggregation Propensity

of Different Batches of Biotherapeutics Using Nile
Red Fluorescence Binding Kinetics
The process of bringing a biotherapeutic to the patient (expression, purification,

formulation, distribution) subjects it to numerous stresses (Cromwell et al. 2006;
Singh 2011). The recombinant protein undergoes steps with high concentration,
high ionic strengths, pH extremes, shear stresses, air–liquid and solid–liquid inter-
faces, freeze/thaw, and long-term storage. Even as it is purified, it is possible for a
batch of material to accumulate small amounts of misfolded or partially unfolded
species that are sufficiently similar to the native desired molecule to be indistin-
guishable in any standard assays but which may carry the seeds for long-term insta-
bility through propensity for higher rates of aggregation or even chemical
degradation. The ability to predict susceptibility to aggregation upon long-term sta-
bility between various batches of material by detecting such aggregation precursor
species would be a very useful tool for the characterization of biotherapeutics.
Nile red is a hydrophobic dye that is sensitive to the polarity of its environment
and binds to hydrophobic regions of proteins (Sackett and Wolff 1987; Demeule
et al. 2007). It fluoresces only upon binding to hydrophobic surfaces such as those
that are created by protein unfolding/aggregation. Its utility in detecting small
amounts of aggregation prone species was explored in this study. It was found that
utilizing Nile red fluorescence in the conventional manner was not sufficiently sen-
sitive, but the sensitivity could be enhanced by thermal incubation and monitoring
the kinetics of change in peak fluorescence intensity.
7.5.1 Methods
Nile red extrinsic dye-based fluorescence was carried out according to the procedure
described by (Demeule et al. 2007), with some modifications. Nile red was dissolved
in DMSO to obtain a 10 mM stock solution. Three microliter of the dye stock solution
was added to 2,997 μL of protein at 20 mg/mL for the fluorescence experiments.
Thermal incubation was conducted by preheating the cuvette to the target temperature
and then adding 3 mL of dye + protein solution into the cuvette at time zero.
An IgG2 mAb was prepared at 20 mg/mL in 20 mM histidine buffer pH 5.5, without
any other excipients or stabilizers. Samples were purposely stressed to induce aggrega-
tion including thermal treatment at 40°C (7 weeks) or 50°C (10 days), or 20 cycles of
freeze/thaw stress. The level of aggregation was measured by SEC, and the samples
were also studied by CD, intrinsic fluorescence, and bis-ANS fluorescence, apart from
the Nile red fluorescence described above. An aggregate sample was also created by
mixing aliquots of 40°C 7 weeks and 5°C control at 1:1 (v/v) ratio.
7.5.2 Results
SEC analysis of the test solutions showed fairly low and similar levels of aggre-
gates (labeled as high molecular mass species, HMMS) (see Table 7.5).
Characterization of these samples by circular dichroism (Fig. 7.13), intrinsic
fluorescence (not shown), bis-ANS fluorescence (Fig. 7.14), and Nile red fluo-
rescence (Fig. 7.15) showed little differences between them. DLS analysis did
not pick up any significant particles beyond the main peak at around 9 nm (data
not shown). However, when samples containing the Nile red dye were incubated
at 63°C (just below the onset of CH2 domain-melting peak by DSC), and moni-
tored for peak fluorescence, an interesting trend was seen. Once at temperature,
the peak Nile red fluorescence intensity of the samples increased in a linear
fashion that was distinct for the different stresses (Fig. 7.16). The freeze/thaw
cycled sample and the control sample were very similar in the evolution of the
peak fluorescence while the thermally stressed samples followed a distinctly
different pattern with a more rapid increase in the peak intensity. The results of
Table 7.5 Aggregates (% HMMS) in stressed samples of

IgG2 mAb measured by size-exclusion chromatography
Sample treatments HMMS (%)
5°C control 0.9
Freeze/thaw cycles FT20 2.0
Heat treated 40 °C 7 weeks + 5 °C 1.0
control 1 + 1 v/v
Heat treated 40 °C 7 weeks 1.1
Heat treated 50 °C 10 days 1.9
Fig. 7.13 Circular dichroism spectra of stressed samples of IgG2 mAb. Sample treatments are as
follows: 5°C control control unstressed sample, FT20 20 cycles of freeze/thaw stress, 40°C 7 weeks
thermal stress at 40°C for 7 weeks, 50°C 10 days thermal stress at 50°C for 10 days
Fig. 7.14 Extrinsic (bis-ANS) fluorescence spectra of stressed samples of IgG2 mAb. Sample treat-
ments are as follows: 5°C control control unstressed sample, FT20 20 cycles of freeze/thaw stress,
40°C 7 weeks thermal stress at 40°C for 7 weeks, 50°C 10 days thermal stress at 50°C for 10 days
Fig. 7.15 Extrinsic (Nile red) fluorescence spectra of stressed samples of IgG2 mAb. Sample treat-
ments are as follows: 5°C control control unstressed sample, FT20 20 cycles of freeze/thaw stress,
40°C 7 weeks thermal stress at 40°C for 7 weeks, 50°C 10 days thermal stress at 50°C for 10 days
Fig. 7.16 Time course of Nile red peak fluorescence for stressed samples of IgG2 mAb under incuba-
tion at 63°C. Sample treatments are as follows: 5°C control control unstressed sample, FT20 20 cycles
of freeze/thaw stress, 40°C 7 weeks thermal stress at 40°C for 7 weeks, 50°C 10 days thermal stress at
50°C for 10 days, 40°C 7 weeks 5°C control 1:1 is a 1:1 by volume mixture of the thermally stressed
and unstressed control samples. Linear regression plots are shown (see Table 7.6)
Table 7.6 Linear regression analysis for time course of Nile red peak fluorescence from stressed
samples of IgG2 mAb shown in Fig. 7.16
Sample treatments HMMS (%) Slope (h−1) Intercept R2
5°C control 0.9 0.0662 22.062 0.9465
Freeze/thaw cycles FT20 2.0 0.0633 20.377 0.9824
Heat treated 40°C 7 weeks + 5°C 1.0 0.0795 51.973 0.9963
control 1 + 1 v/v
Heat treated 40°C 7 weeks 1.1 0.0897 61.477 0.9797
Heat treated 50°C 10 days 1.9 0.1231 48.408 0.9983
a linear model fit are summarized in Table 7.6. The fit parameters do not corre-
late with the measured HMMS values, but they form two distinct groups. It is
clear from the fluorescence data in Fig. 7.15 that the aggregate (levels) or the
applied stress per se cannot be distinguished by their ability to bind the hydro-
phobic dye Nile red. However, when incubated at 63°C, it is likely that mole-
cules that have small structural perturbations and misfolds as a consequence of
the stress history of the samples, begin to unfold in the cuvette. As this occurs,
the Nile red binding and fluorescence response increases. The data in Fig. 7.16
therefore suggests that molecules that underwent freeze/thaw stress, either
aggregate or refold back to the native state when thawed. Therefore, few par-
tially unfolded or misfolded species are created by this stress, and the response
with incubation (in Fig. 7.16) is similar to that of the control sample. However,
when the mAb is thermally stressed, a variety of such species are created which
are not detectable by SEC since they are not aggregates, and are not distinguish-
able by conventional spectroscopic techniques since they are either too few and/
or too similar to the native molecule.But the structural susceptibility of these
molecules is amplified by incubating the sample at 63°C, where the presence of
the Nile red dye is able to capture this phenomenon, which would otherwise be
too small energetically to be detected by DSC (data not shown).
7.5.3 Conclusions
This fluorescent dye-based thermal-scanning method seems to be able to probe and

amplify the presence of small fractions of structurally perturbed species or species
with small amount of structural perturbations in a pool of properly folded mole-
cules. Since it is likely that the presence of perturbed species could lead to further
increases in aggregation over long-term storage at normal (e.g., 2–8°C) temperatures,
this technique offers a valuable tool to characterize various batches of biotherapeutics
for their long-term viability.
A significant limitation of dye-based techniques is that they cannot be employed
on formulated material containing surfactants, since the hydrophobic dye will inter-
act with surfactant (micelles) and fluoresce.
7.6 Case Study 5: Probing Aggregation Propensity

of Biotherapeutics Using Bis-ANS Binding
Thermodynamics and Kinetics
Detection and control of aggregation is a key aspect for the successful develop-
ment of a biotherapeutic product. Exposed hydrophobic patches either by design
or through some degree of misfolding are generally understood to act as seeds for
the nucleation step of aggregation. Such regions may be present in only a small
fraction of the population and would therefore be difficult to detect. An ability to
detect small amounts of such precursors would therefore be useful as a predictor
of long-term behavior of a particular batch of the molecule. Ability to detect small
differences in the exposed hydrophobic regions can also be of value in selecting
among candidate molecules, as an indicator of their aggregation propensity.
Following upon the above concept of exposed hydrophobic regions, it was decided
to use bis-ANS binding as a probe. The proposed hypothesis is that bis-ANS would
bind more efficiently (tighter and faster) to such hydrophobic regions.
7.6.1 Methods
Thermodynamic binding parameters were obtained using an isothermal titration

calorimeter interfaced with a computer for data acquisition and analysis using
Origin 7 software (VP-ITC, GE-Healthcare). Protein and bis-ANS samples were
degassed at 25°C before loading into the ITC cell and syringe, respectively. A typi-
cal injection schedule included the addition of 10 µL of 2 mM bis-ANS solution to
20–30 µM of protein (mAb or HSA) in the calorimetric cell with 25–28 injections
spaced at 3-min intervals. For the first injection, 3 µL of bis-ANS was added and the
corresponding data point was deleted from the analysis. The binding isotherms were
fit to a two-class binding-site model using Origin 7.
The kinetics of change in fluorescence intensity at 25°C was monitored using
stopped-flow instruments (Hi-Tech and Chirascan, Applied Photophysics). The
dead time of mixing to monitoring was about 5 ms. Proteins were mixed at 1:1 vol-
ume ratio with bis-ANS to a final stoichiometric concentration of tenfold molar
excess of ligand, to optimize the timescales. The kinetic profiles were monitored for
different timescales ranging from 200 ms to 200 s, all yielding biphasic kinetics of
bis-ANS binding. The time base was optimized for 20 s, and the data were fit to
biphasic kinetics equation using a nonlinear, least-squares ProK algorithm (Applied
Photophysics) to derive first-order rate constants.
7.6.2 Results
Three monoclonal antibodies were evaluated for their interaction with the hydro-
phobic dye bis-ANS. The titration curves are fit to the model as shown in Table 7.7.
Table 7.7 Thermodynamic parameters of bis-ANS binding to different IgG mAbs and HSA
derived from ITC data at 25°C fit to a two-class binding-site model
Protein pI N1 K1 ΔG1 ΔH1 TΔS1 N2 K2 ΔG2 ΔH2 TΔS2
IgG2 Mol A 8.6 13 11 −9.6 0.18 −9.77 22 0.12 −6.93 −0.20 6.26
IgG2 Mol B 8.0 10 20 −9.9 0.11 −9.98 18 0.16 −7.16 −1.38 5.72
IgG2 Mol C 8.3 3 10 −9.6 0.10 −9.66 17 0.19 −7.19 −1.67 5.54
HSA 4.9 11 82 −10.8 0.29 −11.0 9 0.46 −7.72 −1.91 5.81
The pI of each molecule is also listed. Note the opposite signs of ΔH and –TΔS for the two
classes of binding sites for all proteins
N = ± 1 (SD); K = 1 × 106 M (5% SD); ΔG = kcal/mol (5% SD); ΔH = kcal/mol (5% SD);
−TΔS = kcal/mol (5% SD)
HSA was also included in the study as a reference. The results show that bis-ANS
binds to the molecules with varying stoichiometry and affinity. The data was ana-
lyzed using a two-class binding-site model. Analysis shows that the first class of
binding sites are specific and entropically driven (KD1 = 0.0–0.2 µM; ΔH positive;
TΔS negative; KD1 is inverse of K1, the association constant). This is likely a result
of the interaction of the naphthalene ring structure of bis-ANS molecules with
hydrophobic surfaces/pockets of the protein. The second class of binding sites are
nonspecific, low-affinity sites (KD2 = 10–40 µM; ΔH negative; TΔS positive; KD2 is
inverse of K2, the association constant), probably a result of hydrophilic/ionic
interactions through the sulfonic acid groups. For reference, although the HSA–
bis-ANS N1 value is similar to the mAbs (despite being a smaller protein), the
equilibrium constant (KD1) is significantly larger, indicating a stronger affinity of
bis-ANS for the more hydrophobic HSA protein. Based on the affinity and stoi-
chiometry of the first binding site (N1), the aggregation propensity driven by
hydrophobicity is rank-ordered as HSA > IgG2 MolA > IgG2 MolB > IgG2 MolC.
For aggregation to occur as a consequence of the presence of hydrophobic
patches, the patches must be accessible. It was hypothesized that the patches that are
accessible would be the first to which bis-ANS would bind, and thus the molecule
with more of these accessible patches would have a faster initial binding of bis-
ANS. To distinguish between the molecules on the basis of the (extent of their)
highly accessible hydrophobic patches, binding kinetics was studied by the stopped-
flow technique. Stopped-flow measurements on the four test molecules are shown in
Fig. 7.17 with the model fit data in Table 7.8. A biphasic-binding model was used
for the data analysis. Hydrophobic HSA has the highest k1, indicating the fastest
binding rate, and also correlates with the high aggregate level (% HMMS) measured
by SEC. Among the mAbs, the low general level of aggregation did not correlate
with the kinetics. However, based on k1, the rank order of aggregation propensity is
as follows HSA > IgG2 MolA > IgG2 MolC > IgG2 MolB.
The rank ordering based on stopped-flow kinetics was found to correlate with the
long-term stability behavior of the mAbs.
Fig. 7.17 Stopped-flow binding kinetics for mAbs and HSA with bis-ANS. Raw data is compared
at 20s time scale.
Table 7.8 Bis-ANS binding kinetic parameters derived from a biphasic kinetic model fit to
stopped-flow data
HMMS
Protein (%) (SEC) k1 (s−1) k2 (s−1) A1 (RFES)a A2 (RFES)a R2
IgG2-Mol A 0.3 2.44 ± 0.02 0.048 ± 0.003 3.34 ± 0.23 15.15 ± 0.48 0.99
IgG2-Mol B 1.0 1.41 ± 0.24 0.122 ± 0.010 2.18 ± 0.11 4.49 ± 0.10 0.94
IgG2-Mol C 0.7 1.78 ± 0.09 0.093 ± 0.002 1.71 ± 0.34 40.7 ± 0.18 0.99
HSA 17.3 11.82 ± 0.35 1.32 ± 0.05 3.16 ± 0.54 7.01 ± 0.27 0.98
a
Relative fluorescence emission signal
7.6.3 Conclusions
Novel approaches such as isothermal titration calorimetry and stopped-flow to

extract more information out of extrinsic fluorescence dye-binding studies may
provide newer insight into the aggregation propensity of biotherapeutics.
7.7 Conclusions
Biophysical methods such as differential scanning calorimetry, circular dichroism,

(intrinsic) fluorescence spectroscopy, and AUC are widely applied to assess the
higher-order structure of biotherapeutics and the impact of process and formulation
stresses. The case studies in this chapter illustrate the use of a variety of biophysical
methods in comparability studies as well as in troubleshooting product or process
impact. Some novel applications of dynamic light scattering and fluorescence spec-
troscopy as well as isothermal titration calorimetry and stopped-flow are also pro-
vided. Combinations of the methods are often required to answer any given question
due to the complementary information the methods provide.
References
Carpenter JF, Randolph TW, Jiskoot W et al (2009) Overlooking subvisible particles in therapeutic
protein products: gaps that may compromise product quality. J Pharm Sci 98:1202–1205
Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJA, Middaugh CR, Winter G (2010) Potential
inaccurate quantitation and sizing of protein aggregates by size exclusion chromatography:
essential need to use orthogonal methods to assure the quality of therapeutic protein products.
J Pharm Sci 99:2200–2208
Cromwell MEM, Hilario E, Jacobsen F (2006) Protein aggregation and bioprocessing. AAPS J
8:E572–E579
Dam J, Schuck P (2000) Calculating sedimentation coefficient distributions by direct modeling of
sedimentation velocity concentration profiles. In: Johnson ML, Brand L (eds) Methods in
enzymology, vol 384. Part E: numerical computer methods. Elsevier Academic, New York,
pp 185–212
Demeule B, Gurny R, Arvinte T (2007) Detection and characterization of protein aggregates by
fluorescence microscopy. Int J Pharm 329:37–45
den Engelsman J, Garidel P, Smulders R, Koll H, Smith B, Bassarab S, Seidl A, Hainzl O, Jiskoot
W (2011) Strategies for the assessment of protein aggregates in pharmaceutical biotech product
development. Pharm Res 28:920–933
Horton N, Lewis M (1992) Calculation of the free energy of association for protein complexes.
Protein Sci 1:169–181
ICH Q5E (2005) Comparability of biotechnological/biological products. Available from
www.ICH.org
Kelly SM, Price NC (2000) The use of circular dichroism in the investigation of protein structure
and function. Curr Protein Pept Sci 1:349–384
Kozlowski S, Swann P (2006) Current and future issues in the manufacturing and development of
monoclonal antibodies. Adv Drug Del Rev 58:707–722
Roberts CJ (2006) Non-native protein aggregation: Pathways, kinetics, and shelf-life prediction.
In: Murphy RM, Tsai AM (eds) Misbehaving proteins: protein misfolding, aggregation, and
stability. Springer, New York, pp 17–46
Sackett DL, Wolff J (1987) Nile red as a polarity-sensitive fluorescent probe for hydrophobic pro-
tein surfaces. Anal Biochem 167:228–234
Schuck P (2000) Size distribution analysis of macromolecules by sedimentation velocity ultracen-
trifugation and Lamm equation modeling. Biophys J 78:1606–1619
Schuck P (2003) On the analysis of protein self-association by sedimentation velocity analytical
ultracentrifugation. Anal Biochem 320:104–124
Singh SK (2011) Impact of product-related factors on immunogenicity of biotherapeutics. J Pharm
Sci 100:354–387
Sreerama N, Woody RW (2000) Estimation of protein secondary structure from circular dichroism
spectra: comparison of CONTIN, SELCON, and CDSSTR methods with an expanded refer-
ence set. Anal Biochem 287:252–260
Stafford WF, Braswell EH (2004) Sedimentation velocity, multi-speed method for analyzing poly-
disperse solutions. Biophys Chem 108:273–279
Strickland HE (1974) Aromatic contributions to circular dichroism spectra of proteins. CRC Crit
Rev Biochem 2:113–175
Weiss WFI, Young TM, Roberts CJ (2009) Principles, approaches, and challenges for predicting
protein aggregation rates and shelf life. J Pharm Sci 98:1246–1277
Zou Q, Luo Y (2010) Biophysical characterization for product comparability. BioPharm Int Suppl
12–17
Chapter 8
Biophysical Analysis in Support of Development
of Protein Pharmaceuticals
Sreedhara Alavattam, Barthelemy Demeule, Jun Liu, Sandeep Yadav,

Mary Cromwell, and Steven J. Shire
8.1 Introduction
Proteins as pharmaceuticals, especially monoclonal antibodies, have dominated

recent new drug development and are projected to have a market value in excess of
$140 billion worldwide by 2015 (RNCOS 2012). This development of proteins as
pharmaceuticals has led to increased challenges compared to traditional small mol-
ecule drugs because of the increased complexity of the chemical and physical stabil-
ity of protein pharmaceuticals. Stability of small molecule drugs often involves a
limited number of chemically distinct functional groups which can be assayed using
perhaps one or two methods. Proteins are biopolymers made up of 20 common
amino acids and thus consist of a variety of functional groups susceptible to differ-
ent chemical degradation pathways. In addition to this increased variety of chemical
degradation pathways, the folding of the polypeptide chain is critical for maintain-
ing efficacy and safety of the protein drug. Thus, the overall stability of a protein
involves chemical degradation as well as physical degradation pathways. Although
it is well recognized that proteins can undergo aggregation, it is important to con-
sider that chemical alterations can induce conformational changes as well as pro-
mote formation of aggregates, i.e., the physical stability may be impacted by the
chemistry. The overall stability of a protein is depicted in Fig. 8.1 where the native
properly folded conformer is in equilibrium with the unfolded protein, which is
actually an ensemble of different conformers. The aggregates that form may arise
from an assembly of folded or unfolded forms, and all these forms may or may not
be chemically altered. An important point to consider is that even in the native state,
S. Alavattam • B. Demeule • J. Liu • S. Yadav • M. Cromwell • S.J. Shire (*)

Late Stage Pharmaceutical Development, Genentech, 1 DNA Way,
South San Francisco, CA 94080, USA
e-mail: [email protected]; [email protected]; [email protected];
[email protected]; [email protected]; [email protected]
174 S. Alavattam et al.
Fig. 8.1 Schematic of overall protein stability
amino acid residues, especially those accessible to solvent, are susceptible to

chemical alteration. Thus, development of a stable protein formulation requires sev-
eral assays to monitor the different chemical degradation routes as well as appropri-
ate analytics to assess the physical state including conformation and aggregation of
a protein. This necessitates the use of biophysical methods to investigate the physi-
cal properties of proteins, but this should not be done without the corresponding
analysis of chemical degradation routes since generally the chemical and physical
properties are interlinked and ultimately dictate the stability of the molecule. Often
the optimal formulation conditions for physical stability are not optimal for chemi-
cal stability, and thus the challenge is to create a formulation where the rates for
chemical and physical degradation are optimized to get the desired shelf life for a
commercial drug product.
The choice of which biophysical methods to use is dictated by the types of stud-
ies and specific requirements that support protein drug development. The type of
studies that are often done can be grouped into three general areas: (1) early screen-
ing assessments, (2) intense characterization, and (3) confirmatory studies. Early
screening studies are used to select a molecule from several versions that are being
considered for further development to initiate clinical trials. Once a molecule is
chosen, several formulations need to be analyzed to screen impact on physical sta-
bility of the protein in these conditions. These screening assessments done early in
development require high-throughput as well as small amounts of protein since
there are a large number of samples with limited amount of material. Thus, bio-
physical methods that use small amounts of protein, which are rapid and amenable
8 Biophysical Analysis in Support of Development of Protein Pharmaceuticals 175
to automation technology would be considered. Studies that provide in-depth

characterization can use more labor-intensive biophysical methods such as analyti-
cal ultracentrifugation (AUC). These studies may guide development work or pro-
vide a greater understanding of the physical properties of the protein to ensure
long-term stability. The confirmatory studies provide alternative methods to support
more rapid conventional analysis. There are several reviews that summarize the use
of biophysical analysis in the development of protein biopharmaceuticals (Samra
and He 2012; Kamerzell et al. 2011; Manta et al. 2011), and this chapter is not
intended to be an all-encompassing review of the work that has been done. In this
chapter we cite some key examples from the literature and present some examples
on how biophysical analysis has been used at Genentech supporting protein phar-
maceutical development.
8.2 Early Screening Assessments
8.2.1 Early Molecule Assessments
Early in development there may be several choices of molecule, and it would be

prohibitively expensive to move forward with all the choices. In addition, although
the molecules designed by research may be acceptable in early testing for activity,
there are many attributes such as long-term stability, bioavailability, and manufac-
turability that may dictate the final choice of molecule. Generally as a first pass the
primary structure of the molecule is used to identify particular “hot spots” for poten-
tial chemical degradation. Amino acid residues such as methionine and tryptophan
are labeled as potential sites for oxidation (Ji et al. 2009; Wang 1999), and potential
deamidation and aspartic acid isomerization sites are identified based on N and C
terminal flanking residues (Robinson and Robinson 2004). Assessment of physical
stability can be a more daunting task since it is not obvious how to link physical
stability to primary sequence information. Attempts to assess propensity for aggre-
gation and solubility (Agostini et al. 2012; Tartaglia et al. 2005, 2008) may be use-
ful biophysical computational tools for choice of molecules, but the accuracy of
such methods still needs to be verified for protein molecules in development. A nice
example of the use of computational algorithms is the recent work of Chennamsetty
et al. (2009a, b) where they used spatial aggregation propensity (SAP) using molec-
ular dynamics simulation to predict aggregation prone areas in antibodies and to
make mutants with enhanced stability. Stability of the antibodies was confirmed
using turbidity, size exclusion chromatography (SEC), and differential scanning
calorimetry (DSC) measurements. Recently Lauer et al. (2012) presented a devel-
opability index which is an in silico method for mAbs that uses SAP and the mAbs
net charge to predict aggregation. Wu et al. (2010) have shown that structure-based
engineering identified an aggregation promoting hydrophobic patch that could be
modified with a carbohydrate moiety to provide a more soluble and stable analog of
human recombinant IL13 antibody. All the above-published literature clearly

indicates that computational methods can be used in understanding the physical
behavior of mAbs, especially during candidate selection prior to entering clinical
development. Although computational methods may be useful, the actual screening
of protein molecule choices using biophysical analysis is highly desirable. As stated
earlier since the amount of protein available is limited during such early assess-
ments, assays are required that use small amounts of material. Thus, assays such as
those SEC can be used to assess levels of aggregate in solution. However, assays
that require small amounts of protein to assess particulate formation and physical
properties such as viscosity may need to be developed. This can be very challenging
especially for the development of high-protein concentration formulations such as
those being developed for therapeutic monoclonal antibodies for subcutaneous (SC)
delivery (Shire et al. 2004). As an example, viscosity measurements using standard
rheometers generally require high volumes, but alternative techniques have been
developed such as high-frequency quartz impedance methods (Saluja and Kalonia
2004, 2005; Patel et al. 2009), and several instrument companies are also offering
commercial viscometers that can handle small volumes of proteins. Although these
alternative techniques require smaller volumes, they may not be readily adaptable
for high throughput. Dynamic light scattering (DLS) has been suggested as a way to
predict the viscoelastic properties of proteins by monitoring protein–protein inter-
actions (Saluja et al. 2007; Yadav et al. 2012). In this method the diffusion constant
is determined as a function of protein concentration resulting in an interaction
parameter, kd. A negative value for kd is indicative of attractive protein–protein inter-
actions whereas a positive value represents repulsive interactions. It was shown that
net attractive interactions determined using the kd interaction parameter correlated
with high viscosities measurements using quartz impedance viscometry (Yadav
et al. 2012). Recently this method was adapted for high-throughput screening (HTS)
using a DLS plate reader (Connolly et al. 2012). The results for 29 monoclonal
antibodies showed overall a very good correlation between negative kd values and
viscosity measurements at high concentrations (Fig. 8.2) and demonstrate the utility
of this approach.
8.2.2 Formulation Screening
Gibson et al. (2011) recently have reported the use of HTS methods for protein solu-
bility using an IgG1 mAb. The authors were able to modify a PEG-induced precipi-
tation method in a 96 well format and used ultra violet-visible (UV–Vis) spectroscopy
to screen various buffer compositions and pH that help to maintain mAb solubility.
Relative solubility profiles of both chimeric and human IgG1 mAbs were deter-
mined using this HTS method. He et al. (2010a) reported the use of extrinsic fluo-
rescence in a 96-well format to detect and quantify IgG aggregation. It can be
envisioned that several other analytical techniques, such as turbidity, DLS, and vari-
ous chromatographic methods, can be utilized in the future in an HTS format to help
Fig. 8.2 Bar plots comparing kD (filled square) and mAb solution viscosity (open square) in
(a) 20 mM His-OAc, pH 5.5, (b) 30 mM His-Cl, pH 6.0, (c) 200 mM Arg-Cl, pH 5.0, and
(d) 200 mM Arg-Succ, pH 5.5. Scatter plots display correlation between kd and viscosity for the
corresponding mAbs in (e) 20 mM His-OAc, pH 5.5, (f) 30 mM His-Cl, pH 6.0, (g) 200 mM
Arg-Cl, pH 5.0, and (h) 200 mM Arg-Succ, pH 5.5. Viscosity was measured at 175 mg/mL by cone
and plate rheometry (published previously in Connolly et al. 2012)
determine protein solubility under different formulation conditions, especially

using design of experiments (DOE) principles.
As high concentration mAb formulations become more prevalent, one of the fac-
tors that will need to be addressed is screening for formulations that will mitigate
high solution viscosity in a HTS format. Since several factors such as ionic strength,
protein concentration, buffer composition, and pH influence viscosity, a multivari-
ate analysis for formulation screen in early development would be very helpful.
Recently He et al. (2011) demonstrated the application of a DOE approach with HT
formulation screening to identify the main factors affecting an IgG2 mAb thermo-
stability and solution viscosity. The authors describe the use of DLS measurements
of diffusion coefficients of small volumes of protein samples using an automated
384-well plate reader and correlated the DLS data to viscosity of concentrated pro-
tein solutions as previously reported (He et al. 2010b). In the same study, He et al.
report the thermostability of the mAbs using a relatively new technique, differential
scanning fluorimetry (DSF), instead of the traditionally used DSC, using 96-well
plates and significantly low quantities of the mAb. Several factors as well as the
interactions between factors on both the thermostability and viscosity using a 34 full
factorial design (81 conditions) and robust statistical methods were explored. Full
and reduced models considering main effects of factors and two-way interactions
were appraised, and selection of optimal formulation conditions was evaluated
based on data generated using the JMP® software. Conditions wherein the protein
maintains thermostability and has reasonable viscosity at particular pH and protein
concentrations can be easily viewed using the contour plot and profiler. In addition
to estimating the main factors affecting viscosity and thermostability, the authors
were able to apply prediction formulas to determine formulations that can help meet
predetermined values. They have successfully explored a formulation design space
consisting of different factors using a combination of HTS and statistical methods
that contribute significantly to the concepts of quality by design (QbD) and DOE
approach to formulation development. Such an analysis can be very valuable to
determine optimized formulations for Phase III and commercial development.
Zhao et al. (2010) have recently published formulation development of mAbs
using a robotic system and a high-throughput laboratory. While the individual study
setup has been published before, this is the first of a kind of work that demonstrates
the end to end of formulation screening—from sample preparation to comprehen-
sive sample analysis. Many analytical techniques including UV–vis, DLS, various
chromatographic methods, and turbidity—were performed for optimized formula-
tion screens, and the best formulations were selected in a relatively short amount of
time. The study demonstrates the value of such a high-throughput laboratory in
preformulation development of proteins, especially mAbs.
Lehermayr et al. (2011) compared various biophysical properties of several
monoclonal antibodies using a variety of techniques. Protein–protein interactions,
which define mAb properties of low and high concentration solutions, were assessed
via determination of the second virial coefficient (A2) using SLS and DLS. As mea-
suring A2 values using SLS is tedious, the authors measured the interaction param-
eter kd using DLS. A plot of kd (from DLS) and A2 (from SLS) for eight different
mAbs provided a linear fit and showed the relation as kd = 1.06 A2M − 8.9, where M
is the molecular weight of the mAb. Overall, this methodology using DLS could be
used to analyze A2 using low protein quantities and is amenable to HTS as demon-
strated in this work.
8.3 Intense Characterization Studies
8.3.1 Biophysical Techniques in Intense Characterization
Biophysical measurements that are labor-intensive such as AUC are not amenable to a
high-throughput methodology but can provide important information to guide further
development of a protein therapeutic. As an example, Lu et al. (2008) used sedimenta-
tion velocity AUC to demonstrate its usefulness in monitoring long-term stability and
molecular integrity of antibodies. Specifically, their data was used to support the notion
that a single point mutant in the hinge region of an IgG4 (S241 to P241) led to an
increase in stability of the molecule against freeze–thaw-induced aggregation.
AUC has also been used to characterize the complexes formed in vitro between
an anti IgE–IgG1 antibody and IgE (Liu et al. 1995). Monoclonal antibodies that
bind to free-circulating IgE can be used for the treatment of IgE-induced allergic
asthma, as they prevent the loading of IgE on mast cells or basophils, which can
result in release of inflammatory molecules such as leukotrienes and histamine after
exposure to an allergen (Fig. 8.3). Theoretically since there are two sites on each IgE
where anti IgE can bind and since each anti IgE molecule is bivalent, the complexes
could become very large (Fig. 8.4). Sedimentation velocity (Fig. 8.5) and equilib-
rium measurements were used to determine weight-average molecular weight and
size distribution (Fig. 8.6) to clearly show that this did not happen and that the com-
plexes that formed were of limiting size. The complexes formed are dependent on
the molar ratio of IgE:anti IgE as shown in a schematic diagram (Fig. 8.7). The
formation of these complexes dictates the pharmacokinetics of the drug therapy
since IgE has a clearance time of ~6 h whereas an IgG1, due to the binding to FcRN
neonatal receptor, has a typical half-life of ~2 weeks in serum. The complexes that
form between IgE and anti IgE take on the long half-life typically seen for an IgG1
(Fox et al. 1996). The amount of free IgE in plasma should be related to the clear-
ance rate of anti IgE:IgE complexes, free IgE, unbound anti IgE, and relative bind-
ing affinities of high-affinity receptor for IgE and that of IgE with anti IgE (Fig. 8.8).
Thus, the dose required for effective lowering of free IgE in plasma will be related
to stability of the complexes when interacting with high-affinity receptor. Since it is
possible to detect the formed IgE:anti IgE complexes using sedimentation velocity
AUC, it is possible to perform competitive binding experiments using AUC. AUC
experiments were performed with preformed anti IgE:IgE complexes at a molar
ratio of 6:1 where there is an excess of anti IgE and a predominance of a trimeric
species consisting of two anti IgE molecules bound to one IgE (Fig. 8.7). A soluble
Fig. 8.3 Mechanism of action of an anti IgE mAb in the treatment of IgE allergic-mediated d isease.
The anti IgE mAb can inhibit IgE synthesis and binds free-circulating IgE preventing the IgE from
interacting with the FcεRIα high-affinity receptors on the surface of mast cells or basophils
Fig. 8.4 The theoretical interaction of IgE (dark grey) with an anti IgE mAb (light grey) via bind-
ing of the two high-affinity Fc receptor sites on IgE
form of the high-affinity receptor, sFcεRIα, was then added at several molar ratios,
and the results (Fig. 8.9a) clearly showed that the receptor has a greater affinity for
IgE than anti IgE necessitating excess dosing of anti IgE to effectively lower free-
circulating IgE. In particular, at a molar ratio of IgE:anti IgE of 1:6, all the IgE is
incorporated into a trimeric complex at ~13.3 S, and upon addition of soluble recep-
tor at 0.1:1 of complex, there is a reduction of this 13.3 S peak as well as a slight
increase in the baseline between 8 and 9 S and an increase of the unbound anti IgE
peak at ~7 S. As more soluble receptor is added, there are additional increases in the
anti IgE peak and the appearance of a peak that is likely the receptor:IgE complex,
which was characterized as a dimer (Fig. 8.7) in earlier work using sedimentation
velocity and equilibrium AUC as well as static light scattering (SLS) (Liu et al.
1997). These data were generated using the differential sedimentation method of
Stafford (Stafford 1992), which does not take into account diffusion resulting in a
broadening of the peak and is the likely reason that at lower concentrations it is
difficult to detect a single peak representative of the receptor:IgE dimer. As an
extension of this technique, a competition binding AUC experiment was also done
Fig. 8.5 Differential sedimentation coefficient distribution of IgE (solid line) and anti IgE (dotted
line) monomers at 0.64 mg/mL (a); IgE and anti IgE complexes at various molar ratios (b and c)
in PBS at 10°C. The molar ratios of IgE:anti IgE were as follows: (b) 1:1 (solid line), 1:3 (dash-
dotted line), 1:6 (dashed line), and 1:10 (dotted line): (c) 1:1 (solid line), 3:l (dash-dotted line), 6:1
(dashed line), and 10:1 (dotted line). The sedimentation coefficients have been corrected to the
standard condition of water at 20°C. No faster moving species was observed in early scanning
(previously published in Liu et al. 1995)
using a genetically engineered version of an anti IgE, referred to as anti IgEZ, which
was designed to have a greater affinity for IgE. The competition binding analysis
shows that anti IgEZ does indeed have higher affinity since even at a ratio of soluble
receptor to IgE at 1:1 there is very little disruption of the complex (Fig. 8.9b).
These examples are typical of how biophysics can be used to study the behavior
of biotherapeutics in vitro and help in determining adequate dosing. However, rather
little information on chemical and physical stability is available once the drug is
administered to humans, since characterization under physiological conditions
Fig. 8.6 Sedimentation equilibrium analysis of IgE and anti IgE mAb complex formation in PBS
at 10°C. The weight-average molecular weights of complexes at different molar ratios were
obtained by analyzing the data from three different rotor speeds (5,000, 7,000, and 10,000 rpm) as
a single ideal species simultaneously. The error bars correspond to a 95% confidence interval
(adapted from Liu et al. 1995)
requires specialized tools. Recent advances in hardware have resulted in the capa-
bility to use fluorescence optics in AUC and have been specifically used to detect an
anti IgE molecule binding to IgE in a complex matrix such as human serum
(Demeule et al. 2009a). Two main differences were noticed for the anti IgE binding
to IgE in serum vs. PBS. The absence of the 21 S peak and the presence of an 8.7 S
peak in serum for the anti IgE complex with IgE, instead of a 7.3 S peak in PBS,
were noticeable. The absence of the 21 S peak and the different profile observed in
serum compared to PBS underlines the importance to characterize the molecules
under physiologically relevant conditions. Additionally, the absence of the 7.3 S
peak that is replaced by a 8.7 S peak further confirms the distinct behavior in serum
compared to PBS. The authors hypothesize that affinity of an anti IgE molecule
towards IgE was higher in serum compared to that in PBS and that the largest anti
IgE:IgE complex observed in serum was smaller than expected. AUC equipped with
fluorescence optics was clearly demonstrated to be useful to characterize biophar-
maceuticals under physiological conditions. Direct characterization in serum may
now allow for better drug candidate selection and should be carefully considered
during molecular assessment in early research stages.
Fig. 8.7 Schematic diagram of complex formation by IgE and anti IgE and IgE and soluble
high-affinity Fc receptor, sFcεRIα (adapted from Liu et al. 1997)
Fig. 8.8 Overall scheme for clearance of IgE (purple), anti IgE (red), and formation and clearance
of complexes at excess anti IgE vs. binding of IgE with high-affinity receptor, FcεRIα (grey), on
mast cells and basophils
8.3.2 Higher Order Structure Determinations
Alterations in conformation of proteins obviously require the use of biophysical

analytical tools. Hydrodynamic measurements such as a diffusion coefficient from
DLS or sedimentation coefficient from AUC provide such information as long as
Fig. 8.9 Differential sedimentation coefficient distribution of anti IgE, FcεRIα, and IgE:anti IgE
and IgE:FcεRIα complexes (a) and anti IgEZ, FcεRIα, and IgE:anti IgEZ and IgE:FcεRIα com-
plexes (b): assessment of competition of binding of soluble high-affinity receptor, FcεRIα, with
either preformed IgE:anti IgE or IgE:anti IgEZ complexes using AUC sedimentation velocity
the overall conformational changes are large enough to elicit a significant change in
the determined values. SEC when used with an on-line light scattering detector to
determine weight-average molecular weight can also detect mis-folded conformers
as long as the shape change is large enough to allow for separation from the prop-
erly folded protein monomer (Philo 2006).
Spectrophotometric techniques such as circular dichroism (CD) have often been
used to assess the folded state of a protein. Secondary structures (including α helices
and β sheets) can be identified in the far-UV region, 190–240 nm, of the CD spec-
trum. These ellipticity changes stem from the distinct chiral positioning of the
amide chromophores within different secondary structures (Johnson 1990). In addi-
tion, the local environment of aromatic chromophores such as tryptophan, tyrosine,
and phenylalanine gives rise to a CD signal in the near-UV region, 240–340 nm, that
can be used to infer tertiary structure changes (Johnson 1990). Thus, CD can be
used to determine if the structure of a recombinant DNA-derived protein is that
expected of “natural sourced” proteins or whether the structure is altered when
exposed to stress conditions. As an example, the far-UV spectrum of a mAb refer-
ence material from Genentech showed a minimum at 218 nm indicative of the
β-sheet character consistent with IgG1 antibodies (Brahms and Brahms 1980).
Comparison of both near- and far-UV spectra between reference material and the
stress panel shows no differences in structure by CD, as shown in Fig. 8.10, indicat-
ing that any subtle structural changes due to the applied stress (pH, oxidation, etc.)
may not be picked up in this technique. However, it has been shown recently by Li
et al. (2011) that CD can be used to monitor conformational changes in proteins
during manufacturing process conditions. Li et al. (2011) describe the effect of pH
or denaturants during purification and show a quantitative method to compare CD
spectra using the OMNIC QC compare algorithm. Near-UV CD spectrum of several
protein candidates was assessed either in sodium citrate buffer (pH 3.0) or in PBS
(pH 7.4). Careful evaluation of the near-UV CD spectrum of these candidates
Fig. 8.10 Far- and near-UV CD spectra of a monoclonal antibody after exposure to different stress
conditions
i ndicated that fewer changes were induced by low pH on candidate 1 over candidate
2, and hence the former may be more amenable to manufacturing processes at low
pH. Far-UV CD spectrum of the same two candidates indicated that both proteins
underwent the loss of the native β structure when incubated at pH 3.0; however, the
secondary structure of candidate 1 was found to be relatively more stable than can-
didate 2.
Although CD is often used on its own to probe conformation of proteins, it can
be a very valuable tool when used in conjunction with other biophysical techniques.
The self-association properties and conformation of recombinant DNA-derived
human relaxin, a pregnancy hormone, were studied by sedimentation equilibrium
analytical ultracentrifugation and CD (Shire et al. 1991). The sedimentation equilib-
rium AUC data were consistent with a monomer–dimer self-association model with
an association constant of ~6 × 105 M−1. An approximate five fold increase in weight
fraction of human relaxin monomer elicited by dilution of the protein resulted in no
change in the far-UV CD spectrum at 220 nm.
In contrast, after the same increase in weight fraction of monomer, the near-UV
circular dichroism spectra for human relaxin showed a significant decrease in the
intensity of the CD bands near 277 and 284 nm. Although human relaxin has two
tryptophan residues, the near-UV CD spectra exhibit only a broad shoulder near
295 nm rather than the strong CD bands often found for tryptophan. Moreover, there
is little change in this broad band after dilution of human relaxin to concentrations
that resulted in a five fold increase in the monomer weight fraction (Fig. 8.11).
Fig. 8.11 Near-UV circular dichroism of human relaxin at 0.5 mg/mL (solid line) and 20 μg/mL
(dotted line). Relaxin at 0.5 mg/mL was thermostated at 20 °C in a 1-cm cell, whereas relaxin at
20 μg/mL was in an unthermostated 10-cm cylindrical cuvette. The temperature in the sample
compartment was ~27°C during the data collection process. The CD data were collected at 0.25-
nm intervals at a spectral bandwidth of 0.5 nm and are the result of an average of three scans using
an average time for each single data point collection of 5 s for the 0.5 mg/mL samples and the
result of an average of ten scans using an average time for each single data point collection of 10 s
for the 20 μg/mL sample. The weight fraction of human relaxin monomer estimated from the
determined association constant by sedimentation equilibrium AUC of 100 (g/L)−1 is 0.13 at
0.5 mg/mL and 0.50 at 20 μg/mL (adapted from Shire et al. 1991)
These data suggest that dissociation of the human relaxin dimer to monomer is not
accompanied by large overall changes in secondary structure or alteration in the
average tryptophan environment, whereas there is a significant change in the tyro-
sine environment. This conclusion was affirmed by the X-ray crystal structure of
human relaxin, which crystallized as a dimer with the lone tyrosine from each
monomer at the dimer interface (Eigenbrot et al. 1991). Thus, the solution studies
were in good agreement with the crystal studies, suggesting that the determined
crystal structure is very similar to the structure of the protein in solution.
Another spectroscopic technique widely used to assess conformational changes
in proteins is Fourier transform infrared (FTIR) spectroscopy, which can be used to
probe secondary structural elements including α helices and β sheets in solution as
well as in solid-state dosage forms. These structural features appear in FTIR spectra
as broad, characteristic absorption bands in the regions 1,700–1,620 cm–1 (Amide I)
and 1,600–1,500 cm–1 (Amide II), among others (Byler and Susi 1986; Dong et al.
1990; Jackson and Mantsch 1995). These absorption bands are caused by a combi-
nation of bending and stretching vibrations of bonds along the peptide backbone.
Since the technique can be used to assess conformation of proteins in solution as
Fig. 8.12 Inverted second derivative of the amide I IR spectra of rhDNase I: (a) in aqueous
solution (EGTA treated, no incubation), (b) lyophilized-untreated power (solid line) and lyophi-
lized EGTA-treated power (dotted line) (no incubation for either sample), (c) deamidated form
(aqueous protein incubated in 1 mM CaCl2 for 120 days at 40°C), and (d) aggregated form (aque-
ous EGTA-treated protein incubated in the absence of exogenous calcium ions for 120 days at
40°C) (published previously in Chen et al. 1999)
well as in the solid state, it allows for assessments of excipients used to stabilize the
protein during drying. Studies using omalizumab clearly showed that the native
conformation of a monoclonal antibody could be preserved when a lyoprotectant
such as sucrose was added to the formulation. Freeze drying in the absence of lyo-
protectant resulted in formation of covalent aggregates, which were linked by disul-
fide bonds as shown by nonreducing and reducing SDS polyacrylamide gel
electrophoresis (SDS PAGE) (Andya et al. 2003). FTIR has also been used to inves-
tigate the conformational stability of Pulmozyme in the aqueous and solid states
(lyophilized). Pulmozyme is a recombinant DNA-derived DNase used for the treat-
ment of cystic fibrosis and requires calcium ions for stability and activity (Chen
et al. 1999). Exogenous calcium can be removed by treatment with EGTA leaving
one tightly bound calcium ion per rhDNase molecule. Analysis of the FTIR spectra
in the amide III region in either the aqueous or lyophilized state demonstrated that
removal of exogenous Ca2+ by EGTA treatment had little effect on the secondary
structure (Fig. 8.12, Table 8.1). This result for the aqueous state was confirmed
using CD that also showed that there was no large overall change in the secondary
or tertiary structure upon the removal of calcium. The primary degradation route for
rhDNase in solution is deamidation. For the EGTA-treated protein, there was also
severe covalent aggregation, e.g., formation of intermolecular disulfides facilitated
Table 8.1 FTIR analyses of various rHDNase I samplesa (published previously in Chen et al.
1999)
Secondary structure (%)
Sample α helix β sheet Otherb
Aqueous solution, with Ca 2+,c,d
21 ± 2 23 ± 3 56 ± 6
Aqueous solution, EGTA treated 20 ± 2 26 ± 2 54 ± 3
Lyophilized, with Ca2+,d 13 ± 2 41 ± 3 46 ± 2
Lyophilized, EGTA treated 14 ± 2 45 ± 3 41 ± 3
Deamidated forme 21 ± 1 25 ± 2 54 ± 2
Aggregated formf 10 ± 1 44 ± 4 46 ± 3
a
The secondary structure of rhDNase was calculated by Gaussian curve fitting the original amide
III spectra
b
Other secondary structure includes random coil and turns and extended chains
c
The aqueous solution contained 1 mM calcium chloride
d
Data from Saluja and Kalonia (2004)
e
Aqueous protein incubated in 1 mM Cacl2 for 120 days at 40°C
f
Aqueous EGTA-treated protein incubated in the absence of exogenous calcium ions for 120 days
at 40°C
by the cleavage of Cys173-Cys209, which resulted in complete loss of activity. The

aggregates had a very different secondary structure compared to the native protein.
In particular, the β-sheet band observed at ~1,620 cm−1 wave number in the amide I
second derivative spectra was increased.
For the protein lyophilized in the presence of Ca2+, there was no increase in
deamidated species during solid-state storage; however, some aggregation was
observed. For the lyophilized EGTA-treated protein, aggregation was even more
pronounced, and there was some loss in enzymatic activity upon reconstitution.
Thus, the removal of calcium ions by EGTA treatment decreased the stability of
rhDNase in both the aqueous and solid states even though no large overall calcium-
induced structural changes could be observed by FTIR or CD. This example shows
one of the problems in using spectroscopic methods such as FTIR or CD to assess
structural alterations. Unless there are gross overall changes, it is unlikely that these
techniques can detect the more subtle changes that may occur which still can lead to
large amounts of aggregate and loss of potency. Thus, biophysical techniques that
can detect subtle conformational changes are required, and recent improvement of
an old technique, hydrogen–deuterium (H/D) exchange introduced by Linderstrøm–
Lang in 1954 (Hvidt and Linderstrom-Lang 1954), may address those needs.
8.3.3 H/D Exchange by MS
Hydrogen/deuterium exchange mass spectrometry (HDX-MS) utilizes mass spec-

trometry to identify hydrogen atoms that are exchanged for deuterium and is based
on the principle that the kinetics of H/D exchange is greatly affected by local
structure of the peptide bond. Houde et al. (2009) used HDX-MS to study changes
to IgG1mAb conformation as a result of deglycosylation by comparing the

glycosylated and deglycosylated forms of the antibody. Two regions of the IgG1
(residues 236–253 and 292–308) were found to have altered exchange properties
upon deglycosylation. Previous studies using X-ray crystallography and NMR have
indicated that the residues in the same region are involved with Fc receptor binding
confirming the HDX-MS is a useful tool to show subtle conformation changes.
Burkitt et al. (2010) recently reported the use of HDX-MS in the analysis of
oxidized mAbs. Structural alterations in a number of segments of the Fc region cor-
responding to large changes in the hydrogen exchange profile of residues 247–253
of the heavy chain that contains the peptide LFPPKPKDTL were observed for an
IgG1 mAb after oxidative stress. The peptide is located in the Fc region of the anti-
body adjacent to Met253 and spatially close to Met430, both of which are found to
easily oxidized under various conditions.
Zhang et al. (2012) have recently demonstrated the use of HDX-MS to study the
aggregation phenomenon of bevacizumab. The authors demonstrate that freeze–
thaw-induced aggregation formed native aggregates that increased with F/T cycles
and inversely proportional to protein concentration, whereas, the molecule under-
went nonnative-like aggregation after thermal stress. The three CDRs on the mole-
cule showed reduced H/D exchange indicating that they are involved in strong
intermolecular contacts during F/T-induced aggregation. Structural comparisons of
the aggregates after F/T or thermal stress using intrinsic Trp fluorescence and extrin-
sic ANS fluorescence confirmed that HDX-MS analysis showed that unfolded
structures are predominant during thermal stress-induced aggregation.
8.3.4 A
ssessing Impact of Chemical Changes
on Physical Stability
While chemical changes are examined using various analytical methods, they typi-
cally do not monitor structural changes that may accompany the chemical change.
Liu et al. (2008) have recently reported the structural changes that accompany Met
oxidation in the CH2 domain of an Escherichia coli expressed Fc protein. The
authors report that methionine oxidation led to subtle changes in the protein confor-
mation and used various biophysical techniques such as CD, DSC, and NMR to
confirm their results. Using far-UV CD, it was reported that methionine oxidation
led to an ellipticity decrease of about 10% at 218 nm, indicating a small but detect-
able change in the secondary structure of the protein. Near-UV CD spectra showed
a positive increase in ellipticity indicating that Met oxidation also leads to changes
in the tertiary structure of the protein as well. The authors also used 2D, 1H–15N
HSQC NMR experiments to corroborate their findings on the structural impact of
Met oxidation. The resonances of many of amino acids in close proximity to the
oxidized Met are affected post H2O2 treatment indicating that upon oxidation, the
protein undergoes a dramatic conformational change. Using transgenic mice with
human FcRn, Wang et al. (2011) have recently reported that oxidized Met residues
(Met 252 and Met 428) have significantly shorter serum half-life.
Bertolotti-Ciarlet et al. (2009) have used surface plasmon resonance and cell
binding assays to study the impact of methionine oxidation on the binding of two
humanized IgG1 antibodies to Fcγ receptors and to the neonatal Fc receptor (FcRn).
SPR analysis showed an increase in kd values that was approximately proportional
to level of Met oxidation, reaching a several-fold higher value in highly oxidized
species. The authors conclude that while Met oxidation did not result in substantial
changes to Fcγ binding (except FcγRIIa), binding to FcRn was significantly affected.
Wang et al. (2011) further evaluated the impact of Met oxidation on serum half lives
of two humanized IgG1 mAbs in transgenic mice with human FcRn. Results
obtained from these studies corroborate the fact that Met oxidation leads to signifi-
cant reduction of half-life in serum after Met oxidation in the Fc region of the mAb
and correlates directly to the level of Met oxidation and binding constants as mea-
sured by SPR. The above studies indicate that techniques that help understand and
prevent methionine oxidation during mAb production would be a significant addi-
tion to the arsenal used in characterizing proteins.
8.3.5 B
iophysical Techniques in High-Concentration
Formulation Development
The recent increase in development of monoclonal antibody therapeutics for indica-

tions requiring administration in a clinic, physician’s office, or by self-administration
at home has resulted in the need for liquid subcutaneous (SC) formulations to ensure
user and patient convenience. Since monoclonal antibody therapies are generally
high dosing, on the order of several mg/kg, administration via the SC route with
acceptable dose volumes (<1.5 mL) necessitates development of mAb formulations
over a concentration range from 100 to 200 mg/mL or even higher depending on the
indication. Development of such high-concentration formulations can result in sev-
eral manufacturing, stability, analytical, and delivery challenges (Shire et al. 2004).
Manufacturing of high-concentration protein formulations is typically done using
tangential flow filtration (TFF), which has limitations as the viscosity of the solution
being concentrated increases leading to high back pressure during the TFF process
that may exceed the capacity of the pumps. The high protein concentrations required
for formulations of mAbs often result in a dramatic increase in the viscosity of the
solution as shown by the concentration–viscosity profile for three mAbs (Fig. 8.13).
These three mAbs are constructed from the same human IgG1 Fc framework. mAb2
and mAb3 have very similar behavior, and the solid line is a fit to the extended
Mooney equation, which assumes only repulsive hard-sphere interactions (Ross and
Minton 1977):
 
 [h]c 
h = h0 exp   (8.1)
1 − k [h]c 
 v 
Fig. 8.13 Viscosity of mAbs1–3 as a function of protein concentration. mAb1, mAb2, and mAb3
aqueous samples have a composition of 266 mM sucrose, 16 mM histidine, and 0.03% polysorbate
20 at pH 6.0. mAb2 reconstituted lyophilized samples are at either 100 mg/mL MAb2 in 240 mM
trehalose, 40 mM histidine, and 0.04% polysorbate 20 or 125 mg/mL in 300 mM trehalose, 50 mM
histidine, and 0.05% polysorbate 20, pH 6. The solid curves are the result of a nonlinear regression
of all the data to the modified Mooney equation using a solvent viscosity of 1.1 mPas and intrinsic
viscosity of 6.3 cm3/g (adapted from Liu et al. 1995)
where h0 is the solution viscosity, η is the solvent viscosity, [η] is the intrinsic vis-
cosity of the protein, k is a “crowding factor,” and ν is the Simha parameter related
to the ellipsoid of revolution used to model the protein. The viscosity–concentration
profile for mAb1, on the other hand, cannot be described by the hard-sphere model
(Fig. 8.13, Liu et al. 2005) suggesting that there are other interactions in addition to
those involving excluded volume that mitigate the viscosity behavior of this particu-
lar mAb. It was hypothesized by Liu et al. that net-attractive interactions that result
in reversible concentration-dependent self-association increase the viscosity of a
concentrated monoclonal antibody in aqueous solution. At low concentrations the
three monoclonal antibodies are essentially monomeric in solution as determined by
SEC analysis. Thus, in order to determine if, in fact, mAb1 undergoes self-associa-
tion at high concentration, the analysis needs to be done at high concentration. This
is a challenging task since most analyses are done at low concentrations, typically
at 1 mg/mL. AUC sedimentation equilibrium has been used to investigate protein
self-association at high concentration by using thin plastic gaskets as centerpieces
(Minton and Lewis 1981), but this is not an easy method due to deformations of the
gasket material when torquing the AUC cells. The thin plastic gaskets that result in
a narrow cell pathlength are necessary in order to deal with the high refractive index
that results in a deviation of the light path away from the detecting photomultiplier
in the centrifuge (Gonzalez et al. 2003). An alternative technique previously
described by Minton is preparative AUC (Minton 1989). In this method a prepara-
tive centrifuge is used at low speeds to generate sedimentation equilibrium concen-
tration gradients. After equilibrium is attained the centrifuge tubes are loaded onto
a low-volume fraction collector (Brandel®, Brandel Inc, Gaithersburg, MD), and
5 μL fractions are dispensed into a 96-well plate. All the wells are then diluted to
bring the absorbance reading down to within the dynamic range of the UV plate
reader (Fig. 8.14). The resulting absorbance as a function of radial position is then
used to determine the apparent weight-average molecular weight, Mw, app, at each
radial position using the following equation:
_
 M w,app (1 − v r )(r 2 − r0 2 ) 
C (r ) = C0 exp   (8.2)
 2 RT 
where c(r) is the protein_ concentration at radial position r, c0 is the initial loading
protein concentration, v is the partial specific volume, ρ is the buffer density, ω is
the angular velocity, and r0 is the reference radial position. At these high concentra-
tions there is a huge amount of non-ideality. Since mAb2 and mAb3 viscosity can
be accounted for using the Mooney equation, it is assumed that mAb2 and 3 exist
mainly as monomeric molecules in solution, and the non-ideality correction for the
charge and excluded volume effects can be obtained in the absence of added NaCl
for an IgG1 monomer with 150 kDa molecular weight. These corrections are
obtained from the relationship between apparent molecular weight, Ma, at weight/
volume concentration c and actual molecular weight, M (Chatelier and Minton
1987):
  d ln g  
M a = M 1 + c  (8.3)
  dc  
where γ is the activity coefficient of the monoclonal antibody. Assuming that mAb2
and 3 are essentially monomeric in solution leads to a multiplicative correction fac-
tor as a function of c when M is set equal to 150 kDa. The resulting corrected molec-
ular weight for mAb1 (Fig. 8.14) supports the hypothesis that this monoclonal
antibody undergoes a concentration-dependent reversible self-association. Another
technique, SLS, at high concentrations has essentially corroborated the AUC analy-
sis (Scherer et al. 2010). A rigorous analysis of the SLS data shows that mAb1 self-
associates much more than mAb2. Thus, while literature has focused on
protein–protein interactions in dilute solutions, this work has emphasized the need
to develop newer techniques to understand intra-protein and inter-protein interac-
tions in highly concentrated protein solutions, especially because self-association of
proteins under such conditions appears to be crucial to appreciate the underlying
solution viscosity (Yadav et al. 2010). Kanai et al. proposed that the observed self-
association of mAb1, at pH 6.0, originates from multiple Fab–Fab interactions.
Fig. 8.14 Overview of a preparative AUC experiment. Centrifugation was carried out in an

Beckman XL/A without the installed optics using a swinging bucket rotor. After attainment of
sedimentation, equilibrium samples were fractionated using a Brandell micro fractionator and col-
lected into a 96-well quartz plate and diluted to generate an absorbance vs. radial position gradient
that was used to generate a concentration for analysis to determine weight-average molecular
weight (overview drawing of preparative AUC provided by Sandeep Yadav, corrected MW plot
previously published in Liu et al. 1995)
More recently, Yadav et al. (2011) showed that the presence of specific attractive
interactions at pH 6.0 leads to the self-association and high viscosity of mAb-1 at
high concentrations. They also demonstrate that exposed charged residues, espe-
cially histidyl residues, in the CDR of mAb1 are critical in determining the self-
associating and highly viscous behavior observed at high concentrations. The
authors used various biophysical techniques including CD, sedimentation equilib-
rium, high-frequency ultrasonic rheology, DLS, zeta potential, and net charge
d eterminations to understand the physical properties of several IgG1 molecules

(Yadav et al. 2012).
The recent study reported earlier by Yadav et al. has brought attention that devel-
oping orthogonal methods for mAbs that have potential for self-association, aggre-
gation, precipitation, and/or high viscosity is crucial for successful molecule
development. Saluja et al. (2010) have recently demonstrated that aggregation pro-
pensity of a mAb closely correlated to the second virial coefficient (B2), kd and ks
(interaction parameters and for the diffusion and sedimentation coefficients, respec-
tively). The second virial coefficient is a parameter that is known to be valuable in
predicting protein–protein interactions, while this study also suggests that kd can be
used as a high-throughput predictor of protein aggregation.
8.3.6 B
iophysical Techniques to Support Clinical
In-Use Studies
Pharmacists are responsible for setting a “beyond use” date based on USP 797,
wherein the beyond use date for the compounded sterile preparation (CSP) is
defined as the time by which the compounded preparation must be used to avoid
risks for product degradation, contamination, etc. Physical and chemical stability of
the CSP can be difficult to maintain over extended storage, especially since the
formulation components are diluted within the IV bag contents. Recent published
reports have suggested the use of extended time, beyond that recommended by the
manufacturer, for the storage and administration of CSP. These recommendations
were based on inadequate analytical testing of the CSP. Alavattam et al. 2012 dem-
onstrate that setting of the beyond use date should be carefully assessed using the
appropriate biophysical and analytical methods, given the fact that important excipi-
ents that help stabilize the protein get diluted upon dilution in saline IV bags. This
study shows that as the ratio of mAb to polysorbate 20 increased, the levels of aggre-
gates also increased with agitation as detected by the SEC assay. This suggests that
there is less surfactant to compete with the protein for the air–liquid interface in the
IV bag, thereby allowing mAb to aggregate at the interface. This study also suggests
that while the ratio of mAb to polysorbate 20 is an important contributing factor to
form soluble aggregates, headspace also played a key role. No clear trending was
observed in the subvisible particle analysis using the light obscuration (HIAC–
Royco) method. When performing particle analysis, there may be variability in the
particulate measurements attributed to sample handling and/or instrument accuracy,
and care must be taken to ensure the proper procedures are followed (Cao et al.
2010). However, it was evident that mAb1 and mAb2 with 0% polysorbate had a
larger number of subvisible particles at the ≥10 μm size in bags that contained head-
space, suggesting at least 0.0001% polysorbate may be necessary to protect the
protein at the air–water interface from particulate formation. It appeared that sub-
visible particles increased upon agitation in IV bags with headspace in mAb1 solu-
tions that did not contain polysorbate 20, whereas the subvisible particles were not
significantly different for mAb2 under similar solution c onditions. When mAb3 was
diluted tenfold into PO IV bags and subjected to agitation, a significant increase in
soluble aggregates was observed (13.5% over 1 h). A larger increase in soluble
aggregates was observed for mAb3 than for mAb1 and mAb2 over similar time
scales despite higher polysorbate 20 levels in the IV bags illustrating the high sen-
sitivity of mAb3 to agitation. Visible precipitation of the protein was observed on
further agitation of mAb3 in the IV bag, and no additional testing was performed on
samples that were agitated for longer periods of time. A clear correlation between
subvisible particles and soluble aggregates could not be found in this study.
Kumru et al. (2012) recently reported the physical stability of an IgG4 monoclo-
nal antibody upon dilution into intravenous (i.v.) bags containing 0.9% saline.
Soluble aggregates and subvisible particles were characterized using a variety of
analytical (including SEC) and biophysical methods such as light obscuration,
nanoparticle tracking analysis (NTA), microflow-digital imaging (MFI), and turbid-
ity measurements. Characterization studies with FTIR microscopy and extrinsic
fluorescence spectroscopy demonstrated that isolated particles contained native-like
secondary structure with partially altered tertiary structure, compared with heat-
denatured and non-stressed controls. Transmission electron microscopy (TEM) and
MFI analysis showed particles had an amorphous morphology of varying sizes.
Demeule et al. (2009b) recently reported a case study on characterizing trastu-
zumab samples diluted in either saline or dextrose solutions using asymmetrical
flow field-flow fractionation (FFF), fluorescence spectroscopy, fluorescence micros-
copy, and TEM. When trastuzumab samples were analyzed by FFF using a standard
separation method, no difference could be seen between trastuzumab diluted in
sodium chloride and trastuzumab diluted in dextrose. However, during FFF mea-
surements made with appropriate changes in the protocol, trastuzumab aggregates
were detected in 5% dextrose. This suggests that trastuzumab aggregates are over-
looked if the FFF analysis of trastuzumab diluted in dextrose solution is performed
using 0.9% NaCl. The authors were also able to confirm aggregates using fluores-
cence microscopy and TEM.
All the above studies emphasize the need to perform orthogonal biophysical tests
to clearly understand different mechanisms for protein degradation in IV bags and
support the proper use of protein therapeutics in clinical trials.
8.3.7 B
iophysical Techniques in Antibody Drug Conjugate
Formulation Development
Trastuzumab-DM1 (T-DM1) is one of the most successful antibody drug conjugates

(ADCs) to enter clinical trials in the recent past. It has a non-cleavable linker using
lysine chemistry to link trastuzumab (Herceptin®) to the potent antimicrotubule thiol
containing a maytansine analog (DM1) via a thioether linkage. In addition to devel-
oping analytical techniques to characterize the antibody itself, ADCs need special
chemical and biophysical techniques to understand various properties of the small
Fig. 8.15 Comparison of the DSC thermograms of Tmab, T-MCC, and T-DM1 ([anti-

body] = 1 mg/mL) in 10 mM sodium succinate, 6% (w/v) trehalose dihydrate, and 0.02% (w/v)
polysorbate 20. Inset in the figure depicts reversibility of the first (CH2) transition for Tmab,
T-MCC, and T-DM1 (previously published in Wakankar et al. 2010)
molecule chemotherapeutic drug in order to evaluate drug-to-antibody ratio (DAR),

hydrophobicity due to the small molecule, drug distribution, linker, and drug stability
as well as influence on protein conformation upon drug binding. The challenges in
developing in vivo assays for ADC characterization have recently been reported and
will not be a focus in this chapter (Stephan et al. 2011). Due to the large variance in
physical properties of the antibody itself (typically water soluble) and the small mol-
ecule drug (typically lipophilic), it is often found that aggregation is a major issue in
ADCs. These aggregates seem to reflect the properties of the small molecule drug in
that they are driven by lipophilic association (Hollander et al. 2008). Hollander et al.
have used SEC to identify conditions during conjugation steps in order to identify
reaction conditions that minimize aggregation of calicheamicin conjugated antibod-
ies. They found that a combination of propylene glycol and octanoic acid produced
the least aggregate and high protein recovery with good drug loading. However, fur-
ther characterization of the ADC has not been reported in this chapter. Wakankar
et al. (2010) have recently reported thorough characterization of T-DM1, including
the stability of Tmab (unconjugated IgG1) and the intermediate T-MCC using a vari-
ety of techniques such as DSC, SEC, CE-SDS, and LC-MS. DSC analysis showed
two transitions typical of IgG1 (Fig. 8.15). DSC data from this ADC showed that
conjugation with DM1 leads to a 4.4°C drop in the melting temperature of the first
transition due to CH2 domain (tm is in the order of Tmab>TMCC>T-DM1) whereas

only a 0.8°C drop in the second transition due to CH3 domain was noticed. These
results clearly suggest that conjugation of Tmab has significant impact on the ther-
mal stability of the CH2 domain. Further stability characterization using SEC
showed aggregate formation was in the order of Tmab<T-DM1<TMCC, an order
that is not predicted based on DSC results alone. This indicates that mechanisms
other than the typical hydrophobic interactions may be important, and a plethora of
biophysical and analytical techniques need to be carefully considered during formu-
lation development.
In addition to understanding the effect of conjugation on the melting temperature
and potential aggregation propensity of the ADC, it is important to understand the
effect of the small molecule drug on the higher order structure of the molecule.
Since most of the small molecule cytotoxic drugs exhibit strong absorbance in the
near UV region, analysis using CD spectroscopy may be complicated in this region.
Little to no information has been published in characterizing ADCs using this tech-
nique. Any alterations in higher order structure will probably be difficult to tease out
from variations because of DAR, especially in the near UV region. Far-UV CD,
however, may provide information and can be employed to understand higher order
structure for ADCs.
8.4 Confirmatory Studies
Given that most analytical methods have limitations in characterizing proteins,

alternate methods are used to characterize protein pharmaceuticals. These orthogo-
nal techniques provide vital information that is necessary for clinical development
and are typically done during licensure activities. One area that has been heavily
scrutinized by regulatory agencies is the determination of aggregates and fragments
by SEC. SEC determines molecular size based on the hydrodynamic volume of the
molecules and thus molecules with an altered shape, or extensive glycosylation may
yield incorrect molecular weight when assessed using typical globular protein stan-
dards. This problem can be circumvented using on-line light-scattering detectors to
determine weight-average molecular weight. However, interactions with the col-
umn matrix may result in loss of aggregates as well as decreased resolution. Often
organic modifiers or salts are added to decrease the matrix interactions, but this may
also impact the actual distribution of aggregate size. In addition, if the protein
undergoes reversible self-association that is concentration dependent, the dilution
on typical SEC analysis may result in detection of only lower molecular weight spe-
cies. Regulatory agencies have become aware of such limitations and have been
asking for confirmatory studies using orthogonal biophysical techniques to verify
the accuracy of the SEC determination that are often used to evaluate protein aggre-
gation (Carpenter et al. 2010). Although several biophysical techniques such as
SLS, FFF, and AUC have been used to characterize aggregates and fragments, AUC
Table 8.2 Comparison of AUC and SEC results of a

monoclonal antibody
% HMWS (total)
Samples AUC SEC
A 1.2 (31.4% RSD)a 0.2
B 1.9 1.3
C 4.8 5.5
D 6.4 6.6
E 7.6 (7.9% RSD) 7.2
Sample A consists of a representative mAb drug product batch,
Sample B consists of a mAb batch subjected to light exposure
at 1.2 mlux hours, Sample C consists of a mAb batch subjected
to light exposure at 3.6 mlux hours, Sample D consists of a
mAb batch subjected to acid treatment at pH 3.2, and Sample
E consists of purified basic variants from IE-HPLC
AUC analytical ultracentrifugation, HMWS high-molecular
weight species, RSD relative standard deviation, SEC size
exclusion high-performance liquid chromatography
Samples A and B have HMWS levels that are below the limit
of quantitation of the AUC technique (Gabrielson and Arthur
2011)
has been one of the more extensively used techniques to confirm SEC
analysis(Berkowitz 2006), as described in Chap. 5.
8.4.1 AUC as an Orthogonal Technique to SEC
As an example of the use of AUC in a confirmatory study, a monoclonal antibody

reference standard from Genentech was analyzed by SEC and AUC. The amount of
high-molecular weight species (HMWS), generally dimer, and low-molecular
weight species (LMWS) were determined by SEC to be 0.1–0.2% and 0.0–0.1%,
respectively. There was no difference in HMWS upon dilution, suggesting that the
aggregates are non-dissociable. Measurements were also made on the same sample
at 0.55 mg/mL in formulation buffer using AUC as an orthogonal technique to mea-
sure HMWS. The samples were loaded in Spin Analytical 12-mm 2-sector cells and
spun at 20°C and 40,000 rpm in an XLA/I (Beckman) analytical ultracentrifuge.
The sedimentation coefficient distributions c(s) were calculated with Sedfit v12.1b
(Schuck 2000) using absorbance data collected at 280 nm. In addition, various lev-
els of aggregates were generated by light stress, acid treatment, or purified basic
variants using IE-HPLC (Table 8.2) and analyzed using SEC and AUC. There was
good agreement between AUC and SEC results in terms of HMWS quantitation,
showing no evidence of SEC missing or underestimating major HMWS. A concor-
dance plot demonstrating the agreement between AUC and SEC is shown in
Fig. 8.16.
Fig. 8.16 Concordance plot of AUC sedimentation velocity and SEC analysis of a monoclonal anti-
body. AUC, analytical ultracentrifugation; HMWS, high-molecular weight species; SEC, size exclu-
sion high-performance liquid chromatography. The error bars represent two standard deviations from
n = 3 determination. All other data points denote a single determination. Circles denote samples that
have HMWS levels below the LOQ of the AUC technique (Gabrielson and Arthur 2011)
8.4.2 F
low Field-Flow Fractionation
as an Orthogonal Technique
Another orthogonal technique that has been used to characterize protein aggregates
is FFF (Liu et al. 2006; Rambaldi et al. 2011). This technique is a flow-based sepa-
ration method, whereby separation is achieved by applying an externally generated
field that is perpendicular to laminar flow within a buffer filled open channel.
Although several different external fields have been used (Giddings 2000), charac-
terization of protein aggregates has been done mainly by application of solution
flow as the external perpendicular field and is termed flow FFF. This method,
which covers a wide range of molecular sizes from 0.001 to 50 μm in size, provides
assessments of size and quantity of protein aggregates without the use of standards
or solid-state matrices. Among all the flow-based FFF methods, asymmetrical flow
FFF (AF4), where only the bottom wall of the channel has a semipermeable mem-
brane, is the one most commonly used for therapeutic proteins (Fraunhofer and
Table 8.3 Analysis of a mixture of three proteins (described in the text) using AF4, SEC, and
AUC
% Species
% % Species % Species AUC
MW Species AF4 (n = 5, Accuracya SEC (n = 5, Accuracya (n = 6, Accuracya
Samples (kDa) (actual) mean ± σ) AF4 (%) mean ± σ) SEC (%) mean ± σ) AUC (%)
Protein I 150 32.99 31.97 ± 0.13 1.02 32.80 ± 0.03 0.19 31.1 ± 0.00 1.9
Protein II 100 33.88 35.53 ± 0.08 −1.65 34.31 ± 0.05 −0.43 33.0 ± 0.00 0.9
Protein III 50 33.13 32.50 ± 0.05 0.63 32.87 ± 0.04 0.26 35.9 ± 0.00 −2.8
a
Accuracy was determined by subtracting the experimental values from the actual values
Winter 2004). The applied cross flow drives macromolecules towards the
membrane, and differences in diffusion coefficients create a concentration distribu-
tion in the laminar flow, resulting in different elution times. These elution times
allow for determination of the translational diffusion coefficient, which can be
used to compute the molecular weight assuming spherical structures for all spe-
cies. Although there is no solid-state support for potential interaction of protein,
there have been issues with protein interaction with the channel membrane. Liu
et al. (2012) have investigated this problem and show that with correct choice of
membrane and solvent conditions, it is possible to use this technology to confirm
results by SEC. In particular, they compared the analysis of three mAb prepara-
tions, protein I (a full-length monoclonal antibody), protein II (a single-armed anti-
body, i.e., with only one Fab), and protein III (a Fab fragment) using SEC, AUC,
and AF4. The results show very good agreement between all three techniques
(Table 8.3) demonstrating that AF4 is a viable technique that can be used for con-
firmatory studies.
8.5 Concluding Remarks
This chapter has covered a variety of biophysical techniques, which have been used
in the development of protein biopharmaceuticals, including some perspective of
formulation development and long-term stability. It has not meant to be an all-
encompassing review but rather to demonstrate in what areas of development bio-
physics has been of use to aid in the production of a stable drug product and to
compliment the other chapters in this book. We have not discussed the use of differ-
ent biophysical methods to characterize protein particulates since that was covered
in Chap. 4. We also have highlighted some of the ways we have used biophysical
analysis at Genentech and hope this will be useful for researchers who wish to apply
such technologies.
References
Agostini F, Vendruscolo M, Tartaglia GG (2012) Sequence-based prediction of protein solubility.

J Mol Biol 421(2–3):237–241
Andya JD, Hsu CC, Shire SJ (2003) Mechanisms of aggregate formation and carbohydrate excipi-
ent stabilization of lyophilized humanized monoclonal antibody formulations. AAPS PharmSci
5(2):E10
Alavattam S et al (2012) Stability of IgG1 monoclonal antibodies in intravenous infusion bags
under clinical in-use conditions. J Pharm Sci 101(1):21–30
biopharmaceuticals. AAPS J 8(3):E590–E605
Bertolotti-Ciarlet A et al (2009) Impact of methionine oxidation on the binding of human IgG1 to
FcRn and Fc gamma receptors. Mol Immunol 46(8–9):1878–1882
Brahms S, Brahms J (1980) Determination of protein secondary structure in solution by vacuum
ultraviolet circular dichroism. J Mol Biol 138(2):149–178
Burkitt W, Domann P, O’Connor G (2010) Conformational changes in oxidatively stressed mono-
clonal antibodies studied by hydrogen exchange mass spectrometry. Protein Sci
19(4):826–835
Byler DM, Susi H (1986) Examination of the secondary structure of proteins by deconvolved Ftir
spectra. Biopolymers 25(3):469–487
Cao S, Jiang Y, Narhi LO (2010) A light-obscuration method specific for quantifying subvisible
particles in protein therapeutics. Pharmacopeial Forum 36(3):824–834
Carpenter JF et al (2010) Potential inaccurate quantitation and sizing of protein aggregates by size
exclusion chromatography: essential need to use orthogonal methods to assure the quality of
therapeutic protein products. J Pharm Sci 99(5):2200–2208
Chatelier RC, Minton AP (1987) Sedimentation equilibrium in macromolecular solutions of arbi-
trary concentration. I. Self-associating proteins. Biopolymers 26(4):507–524
Chen B et al (1999) Influence of calcium ions on the structure and stability of recombinant human
deoxyribonuclease I in the aqueous and lyophilized states. J Pharm Sci 88(4):477–482
Chennamsetty N et al (2009a) Aggregation-prone motifs in human immunoglobulin G. J Mol Biol
391(2):404–413
Chennamsetty N et al (2009b) Design of therapeutic proteins with enhanced stability. Proc Natl
Acad Sci USA 106(29):11937–11942
Connolly BD et al (2012) Weak interactions govern the viscosity of concentrated antibody solutions:
high-throughput analysis using the diffusion interaction parameter. Biophys J 103:69–78
Demeule B, Shire SJ, Liu J (2009a) A therapeutic antibody and its antigen form different com-
plexes in serum than in phosphate-buffered saline: a study by analytical ultracentrifugation.
Demeule B et al (2009b) New methods allowing the detection of protein aggregates: a case study
on trastuzumab. MAbs 1(2):142–150
Dong A, Huang P, Caughey WS (1990) Protein secondary structures in water from second-
derivative amide I infrared spectra. Biochemistry 29(13):3303–3308
Eigenbrot C et al (1991) X-ray structure of human relaxin at 1.5 A. Comparison to insulin and
implications for receptor binding determinants. J Mol Biol 221(1):15–21
Fox JA et al (1996) Tissue distribution and complex formation with IgE of an anti-IgE antibody
after intravenous administration in cynomolgus monkeys. J Pharmacol Exp Ther
279(2):1000–1008
Fraunhofer W, Winter G (2004) The use of asymmetrical flow field-flow fractionation in pharma-
ceutics and biopharmaceutics. Eur J Pharm Biopharm 58(2):369–383
Gibson TJ et al (2011) Application of a high-throughput screening procedure with PEG-induced
precipitation to compare relative protein solubility during formulation development with IgG1
Monoclonal antibodies. J Pharm Sci 100(3):1009–1021
Giddings JC (2000) The field-flow fractionation family: underlying principles. In: Schimpf ME,
Caldwell KD, Giddings JC (eds) Field-flow fractionation handbook. Wiley, New York
Gonzalez JM, Rivas G, Minton AP (2003) Effect of large refractive index gradients on the perfor-
mance of absorption optics in the Beckman XL-A/I analytical ultracentrifuge: an experimental
study. Anal Biochem 313(1):133–136
He F et al (2010a) Detection of IgG aggregation by a high throughput method based on extrinsic
fluorescence. J Pharm Sci 99(6):2598–2608
He F et al (2010b) High-throughput dynamic light scattering method for measuring viscosity of
concentrated protein solutions. Anal Biochem 399(1):141–143
He F et al (2011) Screening of monoclonal antibody formulations based on high-throughput ther-
mostability and viscosity measurements: design of experiment and statistical analysis. J Pharm
Sci 100(4):1330–1340
Hollander I, Kunz A, Hamann PR (2008) Selection of reaction additives used in the preparation of
monomeric antibody-calicheamicin conjugates. Bioconjug Chem 19(1):358–361
Houde D et al (2009) Characterization of IgG1 conformation and conformational dynamics by
hydrogen/deuterium exchange mass spectrometry. Anal Chem 81(14):2644–2651
Hvidt A, Linderstrom-Lang K (1954) Exchange of hydrogen atoms in insulin with deuterium
atoms in aqueous solutions. Biochim Biophys Acta 14(4):574–575
Jackson M, Mantsch HH (1995) The use and misuse of FTIR spectroscopy in the determination of
protein structure. Crit Rev Biochem Mol Biol 30(2):95–120
Ji JA et al (2009) Methionine, tryptophan, and histidine oxidation in a model protein, PTH: mecha-
nisms and stabilization. J Pharm Sci 98(12):4485–4500
Johnson WC Jr (1990) Protein secondary structure and circular dichroism: a practical guide.
Proteins 7(3):205–214
Kamerzell TJ et al (2011) Protein-excipient interactions: mechanisms and biophysical character-
ization applied to protein formulation development. Adv Drug Deliv Rev 63(13):1118–1159
Kumru OS et al (2012) Compatibility, physical stability, and characterization of an IgG4 monoclo-
nal antibody after dilution into different intravenous administration bags. J Pharm Sci
101(2):3636–3650
Lauer TM et al (2012) Developability index: a rapid in silico tool for the screening of antibody
aggregation propensity. J Pharm Sci 101(1):102–115
Lehermayr C et al (2011) Assessment of net charge and protein-protein interactions of different
monoclonal antibodies. J Pharm Sci 100(7):2551–2562
Li CH et al (2011) Applications of circular dichroism (CD) for structural analysis of proteins:
qualification of near- and far-UV CD for protein higher order structural analysis. J Pharm Sci
100(11):4642–4654
Liu J et al (1995) Characterization of complex formation by humanized anti-IgE monoclonal anti-
body and monoclonal human IgE. Biochemistry 34(33):10474–10482
Liu J, Ruppel J, Shire SJ (1997) Interaction of human IgE with soluble forms of IgE high affinity
receptors. Pharm Res 14(10):1388–1393
Liu J et al (2005) Reversible self-association increases the viscosity of a concentrated monoclonal
antibody in aqueous solution. J Pharm Sci 94(9):1928–1940
fractionation methods for measuring protein aggregation. AAPS J 8(3):E580–E589
Liu D et al (2008) Structure and stability changes of human IgG1 Fc as a consequence of methio-
nine oxidation. Biochemistry 47(18):5088–5100
Liu J et al (2012) Assessing and improving asymmetric flow field-flow fractionation of therapeutic
proteins. In: Williams KR, Caldwell KD (eds) Field flow fractionation in biopolymer analysis.
Springer, New York
Lu Y et al (2008) The effect of a point mutation on the stability of IgG4 as monitored by analytical
ultracentrifugation. J Pharm Sci 97(2):960–969
Manta B et al (2011) Tools to evaluate the conformation of protein products. Biotechnol
J 6(6):731–741
Minton AP (1989) Analytical centrifugation with preparative ultracentrifuges. Anal Biochem
176(2):209–216
Minton AP, Lewis MS (1981) Self-association in highly concentrated solutions of myoglobin: a
novel analysis of sedimentation equilibrium of highly nonideal solutions. Biophys Chem
14(4):317–324
Patel AR, Kerwin BA, Kanapuram SR (2009) Viscoelastic characterization of high concentration
antibody formulations using quartz crystal microbalance with dissipation monitoring. J Pharm
Sci 98(9):3108–3116
8(3):E564–E571
Rambaldi DC, Reschiglian P, Zattoni A (2011) Flow field-flow fractionation: recent trends in pro-
tein analysis. Anal Bioanal Chem 399(4):1439–1447
RNCOS (2012) Global-protein-therapeutics-market-forecast-to-2015. http://www.pharmaceutical-
market-research.info/research/PMAAAWUF-Global-Protein-Therapeutics-Market-
Forecast-to-2015.shtml
Robinson NE, Robinson AB (2004) Prediction of primary structure deamidation rates of asparagi-
nyl and glutaminyl peptides through steric and catalytic effects. J Pept Res 63(5):437–448
Ross PD, Minton AP (1977) Hard quasispherical model for the viscosity of hemoglobin solutions.
Biochem Biophys Res Commun 76(4):971–976
Saluja A, Kalonia DS (2004) Measurement of fluid viscosity at microliter volumes using quartz
impedance analysis. AAPS PharmSciTech 5(3):e47
Saluja A, Kalonia DS (2005) Application of ultrasonic shear rheometer to characterize rheological
properties of high protein concentration solutions at microliter volume. J Pharm Sci
94(6):1161–1168
Saluja A et al (2007) Ultrasonic storage modulus as a novel parameter for analyzing protein-
protein interactions in high protein concentration solutions: correlation with static and dynamic
light scattering measurements. Biophys J 92(1):234–244
Saluja A et al (2010) Diffusion and sedimentation interaction parameters for measuring the second
virial coefficient and their utility as predictors of protein aggregation. Biophys J 99(8):2657–2665
Samra HS, He F (2012) Advancements in high throughput biophysical technologies: applications
for characterization and screening during early formulation development of monoclonal anti-
bodies. Mol Pharm 9(4):696–707
Scherer TM et al (2010) Intermolecular interactions of IgG1 monoclonal antibodies at high con-
centrations characterized by light scattering. J Phys Chem B 114:12948
Schuck P (2000) Size-distribution analysis of macromolecules by sedimentation velocity ultracen-
trifugation and Lamm equation modeling. Biophys J 78(3):1606–1619
Shire SJ, Holladay LA, Rinderknecht E (1991) Self-association of human and porcine relaxin as
assessed by analytical ultracentrifugation and circular dichroism. Biochemistry
30(31):7703–7711
Shire SJ, Shahrokh Z, Liu J (2004) Challenges in the development of high protein concentration
formulations. J Pharm Sci 93(6):1390–1402
Stafford WF 3rd (1992) Boundary analysis in sedimentation transport experiments: a procedure
for obtaining sedimentation coefficient distributions using the time derivative of the concentra-
tion profile. Anal Biochem 203(2):295–301
Stephan JP, Kozak KR, Wong WLT (2011) Challenges in developing bioanalytical assays for char-
acterization of antibody-drug conjugates. Bioanalysis 3(6):677–700
Tartaglia GG et al (2005) Prediction of aggregation rate and aggregation-prone segments in poly-
peptide sequences. Protein Sci 14(10):2723–2734
Tartaglia GG et al (2008) Prediction of aggregation-prone regions in structured proteins. J Mol
Biol 380(2):425–436
Wakankar AA et al (2010) Physicochemical stability of the antibody-drug conjugate

Trastuzumab-DM1: changes due to modification and conjugation processes. Bioconjug Chem
21(9):1588–1595
Wang W (1999) Instability, stabilization, and formulation of liquid protein pharmaceuticals. Int J
Pharm 185(2):129–188
Wang WR et al (2011) Impact of methionine oxidation in human IgG1 Fc on serum half-life of
monoclonal antibodies. Mol Immunol 48(6–7):860–866
Wu SJ et al (2010) Structure-based engineering of a monoclonal antibody for improved solubility.
Protein Eng Des Sel 23(8):643–651
Yadav S, Shire SJ, Kalonia DS (2010) Factors affecting the viscosity in high concentration solu-
tions of different monoclonal antibodies. J Pharm Sci 99(12):4812–4829
Yadav S et al (2011) Establishing a link between amino acid sequences and self-associating and
viscoelastic behavior of two closely related monoclonal antibodies. Pharm Res
28(7):1750–1764
Yadav S, Shire SJ, Kalonia DS (2012) Viscosity behavior of high-concentration monoclonal anti-
body solutions: correlation with interaction parameter and electroviscous effects. J Pharm Sci
101(3):998–1011
Zhang A et al (2012) Distinct aggregation mechanisms of monoclonal antibody under thermal and
freeze-thaw stresses revealed by hydrogen exchange. Pharm Res 29(1):236–250
Zhao H et al (2010) Formulation development of antibodies using robotic system and high-
throughput laboratory (HTL). J Pharm Sci 99(5):2279–2294
Chapter 9
Case Studies Applying Biophysical Techniques
to Better Characterize Protein Aggregates
and Particulates of Varying Size
Tingting Wang, Sangeeta B. Joshi, Ozan S. Kumru, Srivalli Telikepalli,

C. Russell Middaugh, and David B. Volkin
9.1 Introduction
From a pharmaceutical perspective, protein storage stability is defined by the ana-

lytical techniques used to monitor chemical and physical changes in protein struc-
ture. As new analytical methods with increased sensitivity are introduced, resulting
in an enhanced ability to detect chemical alterations in amino acid residues (oxida-
tion, deamidation, etc.) or physical changes in protein higher-order structure (par-
tial unfolding, aggregation, etc.), questions and concerns can arise about how best
to set specifications to define storage stability. For example, physical changes lead-
ing to protein aggregation have been historically monitored by a combination of
size-exclusion chromatography (SEC) and visual assessments of solution clarity.
As awareness of SEC limitations has grown (Carpenter et al. 2010a), combined with
introduction of improved biophysical methods to detect protein aggregates of vary-
ing size at lower limits of detection (see chapters in this book), it has become appar-
ent that a “new” physical pathway of protein degradation should be monitored as
part of protein stability studies: protein particle formation in the submicron
(0.1–1 µm) and subvisible (1–100 µm) size ranges (Carpenter et al. 2009; Singh
et al. 2010).
The challenge of detecting “new” impurities or degradants, due to development
of improved analytical methods, has been an inherent part of therapeutic protein
drug development for decades. For example, as the ability to detect residual host
cell impurities increased with the introduction of new analytical methods, the need
for new control strategies and specifications for residual host cell protein and DNA
T. Wang • S.B. Joshi • O.S. Kumru • S. Telikepalli • C.R. Middaugh • D.B. Volkin (*)
Department of Pharmaceutical Chemistry, Macromolecule and Vaccine Stabilization Center,
University of Kansas, 2030 Becker Drive, Lawrence, KS 66047, USA
e-mail: [email protected]; [email protected]; [email protected]; [email protected];
[email protected]; [email protected]
206 T. Wang et al.
also increased. This situation leads to improved product quality with lower and
more consistent impurity levels. Similarly, potentially more stable protein formula-
tions may be developed in the future as our ability to monitor and control the forma-
tion of protein aggregates and particles improves with the introduction of more
sensitive biophysical methods (Carpenter et al. 2010b; Mire-Sluis et al. 2011).
The case studies in this chapter are presented as a series of illustrative examples
based on the size of the protein aggregates being examined. This approach permits
better comparisons across laboratories as investigators study different types of
aggregates from a wide range of proteins using a variety of different physicochemi-
cal methods (Narhi et al. 2011). Case studies examining the use of biophysical
methods to better characterize soluble protein aggregates in the size range of
1–100 nm are first presented followed by examples utilizing different biophysical
techniques to better characterize submicron-sized protein particles (0.1–1 µm). The
last section then presents examples of monitoring and evaluating larger subvisible
(1–100 µm) and visible (>100 µm) particles in protein formulations. For each sec-
tion, a brief overview of the methods used in the case studies are presented followed
by descriptions of recent literature examples that have employed a combination of
biophysical methods to better characterize the nature and composition of protein
aggregates and particles formed from different environmental stresses.
9.2 Case Studies Using Biophysical Methods to Count, Size,

and Characterize Protein Soluble Aggregates
9.2.1 Overview of Analytical Methodologies Used to Evaluate

Soluble Protein Aggregates Under Native-Like and
Denaturing Conditions
A variety of analytical approaches have been used to examine the size, amount, and
nature of soluble protein aggregates under both nondenaturing native and denatur-
ing conditions. For example, size-exclusion HPLC (SE-HPLC), sedimentation
velocity analytical ultracentrifugation (SV-AUC), dynamic light scattering (DLS),
and field-flow fractionation (FFF) techniques are commonly performed using non-
denaturing solution conditions to determine the extent and size of soluble protein
aggregates. Under denaturing conditions, that is, in the presence of additives such as
sodium dodecyl sulfate (SDS) or urea, protein aggregates are typically detected by
electrophoresis (SDS-PAGE, capillary SDS) or chromatography (SE-HPLC with
SDS or urea in the mobile phase, d-SEC). These electrophoretic and chromato-
graphic approaches are generally used to better characterize the size, amount, and
nature of soluble protein aggregates (e.g., presence or absence of native and nonna-
tive covalent, disulfide cross-links). More recently, mass spectrometry has been
employed to determine the size and nature of soluble protein aggregates. A brief
overview of the key analytical techniques used in the specific case studies described
9 Case Studies Applying Biophysical Techniques to Better Characterize Protein… 207
in this chapter is provided below, followed by a few illustrative examples from the
literature which have used combinations of these analytical techniques to evaluate
the size, amount, and nature of different types of soluble protein aggregates in
the size range of 1–100 nm.
9.2.1.1 Size-Exclusion High-Performance Liquid Chromatography
Due to its ease of use and high sample throughput, size-exclusion high-performance
liquid chromatography (SE-HPLC) is the most commonly utilized technique for
sizing and quantifying soluble protein aggregates in the molecular weight range of
5–1,000 kDa (Mahler et al. 2009; Philo 2009; Arakawa et al. 2010). Larger protein
aggregates cannot be accurately evaluated for their size since they migrate in the
void volume of the column or detected at all if they cannot pass through the column
matrix. Typically, larger aggregates need to be removed by filtration or centrifuga-
tion prior to injection onto the SE-HPLC column. One disadvantage of SE-HPLC
for use in sizing soluble protein aggregates is that the molecular weight (MW)
calibration standards typically used with this method are globular in nature. In con-
trast, protein aggregates are not necessarily spherical resulting in inaccurate MW
estimates. In addition, during sample preparation or during the chromatography
process, protein aggregates may reversibly associate or dissociate and will therefore
either not be detected or not represent the equilibrium state of interest (Carpenter
et al. 2009; Mahler et al. 2009). Nonetheless, SEC protein stability studies are abun-
dant and are a well-accepted approach (Oliva et al. 2001; Gabrielson et al. 2007;
Kiese et al. 2008, 2010; Tyagi et al. 2009; Van Buren et al. 2009; Bond et al. 2010;
Nayak et al. 2011). Since an often major limitation of SE-HPLC is the potential for
aggregates to interact with the column matrix, orthogonal methods such as analyti-
cal ultracentrifugation, FFF, or DLS, which monitor the hydrodynamic properties of
proteins free in solution, are now being more frequently used to verify SE-HPLC
results (Carpenter et al. 2010a; Bond et al. 2010).
9.2.1.2 Sedimentation Velocity Analytical Ultracentrifugation
An analytical approach often used to confirm SE-HPLC results is SV-AUC (Liu and
Shire 1999; Gabrielson and Arthur 2011). Briefly, an optical centrifugation cell is
loaded in a rotor that contains two chambers, one sample and one reference, with
detection of protein concentration in the cells by UV absorbance. The sample is
centrifuged at high g-force forming a moving boundary that migrates to the outside
of the rotor over time. Lower molecular weight species take longer to sediment than
their higher molecular weight (HMW) counterparts and this behavior is monitored
by absorbance at 280 nm over time. Software data analysis packages (e.g., SEDFIT)
are then used to calculate the percentage of each protein species based on their rela-
tive sedimentation time (Dam and Schuck 2004; Schuck 2000, 2004). Compared to
SE-HPLC, SV-AUC requires no dilution of the protein into a mobile phase and is
208 T. Wang et al.
performed in the absence of any separation matrices. The resolution of SV-AUC is

usually much greater than SE-HPLC and thus can better separate closely related
aggregated species in a protein sample. SE-HPLC is a higher-throughput technique
with straightforward data analysis, while SV-AUC is a more time-consuming,
lower-throughput technique with more care necessary in the analysis and interpreta-
tion of data.
9.2.1.3 Field-Flow Fractionation
An alternative approach to SV-AUC to detect and quantify soluble protein aggre-

gates is FFF (also known as asymmetrical flow field-flow fractionation or FFF) (Cao
et al. 2009). The FFF technique can also be useful for analyzing submicron-sized
(0.1–1 µm) protein particulates based on their molecular mass and size (Mahler
et al. 2009; Reschiglian et al. 2005). FFF methods separate molecules due to differ-
ences in their hydrodynamic radius (i.e., diffusion coefficients) when moving
through different velocity regions of a parabolic flow in a channel (Reschiglian et al.
2005). Recent reviews by Roda et al. (2009) and Yohannes et al. (2011) describe
novel applications and recent developments of the FFF technique with different
biopolymers and bioparticles. Taylor dispersion analysis (TDA), a related technique
based on monitoring dispersion of molecules when injected into a flow of solvent
which is used to determine diffusion coefficients and hydrodynamic radii of mole-
cules, has also recently been evaluated for its ability to monitor and size protein
aggregates (Hawe et al. 2011). The coupling of electrospray mass spectrometer with
differential mobility analyzers (ES-DMA) has been used to measure the size distri-
bution of protein aggregates (3–250 nm) by monitoring their gas phase electropho-
retic mobility (Pease et al. 2008). This technique separates molecules based on their
charge-to-aerodynamic size ratio (vs. ES-MS which measures the charge-to-mass
ratio).
9.2.1.4 Light Scattering
Various light scattering techniques have been employed to monitor protein aggrega-
tion in solution. DLS not only determines the size of protein aggregates in the 1 nm
to 1 µm size range but also can provide information about the polydispersity of the
sample (Kiese et al. 2008, 2010; Mahler et al. 2005). A DLS instrument records the
time-dependent fluctuations in the amount of scattered light when particles are
moving under Brownian motion. From the changes in these intensity fluctuations, it
is possible through autocorrelation analysis to determine a diffusion coefficient of
the particles and, through the Stokes–Einstein equation, their hydrodynamic radius
(Lomakin et al. 1999). Sizes can be estimated by an averaging technique known as
cumulant analysis or from various forms of deconvolution analysis that to a limited
extent resolve individual populations that differ in size by a factor of two or more.
As examples of the utility of DLS for monitoring protein aggregation, Mahler et al.
(2005) and Kiese et al. (2008) have monitored the aggregation behavior of IgG1
monoclonal antibodies during stirring and agitation under a variety of solution con-
ditions. Static light scattering (SLS) can monitor protein aggregation by measuring
changes in the average intensity of scattered light and if performed at different
angles and concentrations can determine changes in a protein molecule’s radius of
gyration (Murphy 1997). The radius of gyration of species in protein solutions can
also be determined by small angle neutron scattering (SANS), for example, in the
detection of microaggregates in different insulin preparations (Heldt et al. 2011).
There are now numerous examples using SE-HPLC in conjunction with UV absor-
bance, refractive index, and multiangle light scattering (MALS) detectors to obtain
absolute molecular weight of soluble, aggregated protein species (Bond et al. 2010;
Wen et al. 1996; Ye 2006). In addition, UV optical density spectroscopy is routinely
used to monitor the presence of aggregation through increases in optical density at
320–350 nm (Mach and Middaugh 2011). This approach has been widely used to
assess the ability of pharmaceutical excipients to minimize protein aggregate for-
mation in a high-throughput format (Bhambhani et al. 2012).
9.2.1.5 SDS-PAGE, cSDS, and d-SEC
The size, extent, and nature of soluble protein aggregates can also be evaluated
under denaturing conditions. Electrophoresis is a standard way to study protein
aggregation, in the molecular weight range of 5–500 kDa (Mahler et al. 2009), using
an electric field to separate macromolecules based on their charge and weight. The
most common electrophoretic technique used to separate and quantify protein
aggregates of different molecular weight is SDS polyacrylamide gel electrophoresis
(SDS-PAGE). SDS-PAGE not only allows size estimation but can also be used to
estimate purity levels and evaluate for the presence of covalent, reducible (e.g.,
disulfide-linked) aggregates (Mahler et al. 2009). Noncovalent protein aggregates,
however, are generally not detected since they are usually disassociated during sam-
ple preparation due to the presence of SDS. Capillary electrophoresis (CE) offers an
attractive alternative to conventional gel electrophoresis due to its robustness, speed,
enhanced resolution, and reproducibility as well as its automation and quantitative
features. CE techniques are now routinely used in the biopharmaceutical industry in
the quality control environments for purity testing as well as assessing stability and
integrity of protein drugs. Although a number of CE modes are available, the most
common technique used is capillary electrophoresis SDS (CE-SDS). A detailed
description of these techniques can be found elsewhere (Kraly et al. 2006; Rustandi
et al. 2008; Kilar 2003). In addition to electrophoresis, SE-HPLC can be run with
SDS, urea, or guanidine hydrochloride in the mobile phase (denaturing SE-HPLC or
d-SEC) to separate disrupted protein aggregates under both reduced and nonreduced
conditions (Michels et al. 2007). Often, dissociation of aggregates occurs prior to
unfolding of the monomeric species as the concentration of the unfolding agent is
increased permitting further analysis. CE-SDS is more commonly used in quality
control environments to quantitatively determine protein aggregate size and purity
210 T. Wang et al.
as well as confirm the presence/absence of disulfide bonds. In contrast, d-SEC is

often utilized during formulation development to rank order protein stability under
different conditions. A number of studies have been published that demonstrate the
utility and feasibility of using these three methods (SDS-PAGE, CE-SDS, d-SEC)
for purity and stability assessment of protein drug candidates under denaturing con-
ditions, especially with monoclonal antibodies (Rustandi et al. 2008; Salas-Solano
et al. 2006; Han et al. 2006; Hunt et al. 1996; Qi et al. 2009; Costello et al. 1992;
Lubiniecki et al. 2010).
9.2.1.6 Mass Spectrometry
High-resolution mass spectrometry techniques are commonly used to characterize

the primary structure of proteins as well as posttranslational modifications (Grillberger
et al. 2009). The ability of MS methods to detect and characterize protein aggregates
is also rapidly progressing and provides high-resolution and site-specific informa-
tion. Mass spectrometry (MS) is used to selectively detect, quantify, and determine
the covalent structure of a given molecule by determining the mass-to-charge (m/z)
ratio (Watson and Sparkman 2008). Analysis by MS requires the molecule to be
ionized and present in the gas phase. A typical mass spectrometer includes four
major components: an ionizer, a m/z analyzer, an ion detector, and a vacuum system.
Until the development of electrospray ionization mass spectrometry (ESI-MS) and
matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), mass
spectrometry had limited use in the analysis of biological molecules (Chait 2011).
The advantages of ESI include (a) ability to ionize fragile molecules intact, (b) ease
of coupling to liquid chromatography, and (c) efficient generation of multiple
charged ions to measure high molecular weight biopolymers. On the other hand, a
major limitation of ESI-MS is its intolerance of impurities and salts. The ability of
ESI-MS to directly monitor soluble protein aggregate formulation during heat treat-
ment of human antithrombin III was recently reported (Wang et al. 2012). While
MALDI-MS possesses high speed and a relatively high tolerance to the presence of
contaminants, it is also relatively nonquantitative. In addition to the selection of
ionization techniques, there are many types of m/z analyzers, including time of
flight, quadruple ion traps, and transmission quadrupole. The specific analytical
application determines the optimal choice of setup and type of mass spectrometer.
Sample preparation for MS analysis of proteins can be divided into two
approaches: bottom-up and top-down. In the bottom-up approach, proteins are
first digested with an enzyme, and the resulting peptide fragments are separated
and analyzed by MS and MS/MS to produce a unique “fingerprint” of an indi-
vidual protein. This “peptide map” of a protein can be compared to the theoreti-
cal gene-derived amino acid sequence to establish or confirm a protein’s identity
as well as to determine if degradation or posttranslational modifications have
occurred. The advantages of the bottom-up approach include its high accuracy in
mass determination and ease of use (Chait 2011). In the top-down approach,
intact proteins are directly analyzed in the mass spectrometer without prior
enzymatic digestion. This approach gives a molecular mass of the intact protein,
their major posttranslational modifications and can potentially detect the pres-
ence of protein aggregates.
Not only have both of these two approaches been used to examine protein aggre-
gates, but they have also been coupled to the analysis of kinetic exchange rates of
hydrogen for deuterium in the protein amide backbone (H/D exchange-MS), to gain
insight into the role of protein flexibility and dynamics in protein aggregation path-
ways. The rate of H/D exchange in the amide backbone hydrogens in a protein dif-
fers in flexible or disordered regions (lacking stable hydrogen bonding) compared to
more tightly folded regions (protected from H/D exchange) (Konermann et al. 2011).
By quenching the H/D exchange reaction at different time points and enzymatically
digesting the protein under quenched conditions, the rate of H/D exchange in differ-
ent regions of the protein can be determined by LC–MS analysis of each peptide over
time (Marcsisin and Engen 2010). Furthermore, regions protected by aggregation
can be identified. Recent advances in chromatographic separations such as UPLC
technology, combined with software advances in the analysis of H/D exchange rates
in hundreds of peptides, has permitted the characterization of the conformational
dynamics of IgG monoclonal antibodies by H/D exchange-MS (Houde et al. 2009).
9.2.1.7 Surface Plasmon Resonance
Surface plasmon resonance (SPR) is a method based on energy transfer from light
(photons) to the electrons on a metal surface (Ahrer et al. 2003). As polarized light
interacts with the surface of a biosensor chip coated with gold, an evanescent wave
is created that is highly sensitive to the changes in the dielectric constant of the adja-
cent medium. This energy transfer can create an angle of minimum reflectance (the
SPR angle) under certain conditions. The change in SPR angle is proportional to the
mass of bound material at the chip surface (Thillaivinayagalingam et al. 2010). By
analyzing real-time changes of the SPR angle (SPR sensorgrams) as fluid passes
over the surface of the biosensor, both the extent of analyte binding to the surface as
well as binding rate kinetics (kon, koff) and binding affinities (KD) can be obtained.
SPR has been extensively applied to the quantitative analysis of protein-ligand and
protein-protein interactions and facilitated by the availability of automated equip-
ment. One requirement for the application of SPR to biological molecules is the
successful immobilization of sufficient amounts of ligand to the surface of the sen-
sor chip. Chemical coupling methods (e.g., through amines, thiols) covalently bind
the targeted ligand onto the biosensor surface. Since a lack of control over the ori-
entation of protein ligands might affect subsequence analyte binding, indirect cou-
pling methods use a second ligand (e.g., an antibody or avidin) in which the target
ligand binds to the immobilized protein creating a more homogenously oriented
bound ligand (Wear et al. 2005).
212 T. Wang et al.
9.2.2 Case Studies Characterizing Size and Amount

of Soluble Protein Aggregates Under Native-Like
and Denaturing Conditions
This section will cover the three case studies illustrating a variety of analytical
approaches to characterize different types of soluble protein aggregates, using
methods under native-like (SE-HPLC and AUC) and denaturing conditions
(SE-HPLC and CE with SDS). The first study uses a novel application of SE-HPLC,
along with orthogonal testing by SV-AUC, to improve the sensitivity of aggregate
detection for a variety of different types of aggregates of an IgG1 monoclonal anti-
body generated from different environmental stresses. The second study examines
similar analytical approaches to characterize a high molecular weight aggregate of
a recombinant therapeutic glycoprotein. The third utilizes analytical methods under
denaturing conditions (SE-HPLC and capillary electrophoresis with SDS) to deter-
mine the level of reducible and nonreducible protein aggregates in stressed samples
of a monoclonal antibody.
In the first case study, an IgG1 monoclonal antibody was subjected to different
environmental conditions such as elevated temperature, decreased pH, and light
exposure to generate different aggregates. Bond et al. were able to accurately, and
with improved sensitivity compared to light scattering or MALS, quantify the
amount of aggregate in force-degraded IgG1 monoclonal antibody solutions using
a dual-wavelength approach (DW-SE-HPLC) with very good recovery (Bond et al.
2010). This optimized SE-HPLC technique monitored aggregate, monomer, and
fragment peaks at two ultraviolet wavelengths (214 and 280 nm). It was determined
that the limit of quantitation for this new method was 0.04%, which was much lower
than the 0.2% found for the previously used single wavelength SE-HPLC method.
In addition, the new approach extended the dynamic range ~five to ten fold. As
shown in Fig. 9.1, different environmental stresses resulted in formation of different
types of oligomeric species and soluble protein aggregates (dimer vs. multimer as
well as covalent cross-links vs. noncovalent interactions). After assay optimization
to minimize interactions with the column matrix, the DW-SE HPLC results corre-
lated well with SV-AUC data in terms of the ability to quantify aggregate levels in
solutions (Fig. 9.1). Interestingly, a high-concentration formulation of the IgG1
monoclonal antibody, when stressed at 40°C for 3 months, showed a low-level
formation of a different type of oligomeric species, a noncovalently associated
dimer (Bond et al. 2010).
Several investigations have now appeared using SV-AUC to confirm the results
obtained with SE-HPLC, specifically to verify a lack of effect of SE-HPLC sample
handling and mobile phase conditions on aggregate formation and dissolution
(Berkowitz 2006; Hughes et al. 2009). In one interesting example with a complex
protein, Hughes and colleagues quantitatively investigated the ability of SE-HPLC
and SV-AUC to detect known amounts of protein aggregates of recombinant human
acid alpha-glucosidase (rhGAA), an enzyme with seven N-lined glycosylation
sites and an apparent molecular weight of 110 kDa. Overall, the authors found a
Fig. 9.1 Analytical characterization of IgG1 monoclonal antibody solutions containing soluble
aggregates measured by orthogonal methods, SE-HPLC (a, b, c) and SV-AUC (d, e, f). Protein
aggregates of varying size and composition were formed by exposure to different environmental
stresses: (a, d) control, (b, e) heating under acidic pH conditions, and (c, f) light exposure. Adapted
from Bond et al. (2010)
good correlation between the two methods, since the amount of the soluble
aggregates measured was not statistically different (Hughes et al. 2009) (Table 9.1).
It should be pointed out, however, that other studies in the literature have found
significantly higher percentages of soluble protein aggregates when measured by
SV-AUC compared to SE-HPLC. This suggests that the SE-HPLC approach may
require additional method development to prevent aggregate adsorption to the col-
umn matrix with certain proteins. For example, in work by Gabrielson et al. with a
214 T. Wang et al.
Table 9.1 Quantitative analysis of aggregate levels as determined by SV-AUC and SE-HPLC for
forced degraded samples of recombinant human acid alpha-glucosidase (rhGAA)
rhGAA forced aggregate (%)
0 0.25 0.5 1.25 2.5 4 5 10 20
% Aggregation (AUC), 0.4 0.9 0.6 1.3 2.5 3.1 4.2 8.5 18.1
N = 6a
SD 0.3 0.7 0.4 0.5 1.1 0.3 0.4 0.7 1.3
RSD (%) 67.0 84.0 69.4 35.7 44.0 8.8 10.4 8.1 7.0
% Aggregation (SEC), N = 9 0.2 0.3 0.5 1.0 1.9 3.0b 4.1 8.4 17.5
SD 0.02 0.03 0.02 0.04 0.03 0.04 0.04 0.04 0.42
RSD (%) 9.2 10.8 3.7 3.7 1.6 1.2 1.0 0.5 2.4
Published in Hughes et al. (2009)
a
Samples were analyzed in a single AUC run using six different cells
b
These SEC data were obtained on six samples
recombinant humanized mAb (Gabrielson et al. 2007), SV-AUC detected one

major peak and three minor peaks, while SE-HPLC only detected one major and
one minor peak. In the same study, FFF and SV-AUC were compared as orthogonal
methods to monitor the amount of soluble aggregates. It was shown that both
SV-AUC and FFF not only detected much higher amounts of aggregates compared
to SE-HPLC but also found additional HMW species. Nonetheless, it should be
pointed out that although FFF has no stationary phase (Fraunhofer and Winter
2004; Liu et al. 2006), protein samples can undergo concentration and dilution
stages in FFF in which protein aggregates may differentially interact with the
medium and affect the amount of reversible aggregate in the sample (Mahler et al.
2009; Reschiglian et al. 2005; Liu et al. 2006).
A third example employs CE-SDS and d-SEC to characterize aggregate forma-
tion in a monoclonal antibody solution under denaturing conditions. One of the first
studies to utilize both CE-SDS and SDS-PAGE for the quantitative analysis of a
recombinant humanized monoclonal antibody (rhuMabHER2) was performed by
Hunt et al. (1996). Both methods detected seven peaks in a nonreduced sample that
correlated well with the seven bands observed for HER2 mAb on SDS-PAGE. The
relative peak areas percent ascribed to high molecular mass aggregate observed on
CE-SDS were consistent with that seen with SEC. The method, however, was found
to be less sensitive than SDS-PAGE with silver staining in detecting minor species
under reducing conditions. To improve sensitivity, precolumn derivatization with fluo-
rogenic reagents coupled with laser-induced fluorescence (LIF) detection are now rou-
tinely used in CE-SDS analysis (Michels et al. 2007; Salas-Solano et al. 2006). Using
this improved methodology, the ability of d-SEC and CE-SDS to determine the level
of nonreducible, cross-linked protein in a degraded monoclonal antibody solution
was evaluated (Fig. 9.2). SE-HPLC was initially used to establish the presence of
monomer, dimer, and HMW aggregates. These species were then evaluated for the
presence and absence of reducible and nonreducible cross-linked proteins under
denaturing conditions by both d-SEC and CE-SDS (Fig. 9.2). Although both meth-
ods showed similar trends, the CE-SDS technique provided better resolution and an
improved ability to quantify the levels of these aggregates (Michels et al. 2007).
Fig. 9.2 Analytical characterization of monomers, dimers, and high molecular weight (HMW)
aggregates of a monoclonal antibody as measured by SE-HPLC (native-like conditions) and d-SEC
and CE-SDS (denaturing conditions). (a) SE-HPLC, (b) d-SEC, and (c) CE-SDS with indicated
levels of percent nonreducible aggregates. Published in Michels et al. (2007)
9.2.3 Case Studies Characterizing Physical Composition

and Biological Activity of Soluble Protein Aggregates
Despite recent advances in the use of biophysical methods to better characterize the
higher-order structure and conformational stability of proteins, especially when
combining multiple techniques with advanced data analysis approaches (Maddux
et al. 2011), the determination of a protein’s biological activity remains the “gold
standard” in terms of monitoring subtle conformational changes. Two case studies
are presented in this section, first demonstrating correlations of certain physical
properties of protein aggregates with biological activity (Remmele et al. 2006) and
second showing how biophysical binding assays using biosensors (SPR) can be
employed to assess the Fc receptor binding activities of different aggregates and
complexes of two different monoclonal antibodies (Luo et al. 2009).
In the case of an IgG1 monoclonal antibody (epratuzumab), samples were shown
to contain both monomers (~150 kDa) and dimers (~300 kDa). The dimers con-
tained ~70% covalent cross-links (as measured by CE-SDS) and consisted of three
different assemblies including Fab-Fab, Fc-Fc, and Fab-Fc complexes. The biologi-
cal activity of the dimers, in terms of relative potency in a cell-based bioassay, was
shown to be twice that of the monomer (i.e., equal activity on a weight basis indicat-
ing that the dimer has twice the number of binding sites as the monomer) (Remmele
et al. 2006). A more recent investigation examined the binding affinities of different
mAb complexes and aggregates using an SPR biosensor (Luo et al. 2009). Similar
to the previous study, monomeric and aggregated forms of the mAbs had similar
antigen-binding properties in the Fab region. It was demonstrated, however, that the
monomers and aggregates showed differences in their Fc region’s ability to bind
different Fc-gamma receptors. As shown in Fig. 9.3a, b, the monomeric form of
216 T. Wang et al.
Fig. 9.3 The Fc receptor binding of the immune complexes and aggregates of mAb1 (a and b) and
mAb2 (c and d) as measured by SPR binding assays. Two different Fc receptors were evaluated:
FcγRIIA (a and c) and FcγRIIIB (b and d). The SPR sensorgrams for the binding of mAb mono-
mer, dimer/HMW, and their immune complexes (containing 1:1 molar ratio of monomeric anti-
body and multivalent antigen) are overlaid. Published in Luo et al. (2009)
mAb1 manifested different binding properties to both the Fc-gamma RIIA and
Fc-gamma RIIIB receptors compared to the dimeric aggregate and a multivalent
immune complex formed between mAb1 and its antigen. Similar results are shown
in Fig. 9.3c, d for monomeric mAb2, high molecular weight aggregates of mAb2,
and a multivalent immune complex formed between the mAb2 and its antigen. The
dimers and multimers of these mAbs showed higher in vitro binding affinities to
Fc-gamma receptors as well (Luo et al. 2009).
9.2.4 Case Studies Characterizing Soluble Protein

Aggregates by Mass Spectrometry
Although the use of mass spectrometry to better characterize soluble protein aggre-
gates formed during storage has been more limited in its applications to date, MS
has been pursued for many years to better characterize well-ordered amyloid fibril
protein aggregates. For example, nanoflow ESI combined with Q-TOF mass
spectrometer has been utilized to study the self-assembly, aggregation, and amyloid
fibril formation of insulin under different conditions (Nettleton et al. 2000). Another
example is the use of ESI-quadrupole MS to quantify β(beta)-amyloid monomers in
the presence of potential therapeutic agents and to identify and rank order the com-
pounds which inhibit the aggregation of β(beta)-amyloid monomers (Cheng and
van Breemen 2005). The initial applications of MS technology to characterize pro-
tein aggregates under pharmaceutical conditions have combined chromatographic
separations with subsequent MS analysis. For example, Van Buren et al. character-
ized dimers of an IgG after isolation and enzymatic digestions (Van Buren et al.
2009). A combination of SE-HPLC, sample dialysis, and ESI-TOF MS was recently
used to identify and characterize different IgG aggregates (Kukrer et al. 2010). In
addition to complex sample handling with limited throughput, these MS approaches
have certain limitations including potential changes in the aggregate profile during
sample dialysis and a bias of MS results toward lower MW species when monitor-
ing a sample containing a mixture of oligomers (Kaltashov et al. 2012).
More sophisticated MS approaches are now being developed to examine and
characterize protein aggregates without these experimental limitations as described
in the following three case studies: (1) use of EDI-MS to directly monitor in real-
time heat-induced protein deformation and aggregation (Wang et al. 2011), (2) use
of ion mobility MS to characterize intact protein assemblies in the gas phase, and
(3) the use of H/D exchange-MS to characterize different proteins (Kheterpal and
Wetzel 2006; Zhang et al. 2011, 2012).
As temperature increases, loss of a protein’s conformational integrity results in
formation of high-charge-density ions and oligomers that can be easily identified by
higher m/z ratios. A new design of a temperature-controlled ESI source not only
permits direct monitoring of protein unfolding and aggregation but also ensures the
efficiency of the heating process and prevents cooling of the protein solution during
the sample introduction to the ESI interface (Wang et al. 2011). With this new
approach, both reversible and irreversible protein unfolding events were captured
by ESI-MS for cytochrome-C and glucocerebrosidase (GCase), respectively.
The ability of ESI mass spectra to directly monitor the formation of dimers, trimers,
tetramers, and pentamers during heat treatment of GCase is shown in Fig. 9.4.
Ion mobility spectrometry (IMS) is an analytical technique used to separate and
identify ionized molecules in the gas phase based on their mobility in a carrier buf-
fer gas. When coupled with mass spectrometry, ion mobility mass spectrometry
(IM-MS) enables the separation and detection of isomers and conformers. More
recently, the technology (also referred to as native MS) has been successfully used
to characterize larger macromolecules, such as proteins and their biological com-
plexes, in the gas phase (Uetrecht et al. 2010). Although Native MS has been
employed to examine the composition and stability of numerous large biological
complexes, confirmation of results with other methods (examining the same protein
complex in the solution state) is typically performed (van Duijn 2010). Kaddis et al.
used ESI-IMS to study the general size dimensions of proteins as protein molecules
were physically transitioned from the solution to the gas phase (Kaddis et al. 2007).
Traveling voltage-wave ion mobility mass spectrometry (TWIMS) is a recent
218 T. Wang et al.
Fig. 9.4 Electrospray ionization (ESI) mass spectra of glucocebrosidase (GCase) in 20 mM

ammonium acetate, pH 4.7 recorded at various solution temperatures. The numbers without paren-
theses indicate charge states of the GCase monomers; the numbers in parentheses indicate the size
of the protein associations. Published in Wang et al. (2011)
advancement in the technology coupling an ion mobility unit to Q-TOF MS. The
application of TWIMS to the characterization of hepatitis B viral capsids (MW of
~3,000 and ~4,000 kDa) as well as GroEL and its complexes with smaller proteins
(MW of 801 and 857 kDa, respectively) has been recently reviewed (Uetrecht et al.
2010).
Hydrogen–deuterium exchange linked to MS (H/D exchange-MS) analysis has
been used to probe the secondary structure of amyloid fibrils of insulin (Kheterpal
and Wetzel 2006). This work combined an ESI-MS system with a bottom-up experi-
mental approach (peptide mapping). First, protonated fibrils were collected by
centrifugation, washed, resuspended, and incubated in deuterated buffer to promote
exchange of labile protons with deuterium. Second, the exchange reaction was
quenched and the fibrils were disaggregated. Finally, the monomers were subjected
to proteolysis and analyzed by MS. The results showed that both the N- and
C-terminal segments (1–19 and 35–40) are readily exchangeable (not involved in
the core structure), whereas the fragment containing residues 20–34 was highly
protected from exchange. H/D exchange-MS has also been used to the study of the
effect of protein flexibility and sequences on protein aggregate formation with bio-
pharmaceutical products including interferon-gamma (Tobler and Fernandez 2002)
and, more recently, with the model enzyme lactate dehydrogenase (LDH) (Zhang
et al. 2011) as well as a monoclonal antibody (Zhang et al. 2012). For smaller pro-
teins such as interferon-gamma and LDH, the rate of H/D exchange can be moni-
tored in the intact protein, as well as at the peptide level (after digestion with pepsin),
to identify sites of increased and decreased flexibility. In contrast, due to their large
size, mAbs are only analyzed after pepsin digestion and H/D exchange reactions
must be carefully examined across hundreds of peptides. The H/D mass spectra of
native and freeze-thaw (F/T)-induced aggregates of LDH are shown in Fig. 9.5. It
can clearly be seen that F/T-induced aggregates consist primarily of unfolded pro-
tein. Peptide level analysis identified several short sequences (13–31, 109–117, and
133–143) with moderate protection from H/D exchange indicating that these
sequences may be the regions of intermolecular interactions within the LDH aggre-
gates induced by freeze-thaw (Zhang et al. 2011). In contrast, in the case of a mono-
clonal antibody, F/T aggregates had similar H/D protection maps compared to
soluble protein, indicating a primarily native-like conformation (Zhang et al. 2012).
9.3 Case Studies Using Biophysical Methods to Count, Size,

and Characterize Protein Submicron Particles
9.3.1 Overview of Analytical Methods Used in Selected Case

Studies to Evaluate Submicron Protein Particulates
The development of improved analytical approaches to examine the size, amount,

and nature of protein particles in the size range of 0.1–1 µm is an ongoing challenge
(Carpenter et al. 2009). Some currently available approaches include DLS
techniques, either as stand-alone instruments or connected to a microscope
[nanoparticle tracking analysis (NTA)]. In addition, more sensitive microscopic
techniques, most frequently used to characterize protein-based amyloid fibrils, are
being evaluated for applicability to better characterize pharmaceutical protein
aggregates. These include atomic force microscopy (AFM), transmission and scan-
ning electron microscopy (TEM and SEM), as well as quartz crystal microbalances
220 T. Wang et al.
Fig. 9.5 Hydrogen–deuterium exchange mass spectrometry (H/D-MS) spectra of native lactate
dehydrogenase (LDH) and freeze/thaw aggregated LDH obtained after 10 s, 10 and 180 min of
deuterium labeling. The dotted spectra in the top and bottom panels are unlabeled and fully labeled
controls, respectively. Published in Zhang et al. (2011)
(QCMs). A brief overview of the key analytical techniques used in the case studies
presented in this chapter is provided below, followed by a few illustrative examples
from the literature which have employed useful combinations of these techniques.
9.3.1.1 Nanoparticle Tracking Analysis
NTA is a laser-illuminated-based microscopy technique that detects and sizes submi-

cron particles in the range of ~50–1,000 nm. NTA individually tracks and sizes
particles moving under Brownian motion. With a knowledge of the viscosity and
temperature of the sample, the instrument calculates the diffusion coefficient which
can then be used via the Stokes–Einstein equation to calculate the hydrodynamic size.
This new analytical technique is currently being assessed by several groups to detect
protein aggregates and particulates that have not been accurately analyzed previously
(Filipe et al. 2010, 2011; Engelsman et al. 2011).
9.3.1.2 Atomic Force Microscopy
An atomic force microscope (AFM) consists of a cantilever with an ultrasharp tip

(probe), a sample stage, and an optical deflection system (Colton et al. 1997).
During AFM measurements, the cantilever is brought into close (atomic) contact
with the sample and scanned across the surface (contact mode). The interaction
between probe and sample, which can include mechanical contact, van der Waals,
capillary, electrostatic, and/or magnetic forces, causes the bending of the cantilever
which is precisely recorded by an optical deflection system. The scanning is done
under feedback control so that the bending of the cantilever remains constant to
maintain a constant force. The up and down motion of the sample is a record of the
sample topography. In the tapping mode, which senses the oscillation amplitude of
cantilever instead of bending, AFM generally causes less damage to sample. For
analysis of biological materials such as protein aggregates, tapping mode AFM is
preferred. One key limitation of AFM is it can only analyze a small fraction of the
total sample volume at any one time.
9.3.1.3 SEM and TEM
Scanning electron microscopy (SEM) and transmission electron microscopy

(TEM) are standard tools used in biological research to examine a wide variety of
samples due to their high resolution (from several nanometers to microns). Both
SEM and TEM can also be used to characterize protein aggregates in the submicron
size range. These techniques use an electron beam to bombard a specimen and
generate various types of electron scattering (Joseph Goldstein et al. 2003; David
Bernard Williams 2009). In the case of transmission electron microscope (TEM),
electrons that go through a specimen (forward scattering) are collected and ana-
lyzed to provide image-based information on sample size and shape. The scanning
electron microscope (SEM) collects the secondary and backscattered electrons to
provide data not only on 2D sample size but also surface topology. Sample prepa-
ration remains a challenge for TEM and SEM (which requires fixation, staining,
and drying) in terms of ensuring the size and morphology of the sample do not
change during preparation and analysis. Fortunately, recent advances in cryotech-
niques (rapid freezing in liquid ethane near liquid nitrogen temperatures) can mini-
mize such problems by analyzing samples in the frozen liquid state. TEM and
SEM analysis is also limited by sampling frequency, requiring acquisition of
numerous images to provide a representative analysis of the sample to determine
particle number and size.
9.3.1.4 Quartz Crystal Microbalance
A new approach to monitor submicron protein particulates in solution is the QCM

(Ferreira et al. 2009). A thin quartz disk is sandwiched between two metal elec-
trodes. When an alternating electric field is applied, a shear deformation will be
222 T. Wang et al.
generated due to the piezoelectric nature of the quartz material. When the mechanical
oscillations are close to the fundamental frequency of the crystal, a resonant oscil-
lation is achieved. Changes in the resonant frequency are related to the mass accu-
mulated on the crystal (Sauerbrey 1959). Commercial systems are designed to
reliably measure mass changes up to ~100 µg (microgram) with a sensitivity of
~1 ng cm−2 (O’Sullivan and Guilbault 1999). Although QCM has been used suc-
cessfully in other fields (e.g., vacuum deposition, thin film deposition control),
application to biological samples has been limited due to its inherent nonspecificity.
Recently, chemical modification of the electrode surface enabled the real-time study
of bovine insulin aggregation to form amyloid fibrils (Knowles et al. 2011).
Immersion of resonators into aqueous solutions, however, significantly reduces
their resolution due to the damping effect of solution viscosity (Burg et al. 2007).
A recently developed nanomechanical resonator, Archimedes, contains a solution
channel inside a hollow resonator that is surrounded by vacuum to overcome this
damping effect. As a result, Archimedes has improved mass resolution (orders of
magnitude) over a commercial QCM for aqueous measurements (Burg et al. 2007).
9.3.2 Case Studies Involving Counting and Sizing

Submicron Protein Particles
In a comparative study by Filipe et al. (2010), the ability of NTA and DLS to size
submicron particulates using polystyrene bead standards of various sizes was initially
examined. Although both techniques can accurately size monodisperse standards,
NTA offers the advantage of also being able to count the number of particles in the
sample. In addition, it was possible to visualize the particles in the samples using
NTA. When particle standards of different sizes were mixed together, NTA can more
easily distinguish their relative amounts compared to DLS. In this study, DLS detected
the largest particle standard in a mixture, while NTA detected each of the standards
and thereby did not skew the reported size distribution. The formation of submicron-
sized protein aggregates was then monitored with heat treated IgG and metal-oxidized
insulin. NTA was again better at analyzing sample polydispersity than DLS. As shown
in Fig. 9.6, NTA can be used to monitor particle counts and size distribution in real
time during the formation of IgG submicron-sized particulates during heating at 50°C.
Recently, Filipe et al. used NTA in conjunction with fluorescence single-particle
tracking analysis to obtain the size of fluorescently labeled monoclonal IgG and HSA
when subjected to heat stress (Filipe et al. 2011).
Nanomechanical resonators, including QCMs, have achieved extraordinary sen-
sitivity in the detection of mass accumulation on surfaces in a vacuum. The detec-
tion of most bimolecular samples, however, requires a solution environment. As
described above, a recently developed nanomechanical resonator, Archimedes,
includes a solution channel inside a hollow resonator that permits solution measure-
ments (Burg et al. 2007). This device is able to measure the mass of objects that are
not attached to the resonator surface, such as floating polystyrene beads, agglutinating
microspheres, and bacteria (Burg et al. 2007; Rumi Chunara et al. 2007; Michel
Godin et al. 2007). The instrument vendor’s website (http://www.affinitybio.com/
applications/protein_formulations.php) provides examples (from currently unpub-
lished data) of the detection of submicron-sized IgG aggregates. Both size and num-
ber distributions of protein particles in solution can be displayed in the form of a
histogram.
9.3.3 Case Studies Which Characterize the Morphology

of Submicron Protein Particles
To better characterize the morphology and size of submicron protein particles,

microscopic techniques such as AFM and TEM have been successfully used.
Tapping AFM has been used for several years to study the in vitro formation of
amyloid fibrils. For example, the process of in vitro generation of amyloid fibrils by
seeded polymerization has been examined (Harper et al. 1997). In addition, a vari-
ety of biophysical techniques including AFM were utilized to study the aggregation
and disaggregation of an amyloidogenic protein, human muscle acylphosphatase
(Calamai et al. 2005). In this work, it was observed that globular aggregates from
100 to 200 nm and larger aggregates (>5 µm) were the dominant species as the
aggregation time increased. Interestingly, upon diluting the protein solution, a vari-
ety of fibrillar structures was observed with AFM indicating disaggregation of the
larger protein complexes.
In terms of pharmaceutical applications, tapping mode AFM has been recently
used to better characterize the nature of the aggregation of monoclonal antibodies
(Cao et al. 2010; Lee et al. 2011; Mach and Arvinte 2011). In one such study, most of
the mAbs manifested similarly sized aggregation intermediates on the order of several
monomers. The subsequent agglomeration of these intermediates into larger particu-
lates, however, was dependent on solution conditions as evaluated by AFM images
(Lee et al. 2011). For example, monomeric mAb1 had an average size of ~15 nm.
Upon heat treatment, the presence of oligomeric structures was seen when the sample
was diluted into water. In contrast, when the heat-treated mAb1 sample was diluted
with saline, agglomerates of the oligomer structures began to appear. The morphology
and size of the agglomerates appeared to differ between the different mAbs examined
suggesting protein-specific and/or stability-dependent mechanisms. For example, as
shown in Fig. 9.7, mAb3 formed submicron-sized particles with amorphous morphol-
ogy containing agglomerates of smaller sized oligomers. These investigators point out
some of the analytical limitations of studying protein particle formation with AFM
including a very small fraction of total sample volume being examined as well as the
sample preparation process potentially producing artifacts. As a result, at least one
orthogonal method was recommended to confirm the results from AFM.
Although numerous groups have used TEM to study and characterize the struc-
ture of amyloid fibrils (Nettleton et al. 2000; Kheterpal and Wetzel 2006), TEM has
not been widely used to characterize protein aggregates and particles from a phar-
maceutical perspective. One group combined quantitative rheology measurements
224 T. Wang et al.
Fig. 9.6 Monitoring of IgG aggregation and submicron particle formation at 50°C in the NTA
(NanoSight) instrument sample chamber. The size distribution (middle panels) with the corre-
sponding NTA video frame (left panels) and 3D graph (size vs. intensity vs. concentration, right
panels) are shown. Published in Filipe et al. (2010)
and negative stain TEM to study the assembly of monoclonal antibody into highly
ordered structures in the presence of multivalent carboxylate ions (Esue et al. 2009).
As shown in Fig. 9.8, TEM images demonstrate smaller bundles that were composed
of straight, single filaments with an average diameter of 4 nm. In addition, the pres-
ence of larger bundles with diameters as large as 200 nm, containing midsized
Fig. 9.7 Atomic force microscopy (AFM) visualization of submicron-size protein particles from
a monoclonal antibody solution: (a) mAb aggregate (in water), (b) the same mAb aggregate in
different viewing area. Published in Lee et al. (2011)
bundles from 29 to 55 nm in diameter, was observed by TEM. This type of detailed

morphological information is currently only available from TEM studies. Sample
preparation, however, involves the drying of the sample, again inducing possible
sample preparation artifacts. The use of cryo-TEM to examine the morphology of
protein particles in the liquid state will be of great interest in future studies.
9.4 Case Studies Using Biophysical Methods to Count, Size

and Characterize Protein Subvisible and Visible Particles
9.4.1 Overview of Analytical Methodologies

from Case Studies Used to Evaluate Subvisible
and Visible Protein Particles
Current analytical methods utilized to count and size subvisible and visible particu-
lates in the size range of 1–100 µm and over 100 µm, respectively, include coulter
counter, light obscuration (LO), microflow digital imaging (MFI), light microscopy,
and visual assessments. Techniques employed to better characterize the morphology
and composition of subvisible and visible protein particulates include zeta potential
measurement, flow cytometry, fluorescence, and FTIR microscopy as well as SEM-
EDX. A brief discussion of each technique will be presented below followed by
relevant case studies that have incorporated these techniques to size, count, and
characterize subvisible and visible protein particles.
226 T. Wang et al.
Fig. 9.8 Transmission electron microscopy (TEM) visualization of submicron and subvisible
protein particles from a monoclonal antibody solution containing citrate. Analysis of micrographs
showed large bundles containing smaller filament bundles. Individual straight filaments were of
varying lengths and had an average diameter of 4 nm. Scale bars are 1 µm and 100 nm. Published
in Esue et al. (2009)
9.4.1.1 Coulter Counter
The Coulter method counts subvisible particles in the size range of ~0.5–50 µm
(micrometer). Unlike LO and MFI, the Coulter method does not rely on the optical
characteristics of the particles, but rather on the perturbation of an electric field.
Single particles flow through a pore and electrical pulses are generated, which pro-
duce a change in conductance. This change is directly proportional to the particle
volume. To measure a change in conductance, the Coulter method requires a con-
ductive buffer to permit an electric field to form. In some cases, this can be accom-
plished by adding a small amount of highly concentrated NaCl (or other electrolyte)
to the buffer (Barnard et al. 2012). Since only single particles can transverse the
pore, the sample must be diluted. Both dilution and the addition of salt during sam-
ple preparation are of concern when monitoring protein particulates since these
steps may either disassociate aggregates or induce their formation. Until recently,
Coulter counter measurements required large sample volumes (10 ml), but newer
instruments have significantly decreased sample volumes to as little as ~0.1 ml
(Barnard et al. 2012; Rhyner 2011; Roberts et al. 2012). These reduced sample vol-
umes are far more practical for use with protein containing samples.
9.4.1.2 Light Obscuration
The principle of light obscuration (LO) is simple: the sample is passed through a
chamber containing a light source. The amount of light that is blocked is directly
proportional to the number and size of the particles. This technique is widely used
as the “gold standard” for measuring subvisible particles in pharmaceutical applica-
tions since both US and EU pharmacopeia methods have been described. The con-
centration and size of particles ranging from 2 to 100 µm (micrometer) can be
determined. Advantages of this technique include speed and the ability to cover the
entire subvisible size range in a single measurement. Disadvantages of LO include
a lack of morphological information and potential artifacts arising from air bubbles.
Therefore, samples are typically degassed prior to analysis (Huang et al. 2009;
Chrai et al. 1987). In addition, LO has been shown to undercount the number of
proteinaceous subvisible particles as described in more detail below.
9.4.1.3 Microflow Digital Imaging
Microflow imaging (MFI) is a flow microscopy technique in which a sample is

drawn through a flow cell by a peristaltic pump. As the sample slowly transverses
the flow cell, bright-field images are illuminated by a 475 nm LED light source,
magnified, captured by a digital camera, and then processed by the instrument soft-
ware. Information about particle counts, size, and morphological characteristics can
all be obtained from MFI measurements. Since MFI does not rely on the blockage
of light, translucent protein particles are more readily detected and included by this
analysis (Huang et al. 2009; Wuchner et al. 2010). Another advantage of MFI is that
it can distinguish between silicone oil droplets, air bubbles, and protein aggregates
based on morphological characteristics (Sharma et al. 2010a; Strehl et al. 2011; Liu
et al. 2011).
9.4.1.4 Light Microscopy and Visual Assessments
During analysis by light microscopy, particles in solution are drawn through a grid-
lined filter, dried, and either observed by a standard light microscope or subsequently
stained with fluorescent or nonfluorescent dyes to enhance resolution (Li et al. 2007;
Demeule et al. 2007a, 2009). Particles can then be counted manually or by automated
methods. Advantages of staining protein aggregates include enhanced detection of
smaller particles and the selective binding of certain dyes to proteinaceous particles
(Li et al. 2007; Rosenberg et al. 2009). Artifacts resulting from sample preparation,
handling, and/or filtering are again the major pitfalls in using microscopy. In addi-
tion, particles <10 µm can be difficult to visualize and care should be taken to ensure
amorphous or fibril shaped particles do not pass through the filter.
228 T. Wang et al.
Visual assessments of protein-containing solutions for visible particles suggest a

straightforward process, but complications can arise due to a variety of factors, that
is, the method setup (lighting, distance, background), training of the analyst (physi-
cal abilities of the individual performing the counting, time spent counting), and
type of container and physical properties of the solution. Furthermore, particle mor-
phology and refractive index can influence the ability to detect and count visible
particles. Ideally, two independent, certified analysts need to count visible particles
in defined illumination conditions using a black and white panel as background.
Blanks and particle standards are also counted in conjunction with the unknowns to
ensure accurate and unbiased counting (Wuchner et al. 2010; Madsen et al. 2009).
9.4.1.5 Zeta Potential
The zeta potential is a property of particles in suspension related to their surface

charge. In an electric field, each particle and its most closely associated ions move
through the solution as a unit, and the potential at the boundary between this unit
and the bulk solution is known as the zeta potential. Zeta potentials can be experi-
mentally determined by measuring the drift velocity of particles in an electrical field
of known strength (Ross and Morrison 2002; Greenwood 2003) or by a method
known as phase analysis light scattering. Since zeta potential reflects the effective
charge on the particles and is thus related to the electrostatic repulsion between
particles in suspension, it has proven to be extremely useful in predicting the stabil-
ity of emulsions, colloids, as well as proteins in solution. A higher zeta potential
value reflects a more stable suspension since the charged particles repel one another
overcoming the natural tendency to aggregate. Conversely, lower values facilitate
interactions between the particles. The zeta potential measurement is often a key to
understanding dispersion and aggregation processes in various applications includ-
ing protein formulation (Sejersen et al. 2007; Uskokovic et al. 2010; Schmitt et al.
2007). The zeta potential can be measured with particles ranging from a few nano-
meters to as large as 30 µm in diameter.
9.4.1.6 Fluorescence-Activated Cell Sorting
Fluorescence-activated cell sorting (FACS), often referred to as flow cytometry, is a

technique used for quantitative measurements on single cells, cellular constituents,
or particles. Flow cytometry has also been used to detect baculovirus particles and
their aggregates (Jorio et al. 2006), monitor platelet aggregation events (Ginsberg
et al. 1990; Harding et al. 2007), and characterize amyloid fibrils (Wall and Solomon
1999). Flow cytometry uses the principles of light scattering, light excitation, and
the emission of fluorochrome molecules to generate specific multiparameter data
from particles and cells in the size range of 0.5–100 µm (Shapiro 1995; Cram 2002).
Using hydrodynamic focusing techniques, cells or particles are presented to a light
source (usually a laser) and the light scattering and fluorescence emission of
thousands of individual particles per second are measured using several detectors
(forward scatter, side scatter and fluorescent detectors). Light scattering and/or fluo-
rescence emission is unique to each cell/particle. Thus, a combination of the two
can be used to distinguish different particles or cells in a heterogeneous sample. The
capability of flow cytometer to characterize individual particles makes it an attrac-
tive technique for use in the study of subvisible particles/aggregates in protein for-
mulations. This technique has been widely used in the field of cellular and molecular
biology, but its potential for use in the study of subvisible particles in protein formu-
lations has only recently been recognized (Ludwig et al. 2011; Mach et al. 2011).
9.4.1.7 Fluorescence Microscopy
This type of microscopy can identify cells, cellular constituents, and particulate
matter with a high degree of specificity and is rapidly expanding in use for pharma-
ceutical applications. The technique is used to study samples that can fluoresce
either by themselves or after treating with fluorescent probes/dyes (Lichtman and
Conchello 2005). Fluorescence microscopy can be very useful for the detection of
subtle changes in the aggregation state of the proteins (David Bernard Williams
2009; Demeule et al. 2007a). Some of the dyes used to selectively monitor
intermolecular β(beta)-sheet interactions often present among proteins that form
β(beta)-amyloid structures are Nile red, Congo red, and Thioflavin T (Hawe et al.
2008a; Khurana et al. 2001; Naiki et al. 1989; Kim et al. 2003). Covalent labeling of
proteins with donor and acceptor fluorophores allows the use of total internal reflec-
tion fluorescence microscopy to image aggregated proteins (Hillger et al. 2007).
9.4.1.8 FTIR Microscopy
By combining Fourier transform infrared (FTIR) detection with optical microscopy,

compositional analysis of particles/aggregates down to 10 µm (micrometer) in size is
possible. FTIR spectroscopy is also a powerful tool to analyze protein secondary
structures and conformational dynamics (Prati et al. 2010; Middaugh and Joshi
2011). When combined with microscopy, this technique provides added information
to analytical results that cannot be obtained by spectroscopy or microscopy alone.
Another advantage of FTIR microscopy over conventional FTIR spectroscopy is that
it permits the collection of an IR spectrum from a small sample area. FTIR micros-
copy is both a sensitive and potentially comprehensive method to help understand the
nature, composition, and morphology of the sample examined and has wide applica-
tions in the areas of forensic (Torraca and Wen 2005), biomedical (Erukhimovitch
et al. 2002; Salman et al. 2002), and pharmaceutical sciences (Wuchner et al. 2010;
Flores-Fernandez et al. 2009; Markovich et al. 1997). One new application of FTIR
microscopy is chemical mapping and imaging. The instrumentation required to
obtain reliable IR chemical images has only recently become available. Using this
technology, multicomponent systems can now be mapped and the location, nature,
230 T. Wang et al.
shape, and size of each material present can be visualized through these chemical
images. These images capture the spectral features unique to each material present
and provide chemical information about the multicomponent system. Creating a
map, however, is a time-consuming process especially if both high spatial resolution
and high signal to noise ratio as required (Prati et al. 2010).
9.4.1.9 Energy-Dispersive X-Ray Spectrometer
Energy-dispersive X-ray spectrometer (EDX or EDS), which utilizes the X-ray gen-
erated when the electron beam bombards the sample to collect information on
chemical composition of a sample, can be connected to both TEM and SEM (Joseph
Goldstein et al. 2003; David Bernard Williams 2009).
9.4.2 Case Studies to Size and Count Subvisible

and Visible Protein Particles
Although MFI is a relatively new technique, there are already many interesting case
studies employing the technology to count and size subvisible particles in protein for-
mulations. Examples include measuring subvisible-sized protein particles induced by
contact with ceramic filling pumps or syringes (Nayak et al. 2011; Majumdar et al.
2011) as well as from exposure to different environmental stresses (Joubert et al. 2011;
Luo et al. 2011; Fradkin et al. 2011; Barnard et al. 2010). Several studies also include
comparisons of MFI or Coulter methods to light obscuration in terms of counting both
protein particles and polystyrene bead standards (Huang et al. 2009; Wuchner et al.
2010; Sharma et al. 2010a, b). The major difference in MFI and LO results for counting
subvisible protein particles is probably due to their translucent nature which results in
undercounting by LO. A recent study by Demule et al. (2010) directly compared count-
ing of protein particles obtained from LO, MFI, and Coulter instruments. One set of
results is shown in Fig. 9.9. These data demonstrate that the smallest amount of protein
subvisible particles detected is obtained with LO, followed by MFI and the Coulter
method. The results were dependent on the dilution of the sample prior to analysis
(Fig. 9.9) as well as the particle size cutoff used during the analysis (Demeule et al.
2010). The authors’ explanation for obtaining higher subvisible counts with the Coulter
method is its insensitivity to differences in the refractive index between the protein par-
ticles and the solvent. On the other hand, a study by Singh and colleagues showed that
the MFI and the Coulter methods generated comparable data except with highly concen-
trated protein samples, although the authors acknowledge that sample handling or dif-
ferences in the measurement methods can potentially affect the final results (Singh et al.
2010). Taken together, these studies again highlight the importance of using orthogonal
techniques to count and size protein particulates in the subvisible range to determine
which methods yield the most reproducible data for an individual protein formulation.
A detailed study by Wuchner and colleagues investigated the effect of long-term
storage of a high-concentration mAb formulation on formation of subvisible and visible
Fig. 9.9 Comparing the performance of the Coulter counter, flow microscopy (MFI), and light
obscuration (LO) for measuring subvisible particles (>3.5 µm) in protein solutions. Results are
reported as average ± standard deviation. Counting was performed at several protein concentrations
with the reported particle numbers corrected for the dilution factor and represent the particle con-
tent of the stock (150 mg/ml) solution. The particle counts of the formulation buffer are included.
Published in Demeule et al. (2010)
protein particles (Wuchner et al. 2010). Direct comparison of LO and MFI revealed five
to tenfold higher subvisible protein particle counts in the MFI measurements, consis-
tent with the results described above with other mAbs. This study proceeded to charac-
terize the count and size of subvisible and visible particles during long-term storage as
measured by MFI. The results found that a shift occurred in the particle size distribution
over time, with larger subvisible particles accumulating during long-term storage
(Fig. 9.10). This result was not observed by LO measurements. The accumulation of
visible particles was also monitored by a combination of visible assessments, inverted
microscopy, and MFI measurements (for protein particles larger than 100 µm; see
Fig. 9.10). The three techniques showed good overall correlation in terms of the count
of visible particle formed during long-term storage (Wuchner et al. 2010).
9.4.3 Case Studies Which Characterize the Morphology

and Composition of Subvisible and Visible Protein Particles
Mach et al. (2011) demonstrated the use of a flow cytometer in a high-throughput

format using a 96-well autosampler, to not only determine the number of subvisible
232 T. Wang et al.
particles in monoclonal antibody formulations but also differentiate protein vs.

nonprotein particles. In this study, SYPRO Orange dye which is capable of binding
to apolar surfaces of conformationally alerted proteins was employed to preferen-
tially stain proteinaceous particles and measure their number. While low fluores-
cence signals were observed in the absence of the dye, the fluorescence increased
almost 100-fold in the presence of SYPRO Orange due to binding of the dye to
protein aggregates. A small number of particles did not stain presumably due to
their nonprotein nature. The authors also demonstrated specificity of the dye for the
protein particles using silicone oil droplets extracted from a commercial plastic
syringe. It was convincingly demonstrated that the flow cytometry method can be
reliably used to screen a large number of stability samples for the presence of sub-
visible particles with good precision and accuracy as well as moderate sample
requirements. In a similar study (Ludwig et al. 2011), flow cytometry was used to
detect and characterize subvisible particles in four different protein formulations
contaminated with silicone oil droplets. The authors showed that this technique had
the ability to distinguish homogenous protein particles/aggregates from heteroge-
neous particles consisting of silicone oil and protein.
Various extrinsic fluorescent dyes in combination with microscopic techniques
have been used in the detection and characterization of protein aggregates and par-
ticulates. Examples include 1-anilinonaphthalene-8-sulfonate (ANS), a dimeric
analog (bis-ANS), SYPRO Orange, Nile red, Congo red, and Thioflavin T. A review
article by Hawe et al. (2008b) provides a detailed overview of solution properties
and underlying mechanisms governing the fluorescent properties of these dyes.
The use of such dyes in various fields of protein analysis has been discussed
elsewhere (Filipe et al. 2011; He et al. 2010; Li et al. 2011). In general, they are
extremely sensitive to polarity of their microenvironments and hence have become
important tools in protein characterization studies. These apolar probes are typically
essentially nonfluorescent in aqueous solution but become strongly fluorescent in
less polar environments such as when a protein’s hydrophobic regions became more
exposed due to conformational alterations and/or formation of aggregates.
Fluorescence microscopy permits the evaluation of high-concentration protein
formulations without dilution and with minimal impact on the local environment of
the protein (Demeule et al. 2009). Therefore, it can be an extremely useful tool for
the detection of protein aggregates/particulates. The use of Nile red for the early
detection of small amounts of immunoglobulin aggregates using fluorescence
microscopy has been shown by Demeule et al. (2007b). These studies were per-
formed on an inverted microscope equipped with a mercury discharge lamp for fluo-
rescence microscopy and a tungsten lamp for bright-field microscopy. Suspensions
of a recombinant humanized monoclonal antibody were stained with Nile red and
visualized by fluorescence microscopy. The same sample was examined by bright-
field microscopy for comparison. A more efficient visualization of the protein
aggregates was observed following addition of Nile red and subsequent observation
by fluorescence microscopy compared to the samples examined with bright-field
microscopy (see Fig. 9.11). The high specificity of Nile red for protein aggregates
eliminated false positives due to the presence of impurities/dust. Multiple other
Fig. 9.10 Differential particle levels (subvisible- and visible-sized categories) in a 90 mg/ml IgG1
monoclonal antibody solution during storage for 18 months at 2–8°C as measured by MFI. Data
were collected from two to four preparations and presented as the average of differential particle
number per particle size range (ECD) with error bars of ±1 SD. Results from a frozen retain stored
for 12 months at −70°C were used as a surrogate for time 0. Published in Wuchner et al. (2010)
publications have shown the use of Nile red staining and fluorescence microscopy
for the detection and characterization of protein aggregates and particulates
(Demeule et al. 2007a, 2009; Sutter et al. 2007). Caution, however, needs to be
taken while using fluorescent dyes for protein characterization since these dyes may
also induce formation of aggregates and particulates because of their tendency to
perturb protein structure (Engelhard and Evans 1995).
234 T. Wang et al.
Fig. 9.11 Bright-field image (left) of unstained antibody A aggregates (in a 10 mM phosphate
buffer pH 7 solution containing 0.8 mg/ml protein) and a fluorescence image (right) of the same
antibody aggregates stained with Nile red. Both methods showed aggregates similar in size and
shape. Published in Demeule et al. (2007a, b)
FTIR microscopy is another commonly used analytical technique for particle

characterization which is now being employed to detect and characterize the nature
and composition of protein aggregates/particles. This technique can also help deter-
mine if the particles encountered during manufacturing or storage are proteinaceous
in nature (Cao et al. 2010; Liu et al. 2010). FTIR microscopy has also been employed
to investigate the structural changes leading to aggregation as induced by exposure of
lyophilized insulin to high levels of humidity (Giselle et al. 2009). In addition, this
method has been used for the easy and rapid discrimination between bacteria and
fungi by obtaining unique spectral peaks from these organisms (Erukhimovitch et al.
2002) as well as for detection of cells infected with Herpes virus (Salman et al. 2002).
Another experimental approach is capturing protein particles on gold-coated
membrane filters. Using FTIR microscopy, aggregates can then be visualized and
chemically mapped. Wuchner et al. (Wuchner et al. 2010) used a combination of
FTIR microscopy and SEM-EDX to assess the composition of protein particulates
formed during long-term storage of a high-concentration IgG1 monoclonal anti-
body formulation. Protein particles were isolated on a gold-coated membrane filter
and the IR spectra were acquired by attenuated total reflectance (ATR) infrared
spectroscopy employing a microscope coupled to an FTIR spectrometer. The
authors were able to successfully show that the key amide bands (typical of proteins)
Fig. 9.12 Representative ATR FTIR spectrum of a protein particle (b), overlaid with IgG1 refer-
ence spectrum (a), and polydimethylsiloxane reference spectrum (c). Protein stability sample con-
tained 90 mg/mL IgG1 stored for 16 months at 2–8°C. Published in Wuchner et al. (2010)
in the FTIR spectrum of protein particles at a 16-month stability time point were
comparable to those obtained for a reference IgG1 preparation (Fig. 9.12). Additional
bands characteristic of silicone were detected for most of the particles examined.
The protein particulates were further characterized by SEM-EDX to study their
composition (Wuchner et al. 2010) as shown in Table 9.2. In addition to silicon,
some of the protein particles contained aluminum and fluorine. These compositional
data from FTIR microscopy and SEM-EDX analysis indicated heterogeneity across
the particles examined, suggesting protein particle formation is a heterogeneous
process (Table 9.2).
9.5 Conclusion and Future Directions
This chapter has described recent case studies on the use of different analytical tech-
niques to better characterize the nature of protein aggregates and particles. The focus
was on recent work from different investigators who are exploring both new tech-
niques as well as combinations of analytical approaches, to better understand the
number, size range, and composition of protein aggregates and particles in protein
formulations. It is well recognized that no one analytical measurement can detect all
types and sizes of protein aggregates (Philo 2006). Some of the newer biophysical
approaches used to monitor protein aggregate and particle formation have come
from the field of protein aggregation diseases (Aguzzi and O’Connor 2010).
236 T. Wang et al.
Table 9.2 Analytical characterization of protein particles isolated from long-term stability
samples of a 90 mg/ml IgG1 monoclonal antibody solution formulation stored at 2–8°C
Age of
Sample material Inverted microscopy FTIR SEM-EDX
Batch 1 11 mo Some large and many small Protein (6/6), C,N,O (4/4), Si
translucent particles of typical silicone (6/6) (4/4)
proteinaceous morphology
translucent particles of typical silicone (5/5) (3/3), Al
proteinaceous morphology (1/3), F (1/3)a
translucent particles of typical silicone (4/7) (4/6), Al (4/6)
proteinaceous morphology
translucent particles of typical silicone (3/4)
proteinaceous morphology (10/10)
Published in Wuchner et al. (2010)
mo, months at 2–8°C
Inverted microscopy: large particles through microscope approximately >50 µm, small particles
approximately <20 µm
FTIR microscopy: between 5 and 10 particles per sample were analyzed. The frequencies of main
components (spectral matches) are indicated in parentheses (number of particles with component
per total number or particles analyzed)
SEM-EDX: between 3 and 6 particles per sample were analyzed. The frequencies of the detected
elements are indicated in parentheses (number of particles with element per total number of par-
ticles analyzed)
a
Fluorine was detected at trace level
As new case studies are published demonstrating a further enhanced ability to detect
protein aggregates and particles under varying conditions, our mechanistic under-
standing of the interrelationships of protein aggregation pathways (Philo and
Arakawa 2009; Wang et al. 2010) as well as specific and nonspecific protein-protein
interactions leading to protein self-association (Saluja and Kalonia 2008) will con-
tinue to improve. The interrelationships between chemical and physical stabilities
are also emerging as an important new direction since some aggregates may contain
an altered nature (e.g., mixed disulfide formation and oxidation) as well as the pres-
ence of nonprotein nucleating species such as metals (Carpenter et al. 2009; Narhi
et al. 2011; Joubert et al. 2011; Luo et al. 2011). Similarly, although not the focus of
this chapter, our mechanistic understanding of the interrelationships of protein
aggregate composition and size to immunogenic potential is increasing (Carpenter
et al. 2010b; Mire-Sluis et al. 2011; Rosenberg 2006; Maas et al. 2007). Based on
this enhanced understanding, more rational approaches to the selection and optimi-
zation of stabilizing excipients to minimize protein aggregation and particle forma-
tion should result (Kamerzell et al. 2011). The combination of utilizing better
analytical methods along with an enhanced understanding of protein aggregation
pathways should ultimately lead to important improvements in the impurity levels
and long-term stability of protein formulations.
References
Aguzzi A, O’Connor T (2010) Protein aggregation diseases: pathogenicity and therapeutic

perspectives. Nat Rev Drug Discov 9(3):237–248
Ahrer K, Buchacher A, Iberer G, Josic D, Jungbauer A (2003) Analysis of aggregates of human
immunoglobulin G using size-exclusion chromatography, static and dynamic light scattering.
J Chromatogr A 1009(1–2):89–96
Arakawa T, Ejima D, Li T, Philo JS (2010) The critical role of mobile phase composition in size
exclusion chromatography of protein pharmaceuticals. J Pharm Sci 99(4):1674–1692
Barnard JG, Singh S, Randolph TW, Carpenter JF (2010) Subvisible particle counting provides a
sensitive method of detecting and quantifying aggregation of monoclonal antibody caused by
freeze-thawing: insights into the roles of particles in the protein aggregation pathway. J Pharm
Sci 100(2):492–503
Barnard JG, Rhyner MN, Carpenter JF (2012) Critical evaluation and guidance for using the coul-
ter method for counting subvisible particles in protein solutions. J Pharm Sci 101(1):140–153
biopharmaceuticals. AAPS J 8(3):E590–E605
Bhambhani A, Kissmann JM, Joshi SB, Volkin DB, Kashi RS, Middaugh CR (2012) Formulation
design and high-throughput excipient selection based on structural integrity and conforma-
tional stability of dilute and highly concentrated IgG1 monoclonal antibody solutions. J Pharm
Sci 101(3):1120–1135
Bond MD, Panek ME, Zhang Z et al (2010) Evaluation of a dual-wavelength size exclusion HPLC
method with improved sensitivity to detect protein aggregates and its use to better characterize
degradation pathways of an IgG1 monoclonal antibody. J Pharm Sci 99(6):2582–2597
Burg TP, Godin M, Knudsen SM et al (2007) Weighing of biomolecules, single cells and single
nanoparticles in fluid. Nature 446(7139):1066–1069
Calamai M, Canale C, Relini A, Stefani M, Chiti F, Dobson CM (2005) Reversal of protein aggre-
gation provides evidence for multiple aggregated States. J Mol Biol 346(2):603–616
Cao S, Pollastrini J, Jiang Y (2009) Separation and characterization of protein aggregates and
particles by field flow fractionation. Curr Pharm Biotechnol 10(4):382–390
Cao X, Masatani P, Torraca G, Wen ZQ (2010) Identification of a mixed microparticle by com-
bined microspectroscopic techniques: a real forensic case study in the biopharmaceutical
industry. Appl Spectrosc 64(8):895–900
Carpenter JF, Randolph TW, Jiskoot W et al (2009) Overlooking subvisible particles in therapeutic
protein products: gaps that may compromise product quality. J Pharm Sci 98(4):1201–1205
Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJ, Middaugh CR, Winter G (2010a) Potential
J Pharm Sci 99(5):2200–2208
Carpenter J, Cherney B, Lubinecki A et al (2010b) Meeting report on protein particles and immu-
nogenicity of therapeutic proteins: filling in the gaps in risk evaluation and mitigation.
Biologicals 38(5):602–611
Chait BT (2011) Mass spectrometry in the postgenomic era. Annu Rev Biochem 80:239–246
Cheng X, van Breemen RB (2005) Mass spectrometry-based screening for inhibitors of beta-
amyloid protein aggregation. Anal Chem 77(21):7012–7015
Chrai S, Clayton R, Mestrandrea L et al (1987) Limitations in the use of HIAC for product particle
counting. J Parenter Sci Technol 41(6):209–214
Colton RJ, Baselt DR, Dufrene YF, Green JB, Lee GU (1997) Scanning probe microscopy. Curr
Opin Chem Biol 1(3):370–377
Costello MA, Woititz C, Defeo J et al (1992) Characterization of humanized anti-tac monoclonal-
antibody by traditional separation techniques and capillary electrophoresis. J Liq Chromatogr
15(6–7):1081–1097
Cram LS (2002) Flow cytometry, an overview. Methods Cell Sci 24(1–3):1–9
238 T. Wang et al.
Dam J, Schuck P (2004) Calculating sedimentation coefficient distributions by direct modeling of

sedimentation velocity concentration profiles. Meth Enzymol 384:185–212
David Bernard Williams CBC (2009) The transmission electron microscope. Springer, New York
Demeule B, Gurny R, Arvinte T (2007a) Detection and characterization of protein aggregates by
fluorescence microscopy. Int J Pharm 329(1–2):37–45
Demeule B, Lawrence MJ, Drake AF, Gurny R, Arvinte T (2007b) Characterization of protein
aggregation: the case of a therapeutic immunoglobulin. Biochim Biophys Acta
1774(1):146–153
Demeule B, Palais C, Machaidze G, Gurny R, Arvinte T (2009) New methods allowing the detec-
tion of protein aggregates: a case study on trastuzumab. MAbs 1(2):142–150
Demeule B, Messick S, Shire SJ, Liu J (2010) Characterization of particles in protein solutions:
reaching the limits of current technologies. AAPS J 12(4):708–715
Engelhard M, Evans PA (1995) Kinetics of interaction of partially folded proteins with a hydro-
phobic dye: evidence that molten globule character is maximal in early folding intermediates.
Protein sci 4(8):1553–1562
Engelsman J, Garidel P, Smulders R et al (2011) Strategies for the assessment of protein aggregates
in pharmaceutical biotech product development. Pharm Res 28(4):920–933
Erukhimovitch V, Talyshinsky M, Souprun Y, Huleihel M (2002) Spectroscopic characterization
of human and mouse primary cells, cell lines and malignant cells. Photochem Photobiol
76(4):446–451
Esue O, Kanai S, Liu J, Patapoff TW, Shire SJ (2009) Carboxylate-dependent gelation of a mono-
clonal antibody. Pharm Res 26(11):2478–2485
Ferreira GN, da-Silva AC, Tome B (2009) Acoustic wave biosensors: physical models and biologi-
cal applications of quartz crystal microbalance. Trends Biotechnol 27(12):689–697
Filipe V, Hawe A, Jiskoot W (2010) Critical evaluation of Nanoparticle Tracking Analysis (NTA)
by NanoSight for the measurement of nanoparticles and protein aggregates. Pharm Res
27(5):796–810
Filipe V, Poole R, Kutscher M, Forier K, Braeckmans K, Jiskoot W (2011) Fluorescence single
particle tracking for the characterization of submicron protein aggregates in biological fluids
and complex formulations. Pharm Res 28(5):1112–1120
Flores-Fernandez GM, Sola RJ, Griebenow K (2009) The relation between moisture-induced
aggregation and structural changes in lyophilized insulin. J Pharm Pharmacol
61(11):1555–1561
Fradkin AH, Carpenter JF, Randolph TW (2011) Glass particles as an adjuvant: a model for
adverse immunogenicity of therapeutic proteins. J Pharm Sci 100(11):4953–4964
ceutics and biopharmaceutics. Eur J Pharm Biopharm 58(2):369–383
Gabrielson JP, Brader ML, Pekar AH et al (2007) Quantitation of aggregate levels in a recombinant
humanized monoclonal antibody formulation by size-exclusion chromatography, asymmetrical
flow field flow fractionation, and sedimentation velocity. J Pharm Sci 96(2):268–279
Ginsberg MH, Frelinger AL, Lam SC et al (1990) Analysis of platelet aggregation disorders based
on flow cytometric analysis of membrane glycoprotein IIb-IIIa with conformation-specific
monoclonal antibodies. Blood 76(10):2017–2023
Giselle M, Flores-Fernandez G, Solá R, Griebenow K (2009) The relation between moisture-
induced aggregation and structural changes in lyophilized insulin. J Pharm Pharmacol
61(11):1555–1561
Greenwood R (2003) Review of the measurement of zeta potentials in concentrated aqueous sus-
pensions using electroacoustics. Adv Colloid Interface Sci 106:55–81
Grillberger L, Kreil TR, Nasr S, Reiter M (2009) Emerging trends in plasma-free manufacturing of
recombinant protein therapeutics expressed in mammalian cells. Biotechnol J 4(2):186–201
Han M, Phan D, Nightlinger N et al (2006) Optimization of CE-SDS method for antibody separa-
tion based on multi-users experimental practices. Chromatographia 64:334–342
Harding SA, Din JN, Sarma J et al (2007) Flow cytometric analysis of circulating platelet-
monocyte aggregates in whole blood: methodological considerations. Thromb Haemost
98(2):451–456
Harper JD, Lieber CM, Lansbury PT Jr (1997) Atomic force microscopic imaging of seeded fibril
formation and fibril branching by the Alzheimer’s disease amyloid-beta protein. Chem Biol
4(12):951–959
Hawe A, Friess W, Sutter M, Jiskoot W (2008a) Online fluorescent dye detection method for the
characterization of immunoglobulin G aggregation by size exclusion chromatography and
asymmetrical flow field flow fractionation. Anal Biochem 378(2):115–122
Hawe A, Sutter M, Jiskoot W (2008b) Extrinsic fluorescent dyes as tools for protein characteriza-
tion. Pharm Res 25(7):1487–1499
Hawe A, Hulse WL, Jiskoot W, Forbes RT (2011) Taylor dispersion analysis compared to dynamic
light scattering for the size analysis of therapeutic peptides and proteins and their aggregates.
Pharm Res 28(9):2302–2310
He F, Phan DH, Hogan S et al (2010) Detection of IgG aggregation by a high throughput method
based on extrinsic fluorescence. J Pharm Sci 99(6):2598–2608
Heldt CL, Sorci M, Posada D, Hirsa A, Belfort G (2011) Detection and reduction of microaggre-
gates in insulin preparations. Biotechnol Bioeng 108(1):237–241
Hillger F, Nettels D, Dorsch S, Schuler B (2007) Detection and analysis of protein aggregation
with confocal single molecule fluorescence spectroscopy. J Fluoresc 17(6):759–765
Houde D, Arndt J, Domeier W, Berkowitz S, Engen JR (2009) Characterization of IgG1 conforma-
tion and conformational dynamics by hydrogen/deuterium exchange mass spectrometry. Anal
Chem 81(14):5966
Huang C-T, Sharma D, Oma P, Krishnamurthy R (2009) Quantitation of protein particles in paren-
teral solutions using micro-flow imaging. J Pharm Sci 98(9):3058–3071
Hughes H, Morgan C, Brunyak E et al (2009) A multi-tiered analytical approach for the analysis
and quantitation of high-molecular-weight aggregates in a recombinant therapeutic glycopro-
tein. AAPS J 11(2):335–341
Hunt G, Moorhouse KG, Chen AB (1996) Capillary isoelectric focusing and sodium dodecyl
sulfate-capillary gel electrophoresis of recombinant humanized monoclonal antibody HER2.
J Chromatogr A 744(1–2):295–301
Jorio H, Tran R, Meghrous J, Bourget L, Kamen A (2006) Analysis of baculovirus aggregates
using flow cytometry. J Virol Methods 134(1–2):8–14
Joseph Goldstein DN, Joy D, Lyman C, Echlin P, Lifshin E, Sawyer L, Michael J (2003) Scanning
electron microscopy and X-ray microanalysis, 3rd edn. Plenum, New York
Joubert MK, Luo Q, Nashed-Samuel Y, Wypych J, Narhi LO (2011) Classification and character-
ization of therapeutic antibody aggregates. J Biol Chem 286(28):25118–25133
Kaddis CS, Lomeli SH, Yin S et al (2007) Sizing large proteins and protein complexes by electrospray
ionization mass spectrometry and ion mobility. J Am Soc Mass Spectrom 18(7):1206–1216
Kaltashov IA, Bobst CE, Abzalimov RR, Wang G, Baykal B, Wang S (2012) Advances and chal-
lenges in analytical characterization of biotechnology products: mass spectrometry-based
approaches to study properties and behavior of protein therapeutics. Biotechnol Adv
30(1):210–222
Kamerzell TJ, Esfandiary R, Joshi SB, Middaugh CR, Volkin DB (2011) Protein-excipient interac-
tions: mechanisms and biophysical characterization applied to protein formulation develop-
ment. Adv Drug Deliv Rev 63(13):1118–1159
Kheterpal I, Wetzel R (2006) Hydrogen/deuterium exchange mass spectrometry—a window into
amyloid structure. Acc Chem Res 39(9):584–593
Khurana R, Uversky VN, Nielsen L, Fink AL (2001) Is Congo red an amyloid-specific dye? J Biol
Chem 276(25):22715–22721
Kiese S, Papppenberger A, Friess W, Mahler HC (2008) Shaken, not stirred: mechanical stress
testing of an IgG1 antibody. J Pharm Sci 97(10):4347–4366
Kiese S, Pappenberger A, Friess W, Mahler H-C (2010) Equilibrium studies of protein aggregates
and homogeneous nucleation in protein formulation. J Pharm Sci 99(2):632–644
240 T. Wang et al.
Kilar F (2003) Recent applications of capillary isoelectric focusing. Electrophoresis

24(22–23):3908–3916
Kim YS, Randolph TW, Manning MC, Stevens FJ, Carpenter JF (2003) Congo red populates par-
tially unfolded states of an amyloidogenic protein to enhance aggregation and amyloid fibril
formation. J Biol Chem 278(12):10842–10850
Knowles TPJ, Devlin GL, Dobson CM, Welland ME (2011) Probing protein aggregation with
quartz crystal microbalances. In: Hill AF, Barnham KJ, Bottomley SP, Cappai R (eds) Protein
folding, misfolding, and disease, vol 752, 1st edn. Humana Press, New York
Konermann L, Pan J, Liu YH (2011) Hydrogen exchange mass spectrometry for studying protein
structure and dynamics. Chem Soc Rev 40(3):1224–1234
Kraly J, Fazal MA, Schoenherr RM et al (2006) Bioanalytical applications of capillary electropho-
resis. Anal Chem 78(12):4097–4110
Kukrer B, Filipe V, van Duijn E et al (2010) Mass spectrometric analysis of intact human mono-
clonal antibody aggregates fractionated by size-exclusion chromatography. Pharm Res
27(10):2197–2204
Lee H, Kirchmeier M, Mach H (2011) Monoclonal antibody aggregation intermediates visualized
by atomic force microscopy. J Pharm Sci 100(2):416–423
Li B, Flores J, Corvari V (2007) A simple method for the detection of insoluble aggregates in
protein formulations. J Pharm Sci 96(7):1840–1843
Li Y, Mach H, Blue JT (2011) High throughput formulation screening for global aggregation
behaviors of three monoclonal antibodies. J Pharm Sci 100(6):2120–2135
Lichtman JW, Conchello JA (2005) Fluorescence microscopy. Nat Methods 2(12):910–919
Liu J, Shire SJ (1999) Analytical ultracentrifugation in the pharmaceutical industry. J Pharm Sci
88(12):1237–1241
fractionation methods for measuring protein aggregation. AAPS J 8(3):E580–E589
Liu W, Swift R, Torraca G et al (2010) Root cause analysis of tungsten-induced protein aggrega-
tion in pre-filled syringes. PDA J Pharm Sci Technol 64(1):11–19
Liu L, Ammar DA, Ross LA, Mandava N, Kahook MY, Carpenter JF (2011) Silicone oil microdro-
plets and protein aggregates in repackaged bevacizumab and ranibizumab: effects of long-term
storage and product mishandling. Invest Ophthalmol Vis Sci 52(2):1023–1034
Lomakin A, Benedek GB, Teplow DB (1999) Monitoring protein assembly using quasielastic light
scattering spectroscopy. Methods Enzymol 309:429–459
Lubiniecki A, Volkin DB, Federici M et al (2010) Comparability assessments of process and prod-
uct changes made during development of two different monoclonal antibodies. Biologicals
39(1):9–22
Ludwig DB, Trotter JT, Gabrielson JP, Carpenter JF, Randolph TW (2011) Flow cytometry:
a promising technique for the study of silicone oil-induced particulate formation in protein
formulations. Anal Biochem 410(2):191–199
Luo Y, Lu Z, Raso SW, Entrican C, Tangarone B (2009) Dimers and multimers of monoclonal
IgG1 exhibit higher in vitro binding affinities to Fcgamma receptors. MAbs 1(5):491–504
Luo Q, Joubert MK, Stevenson R, Ketchem RR, Narhi LO, Wypych J (2011) Chemical modifica-
tions in therapeutic protein aggregates generated under different stress conditions. J Biol Chem
286(28):25134–25144
Maas C, Hermeling S, Bouma B, Jiskoot W, Gebbink MF (2007) A role for protein misfolding in
immunogenicity of biopharmaceuticals. J Biol Chem 282(4):2229–2236
Mach H, Arvinte T (2011) Addressing new analytical challenges in protein formulation develop-
ment. Eur J Pharm Biopharm 78(2):196–207
Mach H, Middaugh CR (2011) Ultraviolet spectroscopy as a tool in therapeutic protein develop-
ment. J Pharm Sci 100(4):1214–1227
Mach H, Bhambhani A, Meyer BK et al (2011) The use of flow cytometry for the detection of
subvisible particles in therapeutic protein formulations. J Pharm Sci 100(5):1671–1678
Maddux NR, Joshi SB, Volkin DB, Ralston JP, Middaugh CR (2011) Multidimensional methods
for the formulation of biopharmaceuticals and vaccines. J Pharm Sci 100(10):4171–4197
Madsen R, Cherris R, Shabushnig JG, Hunt DG (2009) Visible particulates in injections—a history
and a proposal to revise USP general chapter injections (1). Pharmacop Forum 35:1383–1387
Mahler HC, Muller R, Friess W, Delille A, Matheus S (2005) Induction and analysis of aggregates
in a liquid IgG1-antibody formulation. Eur J Pharm Biopharm 59(3):407–417
Mahler H-C, Friess W, Grauschopf U, Kiese S (2009) Protein aggregation: pathways, induction
factors and analysis. J Pharm Sci 98(9):2909–2934
Majumdar S, Ford BM, Mar KD, Sullivan VJ, Ulrich RG, D’souza AJM (2011) Evaluation of the
effect of syringe surfaces on protein formulations. J Pharm Sci 100(7):2563–2573
Marcsisin SR, Engen JR (2010) Hydrogen exchange mass spectrometry: what is it and what can it
tell us? Anal Bioanal Chem 397(3):967–972
Markovich RJ, Taylor AK, Rosen J (1997) Drug migration from the adhesive matrix to the polymer
film laminate facestock in a transdermal nitroglycerin system. J Pharm Biomed Anal
16(4):651–660
Michel Godin AKB, Burg TP, Babcock K, Manalis SR (2007) Measuring the mass, density, and
size of particles and cells using a suspended microchannel resonator. Appl Phys Lett 91:3
Michels DA, Brady LJ, Guo A, Balland A (2007) Fluorescent derivatization method of proteins for
characterization by capillary electrophoresis-sodium dodecyl sulfate with laser-induced fluo-
rescence detection. Anal Chem 79(15):5963–5971
Middaugh CR, Joshi SB (2011) Spectroscopy of vaccines. In: Singh M, Srivastava IK (eds)
Development of vaccines, from discovery to clinical testing. Wiley, Hoboken, pp 263–292
Mire-Sluis A, Cherney B, Madsen R, Polozova A, Rosenberg A, Smith H, Arora T, Narhi L (2011)
Analysis and immunogenic potential of aggregates and particles. BioProcess Int 9(10):38–43
Murphy RM (1997) Static and dynamic light scattering of biological macromolecules: what can
we learn? Curr Opin Biotechnol 8(1):25–30
Naiki H, Higuchi K, Hosokawa M, Takeda T (1989) Fluorometric determination of amyloid fibrils
in vitro using the fluorescent dye, thioflavin T1. Anal Biochem 177(2):244–249
Narhi LO, Schmit J, Bechtold-Peters K, Sharma D (2011) Classification of protein aggregates.
J Pharm Sci 101(2):493–498
Nayak A, Colandene J, Bradford V, Perkins M (2011) Characterization of subvisible particle for-
mation during the filling pump operation of a monoclonal antibody solution. J Pharm Sci
100(10):4198–4204
Nettleton EJ, Tito P, Sunde M, Bouchard M, Dobson CM, Robinson CV (2000) Characterization
of the oligomeric states of insulin in self-assembly and amyloid fibril formation by mass spec-
trometry. Biophys J 79(2):1053–1065
O’Sullivan CK, Guilbault GG (1999) Commercial quartz crystal microbalances – theory and appli-
cations. Biosens Bioelectron 14(8–9):663–670
Oliva A, Llabres M, Farina JB (2001) Comparative study of protein molecular weights by size-
exclusion chromatography and laser-light scattering. J Pharm Biomed Anal 25:833–841
Pease LF 3rd, Elliott JT, Tsai DH, Zachariah MR, Tarlov MJ (2008) Determination of protein
aggregation with differential mobility analysis: application to IgG antibody. Biotechnol Bioeng
101(6):1214–1222
Philo JS (2006) Is any measurement method optimal for all aggregate sizes and types? AAPS
J 8(3):E564–E571
aggregates. Curr Pharm Biotechnol 10(4):359–372
Philo JS, Arakawa T (2009) Mechanisms of protein aggregation. Curr Pharm Biotechnol
10(4):348–351
Prati S, Joseph E, Sciutto G, Mazzeo R (2010) New advances in the application of FTIR micros-
copy and spectroscopy for the characterization of artistic materials. Acc Chem Res
43(6):792–801
Qi P, Volkin DB, Zhao H et al (2009) Characterization of the photodegradation of a human IgG1
monoclonal antibody formulated as a high-concentration liquid dosage form. J Pharm Sci
98(9):3117–3130
242 T. Wang et al.
Remmele RL Jr, Callahan WJ, Krishnan S et al (2006) Active dimer of Epratuzumab provides
insight into the complex nature of an antibody aggregate. J Pharm Sci 95(1):126–145
Reschiglian P, Zattoni A, Roda B, Michelini E, Roda A (2005) Field-flow fractionation and bio-
technology. Trends Biotechnol 23(9):475–483
Rhyner MN (2011) The Coulter principle for analysis of subvisible particles in protein formula-
tions. AAPS J 13(1):54–58
Roberts GS, Yu S, Zeng Q et al (2012) Tunable pores for measuring concentrations of synthetic
and biological nanoparticle dispersions. Biosens Bioelectron 31(1):17–25
Roda B, Zattoni A, Reschiglian P et al (2009) Field-flow fractionation in bioanalysis: a review of
recent trends. Anal Chim Acta 635(2):132–143
Rosenberg AS (2006) Effects of protein aggregates: an immunologic perspective. AAPS
J 8(3):E501–E507
Rosenberg E, Hepbildikler S, Kuhne W, Winter G (2009) Ultrafiltration concentration of monoclo-
nal antibody solutions: development of an optimized method minimizing aggregation. J Membr
Sci 342(1–2):50–59
Ross S, Morrison ID (2002) Colloidal dispersions. Wiley Interscience, New York
Rumi Chunara MG, Knudsen SM, Manalis SR (2007) Mass-based readout for agglutination
assays. Appl Phys Lett 91:3
Rustandi RR, Washabaugh MW, Wang Y (2008) Applications of CE SDS gel in development of
biopharmaceutical antibody-based products. Electrophoresis 29(17):3612–3620
Salas-Solano O, Tomlinson B, Du S, Parker M, Strahan A, Ma S (2006) Optimization and valida-
tion of a quantitative capillary electrophoresis sodium dodecyl sulfate method for quality con-
trol and stability monitoring of monoclonal antibodies. Anal Chem 78(18):6583–6594
Salman A, Erukhimovitch V, Talyshinsky M, Huleihil M, Huleihel M (2002) FTIR spectroscopic
method for detection of cells infected with herpes viruses. Biopolymers 67(6):406–412
Saluja A, Kalonia DS (2008) Nature and consequences of protein-protein interactions in high
protein concentration solutions. Int J Pharm 358(1–2):1–15
Sauerbrey G (1959) Use of quartz vibration for weighing thin films on a microbalance. J Phys
155:206–212
Schmitt C, Bovay C, Rouvet M, Shojaei-Rami S, Kolodziejczyk E (2007) Whey protein soluble
aggregates from heating with NaCl: physicochemical, interfacial, and foaming properties.
Langmuir 23(8):4155–4166
Schuck P (2000) Size-distribution analysis of macromolecules by sedimentation velocity
ultracentrifugation and lamm equation modeling. Biophys J 78(3):1606–1619
Schuck P (2004) A model for sedimentation in inhomogeneous media. II. Compressibility of aque-
ous and organic solvents. Biophys Chem 108(1–3):201–214
Sejersen M, Salomonsena T, Ipsen R, Clark R, Rolin C, Engelsen S (2007) Zeta potential of pectin-
stabilised casein aggregates in acidified milk drinks. Int Dairy J 17:302–307
Shapiro HM (1995) Practical flow cytometry. Wiley-Liss, New york
Sharma DK, Oma P, Pollo MJ, Sukumar M (2010a) Quantification and characterization of subvis-
ible proteinaceous particles in opalescent mAb formulations using micro-flow imaging.
J Pharm Sci 99(6):2628–2642
Sharma DK, King D, Oma P, Merchant C (2010b) Micro-flow imaging: flow microscopy applied
to sub-visible particulate analysis in protein formulations. AAPS J 12(3):455–464
Singh SK, Afonina N, Awwad M et al (2010) An industry perspective on the monitoring of subvis-
ible particles as a quality attribute for protein therapeutics. J Pharm Sci 99(8):3302–3321
Strehl R, Rombach-Riegraf V, Diez M et al (2011) Discrimination between silicone oil droplets
and protein aggregates in biopharmaceuticals: a novel multiparametric image filter for sub-
visible particles in microflow imaging analysis. Pharm Res 27:1–9
Sutter M, Oliveira S, Sanders NN et al (2007) Sensitive spectroscopic detection of large and dena-
tured protein aggregates in solution by use of the fluorescent dye Nile red. J Fluoresc
17(2):181–192
Thillaivinayagalingam P, Gommeaux J, McLoughlin M, Collins D, Newcombe AR (2010)

Biopharmaceutical production: Applications of surface plasmon resonance biosensors.
J Chromatogr B Analyt Technol Biomed Life Sci 878(2):149–153
Tobler SA, Fernandez EJ (2002) Structural features of interferon-gamma aggregation revealed by
hydrogen exchange. Protein sci 11(6):1340–1352
Torraca G, Wen Z (2005) Forensic investigation of biopharmaceutical manufacturing incidents by
light microscopy, ftir microscopy, scanning electron microscopy and energy dispersive X-ray
spectrometry. Microsc Microanal 11:1224–1225
Tyagi AK, Randolph TW, Dong A, Maloney KM, Hitscherich C Jr, Carpenter JF (2009) IgG par-
ticle formation during filling pump operation: a case study of heterogeneous nucleation on
stainless steel nanoparticles. J Pharm Sci 98(1):94–104
Uetrecht C, Rose RJ, van Duijn E, Lorenzen K, Heck AJ (2010) Ion mobility mass spectrometry
of proteins and protein assemblies. Chem Soc Rev 39(5):1633–1655
Uskokovic V, Castiglione Z, Cubas P, Zhu L, Li W, Habelitz S (2010) Zeta-potential and particle
size analysis of human amelogenins. J Dent Res 89(2):149–153
Van Buren N, Rehder D, Gadgil H, Matsumura M, Jacob J (2009) Elucidation of two major aggre-
gation pathways in an IgG2 antibody. J Pharm Sci 98(9):3013–3030
van Duijn E (2010) Current limitations in native mass spectrometry based structural biology. J Am
Soc Mass Spectrom 21(6):971–978
Wall J, Solomon A (1999) Flow cytometric characterization of amyloid fibrils. Methods Enzymol
309:460–466
Wang W, Nema S, Teagarden D (2010) Protein aggregation–pathways and influencing factors. Int
J Pharm 390(2):89–99
Wang G, Abzalimov RR, Kaltashov IA (2011) Direct monitoring of heat-stressed biopolymers
with temperature-controlled electrospray ionization mass spectrometry. Anal Chem
83(8):2870–2876
Wang G, Johnson AJ, Kaltashov IA (2012) Evaluation of electrospray ionization mass spectrom-
etry as a tool for characterization of small soluble protein aggregates. Anal Chem
84(3):1718–1724
Watson JT, Sparkman OD (2008) Introduction to mass spectrometry: instrumentation, applica-
tions, and strategies for data interpretation. Wiley, Chichester
Wear MA, Patterson A, Malone K, Dunsmore C, Turner NJ, Walkinshaw MD (2005) A surface
plasmon resonance-based assay for small molecule inhibitors of human cyclophilin A. Anal
Biochem 345(2):214–226
absorbance and refractive index detectors for studying proteins and their interactions. Anal
Biochem 240:155–166
Wuchner K, Büchler J, Spycher R, Dalmonte P, Volkin DB (2010) Development of a microflow
IgG1 monoclonal antibody formulation. J Pharm Sci 99(8):3343–3361
Ye H (2006) Simultaneous determination of protein aggregation, degradation, and absolute molec-
ular weight by size exclusion chromatography-multiangle laser light scattering. Anal Biochem
356(1):76–85
Yohannes G, Jussila M, Hartonen K, Riekkola ML (2011) Asymmetrical flow field-flow fraction-
ation technique for separation and characterization of biopolymers and bioparticles.
J Chromatogr A 1218:4104–4116
Zhang A, Qi W, Singh SK, Fernandez EJ (2011) A new approach to explore the impact of freeze-
thaw cycling on protein structure: hydrogen/deuterium exchange mass spectrometry (HX-MS).
Pharm Res 28(5):1179–1193
Zhang A, Singh SK, Shirts MR, Kumar S, Fernandez EJ (2012) Distinct aggregation mechanisms
of monoclonal antibody under thermal and freeze-thaw stresses revealed by hydrogen exchange.
Pharm Res 29(1):236–250
Chapter 10
Investigation of Nonconformance During
Protein Therapeutic Manufacturing
Zai-Qing Wen, Gianni Torraca, Guiyang Li, Xiaolin Cao,

and Chanel K. Yee
10.1 Introduction
Development and manufacturing of protein therapeutics involves complicated pro-

cesses from molecular cloning, cell banking, cell culture fermentation, protein puri-
fication, drug product formulation, fill/finishing, packaging, and distribution (Nail
and Akers 2002). Manufacturing nonconformances (NC) that involve particulate
matter can happen during any step of the processes. However, the particulate matter
in the upstream processes (fermentation and purification) may have relatively less
product impact than those observed in the downstream processes (formulation and
fill and finishing). In particular, manufacturing NC at the fill and finishing step can
have direct impact on the final drug product quality. According to the USP compen-
dia, product containers such as glass vials and syringes should be practically free of
particles (United States Pharmacopeia 29-Formulary 24, 2012). If unexpected visi-
ble particulate matter is observed during the visual inspection, the container is
rejected as part of quality control. The potential concern around microparticles in
primary containers is patient safety. If microparticles in drug product were injected
into patient, they could potentially cause inflammation depending on the physical
size and chemical composition of the particles (Barber 2000). Another concern is
the potential immunogenicity of microparticles, if they contain partially denatured
protein (Rosenberg 2006).
Identification of the microparticles and material residues is the first step in these
investigations. The analytical work should first determine the physical size and
chemical identity of these particles, along with the classification of the particles as
intrinsic or extrinsic. The intrinsic particles are normally generated by the product
Z.-Q. Wen (*) • G. Torraca • G. Li • X. Cao • C.K. Yee

Department of Process and Product Development, Amgen Inc.,
One Amgen Center Drive, Thousand Oaks, CA 91320, USA
e-mail: [email protected]; [email protected]; [email protected];
246 Z.-Q. Wen et al.
and the container per se. They include both visible (>100 μm) protein aggregates
growing from single protein molecule either through self-association or partial
denaturation and then aggregation (Narhi 2012) and/or particles generated from the
materials of the primary container such as glass (Iacocca and Allgeier 2007), and
silicone oil (Wen et al. 2009), which is used as a lubricant in prefilled syringes
(PFSs). Extraneous particulate matter or foreign contaminants are material residues
carried over into the primary container that were generated during the manufactur-
ing process of the primary container by the vendor, as well as unexpected foreign
particles from manufacturing environment occasionally fall into the container dur-
ing filling process. For instance, tungsten oxide (Wen et al. 2007) and polymer resi-
due left inside of PFSs (Liu et al. 2010) were from a tungsten pin and nylon pin used
for PFS manufacturing. The identification of these microparticles plays a critical
role in determining the root cause of the nonconformance, including the exact
original source of the particle and the mechanism of particulate generation. This in
turn enables process engineers and formulation scientists to take preventive and cor-
rective action and to help quality and toxicological scientists to assess the product
and safety impact, ultimately determining the disposition or rejection of that par-
ticular lot.
There are many analytical techniques that are employed to determine the physi-
cal size and chemical identities of the microparticles. The most frequently employed
are optical microscopy to determine the morphology and physical size, vibrational
microspectroscopies including Fourier transform infrared (FTIR) and Raman for
the determination of the chemical identity (Humecki 1995; Li et al. 2009; Wen
2007; Wen et al. 2008), and scanning electron microscope (SEM) with energy dis-
persive spectroscopy (EDS) for elemental composition (Goldstein et al. 1992). In
addition, FTIR and Raman microscopy can provide higher-order structural informa-
tion of protein the microparticle may contain (Wen 2007; Wen et al. 2008). Moreover,
information on whether the protein particles are denatured and/or aggregated can be
assessed by the derivative vibrational spectra as well. Using SEM/EDS analysis,
one can further provide a mapping of the distribution of individual elements
throughout the microparticle to determine if the specific element is evenly distrib-
uted in the microparticle or merely heterogeneous among other components in the
particle (Liu et al. 2010).
In this chapter, we will give a brief overview of microparticle analysis and thera-
peutic protein manufacturing NC investigation as currently performed in our labora-
tory and throughout the industry. The general approach to microparticle analysis
and NC incident investigation will be outlined, and the principle of the analytical
procedures will be presented. The goal of this chapter is to give insight into mic-
roparticle NC investigation, the challenges it presents, and how it fits into the overall
application of biophysical and analytical methods to biopharmaceutical develop-
ment and manufacturing. All of the cases presented in the chapter are from previous
investigations that involve protein therapeutics contained in primary containers
including prefilled glass syringes and glass vials. These incidents cover a broad
range of materials that could be encountered during the fill and finishing processes
during manufacturing of protein therapeutics.
10 Investigation of Nonconformance During Protein Therapeutic Manufacturing 247
10.2 Particle Analysis Methods
10.2.1 Microparticle Analytical Scheme
Figure 10.1 shows the flow chart of the scheme of microparticle analysis. The
majority of the microparticle samples are obtained from manufacturing sites and
quality control labs that conducted visual inspection of products contained in pre-
filled glass syringe or glass vials. The primary containers that have microparticles
are normally sent to the forensic lab for analysis at a controlled temperature of
4–8°C. At the beginning of each investigation, a picture of this primary container is
taken with a digital camera or a stereomicroscope to record the particles in the
original state. It is recommended to record images of the entire container as well as
Fig. 10.1 Microparticle analysis procedure

the featured area that has the particles to make sure that their integrity was not
compromised during transportation before the container is opened. In particular,
for glass vials, a picture of the vial cap should be recorded before opening it. This
is important to avoid potential controversy regarding if the vial had been punctured
or not. This is particularly important for vials containing commercial drug product
and returned as part of a customer complaint from patients or clinical centers. If the
vial has been punctured by a needle, the source of the microparticles may not be
caused by the manufacturing process.
The majority of the microparticles can be isolated through filtration using filter
appropriate for the planned analysis. A gold-coated membrane filter (Rapid-ID,
Berlin, Germany) facilitates FTIR-microscopic analysis of the particle directly on
the filter without the need to pick up the particle and transfer it to a potassium bro-
mide (KBr) disc. A polycarbonate membrane filter (Millipore Inc. Billerica) is bet-
ter for SEM/EDS analysis, as it has no strong gold signal to interfere with other
element analysis on the EDS spectrum. If the microparticle is larger than 100 µm, it
may be retrieved directly with an appropriate probe (tungsten probe or plastic spat-
ula) under a stereomicroscope in a laminar flow hood in a clean room (Class 100).
The microparticles that are retained on the filters are then examined with a stereo-
microscope to record their size and morphology, followed by FTIR-microscopic
analysis. If they can be identified definitively by FTIR at this step (especially if they
are organic in nature), further SEM/EDS analysis may not be required. However, for
metallic and inorganic materials, SEM/EDS analysis is a must as these materials
may not exhibit appropriately detailed FTIR or Raman spectra. Even for organic or
polymeric materials, SEM/EDS analysis can provide additional information regard-
ing the elemental composition of the materials to substantiate the vibrational spec-
troscopic analysis. In some cases the microparticles inside the glass container may
be identified using in situ Raman microscopy (Cao et al. 2009, 2010). If the samples
are amenable to in situ Raman analysis, this will be the ideal situation as Raman
microscopy can provide identification without the need to open the container to
isolate the microparticle and is a nondestructive technique.
For complex microparticles that are composed of multiple components, the above-
mentioned techniques may not be sufficient to provide a definitive determination,
and additional chromatographic and spectroscopic techniques such as inductively
coupled plasma mass spectrometry (ICP-MS), gas chromatography (GC), high-
performance liquid chromatography (HPLC/MS), and nuclear magnetic resonance
(NMR) spectroscopy may be required. This may eventually lead to a full-scale com-
prehensive trace material analysis, which is beyond the scope of this chapter.
10.2.2 Optical Microscopy
Optical microscopy is the principal technique to observe and examine microparticles

contained in a primary container. Optical microscopy is used to visualize the micron
particles, in order to manipulate, pick up, and then transfer them for further FTIR
and Raman analysis. Three major optical microscopes are commonly employed as
the first step for microparticle analysis. A Carl Zeiss Stemi 2000C stereomicroscope
can be employed to take pictures of microparticles that have been isolated onto fil-
ters. An Axioskop 2 MAT polarized microscope (Carl Zeiss Inc, Peabody) is
employed to examine microparticles that may have birefringence such as liquid crys-
tal polymer fiber. The polarized microscope can help to identify some materials that
have a distinct birefringence such as cotton cellulose fibers, based on their optical
properties. For in situ examination of microparticles contained inside the prefilled
glass syringe and large-size glass vials, a long working distance and high-resolution
optical microscope is required. These requirements are fulfilled with a Keyence
high-magnification optical microscope VHX-600 or similar instrument (Keyence,
Itasca), which can take high-magnification-quality images of very tiny particles in
situ. The three microscopes are selectively employed according to the specific needs
of investigations and the type of containers and samples being analyzed.
10.2.3 Vibrational Microspectroscopy
FTIR and Raman microscopy are two vibrational microspectroscopic techniques that
provide complementary information on the molecular structure of materials (Humecki
1995; Li et al. 2009; Wen 2007; Wen et al. 2008). They are the result of a marriage
between vibrational spectroscopy and optical microscopy. The combination of a
microscope and a vibrational spectrometer offers the power to focus easily on a spe-
cific microparticle using the microscopic unit and then to record the vibrational spec-
trum of this selected particle with the spectrometer. They are the primary techniques
for determining the identities of microparticles composed of polymeric or organic
compounds at the molecular level. Some inorganic materials can also be examined by
FTIR and Raman microscopy if they exhibit strong vibrational spectra.
The principles of FTIR and Raman spectroscopy are detailed in Chap. 3. FTIR-
microscopy is generally preferred as the analytical tool for microparticle analysis for
particles that have been isolated on a filter as the FTIR-microscope has been devel-
oped into a very sophisticated tool that is relatively easy to operate. Moreover, a large
FTIR database of reference spectra with more than a quarter of a million compounds
is now available from Bio-Rad Know-it-all (Shermann and Brodbelt 2002) that can
be used to aid in particle identification. For microparticles composed of a single
component, the FTIR spectrum of the microparticle is compared with the reference
FTIR spectral library; if a perfect match is found to a reference spectrum from the
library, the identity of the microparticle can be established. In cases where there is no
perfect match, the chemical and structural information from a set of best matching
reference spectra could still provide information on the category or structural family
of the material, and to narrow down the possibilities and the direction the investiga-
tion should take. Further analysis relies on the experienced vibrational spectroscopist
to interpret the FTIR spectra based on the structural chemistry using the empirical
group frequency analysis approach, to determine the most likely composition based
on the combination of functional groups and any knowledge relevant to the particles
(Li et al. 2009; Smith 1999). In many cases, an identification of the type or category
of a material can be very helpful in finding the root cause of the microparticles. This
type of information can also be used to assess the potential chemical toxicity of the
particles and for making decisions on product quality impact.
Raman microscopy is employed as a complementary technique to FTIR-
microscopy. However, it has the advantage over FTIR-microscopy in that it can
perform in situ analysis without the particle isolation procedure (Goldstein et al.
1992; Cao et al. 2009) for microparticles contained in glass vials and prefilled glass
syringes. This is because the visible laser beam used in Raman can penetrate through
the container without any interference from the glass or the aqueous solution. The
glass (or plastic) containers strongly absorb infrared radiation and thus prevent
FTIR measurements in situ. Another unique feature of Raman microscopy is that
molecules that have rich electron density, such as conjugated aromatic compounds,
show strong Raman bands (Chaps. 3, 13, and 16). Inorganic crystal material can
exhibit strong Raman scattering in the low-frequency region as well due to the
vibrational modes of the crystal lattice. The weakness of Raman microscopy is rela-
tively low instrument sensitivity compared to FTIR-microscopy, resulting in longer
acquisition time of a spectrum. Samples with a strong fluorescent background can
also interfere with the ability to obtain a Raman spectrum. Currently the reference
database for Raman spectra is also much smaller (15,000 Raman spectra) than the
FTIR spectral database (a quarter of a million). Fortunately, the Raman database of
organic and polymeric materials has been expanding rapidly in recent years with the
continued improvement of the instrumentation, the increased ease of use, and the
accumulation of Raman spectra of various materials in labs around the world.
Raman microscopy will play an even greater role for microparticle analysis in the
pharmaceutical industry in the near future.
10.2.4 Scanning Electron Microscope/Energy

Dispersive Spectroscopy
SEM coupled with the energy dispersive X-ray spectroscopy (EDS) is a chemical
microanalysis technique that utilizes X-rays emitted from the sample upon
bombardment with an electron beam to characterize the elemental composition of
the materials (Goldstein et al. 1992). The SEM function provides a high-resolution
micrograph of the microparticle and the EDS provides a spectrum of the elementary
composition. When the microparticle is irradiated by the electron beam, electrons
are ejected from the atoms of the sample. The resulting electron vacancy is filled by
electrons from higher orbit shells, and the characteristic X-rays of the element are
emitted. The EDS spectrum consists of the characteristic X-rays of the elements,
which is used for qualitative determinations of the elements in the sample (Goldstein
et al. 1992). SEM/EDS is particularly useful for identification of microparticles
from metallic and inorganic materials that do not produce molecular vibrational
spectra. It is also a powerful tool for identifying composite materials of multiple

elements. Another excellent feature of SEM/EDS instruments is the Cameo map-
ping function of an individual element in a complex microparticle composed of
multiple elements. This feature can generate a distribution map of a selected ele-
ment in the microparticle, showing whether the specific element is evenly distrib-
uted throughout the entire microparticle or concentrated in certain local areas. The
combined use of optical microscopy, FTIR, and Raman microscopy as well as SEM/
EDS offers a comprehensive toolbox that can be used for the analysis of the
physical–chemical properties of the microparticles, in particular, for multicomponent
materials.
10.3 Cases Studies
In the following sections, we will describe a few case studies of manufacturing NC

investigation where visible microparticles were observed in primary containers of
protein therapeutics.
10.3.1 Tungsten/Protein Particulates
Microparticles were initially observed in a protein drug product in prefilled glass

syringes during a stability quality inspection, and the PFSs containing the particu-
lates were then sent to the forensic lab for analysis. The top panel of Fig. 10.2 shows
a picture of the microparticles retained on a membrane filter after isolation. The
largest microparticle is about 100 µm and others are smaller. The bottom panel of
Fig. 10.2 shows the FTIR spectrum of one representative particle, which is typical
of protein FTIR spectrum. However, the FTIR spectrum exhibited a few extra dis-
tinctive bands at 954, 906, and 808 cm−1, respectively; the normal protein FTIR
spectrum does not have sharp bands in this region. This suggested that there might
be other components in the microparticle in addition to the protein. Subsequent
analysis of the microparticle by SEM/EDS revealed that there was tungsten evenly
distributed throughout the particle. Figure 10.3 displays the SEM micrograph and
the EDS spectrum of a typical microparticle observed in the sample, showing the
characteristic tungsten peaks at 1.775 and 8.396 keV, respectively. The identification
of tungsten in the particle helped to explain the presence of the three small peaks at
954, 906, and 808 cm−1 in the FTIR spectrum; they can be assigned to tungsten oxide
(Iacocca and Allgeier 2007). However, it was surprising to observe tungsten in pro-
tein particles in a PFS which is made of glass with no tungsten as its material com-
ponents. The immediate question was, where did the tungsten come from and what
was the role of tungsten in the formation of these protein particulates?
Consulting with the prefilled glass syringe manufacturer revealed that tungsten
pins were employed during the manufacturing of the prefilled glass syringe barrel at
Fig. 10.2 The micrograph of microparticles (top panel) and the FTIR spectrum (up trace) of a
microparticle in a pre-filled syringe and the reference FTIR spectrum of a protein (lower trace)
Fig. 10.3 The SEM micrographs and the EDS spectra of a microparticle and a tungsten chip found
in a pre-filled syringe
high temperatures in the range of 1000°C. The tungsten pin was used to create the
hole of the glass barrel where the needle would later be inserted as the glass was
heated at temperatures approaching the glass melting point. Oxidization of the tung-
sten pin surface can generate tungsten oxide, since the glass syringe barrel was
made in an open air environment at this high temperature. These tungsten-rich com-
pounds can deposit (most likely via sublimation) on the interior of the glass syringe,
primarily near the tip of the barrel (Liu et al. 2010). During the manufacturing pro-
cess of the PFSs, the barrels were rinsed with jets of 80°C water for injection prior
to final packaging and sterilization; however, the tungsten deposits may not be suf-
ficiently soluble to be completely removed by the washing process (Liu et al. 2010).
The possibility of the presence of tungsten residues in the syringe was confirmed by
the analysis of white deposits seen in the funnel area of empty syringe barrels. The
root cause of tungsten element in the microparticles was thus attributed to the tung-
sten pins used at high temperature during the manufacturing process of the glass
barrel (Wen et al. 2007). The residual tungsten species remained in the barrel and
were carried over to the product during the fill and finishing process, at which point
they came in contact with the protein therapeutic. This resulted in particle formation
on stability which was then found by visual inspection.
To understand the role of tungsten compounds in protein particulation, subse-
quent experiments spiking tungsten oxide compounds and used tungsten pin extracts
into a protein solution were conducted (Jiang et al. 2009). These studies demon-
strated that some proteins under specific solution conditions indeed precipitated into
microparticles with the addition of tungsten compounds at a level above 1 ppm of
used tungsten pin extractables. The induced microparticles contained both proteins
and tungsten as confirmed by SEM/EDS (Jiang et al. 2009). Moreover, the tungsten
was also distributed evenly throughout these microparticles as revealed by the EDS
Cameo mapping of tungsten (see Fig. 10.4) (Liu et al. 2010). Biophysical character-
ization of the protein–tungsten aggregates in solution using Raman, FTIR, and DLS
confirmed that the protein particles induced by tungsten in this case were reversible.
They were associated with the formation of tungsten oxide colloids, which require
specific solution conditions and a stoichiometry ratio (1:1) between protein and
tungsten (Jiang et al. 2009). To prevent microparticle formation in protein drug
products induced by tungsten residues, a specification of below 0.5 ppm tungsten
was established for the lots of PFSs. The manufacturer of the PFSs also changed
their manufacturing processes to minimize the tungsten residues in PFSs.
10.3.2 Glass Gel-Like Particulates
In this case, visible particles were initially observed in a few frozen placebo vials of
the protein therapeutic. The particles in the placebo vials were visible to the naked
eye, appeared translucent to white, and amorphous. However, no similar particles
were seen in the protein therapeutic vials which had the exact same formulation as
the placebo. The particles were first isolated on a membrane filter and examined
with optical microscopy followed by FTIR-microscopic analysis.
Fig. 10.4 The cameo mapping of Tungsten element in a microparticle and its EDS spectrum,
which confirms that the tungsten element is distributed evenly on the entire microparticle, which is
a mixture of protein and tungsten element
Figure 10.5 shows the results of the optical microscopic and FTIR-microscopic
analyses. The top panel of Fig. 10.5 shows the micrograph of the microparticles in
the size range of 20–250 µm. The bottom panel of Fig. 10.5 is the FTIR spectra of
the particles together with a reference FTIR spectrum of polysorbate-20 and a refer-
ence FTIR spectrum of silica gel from the Bio-Rad FTIR spectral library (Shermann
and Brodbelt 2002). The FTIR spectrum of the particle showed bands at 1,736 cm−1
(C═O stretching), 2,926 and 2,853 cm−1 (CH2 stretching), 1,460 cm−1 (CH2 bend-
ing), and 1,070 cm−1 (CO stretching), respectively. These FTIR features of the par-
ticles are very similar to the FTIR reference spectrum of polysorbate-20 except that
the strongest band is centered at 1,070 cm−1 in the particle spectrum, while polysor-
bate-20 shows two strong bands at 1,100 and 2,800–2,950 cm−1 region. Since poly-
sorbate-20 was part of the placebo formulation, it was initially suspected to be the
major component and the source of the particles. However, careful examination of
the FTIR spectrum of the particle found that there are a few bands of medium
intensity at 958 and 802 cm−1 that could not be attributed to polysorbate-20.
Moreover, the CH2 stretching bands at 2,925 and 2,853 cm−1 had much weaker signals
than those of polysorbate-20, which has many CH2 groups from its polyoxyethylene
Fig. 10.5 The optical micrograph of the microparticles (top) and FTIR spectra of a microparticle,
the reference FTIR spectra of Polysorbate-20 and of a silica gel
chain and the long fatty acid ester moiety. This suggested that the particulates may
have contained polysorbate-20 but may have other components which contribute to
the FTIR bands that cannot be assigned to surfactant (Fig. 10.6).
Subsequent SEM/EDS analysis of the particles revealed an extra element Si in
addition to the two major expected organic elements of C and O. This indicated
that the particles included silicon-containing material. Trace aluminum was also
found in some of the particles (data not shown here) by SEM/EDS. Moreover, a
Cameo mapping of the particle with Si element showed that the silicon was distrib-
uted evenly throughout the particle (see Fig. 10.5). The vial was made of borosili-
cate glass, and silicon can be dissolved from the glass vial in aqueous solution
(Perera and Doremus 1991). Quantitative comparative analysis of the silicon and
other glass elements (boron, aluminum, sodium) by ICP-MS revealed that the pla-
cebo glass vial containing visible particles had 30 times silicon in solution than a
control placebo plastic vial (below the detection limit of 150 ppb) (Ricci et al. 2011).
These analyses strongly suggested that the visible particles were composed of silica
Fig. 10.6 Secondary electron image of a representative particle on gold coated filter (Left) the sili-
con element map displays Si in blue (Right). Note that Si is evenly distributed throughout the
particle and it matches the particle shape well shown on the left
and polysorbate-20 and that the glass container itself was the source of the silica
particles. The root cause analysis then focused on why the placebo vials had the
visible particles, but the product vials did not, and what caused the placebo vial to
generate visible particulates that contain silicon.
Reviewing the manufacturing history of the protein product and placebo vials
revealed some subtle differences in the post-fill handling and storage temperatures
between the product and placebo vials. To maximize the stability of drug product, it
was frozen at −30°C immediately after fill. In contrast, the placebo was considered
to be robust and therefore their storage temperature was not deemed critical. For
practical and logistical considerations, the placebo vials were stored at 2–8°C after
fill until they were needed for clinical labeling and packaging. Since the demand
varied over the course of the clinical studies, the time of storage at 2–8°C varied
across the placebo lots by months. On the other hand, the product vials were frozen
and stored at −30°C both after fill and packaging and at the clinical trial sites.
Moreover, it was realized that the product and placebo were formulated at neutral
pH 7, a pH that can accelerate glass dissolution (Perera and Doremus 1991). Reports
in the literature indicated that silica, the dominant component of the glass vials, is
less stable in neutral pH than under relatively acidic conditions (Perera and Doremus
1991). The combination of neutral pH and storage of the placebo vials at 2–8°C for
a period apparently increased the dissolution of glass into solution in the form of
silicic acids, which then polymerized into silica gel (Hsu-Chiang et al. 2005). The
polysorbate-20 molecules were merely associated to the silica gel particulates, but
not the cause. Scanning of the inner surface of the placebo glass vials by SEM
revealed that there were defects in the glass surface of the liquid contact area, but
not on the nonliquid contact area above the fill line (data not shown here). This fur-
ther reinforces the hypothesis that the silica gel particles were indeed formed from
the glass vial through glass dissolution. Since the protein drug product was frozen
at −30°C immediately after fill and was stored at −30°C, the silica dissolution was
greatly minimized at lower temperature for the product container (Ricci et al. 2011).
Therefore, the storage condition of the placebo was changed to match that of the
protein therapeutic product, reducing the storage time of the placebo at 2–8°C to
less than 48 h.
Fig. 10.7 Optical micrograph, FTIR spectra of the bubble-like particle (top trace FTIR is the
bubble like microparticle and the bottom trace FTIR is the reference spectrum of silicone oil
10.3.3 Silicone Oil/Protein Particulates
During a product comparability study, bubble-like visible particles were observed in

a product that was contained in PFSs. FTIR spectra, along with the corresponding
optical images (insets), of the isolated particles are shown in Fig. 10.7. The parti-
cles, varying from microns to millimeters, appeared to be “bubble-like” with most
of the particles dried on the edge. The infrared spectra of the particles revealed two
different spectral characteristics. The particle shown in Fig. 10.7a exhibits a typical
protein IR spectrum with characteristic bands of N–H stretching (3,292 cm−1), the
CH2 asymmetric stretching (2,937 cm−1), the CH3 symmetric (2,876 cm−1) and
asymmetric (2,966 cm−1) stretching modes, and the amide I (1,646 cm−1) and amide
II (1,539 cm−1) bands. The FTIR spectrum of the particle shown in Fig. 10.7b, on the
other hand, revealed the spectral characteristics of a mixture containing protein and
another component. In addition to the major protein bands mentioned above, the IR
spectrum showed typical silicone oil spectral signatures, indicated by the CH3
asymmetric stretching (2,965 cm−1) band, the Si-CH3 deformation band at
1,261 cm−1, and the twin peaks at 1,094 and 1,026 cm−1 associated with the Si–O–Si
and the Si–O–C stretching vibrations, respectively (Smith 1999). Also present were
Fig. 10.8 The two SEM/EDS spectra are obtained from the bubble like particles. The top panel of
the EDS spectrum was obtained on the protein component of the bubble-like particle. The bottom
panel of the EDS spectrum was obtained on the bubble-like micro-particle that containing silicone
oil and protein
the Si–C stretching bands at 865 and 800 cm−1. The FTIR spectra were consistent
with a protein–silicone oil mixture.
The SEM/EDS analyses, shown in Fig. 10.8, supported the FTIR results as the
EDS spectra of the two particles differed only in the presence of Si, which is attrib-
uted to the presence of silicone oil in the particle shown in Fig. 10.7b. Other than Si,
both particles were found to contain C, N, O, and S. This is consistent with the FTIR
results as a proteinaceous material containing the basic elements of sulfur, carbon,
oxygen, and nitrogen, although some C and O signals could also be contributed by
the underlying polycarbonate membrane filter. The EDS spectra also showed P, Na,
and Cl, which can be attributed to the formulation buffer. The SEM micrographs
demonstrated morphological similarities of both “dried bubble” particles. The
source of the silicone oil was the PFS, as silicone oil is coated in a very thin layer of
a few hundred nanometers on the glass barrel surface as a lubricant (Wen et al.
2009). The generation of these bubble-like particles was attributed to the silicone
oil/protein/buffer interaction during the transportation (vibration) of the products
via dewetting processes as silicone oil is hydrophobic, but is in contact with aqueous
solution. Tiny silicone oil droplets on the glass surface can migrate together to form
relatively large visible bubble-like particles associated with protein molecules.
10.4 Conclusions
NC investigations and root cause analyses are critical to protein therapeutic devel-
opment and commercialization. Physical particulate contaminants are one of the
types of NC that can occur during the fill and finishing process of biopharmaceutical
products. When this occurs, the physical–chemical nature of the particulate matter
must be identified as one of the first steps in determining the origin and the root
cause to allow remediation. This requires the use of the appropriate tools from the
biophysical toolbox as described above.
It is worthwhile to point out that many physical–chemical contaminants found in
the final biopharmaceutical products were originally generated during the process-
ing of the drug product container and were associated with the container manufac-
turing equipment, being carried over into the final commercial drug product. Other
times, it is the glass container per se that interacted with the formulation compo-
nents of the protein products to generate particulate issues such as glass delamina-
tion. To quickly resolve NC investigations, detailed analysis of materials present,
and determination of their root cause and source are mandatory. Moreover, close
collaboration among the raw material vendor, primary container manufacturer, and
the protein pharmaceutical manufacturer is required for resolution of the NCs and
to take appropriate preventive measures.
References
Barber TA (2000) Control of particulate matter contamination in healthcare manufacturing.

Interpharm, Denver
Cao X, Wen ZQ, Vance A, Torraca G (2009) Raman microscopic applications in the biopharma-
ceutical industry: in situ identification of foreign particulates inside glass containers with aque-
ous formulated solutions. Appl Spectrosc 63:830–834
Cao X, Masatani P, Torraca G, Wen ZQ (2010) Identification of a mixed microparticle by com-
bined microspectroscopic techniques: a real forensic case study in the biopharmaceutical
industry. Appl Spectrosc 64:895–900
Goldstein JI, Newbury DE, Echlin P, Joy DC, Romig AD Jr, Lyman CE, Fiori C, Lifshin E (1992)
Scanning electron microscopy and X-ray microanalysis, 2nd edn. Plenum, New York
Humecki HJ (1995) Practical guide to infrared microspectroscopy. CRC, New York
Iacocca RG, Allgeier M (2007) Corrosive attack of glass by a pharmaceutical compound. J Mater
Sci 42:801–811
Jiang Y, Nashed-Samuel Y, Li C, Liu W, Pollastrini J, Mallard D, Wen ZQ, Fujimori K, Pallitto M,
Donahue L, Chu G, Torraca G, Vance A, Mire-Sluis T, Freund E, Davis J, Narhi LO (2009)
Tungsten-induced protein aggregation: solution behavior. J Pham Sci 98:4695–4710
Kuan H-C, Su H-Y, Ma C-C (2005) Synthesis and characterization of polysilicic acid nanoparti-
cles/waterborne polyurethane nanocomposite. J Mater Sci 40(23):6063–6070
Li G, Torraca G, Jing W, Wen Z-Q (2009) Applications of FTIR in identification of foreign materi-
als for biopharmaceutical clinical manufacturing. Vib Spectrosc 50:152–159
Liu W, Swift R, Torraca G, Nashed-Samuel Y, Wen ZQ, Jiang Y, Vance A, Mire-Sluis A, Freund
E, Davis J, Narhi LO (2010) Root cause analysis of tungsten-induced protein aggregation in
pre-filled syringes. PDA Pharm J Sci Technol 64:11–19
Nail SL, Akers MJ (2002) Development and manufacture of protein pharmaceuticals. Springer,
New York
Narhi LO, Schmit J, Bechtold-Peters K, Sharma D (2012) Classification of protein aggregates.
J Pharm Sci 101:493–498
Perera G, Doremus RH (1991) Dissolution rates of silica glasses in water at pH 7. J Am Ceram Soc
74(6):1269–1274
Ricci MS, Hair A, Thirumangalath R, Dharmavaram V, Piedmonte D, Latypov R, Lau H, Bjorklund

J, Callahan W, Purtell J, Abel J, Li G, Fujimori K, Torraca G, Jing W, Wen Z-Q, Nashed-
Samuel Y, Ratnaswamy G, Swift R (2011) Glass dissolution and delamination in formulation
development: implications for neutral pH product development. In: American Chemical
Society Division of Biochemical Technology 241st ACS national meeting, Anaheim, 27–31
March 2011
Rosenberg AA (2006) Effects of protein aggregates: an immunologic perspective. AAPS J
8(3):E501–E507
Shermann C, Brodbelt J (2002) Knowitall analytical system. J Am Chem Soc 124:7
Smith B (1999) Infrared spectral interpretation: a systematic approach. CRC, New York
United States Pharmacopeia 29-Formulary 24 (2012), p 2722 (Chap. 788) USP 35-NF30
Wen ZQ (2007) Raman spectroscopy of protein pharmaceuticals. J Pharm Sci 96:2861–2878
Wen Z-Q, Torraca G, Yee C, Li G (2007) Investigation of contaminants in protein pharmaceuticals
in pre-filled syringes by multiple microspectroscopies. Am Pharm Rev 10:101–109
Wen ZQ, Cao X, Vance A (2008) Conformation and side chains environments of recombinant
human interleukin-1 receptor antagonist (rh-IL-1ra) probed by Raman, Raman optical activity,
and UV-resonance Raman spectroscopy. J Pharm Sci 97:2228–2241
Wen ZQ, Vance A, Vega F, Cao X, Eu B, Schulthesis R (2009) Distribution of silicone oil in pre-
filled glass syringes probed with optical and spectroscopic methods. PDA Pharm J Sci Technol
63:149–158
Chapter 11
Higher-Order Structure and Protein Aggregate
Characterization of Protein Therapeutics:
Perspectives from Good Manufacturing
Practices and Regulatory Guidance
Evi B. Struble, John F. Cipollo, Chava Kimchi-Sarfaty,

Zuben E. Sauna, Jack A. Ragheb, and Ewa Marszal
11.1 Introduction
Proteins are complex molecules that affect gene expression, cellular structure,
metabolism, reproduction, inter-/intracellular signaling, immune responses, and virtu-
ally every aspect of life. Function of proteins depends on the orientation of the func-
tional groups of their building blocks. For example, the presence and orientation of
the three critical amino acids (Ser, His, and Asp) in the catalytic triad of serine prote-
ases, such as coagulation factors, determines their ability to perform their biologic
role, i.e., cleave a peptide bond and, in the case of the coagulation cascade, activate
the substrate. Furthermore, nearby regions (secondary and tertiary structures) are
Disclaimer The findings and conclusions in this chapter have not been formally disseminated by
the Food and Drug Administration and should not be construed to represent any agency determina-
tion or policy.
E.B. Struble • C. Kimchi-Sarfaty • Z.E. Sauna • E. Marszal (*)
Division of Hematology, Office of Blood Research and Review,
Administration (FDA), 29 Lincoln Drive, Bethesda, MD 20892, USA
e-mail: [email protected]; [email protected];
J.F. Cipollo
Division of Bacterial, Parasitic and Allergenic Products,
Office of Vaccine Research and Review, CBER, FDA, 29 Lincoln Drive,
Bethesda, MD 20892, USA
J.A. Ragheb
Division of Therapeutic Proteins, Office of Biotechnology Products, Center for Drug Evaluation
and Research, FDA, 29B Lincoln Drive, Bethesda, MD 20892, USA
262 E.B. Struble et al.
essential for substrate specificity and, in the case of coagulation factors, enhancement
of the biologic activity by cofactors and attenuation of the effect by inhibitors.
Unlike the bonds that determine the structures of small molecules, the interactions
driving the complex structure of proteins—hydrogen bonds, electrostatic interac-
tions, and van der Waals forces—are noncovalent in nature. As such, the energy that
maintains the stability of a protein’s three-dimensional structure is inherently low,
in many cases not much greater than the thermal energy at room temperature (Pfeil
and Privalov 1976). Thus, increases in temperature result in protein unfolding
(Griko et al. 1988). Other environmental conditions such as pH (Anderson et al.
1990), solvents (Grothe et al. 2009), salts, small molecule additives, impurities
(Bolen and Baskakov 2001; Maity et al. 2009; Rosgen 2007; England and Haran
2011), and even other proteins (Hartl and Hayer-Hartl 2009) can influence the three-
dimensional structure and result in complete or partial unfolding of the protein.
Destabilized proteins often form aggregates, which may vary from dimers with a
size in the low nanometer range to visible agglomerates of thousands of protein
molecules reaching a size of several millimeters. Aggregated proteins lose physio-
logic activity (Meager et al. 2011; Carpenter et al. 2010a) and may acquire new,
toxic attributes. Examples include diseases associated with protein aggregation such
as systemic amyloidoses (Obici et al. 2005), neurological disorders (Aguzzi and
O’Connor 2010), and, as recently shown, cancer (Xu et al. 2011). Interestingly, it has
been postulated that in some diseases toxicity is associated with oligomeric interme-
diates rather than the large aggregates (Haass and Selkoe 2007; Klein et al. 2001).
The formation of protein aggregates is a major concern in the manufacture and
clinical use of protein therapeutics. Protein aggregates have been associated with
enhanced immune responses both in animal studies and clinical trials (Dresser
1962; Braun et al. 1997; Moore and Leppert 1980). Aggregates may stimulate the
innate immune system to promote an adaptive immune response by enhancing anti-
gen presentation and the release of cytokines, leading to immunogenicity. Cytokine
release may also directly induce more immediate infusion-associated reactions.
The effects of the resulting antidrug antibodies (ADA) can range from a small
alteration in the pharmacokinetic properties of the therapeutic protein to a decrease
or complete loss of efficacy due to immune-mediated clearance or neutralization of
the biologic activity of the drug. When both the ADA and the therapeutic protein are
present at sufficiently high levels, such as may occur with therapeutic monoclonal
antibodies, toxicity can occur due to deposition of circulating immune complexes.
In the most severe cases, the immune reaction to the offending therapeutic protein
(e.g., erythropoietin) can cause a breakdown of immune tolerance, resulting in neu-
tralization of the patient’s endogenous counterpart to the therapeutic.
Given the consequences to both activity and toxicity, conformational changes
and protein aggregation affect product quality and can affect product safety. Product
safety and efficacy are determined during clinical trials and must not decline over
the shelf life of a given lot or the market life of a product. It is, thus, essential that
the structural and functional features of the product be characterized to an accept-
able extent during product development and that each batch of the product has simi-
lar characteristics at release and throughout its shelf life. In part, this is ensured by
11 Higher-Order Structure and Protein Aggregate Characterization of Protein… 263
product release testing, but because this testing may not be capable of detecting
subtle changes that can occur in the product due to variations in manufacturing
processes, product consistency over time must be ensured by tight control of the
manufacturing process. In addition, any major manufacturing change should be
supported by detailed product characterization and comparability studies, including
a comparison of product stability profiles.
Factors that affect protein structure, aggregation, and stability are numerous,
and, while we acknowledge that not all of them can be controlled during the manu-
facture of biotherapeutic products, the principles outlined in this chapter are meant
to be science-based good manufacturing considerations to be applied when and if
appropriate and feasible for a given specific product.
It cannot be overemphasized that assuring the quality of biologic products is a
complex process that can be achieved only through tight control of process parame-
ters and in-depth characterization beginning with product development and continu-
ing through the lifecycle of the product, including routine testing of every batch and
long-term monitoring of product stability. Recent technical developments allow for
the manufacture of biosynthetic protein therapeutics under tightly defined conditions.
This has increased our potential to limit impurities and control biosynthetic protein
characteristics to a greater extent than is possible for naturally derived products.
When coupled with extensive characterization of protein structure, these advance-
ments better ensure protein comparability after a manufacturing change.
11.2 Intrinsic Factors Affecting Protein Higher-Order

Structure and Stability
11.2.1 Protein Sequence and DNA Sequence
Since Anfinsen’s Nobel Prize winning work (Anfinsen 1973), it is well established
that the primary sequence of the polypeptide chain determines a protein’s three-
dimensional structure. Since then, the discovery of the protein basis of “inborn
errors of metabolism” (Garrod 1908; McCarthy 2011; Benoit et al. 2010; Elia and
Albanese 2010; Percy and Rumi 2009; Bentham and Bhattacharya 2008; Scriver
2007; Pasinelli and Brown 2006; Harrison et al. 2008) and the experience acquired
in the development and clinical use of engineered proteins (Alexander et al. 2007;
Meeker et al. 1996; Stepanenko et al. 2008) have underscored that even seemingly
small changes in protein primary sequence can affect structure and stability. In the
case of therapeutic proteins made from recombinant sources, this understanding
serves to reinforce the importance of the amino acid sequence, its implications for
aggregation propensity, and, as a result, product efficacy and safety.
However, the amino acid sequence is not the only determinant of the protein
three-dimensional structure. Indeed, folding kinetics and the structure of proteins
can also be affected by synonymous mutations resulting in phenotypic changes

such as an increased resistance to anticancer drugs (Kimchi-Sarfaty et al. 2007).
One of the possible mechanisms for this effect is via codon usage and cognate
tRNAs frequency (Sharp et al. 1988; Ikemura 1985). Organism and cell-specific
distribution of tRNAs, together with biased codon usage, could provide pauses dur-
ing translation at critical sites, such as at the end of individual domains. This may
allow a protein to fold correctly by curtailing disruptive interactions between
unfolded regions of the protein. Thus, in order to achieve the desired fold in a
recombinant protein, the conditions for its synthesis should be optimized with
respect to codon usage and tRNA abundance. Other parameters of its synthesis,
which are addressed in more detail in Sect. 3.1 of this chapter, should also be strin-
gently controlled, especially since minor structural changes that may result from the
suboptimal rate of translation and the composition of the cellular milieu may be
difficult to identify.
11.2.2 Protein Glycosylation
Glycosylation is another determinant in the formation of protein higher-order

structure and can help to improve protein stabilization. The added stability that
glycosylation may provide can serve as a counterbalance to many of the degrading
forces that might be encountered during therapeutic protein manufacturing and
clinical application, such as differences in pH, temperature, or chemical condi-
tions. The stabilizing effects include reduced short- and long-range mobility (Sola
et al. 2007; Liu et al. 2006), increased intramolecular electrostatic interactions,
formation of additional hydrogen bonds and van der Waals interactions, and
increased solubility (Mitra et al. 2006; Rademacher et al. 1988). Increased solvent
contacts also decrease intermolecular electrostatic interactions, thus further dimin-
ishing the tendency toward aggregation. Glycosylation can also act as a shield
against some chemical degradative processes such as oxidation and cross-linking.
Examples of these include free radical scavenging in erythropoietin (Uchida et al.
1997) and suppression of transamidation in glyco-engineered insulin (Baudys
et al. 1995). In general, glycosylation diminishes the immunogenicity of therapeu-
tic proteins.
Pharmacokinetics can be positively or negatively affected by changes in glyco-
sylation. The addition of glycosylation sites or greater glycosylation site occupancy
has been reported to add considerably to serum half-life and clearance times (Fares
et al. 1992; LaPolt et al. 1992). However, loss of glycan or inappropriate glycan
composition can negatively impact clearance (Appa et al. 2010; Wright and
Morrison 1998). As we will discuss in the next section, manufacturing conditions
can dramatically affect glycan composition and therapeutic efficiency. Therefore,
both glycan characterization and rigorous control over manufacturing conditions
that influence this parameter are necessary for stable therapeutic protein pharmaco-
kinetic profiles.
11.3 Extrinsic Factors Affecting Protein Higher-Order

Structure and Stability
11.3.1 Recombinant Protein Production Process
Protein folding and its glycosylation can be affected by conditions encountered during
the upstream manufacturing process, including both the intra- and extracellular
milieu. Chaperones such as the Hsp70s play an essential role in correct folding and
secretion of proteins (Kramer et al. 2009). During protein synthesis and
posttranslational modifications, binding to chaperones protects the polypeptide
chain from collapse and incorrect folding that could occur in the absence of the yet
to be translated, downstream sections or from aggregation in the highly crowded cell
environment. Although chaperones are found across all organisms, there are mecha-
nistic differences as well as differences in expression levels amongst commonly
used recombinant protein production systems and cell lines (Hartl and Hayer-Hartl
2009). To add to this complexity, the availability of chaperones depends on other
factors such as temperature, nutrient state (Neuhofer et al. 1999), and the expression
level of the recombinant protein. Chaperone availability in turn can affect folding,
aggregation, and yield of the product. These factors should be carefully analyzed not
only when the process is being developed but also when changes in media formula-
tion and other process parameters are being considered or implemented.
Many cell culture process variables can affect glycosylation. Manufacturing
scale, fermentation type, and fermentation conditions all must be considered.
Bioreactor pH, manganese ions, dissolved oxygen, ammonia concentration, and
temperature have all been shown to affect glycosylation patterns (Hossler et al.
2009). The three major modes of production—batch, fed-batch, and perfusion—all
have different waste product accumulation and nutrient depletion profiles, which
can have profound effects on glycosylation profiles. Accumulation of ammonia,
principally through glutamine and asparagine metabolism, raises media pH, which
can diminish the activity of the late Golgi glycosyltransferases (Thorens and Vassalli
1986). Manganese is required for the activity of some glycosyltransferases, includ-
ing those involved in late carbohydrate modifications such as the addition of
N-acetylneuraminic acid and galactose. Nucleotide sugar precursors and dolichol,
which are substrates for glycosylation biosynthesis, can become depleted, thus
affecting glycosylation efficiency (Castro et al. 1995; Jenkins et al. 1994;
Kochanowski et al. 2008). The mode of production dictates the necessary level of
control over these factors, and therefore, its choice must be weighed with care.
Shear stress has also been shown to affect late Golgi processing (Senger and Karim
2003) and should be taken into account, especially when the scale of production is
changed (Hossler et al. 2009). All of these factors should be considered early in the
development of the manufacturing process as differences in cellular stresses due to
culture scale and production mode can have profound effects on glycosylation pro-
files and thus the conformation, efficacy, and safety of the protein product.
11.3.2 Purification Process
Protein therapeutic products are derived from highly complex starting materials—human
plasma or blood, transgenic animal milk, bacterial, plant, insect, yeast, and mam-
malian cells or tissues. Multistep purification protocols are often used that may
include precipitation and filtration steps, chromatography steps, and freeze/thawing.
In addition, viral inactivation operations involving filtration, elevated temperatures,
low pH exposure, and/or the use of solvent/detergents are frequently employed.
Virtually every step of the manufacturing process may have an effect on protein
structure (Vazquez-Rey and Lang 2011; Cromwell et al. 2006). Therapeutic pro-
teins are often purified from a mixture containing cell and/or tissue components
based on different surface characteristics that affect solubility and interactions with
solid media. Conditions selected for precipitation operations and chromatography
steps involve pH changes, addition of a nonpolar solvent, or a change in ionic
strength. These changes may alter the distribution of charges on the protein surface
and the properties of the solvent, affecting the interactions of the protein with both
itself and the solvent and thus potentially impact the integrity of the protein
(Zimmerman 2006; Lewis and Nail 1997; Shukla et al. 2007). Temperature also has
an effect on the strength of these interactions, with a rise in temperature strengthen-
ing hydrophobic and reducing ionic interactions. However, ultimately, an increase
in temperature results in higher kinetic energy and will eventually lead to partial
unfolding. Unfolded protein species are highly susceptible to formation of aggre-
gates which are more stable in common formulation media than the native form of
the protein. Additionally, low temperatures may induce cold denaturation, and inter-
action with ice during lyophilization, an operation that involves freezing and dehy-
dration, may also destabilize protein structure (Rathore and Rajan 2008; Tang and
Pikal 2004). Mechanical stress, which affects therapeutics when the protein solu-
tions are pumped, mixed, filtered, or filled, may lead to structural distortions and
aggregation of the protein product (Rathore and Rajan 2008). In addition, chemical
modifications (e.g., oxidation and deamidation) occurring during the purification
process or upon storage may lead to conformational changes, activity reduction, and
aggregation (Luo et al. 2011; Patel et al. 2011). Extractables and leachables are
other potential sources of chemical modifications that may impact protein stability
(Huang et al. 2011). Lastly, lyophilization, used in the manufacture of many biolog-
ics, can result in local concentration and pH changes that may induce conforma-
tional changes in proteins leading to aggregation.
11.3.3 Storage and Transportation
Protein stability upon storage and transportation greatly depends on the protein’s
physical state, i.e., on whether the protein is dissolved or lyophilized and on the
composition of the product formulation. The formulation should be carefully
optimized not only to ensure long shelf life but also to enable successful completion
of the last steps of manufacturing (Jorgensen et al. 2009; Shire 2009). Other impor-
tant factors include temperature (and potential temperature excursions), presence of
interfaces such as a liquid–air interface, and interactions with the container-closure
system, including leachables, extraneous particles such as glass, metal, and silicone
oil; the latter of which is commonly used as a lubricant in syringes and on stoppers
(Bee et al. 2011). Product agitation during transport and handling may also induce
aggregation of the protein product (Kiese et al. 2008; Thirumangalathu et al. 2009).
Thus, in addition to thorough characterization of the final container product, careful
analysis of the effect of transportation parameters on the protein conformation and
aggregation state of the product is important.
11.3.4 Aggregation During Reconstitution and Administration
Lyophilized biotherapeutics are reconstituted before administration. Reconstitution

may involve the use of transfer devices that help to transfer sterile water for injec-
tion to vials containing the lyophilized product. If siliconized, the elements of infu-
sion sets, such as transfer devices, needles, and syringes, may contribute to formation
of protein aggregates in the product. Different devices may carry different amounts
of silicone. Therefore, similar devices from different suppliers may not be equiva-
lent or interchangeable.
Patient exposure to protein aggregates may be limited by the use of in-line filters,
which are recommended for many protein therapeutics. Nonetheless, to ensure high
quality of the product, manufacturers should minimize the level of aggregates in the
final container.
11.4 Product Control and Process Control:

Elements of Product Quality Assurance
Protein therapeutics, in contrast to small molecule drugs, can never be fully charac-
terized. Their structure is complex, small structural changes are difficult to recog-
nize, and contaminating factors that co-purify with the protein product and which
exist in minute amounts may not be well defined. The quality of biological materials
is verified by final lot release testing of the product and by ensuring the consistency
of manufacture. The latter is achieved by establishing in-process parameters and
product limits that have been demonstrated to deliver a defined quality product.
Extended product characterization is performed for licensure and at the time of any
major manufacturing change. As product quality must be maintained throughout
shelf life, quality attributes are evaluated during stability studies.
11.4.1 Analytical Methods for Characterization of the Primary,

Secondary, and Higher-Order Structure
In addition to functional characterization, to help ensure process control, comparability,

and manufacturing consistency, recombinant therapeutic proteins should be charac-
terized to an acceptable extent by assessing the protein’s primary, secondary, and
higher-order structure. Analytical techniques used to assess protein structure
should be scientifically sound, reflect the current state of scientific understanding,
and have high sensitivity and precision. Considering the complexity of protein
structure and the inherent limits of any given analytical tool, at least two orthogonal
methods should be used for structural characterization. For example, the primary
sequence can be determined by using mass spectrometric analysis, which is based
on fragmentation of the polypeptide chain and the rate of fragment migration in an
electromagnetic field, and by N-terminal sequencing following the Edman degrada-
tion reaction or by C-terminal sequencing following chemical degradation with
thiocyanate.
The secondary structure can be characterized using circular dichroism (CD),
which is based on the differences in absorption of polarized light in proteins, or by
analyzing the spectra of frequencies of vibration for a molecule using infrared and
Raman spectroscopy. The secondary and tertiary structure can also be assessed by
NMR spectroscopy, which measures chemical shifts of the magnetically active nuclei
(1H, 13C, 15N, etc.) when the protein solution is placed in a magnetic field. Tertiary
and quaternary structure can be probed with methods such as intrinsic and dye fluo-
rescence spectroscopy which respectively measure fluorescence of aromatic amino
acids such as tryptophan and fluorescent reporter molecules bound to the protein,
thus providing information on the local environment of the fluorescing moiety. X-ray
crystallography, based on calculating the electron density of atoms from the diffrac-
tion pattern of the crystallized protein, is a powerful technique for determining the
secondary, tertiary, and quaternary structure of a protein. Hydrogen/deuterium (H/D)
exchange mass spectrometry enables detection of conformational differences based
on solvent accessibility to the protein amino acid residues. Chromatography tech-
niques, measuring the mobility in a chromatographic column, can also provide infor-
mation on changes in protein structure and surface characteristics.
Protein aggregate content can be detected by means of visual inspection, microscopy
(optical, fluorescence, electron, atomic force), light obscuration, flow imaging, con-
ductivity-based counting (Coulter method), static and dynamic light scattering,
nanoparticle tracking analysis, turbidity, nephelometry, X-ray scattering, and mass
spectrometry. Separation methods such as size exclusion chromatography, field-
flow fractionation, polyacrylamide gel electrophoresis, capillary electrophoresis,
and analytical ultracentrifugation can also be used. Recently, fluorescence single-
particle tracking was used to size submicron matter in undiluted whole blood
(Braeckmans et al. 2010), and the technique has been adopted to evaluate its poten-
tial to detect and determine the size of submicron protein aggregates in serum,
plasma, and pharmaceutical formulations containing albumin (Filipe et al. 2011).
The characterization, utility, and limitations of these and other techniques have been
discussed in numerous reports (see, e.g., Philo 2009; Huang et al. 2009; Fraunhofer
and Winter 2004; Filipe et al. 2010 and Chap. 3 and 4 of this volume). Although the
techniques have evolved in sophistication and sensitivity, analyzing therapeutic pro-
teins or their aggregates in biological media remains a major challenge.
There is an interest in industry to use high throughput screens to assess higher-
order structure. Among the established methods of measuring protein stability, ther-
mal and chemical denaturation monitored by intrinsic fluorescence were the earliest
to be automated (Stites et al. 1995) and are constantly being improved upon (Edgell
et al. 2003; Aucamp et al. 2005). This technology has been reduced to the nanoliter
scale (Gaudet et al. 2010).
A technique that does not depend on the intrinsic fluorescence of a protein, which
can vary widely depending on the local environment and number of fluoroactive
residues, is differential scanning fluorimetry (DSF). This method, based on moni-
toring fluorescence of a specific dye that has increased affinity for the hydrophobic
regions of protein surfaces, has been found to be applicable to a broad range of
proteins (Ericsson et al. 2006, Chap. 2 of this volume), and good correlations have
been found between melting temperature values determined from CD thermal dena-
turation and DSF (Lavinder et al. 2009).
One goal of adopting high throughput assays for determining protein stability is
to sort out the stable variants in a library. To reduce the search space, considerable
effort has been expended in developing first-principles computational methods to
predict the stability of proteins. These programs have relied, to a large extent, on
knowledge-based potentials based on distances between residue pairs or backbone
torsion angles (Gilis and Rooman 2000), statistical potentials (Parthiban et al.
2006), empirical potentials that describe the physical interactions that contribute to
protein stability (Yin et al. 2007), and different types of machine learning tools
(Capriotti et al. 2005; Cheng et al. 2006; Masso and Vaisman 2008). While most
programs require knowledge of the structure of the target protein, some methods
can predict stability changes in proteins of unknown structure with reasonable accu-
racy (Capriotti et al. 2005). While these algorithms are unlikely to produce a com-
prehensive predictive model in the foreseeable future, they can be used in conjunction
with experimental methods.
The above method enumeration is by no means exhaustive. A recent industry
effort to review and evaluate analytical methods available for protein aggregates
detection and characterization was reported upon by den Engelsman et al. (den
Engelsman et al. 2011). Furthermore, there are a large number of publications on
biophysical methods that can be used as a guide to select an appropriate set of
methods. There are a few methods which, although at present are not used widely
by industry, show great potential for evaluating therapeutic biologics and should
be explored further. For example, H/D exchange in conjunction with MALDI
mass spectrometry, another robust and well-characterized method, has recently
been adapted to measure protein stabilities in living cells (Ghaemmaghami and
Oas 2001). It was also recently employed to detect conformational changes asso-
ciated with posttranslational modifications of a monoclonal antibody preparation.
Interestingly, the detected changes correlated with changes in receptor binding

(Houde et al. 2010).
Due to inherent differences in method sensitivity, complexity of the measure-
ment, and other considerations such as inter-operator precision, different methods
to determine higher-order structure may give vastly different results. For example,
slight changes in the amino acid composition of proteins can greatly affect NMR
spectra but do not affect the overall structure as determined by X-ray crystallogra-
phy (Struble et al. 2008). While some differences in results are to be expected,
method validation and the use of appropriate controls and standards should reduce
variability. Furthermore, for many of these methods, data fitting by using mathemat-
ical models and interpretation still poses a challenge and is an area where an inte-
grated effort is needed to overcome existing limitations. This need will intensify in
the coming years with increased demand for optimized bioengineered products that
are safe and efficacious, as well as convenient for patients. There has been consider-
able progress in the last decade, but there continues to be unmet needs with respect
to this important issue for the biotechnology and pharmaceutical industries.
11.5 Industry Standards and Regulatory Considerations
Several guidances promulgated by the FDA1 or the International Conference on

Harmonization of Technical Requirements for Registration of Pharmaceuticals for
Human Use (ICH),2 to which the FDA is a signing party, address issues central to
ensuring the structural and functional integrity (in short the “quality”) of a biologi-
cal therapeutic product. The long-standing emphasis on understanding the manufac-
turing process and the biological product is critical for successful and controlled
manufacture of biologics, including protein therapeutics. The recent Quality by
Design paradigm provides a structured, systematic approach to understanding the
process and product, which can provide substantial knowledge and understanding
that is useful throughout the product lifecycle (ICH Q8; ICH Q9; ICH Q10).
Although the control of the process is an important and inseparable component of
the product quality assurance, it is beyond the scope of this chapter. Below, we focus
solely on the product characterization and quality consistency assessment.
1
FDA guidance documents contain recommendations that reflect FDA’s current thinking on given
issues and are intended to assist the industry in carrying out its obligations under statutes and regu-
lations. They do not create or confer any rights for or on any person and do not operate to bind FDA
or the public. They are publically available on the FDA website. http://www.fda.gov/Drugs/
GuidanceComplianceRegulatoryInformation/Guidances/default.htm.
2
ICH is an organization that brings together regulators and pharmaceutical industry representatives
from Europe, Japan, and the USA. The ICH mission is to facilitate development of safe, effective,
and high-quality drugs in the most resource-efficient manner by harmonization of guidelines and
requirements for product registration.
11.5.1 Structural Characterization, Purity, Specifications,

Stability, and Product Consistency Throughout Lifecycle
For approval of a new pharmaceutical product, the ICH M4Q(R1) guidance for
industry ICH M4Q(R1) recommends the submission of details on primary, second-
ary, and higher-order structure for “the desired product and product-related sub-
stances.” In addition to information on biological activity of the product, the
guidance recommends the submission of information on the molecular mass, details
of “posttranslational forms (e.g., glycoforms),” and a schematic amino acid sequence
showing glycosylation sites or other posttranslational modifications.
Part B of ICH Q5 guidance discusses the need for characterization of the nucleic
acid coding sequence or the transcription products, as appropriate. This guidance,
published in 1995, although it does not specifically address the effect of the sequence
on protein folding, points out that “the genetic sequence of recombinant proteins
produced in living cells can undergo mutations that could alter the properties of the
protein with potential adverse consequences to patients.” As we know today, even
an alteration in nucleotide sequence that has no effect on the amino acid sequence
may affect the rate of translation and folding of a protein.
ICH Q6B guidance focuses on product physicochemical characterization, which
generally includes determination of the composition, physical properties, and pri-
mary structure of the product. The guidance provides examples of product attributes
that can be considered for structural characterization, such as amino acid composi-
tion, terminal sequencing, peptide mapping, determination of the primary sequence,
location of disulfide bonds, and characterization of the carbohydrate content and
structure. Physicochemical characterization includes determination of the molecular
weight, isoform pattern, extinction coefficient, electrophoretic pattern, liquid chro-
matography patterns, and spectroscopic profiles. The guidance emphasizes that opti-
mal physicochemical characterization varies from product to product and that new
technologies are still being developed and should be applied when appropriate.
Demonstration of product purity presents a challenge, and the results may depend
on the method used (ICH Q6B). The relative purity of the product is often deter-
mined in terms of specific activity expressed in units of biological activity present
in a mass unit of the product. Specific activity determined in such a way is also
highly method dependent. Thus, the purity is assessed by using a combination of
analytical techniques. However, small conformational changes in the protein struc-
ture may be difficult to identify, and guidance on how to efficiently select a method(s)
that identifies such changes in an individual protein is lacking.
ICH Q6B also provides guidance on the types of impurities that may be found in
a product. Impurities may be either process related (i.e., derived from the manufac-
turing process) or product related, e.g., degradation products that may be inactive and
affect the safety profile of the product. Protein aggregates are included in the latter
category. Impurities can have a known structure(s), may be partially characterized, or
remain unidentified. When adequate quantities of impurities are present or can be
generated by stressing the protein, the guidance advises that they be characterized to
the greatest possible extent and, where feasible, their biological activity evaluated.
Product characterization sets a basis for establishing specifications, which repre-
sent a set of test methods and acceptance criteria that are used in routine product
testing and are chosen not to characterize the product but rather to confirm its
quality. In contrast to methods used for product characterization that need to be
“scientifically sound and provide results that are reliable” (ICH Q5E) (i.e., quali-
fied), methods used in determining specifications and stability should be validated
(ICH Q6B; ICH Q5C; ICH Q2 (R1)). This limits the selection of the methods that
can be considered for routine product testing because some of the biophysical meth-
ods are difficult to validate.
The structural and functional properties of a protein product can change with
time. Thus, it is essential to establish criteria for an acceptable product stability pat-
tern and to determine the product shelf life. The stability of a biological product is
evaluated following ICH Q1A and ICH Q5C guidances. It is recommended that
attributes at risk for change during storage and likely to influence quality, safety,
and/or efficacy be monitored. The manufacturer should develop specifications that
provide assurance that changes in the purity and potency of the product are detected.
ICH Q1A states that “the testing should cover, as appropriate, the physical, chemi-
cal, biological, and microbiological attributes” and that “validated stability-
indicating analytical procedures should be applied.”
The ICH Q6B guidance notes that the inherent degree of structural heterogeneity
in proteins due to the biosynthetic processes used by living organisms or resulting
from manufacture and/or storage necessitates that the pattern of heterogeneity be
defined and lot-to-lot consistency and comparability between lots used in preclini-
cal and clinical studies be demonstrated. The guidance states that “if a consistent
pattern of product heterogeneity is demonstrated, an evaluation of the activity, effi-
cacy, and safety (including immunogenicity) of individual forms may not be neces-
sary.” When process changes and degradation products result in heterogeneity
patterns which differ from those observed in the material used during preclinical
and clinical development, the guidance recommends evaluation of the significance
of such alterations. The ICH Q6B guidance also notes that under certain circum-
stances, physicochemical tests may replace a biological assay to measure the bio-
logical activity. However, such instances are limited to cases where sufficient
physicochemical information about the drug exists, including when the higher-order
structure can be thoroughly established by qualified methods, a relevant physico-
chemical correlation with biologic activity is demonstrated, and a well-established
manufacturing history exists.
Changes to the manufacturing process may have an effect on product composition
and its structural properties. A determination of comparability of the product pre- and
post-manufacturing change can be based on a combination of analytical testing and,
in some cases, nonclinical and clinical data when physicochemical or biological
assays are not considered adequate to confirm that there has been no adverse effect
on the product. ICH Q5E guidance states that “generally, quality data on the pre- and
post-change product are generated, and a comparison is performed that integrates and
evaluates all data collected, e.g., routine batch analyses, in-process control, process
validation/evaluation data, characterisation and stability, if appropriate. The compari-
son of the results to the predefined criteria should allow an objective assessment of
whether or not the pre- and post-change product are comparable.”
Since manufacturing process changes may have an impact on protein structure,
ICH Q5E advises manufacturers to “attempt to determine that higher-order struc-
ture (secondary, tertiary, and quaternary structure) is maintained.” If the appropriate
higher-order structural information cannot be obtained, the guidance indicates that
a relevant biological activity assay with appropriate precision and accuracy could
indicate that changes in conformational structure have not occurred. However, it
should be noted that the mass of protein aggregates in the product can be so small
that their presence has no meaningful impact on the functional test results. The
small mass of protein aggregates, especially in the subvisible size range, and their
heterogeneity present a significant analytical challenge.
ICH Q5E advises further that to address the full spectrum of physicochemical
properties, more than one analytical technique may be needed to evaluate the same
quality attribute, e.g., the secondary and tertiary structures and presence of impuri-
ties. In such cases, each method should be based on different physicochemical or
biological principles to maximize the chance that differences in the product caused
by a change in the manufacturing process are detected.
Proteins may be sensitive to changes in buffer composition, processing and hold-
ing conditions, and the use of organic solvents. Even slight modifications to manu-
facturing, storage, and handling may affect the stability of the product. For example,
the presence of trace amounts of metal ions might activate proteases or catalyze
chemical modifications of the product that may lead to protein aggregation.
Therefore, ICH Q5E recommends initiation of real-time/real temperature stability
studies on the product potentially affected by the change. Stability studies per-
formed under accelerated conditions, such as increased temperature, light intensity,
agitation, and freeze–thaw cycles, may be particularly helpful in identifying subtle
changes that are not detectable by routine product characterization, such as may
occur following a manufacturing change.
11.5.2 Characterization of Protein Aggregates
The inherent instability of proteins makes them susceptible to conformational

changes and aggregation during manufacture, storage, and any changes in the man-
ufacturing conditions. Protein aggregates in biologics present a highly heteroge-
neous group of protein agglomerates which differ in solubility, size, and structure,
all of which makes comprehensive characterization difficult if not impossible given
the current state of the available analytical resources. The size of protein aggregates
ranges from nanometers to micrometers. The smallest aggregates, with sizes below
100 nm, are often quantified by using SE-HPLC with UV detection. While such an
assay is often used for product release, it is not free from limitations (Carpenter
et al. 2010b). These limitations include sample dilution, which has the potential to
shift the equilibrium for soluble aggregate formation, and exclusion/adsorption of
aggregates on the SE-HPLC matrix. Also, extinction coefficients may be conforma-
tion dependent, and using the same extinction coefficient for different protein
species may lead to erroneous results. Solutions proposed include the use of orthog-
onal methods, determination of mass recovery, confirmation of method suitability
by using stressed protein samples, and development of new media. For HPLC and
other methods used for protein aggregates characterization, development of protein
standards would be a useful and important advance toward standardizing various
characterization methods.
Until recently, protein aggregates with sizes between 100 nm and 10 μm were
neither quantified nor characterized (Carpenter et al. 2009), in part due to the lack of
relevant analytical techniques. Recently, there is increasing interest in characteriza-
tion of protein aggregates in this size range. It has been postulated that aggregates of
this size may be of particular importance due to risk of an immune response stimu-
lated by their presence. Fortunately, recent advances in the analytical field, including
the development of new methods (e.g., resonant mass measurement-based tech-
niques, nanoparticle tracking analysis) and expanding the dynamic range of existing
methods (flow microscopy), may eventually allow this gap in our understanding of
the impact of subvisible particles on product quality and safety to be bridged.
Standards for the amount of particulate matter in the final container of injectable
therapeutics are recommended by the pharmacopeias (ICH Q6B). United States
Pharmacopeial Convention (USP)3 requirements for particulate matter in parenteral
products are described in general chapters “Injections” (USP <1>) and “Particulate
Matter in Injections” (USP <788>). According to the USP, therapeutic protein solu-
tions need to be “essentially free from visible particles.” Each container of the prod-
uct should be visually inspected for the presence of observable particulate and other
foreign matter (“visible particulates”). Every container showing evidence of visible
particulates should be rejected. However, this standard was set for “particulate mat-
ter” defined as “extraneous mobile undissolved particles, other than gas bubbles,
unintentionally present in the solutions.” As such, it cannot be directly applied to all
biological products because protein aggregates are often unavoidable in protein ther-
apeutics, some of which may contain protein aggregates even in the visible size range.
The meaning of the term “essentially free” has been under discussion for an
extended period of time, and attempts to implement some numerical standards have
been made (Madsen et al. 2009). Recently, USP has proposed a new general chapter
“Visible particulates in injections” (USP <790>), which includes a specification for
determining whether a product meets the criterion of being “essentially free” of par-
ticles. The chapter also states that “some products, such as those derived from pro-
teins, may contain intrinsic particles of agglomerates” and in such cases recommends
meeting the requirements of individual monographs.
3
USP is a standard setting organization that can aid in ensuring the quality and safety of drugs,
dietary supplements, and food ingredients.
Particles with sizes ≥10 μm have been quantified by using the USP chapter
<788> methods, light obscuration, and/or microscopic test. The methods and the
acceptance criteria were developed for extraneous foreign particles and should not
be applied to biologics without critical analysis. The microscopy method, which is
applicable to non-proteinaceous products, should not be used for characterizing
protein aggregates because they are fragile and translucent, may not be visible on
the membrane used for filtration, or may break down and be filtered out.
The limitations of the light obscuration method include the need for large sample
volume, which given biological product final container volumes may render the
method very costly. The details of sample preparation are also problematic for bio-
logical materials; protein aggregates may dissociate or form under the sample
degassing conditions (sonication) recommended in chapter <788>. Also, the method
cannot distinguish between protein aggregates and extraneous, non-proteinaceous
matter (e.g., silicone oil). More objective and data-driven ways to determine the
acceptance criteria for particulates in biologic products are being developed based
on process performance together with the product safety profile.
The exceptions that are not subject to the requirements of chapter <788> include
parenteral products for which the labeling specifies the use of a final filter prior to
administration, provided scientific data exist to justify the exemption, such as in-use
qualification of the filter. However, we note that the presence of protein aggregates
represents not only a potential safety problem but is also a product quality issue.
In-line filters may become clogged by protein aggregates, exposing patients to the
inconvenience and potential safety risk of having filters changed during infusion of
a large volume biologics; thus, the quantity of the aggregates should be controlled
even if the product is filtered.
Some of the above limitations were considered during the development of a new
USP general chapter “Subvisible particulate matter in therapeutic protein injec-
tions” (USP <787>), which addresses the presence of subvisible particles in bio-
therapeutics and has been published in the Pharmacopeial Forum for comments.
Also of note are two recent improvements in methodology. Due to technical
advancements, the light obscuration method can be validated to monitor particle
size down to 2 μm. In addition, new and promising flow imaging methods may help
characterize particulate matter content in the size range that earlier was mainly
assessed by light obscuration.
Quantification and characterization of protein aggregates is not an easy task. The
complexity is related to differences in protein propensity to aggregate, heterogene-
ity of aggregation, and unique aggregate-associated potential risk. As discussed
above, even with the existing gaps, many analytical methods are available for use,
although identifying the relevant test(s) for a specific product is not trivial given the
complexities of the protein therapeutic products. It is extremely encouraging that
members of the biologics community, which includes industry, academic, and gov-
ernment scientists and regulators, are working together to develop a better under-
standing of protein aggregates in protein therapeutics and to devise effective
methods to identify and mitigate the associated risks (Marszal and Fowler 2012).
11.6 Conclusion
Protein therapeutic products are characterized by a complex three-dimensional

structure. This structure both defines and determines a highly specific biologic
activity while itself being defined and dependent on the parameters and conditions
used for protein production, purification, handling, and storage. Structural changes
that may occur during the production and distribution processes will affect product
quality and may affect product safety, necessitating the need for a tight control of
these processes.
The safety and efficacy of marketed products are established in clinical trials.
When changes to the manufacturing process are being implemented, detailed struc-
tural characterization, determination of product stability, and demonstration of
product structural comparability are necessary to ensure that product quality, and
thus safety and efficacy, are comparable to that material used in the clinical trials.
Structural characterization is complex; it requires the use of numerous methods
that look at different structural features and are based on different physical princi-
ples. Understanding the advantages and limitations of different methods is essential
for obtaining meaningful structural information. New methods and improvements
to existing methods are being developed to fill the gaps resulting from the limita-
tions of current methods.
Various publications discussed in this chapter, including those promulgated or
used by regulatory agencies, provide an overview of the scientific and good manu-
facturing principles that can be used to facilitate the selection of product parameters
to be characterized. However, given the complexity and dynamic nature of proteins,
structural characterization can never be complete. It is knowledge of the manufac-
turing process design space and tight control over process parameters that, in addi-
tion to testing and characterization, ensure product quality.
References
Aguzzi A, O’Connor T (2010) Protein aggregation diseases: pathogenicity and therapeutic per-
spectives. Nat Rev Drug Discov 9:237–248
Alexander PA, He Y, Chen Y, Orban J, Bryan PN (2007) The design and characterization of two
proteins with 88 % sequence identity but different structure and function. Proc Natl Acad Sci
USA 104:11963–11968
Anderson DE, Becktel WJ, Dahlquist FW (1990) pH-induced denaturation of proteins: a single
salt bridge contributes 3–5 Kcal/Mol to the free energy of folding of T4 lysozyme. Biochemistry
29:2403–2408
Anfinsen CB (1973) Principles that govern the folding of protein chains. Science 181:223–230
Appa RS, Theill C, Hansen L, Moss J, Behrens C, Nicolaisen EM, Klausen NK, Christensen MS
(2010) Investigating clearance mechanisms for recombinant activated factor VII in a perfused
liver model. Thromb Haemost 104:243–51
Aucamp JP, Cosme AM, Lye GJ, Dalby PA (2005) High-throughput measurement of protein sta-
bility in microtiter plates. Biotechnol Bioeng 89:599–607
Baudys M, Uchio T, Mix D, Wilson D, Kim SW (1995) Physical stabilization of insulin by

glycosylation. J Pharm Sci 84:28–33
Bee JS, Randolph TW, Carpenter JF, Bishop SM, Dimitrova MN (2011) Effects of surfaces and
leachables on the stability of biopharmaceuticals. J Pharm Sci 100:4158–4170
Benoit G, Machuca E, Heidet L, Antignac C (2010) Hereditary kidney diseases: highlighting the
importance of classical Mendelian phenotypes. Ann N Y Acad Sci 1214:83–98
Bentham J, Bhattacharya S (2008) Genetic mechanisms controlling cardiovascular development.
Ann N Y Acad Sci 1123:10–19
Bolen DW, Baskakov IV (2001) The osmophobic effect: natural selection of a thermodynamic
force in protein folding. J Mol Biol 310:955–963
Braeckmans K, Buyens K, Bouquet W, Vervaet C, Joye P, De VF, Plawinski L, Doeuvre L, Angles-
Cano E, Sanders NN et al (2010) Sizing nanomatter in biological fluids by fluorescence single
particle tracking. Nano Lett 10:4435–4442
Braun A, Kwee L, Labow MA, Alsenz J (1997) Protein aggregates seem to play a key role among
the parameters influencing the antigenicity of interferon alpha (IFN-Alpha) in normal and
transgenic mice. Pharm Res 14:1472–1478
Capriotti E, Fariselli P, Casadio R (2005) I-Mutant2.0: predicting stability changes upon mutation
from the protein sequence or structure. Nucleic Acids Res 33:W306–W310
Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJ, Middaugh CR, Winter G, Fan YX,
Kirshner S, Verthelyi D, Kozlowski S et al (2009) Overlooking subvisible particles in therapeu-
tic protein products: gaps that may compromise product quality. J Pharm Sci 98:1201–1205
Carpenter J, Cherney B, Lubinecki A, Ma S, Marszal E, Mire-Sluis A, Nikolai T, Novak J, Ragheb
J, Simak J (2010a) Meeting report on protein particles and immunogenicity of therapeutic
proteins: filling in the gaps in risk evaluation and mitigation. Biologicals 38:602–611
Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJ, Middaugh CR, Winter G (2010b) Potential
J Pharm Sci 99:2200–2208
Castro PM, Ison AP, Hayter PM, Bull AT (1995) The macroheterogeneity of recombinant human
interferon-gamma produced by Chinese-hamster ovary cells is affected by the protein and lipid
content of the culture medium. Biotechnol Appl Biochem 21:87–100
Cheng J, Randall A, Baldi P (2006) Prediction of protein stability changes for single-site mutations
using support vector machines. Proteins 62:1125–1132
Cromwell ME, Hilario E, Jacobson F (2006) Protein aggregation and bioprocessing. AAPS J
8:E572–E579
den Engelsman J, Garidel P, Smulders R, Koll H, Smith B, Bassarab S, Seidl A, Hainzl O, Jiskoot
W (2011) Strategies for the assessment of protein aggregates in pharmaceutical biotech product
development. Pharm Res 28:920–933
Dresser DW (1962) Specific inhibition of antibody production. II. Paralysis induced in adult mice
by small quantities of protein antigen. Immunology 5:378–388
Edgell MH, Sims DA, Pielak GJ, Yi F (2003) High-precision, high-throughput stability determina-
tions facilitated by robotics and a semiautomated titrating fluorometer. Biochemistry 42:
7587–7593
Elia AE, Albanese A (2010) Emerging parkinsonian phenotypes. Rev Neurol 166:834–840
England JL, Haran G (2011) Role of solvation effects in protein denaturation: from thermodynam-
ics to single molecules and back. Annu Rev Phys Chem 62:257–277
Ericsson UB, Hallberg BM, Detitta GT, Dekker N, Nordlund P (2006) Thermofluor-based high-
throughput stability optimization of proteins for structural studies. Anal Biochem
357:289–298
Fares FA, Suganuma N, Nishimori K, LaPolt PS, Hsueh AJ, Boime I (1992) Design of a long-
acting follitropin agonist by fusing the C-terminal sequence of the chorionic gonadotropin beta
subunit to the follitropin beta subunit. Proc Natl Acad Sci USA 89:4304–8
Filipe V, Hawe A, Jiskoot W (2010) Critical evaluation of nanoparticle tracking analysis (NTA) by
nanosight for the measurement of nanoparticles and protein aggregates. Pharm Res 27:796–810
Filipe V, Poole R, Kutscher M, Forier K, Braeckmans K, Jiskoot W (2011) Fluorescence single

particle tracking for the characterization of submicron protein aggregates in biological fluids
and complex formulations. Pharm Res 28:1112–1120
ceutics and biopharmaceutics. Eur J Pharm Biopharm 58:369–383
Garrod A (1908) The Croonian lectures on inborn errors of metabolism, lecture II: alkaptonuria.
Lancet 2:73–79
Gaudet M, Remtulla N, Jackson SE, Main ER, Bracewell DG, Aeppli G, Dalby PA (2010) Protein
denaturation and protein: drugs interactions from intrinsic protein fluorescence measurements
at the nanolitre scale. Protein Sci 19:1544–1554
General Chapter <1 > Injections (2012) USP 35-NF 30, pp 33–37
General Chapter <787 > Subvisible particulate matter in therapeutic protein injections [New]
(USP36-NF31 1S) (2012) Pharmacopeial Forum 38(3)
General Chapter <788 > Particulate Matter in Injections (2012) USP 35-NF 30, pp 339–342
General Chapter <790 > Visible particulates in injections [New] (USP36-NF31 1S) (2012)
Pharmacopeial Forum 38(2)
Ghaemmaghami S, Oas TG (2001) Quantitative protein stability measurement in vivo. Nat Struct
Biol 8:879–882
Gilis D, Rooman M (2000) PoPMuSiC, an algorithm for predicting protein mutant stability
changes: application to prion proteins. Protein Eng 13:849–856
Griko YV, Privalov PL, Venyaminov SY, Kutyshenko VP (1988) Thermodynamic study of the
apomyoglobin structure. J Mol Biol 202:127–138
Grothe HL, Little MR, Cho AS, Huang AJ, Yuan C (2009) Denaturation and solvent effect on the
conformation and fibril formation of TGFBIp. Mol Vis 15:2617–2626
Haass C, Selkoe DJ (2007) Soluble protein oligomers in neurodegeneration: lessons from the
Alzheimer’s amyloid beta-peptide. Nat Rev Mol Cell Biol 8:101–112
Harrison M, Maresso K, Broeckel U (2008) Genetic determinants of hypertension: an update. Curr
Hypertens Rep 10:488–495
Hartl FU, Hayer-Hartl M (2009) Converging concepts of protein folding in vitro and in vivo. Nat
Struct Mol Biol 16:574–581
Hossler P, Khattak SF, Li ZJ (2009) Optimal and consistent protein glycosylation in mammalian
cell culture. Glycobiology 19:936–49
Houde D, Peng Y, Berkowitz SA, Engen JR (2010) Post-translational modifications differentially
affect IgG1 conformation and receptor binding. Mol Cell Proteomics 9:1716–1728
teral solutions using micro-flow imaging. J Pharm Sci 98:3058–3071
Huang M, Horwitz TS, Zweiben C, Singh SK (2011) Impact of extractables/leachables from filters
on stability of protein formulations. J Pharm Sci 100:4617–4630
ICH Harmonized Tripartite Guideline M4Q(R1) (2002) The common technical document for the
registration of pharmaceuticals for human use: quality
ICH Harmonized Tripartite Guideline Q10 (2008) Pharmaceutical quality system
ICH Harmonized Tripartite Guideline Q1A(R2) (2003) Stability testing of new drug substances
and products
ICH Harmonized Tripartite Guideline Q2(R1) (2005) Validation of analytical procedures: text and
methodology
ICH Harmonized Tripartite Guideline Q5B (1995) Quality of biotechnological products: analysis
of the expression construct in cells used for production of r-DNA derived protein products
ICH Harmonized Tripartite Guideline Q5C (1995) Quality of biotechnological products: stability
testing of biotechnological/biological products
ICH Harmonized Tripartite Guideline Q5E (2004) Comparability of biotechnological/biological
products subject to changes in their manufacturing process
ICH Harmonized Tripartite Guideline Q6B (1999) Specifications: test procedures and acceptance
criteria for biotechnological/biological products
ICH Harmonized Tripartite Guideline Q8(R2) (2009) Pharmaceutical development
ICH Harmonized Tripartite Guideline Q9 (2005) Quality risk management

Ikemura T (1985) Codon usage and tRNA content in unicellular and multicellular organisms. Mol
Biol Evol 2:13–34
Jenkins N, Castro P, Menon S, Ison A, Bull A (1994) Effect of lipid supplements on the production
and glycosylation of recombinant interferon-gamma expressed in CHO cells. Cytotechnology
15:209–15
Jorgensen L, Hostrup S, Moeller EH, Grohganz H (2009) Recent trends in stabilising peptides and
proteins in pharmaceutical formulation—considerations in the choice of excipients. Expert
Opin Drug Deliv 6:1219–1230
Kiese S, Papppenberger A, Friess W, Mahler HC (2008) Shaken, not stirred: mechanical stress
testing of an IgG1 antibody. J Pharm Sci 97:4347–4366
Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, Ambudkar SV, Gottesman MM
(2007) A "silent" polymorphism in the MDR1 gene changes substrate specificity. Science
315:525–528
Klein WL, Krafft GA, Finch CE (2001) Targeting small abeta oligomers: the solution to an
Alzheimer’s disease conundrum? Trends Neurosci 24:219–224
Kochanowski N, Blanchard F, Cacan R, Chirat F, Guedon E, Marc A, Goergen JL (2008) Influence
of intracellular nucleotide and nucleotide sugar contents on recombinant interferon-gamma
glycosylation during batch and fed-batch cultures of CHO cells. Biotechnol Bioeng 100:
721–33
Kramer G, Boehringer D, Ban N, Bukau B (2009) The ribosome as a platform for co-translational
processing, folding and targeting of newly synthesized proteins. Nat Struct Mol Biol
16:589–597
LaPolt PS, Nishimori K, Fares FA, Perlas E, Boime I, Hsueh AJ (1992) Enhanced stimulation of
follicle maturation and ovulatory potential by long acting follicle-stimulating hormone ago-
nists with extended carboxyl-terminal peptides. Endocrinology 131:2514–2520
Lavinder JJ, Hari SB, Sullivan BJ, Magliery TJ (2009) High-throughput thermal scanning: a gen-
eral, rapid dye-binding thermal shift screen for protein engineering. J Am Chem Soc
131:3794–3795
Lewis JD, Nail SL (1997) The influence of ion exchange chromatography conditions on aggrega-
tion of equine IgG. Process Biochem 32:279–283
Liu H, Bulseco GG, Sun J (2006) Effect of posttranslational modifications on the thermal stability
of a recombinant monoclonal antibody. Immunol Lett 106:144–153
Luo Q, Joubert MK, Stevenson R, Ketchem RR, Narhi LO, Wypych J (2011) Chemical modifica-
tions in therapeutic protein aggregates generated under different stress conditions. J Biol Chem
286:25134–25144
Madsen RE, Cherris RT, Habushnig JG, Hunt DG (2009) Visible particulates in injections-a his-
tory and a proposal to revise USP general chapter injections <1> Pharmacopeial Forum
35:1383–1387
Maity H, O'Dell C, Srivastava A, Goldstein J (2009) Effects of arginine on photostability and
thermal stability of IgG1 monoclonal antibodies. Curr Pharm Biotechnol 10:761–766
Marszal E, Fowler E (2012) Overview: workshop on predictive science of the immunogenicity
aspects of particles in biopharmaceutical products. J Pharm Sci 101:3555–9
Masso M, Vaisman II (2008) Accurate prediction of stability changes in protein mutants by com-
bining machine learning with structure based computational mutagenesis. Bioinformatics
24:2002–2009
McCarthy EF (2011) Genetic diseases of bones and joints. Semin Diagn Pathol 28:26–36
Meager A, Dolman C, Dilger P, Bird C, Giovannoni G, Schellekens H, Thorpe R, Wadhwa M
(2011) An assessment of biological potency and molecular characteristics of different innova-
tor and noninnovator interferon-beta products. J Interferon Cytokine Res 31:383–392
Meeker AK, Garcia-Moreno B, Shortle D (1996) Contributions of the ionizable amino acids to the
stability of staphylococcal nuclease. Biochemistry 35:6443–6449
Mitra N, Sinha S, Ramya TN, Surolia A (2006) N-linked oligosaccharides as outfitters for glyco-
protein folding, form and function. Trends Biochem Sci 31:156–163
Moore WV, Leppert P (1980) Role of aggregated human growth hormone (hGH) in development
of antibodies to hGH. J Clin Endocrinol Metab 51:691–697
Neuhofer W, Muller E, Grunbein R, Thurau K, Beck FX (1999) Influence of NaCl, urea, potassium
and pH on HSP72 expression in MDCK cells. Pflugers Arch 439:195–200
Obici L, Perfetti V, Palladini G, Moratti R, Merlini G (2005) Clinical aspects of systemic amyloid
diseases. Biochim Biophys Acta 1753:11–22
Parthiban V, Gromiha MM, Schomburg D (2006) CUPSAT: prediction of protein stability upon
point mutations. Nucleic Acids Res 34:W239–W242
Pasinelli P, Brown RH (2006) Molecular biology of amyotrophic lateral sclerosis: insights from
genetics. Nat Rev Neurosci 7:710–723
Patel J, Kothari R, Tunga R, Ritter NM, Tunga BS (2011) Stability considerations for biopharma-
ceuticals, Part 1. Overview of protein and peptide degradation pathways. BioProcess Int
9:20–31
Percy MJ, Rumi E (2009) Genetic origins and clinical phenotype of familial and acquired erythro-
cytosis and thrombocytosis. Am J Hematol 84:46–54
Pfeil W, Privalov PL (1976) Thermodynamic investigations of proteins. III. Thermodynamic
description of lysozyme. Biophys Chem 4:41–50
Rademacher TW, Parekh RB, Dwek RA (1988) Glycobiology. Annu Rev Biochem 57:785–838
Rathore N, Rajan RS (2008) Current perspectives on stability of protein drug products during
formulation, fill and finish operations. Biotechnol Prog 24:504–514
Rosgen J (2007) Molecular basis of osmolyte effects on protein and metabolites. Methods Enzymol
428:459–486
Scriver CR (2007) The PAH gene, phenylketonuria, and a paradigm shift. Hum Mutat
28:831–845
Senger RS, Karim MN (2003) Effect of shear stress on intrinsic CHO culture state and glycosyl-
ation of recombinant tissue-type plasminogen activator protein. Biotechnol Prog 19:1199–209
Sharp PM, Cowe E, Higgins DG, Shields DC, Wolfe KH, Wright F (1988) Codon usage patterns
in Escherichia coli, Bacillus subtilis, Saccharomyces cerevisiae, Schizosaccharomyces pombe,
Drosophila melanogaster and Homo sapiens; a review of the considerable within-species diver-
sity. Nucleic Acids Res 16:8207–8211
Shire SJ (2009) Formulation and manufacturability of biologics. Curr Opin Biotechnol
20:708–714
Shukla AA, Gupta P, Han X (2007) Protein aggregation kinetics during Protein A chromatography:
case study for an Fc fusion protein. J Chromatogr A 1171:22–28
Sola RJ, Rodriguez-Martinez JA, Griebenow K (2007) Modulation of protein biophysical proper-
ties by chemical glycosylation: biochemical insights and biomedical implications. Cell Mol
Life Sci 64:2133–2152
Stepanenko OV, Verkhusha VV, Shavlovsky MM, Kuznetsova IM, Uversky VN, Turoverov KK
(2008) Understanding the role of Arg96 in structure and stability of green fluorescent protein.
Proteins 73:539–551
Stites WE, Byrne MP, Aviv J, Kaplan M, Curtis PM (1995) Instrumentation for automated deter-
mination of protein stability. Anal Biochem 227:112–122
Struble EB, Ladner JE, Brabazon DM, Marino JP (2008) New crystal structures of ColE1 Rom and
variants resulting from mutation of a surface exposed residue: implications for RNA-
recognition. Proteins 72:761–768
Tang X, Pikal MJ (2004) Design of freeze-drying processes for pharmaceuticals: practical advice.
Pharm Res 21:191–200
Thirumangalathu R, Krishnan S, Ricci MS, Brems DN, Randolph TW, Carpenter JF (2009)
Silicone oil- and agitation-induced aggregation of a monoclonal antibody in aqueous solution.
J Pharm Sci 98:3167–3181
Thorens B, Vassalli P (1986) Chloroquine and ammonium chloride prevent terminal glycosylation
of immunoglobulins in plasma cells without affecting secretion. Nature 321:618–620
Uchida E, Morimoto K, Kawasaki N, Izaki Y, Abdu Said A, Hayakawa T (1997) Effect of active
oxygen radicals on protein and carbohydrate moieties of recombinant human erythropoietin.
Free Radic Res 27:311–323
Vazquez-Rey M, Lang DA (2011) Aggregates in monoclonal antibody manufacturing processes.
Biotechnol Bioeng 108:1494–1508
Wright A, Morrison SL (1998) Effect of C2-associated carbohydrate structure on Ig effector func-
tion: studies with chimeric mouse-human IgG1 antibodies in glycosylation mutants of Chinese
hamster ovary cells. J Immunol 160:3393–3402
Xu J, Reumers J, Couceiro JR, De SF, Gallardo R, Rudyak S, Cornelis A, Rozenski J, Zwolinska
A, Marine JC et al (2011) Gain of function of mutant P53 by coaggregation with multiple
tumor suppressors. Nat Chem Biol 7:285–295
Yin S, Ding F, Dokholyan NV (2007) Modeling backbone flexibility improves protein stability
estimation. Structure 15:1567–1576
Zimmerman TP (2006) Yield improvement for manufacture of α1-proteinase inhibitor. Vox Sang
91:309–315
Index
A application, 155
Absorbance spectroscopy, 36–38 Fc-fusion protein, 158
ADA. See Antidrug antibodies (ADA) SEC, 156
ADCs. See Antibody drug Antibody drug conjugates (ADCs)
conjugates (ADCs) chemical and biophysical techniques,
AFM. See Atomic force microscope (AFM) 195–196
Aggregation DSC analysis, 196–197
CDR loops, 129 in vivo assays and DAR, 196
conformational and colloidal stability, 135 SEC, 196, 197
in silico protein, 131 small molecule cytotoxic drugs, 197
Lumry–Eyring framework, 134–135 T-DM1, 195
soluble protein, 206–219 Antidrug antibodies (ADA), 262
thermal stability, 135–136 Atomic force microscope (AFM)
Aggregation propensity cantilever, 221
circular dichroism, 164–165 measurements, 221
extrinsic fluorescence, 164, 165 morphology, submicron protein
high molecular mass species, 164 particulates, 223, 225
linear regression analysis, 167 AUC. See Analytical ultracentrifugation
Nile red, 163 (AUC)
stressed samples, 166
thermal incubation, 164
Analytical ultracentrifugation (AUC) B
advantages and limitations, 89 Biopharmaceuticals development, protein
orthogonal technique to SEC HOS. See Higher-order structure
concordance plot, 198, 199 (HOS)
HMWS and LMWS, 198 Biophysical analysis, protein
mAbs, 198 pharmaceuticals
SE-AUC, 70–71 AUC, 198–199
sedimentation velocity and sedimentation characterization (see Biophysical
equilibrium, 88–89 characterization)
SV-AUC, 66–70 description, 174–175
use, 88–89 FFF, 199–200
Analytical ultracentrifugation- orthogonal techniques and SEC, 197
gravity-sweep, 148 protein stability development, 173–174
Analytical ultracentrifugation-sedimentation screening assessments (see Screening
velocity (AUC-SV) assessments, biophysical analysis)
for the Life Sciences 4, DOI 10.1007/978-1-4614-4316-2,
284 Index
Biophysical characterization Biophysical techniques

ADCs, 195–197 analytical methods, 205
chemical changes, physical stability, and chemical interrelationships, 236
189–190 impurities/degradants, 205–206
clinical in-use studies protein storage stability, 205
CSP, 194 protein subvisible and visible particles
physical stability, IgG4 mAbs, 195 (see Subvisible and visible
trastuzumab aggregation, 195 protein particles)
HDX-MS, 188–189 soluble protein aggregates (see Soluble
high-concentration formulation protein aggregation)
development submicron protein particulates (see
AUC sedimentation equilibrium, 191 Submicron protein particulates)
biophysical techniques, 193–194 Biophysics
mAbs therapeutics, 190 AUC, 66
Mooney equation, 192 DLS, 63
preparative AUC experiment, 192, 193 therapeutic proteins (see High-throughput
rigorous analysis, SLS data, 192 biophysical approaches, therapeutic
TFF, 190 proteins)
thin plastic gaskets, 191–192 Biotherapeutics
viscosity, mAbs1–3, 190–191 lyophilized, 267
higher order structure determinations manufacture products, 263
hydrodynamic measurements, 183–184
spectrophotometric techniques,
184–188 C
intense Calorimetry, 59–60
anti IgE mAb, 179, 180 Capillary electrophoresis SDS (CE-SDS),
AUC, 179 209–210
complex formation, IgE:anti CD. See Circular dichroism (CD)
IgE, 179, 183 Chemistry, manufacturing and controls
differential sedimentation coefficient (CMC), 129–130
distribution, 180–181, 184 Circular dichroism (CD)
mAbs, 179 biological molecules, 101
sedimentation velocity and equilibrium description, 100
measurements, 179, 181, 182 HOS, 49–50
serum vs. PBS, 182 near-UV CD spectrum, 184–185
theoretical interaction, IgE with protein structure, 104
anti IgE mAb, 179, 180 SRCD, 100
Biophysical methods UV CD analysis, 102–103
CD (see Circular dichroism (CD)) CMC. See Chemistry, manufacturing and
description, 99–100 controls (CMC)
difficult to validate, 272 Comparability
DSC, 107–109 antibody, 152
FTIR, 104–107 AUC-SV profiles, 151
light scattering, 113 biophysical characterization, 147–148
LO and MFI, 118–120 biotherapeutic, 154
LS and UV detectors, 114–115 CD spectra, 148
LS-RI detectors, 115 drug substance, 153
LS-UV-RI detectors, 115–116 DSC thermogram, 152–153
molecular weight fractional error, 116–117 tryptophan emission fluorescenc, 149–150
molecular weight measurement, 114 Compounded sterile preparation (CSP), 194
signal and noise, 113–114 Coulter counter, 226
SV-AUC, 109–112 CSP. See Compounded sterile preparation
therapeutic biologics, 269 (CSP)
Index 285
D protein–protein interactions, 65–66

DAR. See Drug-to-antibody ratio (DAR) size distributions, biopharmaceutical
Denaturing SEC (d-SEC), 210 proteins, 89–90
Design of experiment (DOE), 20–21 Stokes–Einstein relation, 65
Developability, drug candidates. See Drug
candidates, developability
Differential scanning calorimetry (DSC) E
analysis, 196–197 EDX/EDS. See Energy-dispersive X-ray
autosampling instruments, 15–16 spectrometer (EDX/EDS)
description, 59–60, 107 Electrical sensing zone (ELS) analysis, 91
immunoglobulin, 16 Electron microscopy (EM), 91–92
measurements, 175 Electrospray ionization (ESI), 217, 218
purity analysis, 108 Electrospray mass spectrometer with
qualification, 108 differential mobility analyzers
thermal shift assay, 60 (ES-DMA), 208
thermodynamics, reversible reaction, 60 ELS analysis. See Electrical sensing zone
validation, 109 (ELS) analysis
Differential scanning fluorometry (DSF) EM. See Electron microscopy (EM)
formulation screening, 16–17 Empirical phase diagram (EPD)
ProteoStat™ assay, 17 aggregation, 25
unfolding temperature, protein, 16 aldolase, BSA, chymotrypsin and
DLS. See Dynamic light scattering (DLS) lysozyme, 22, 23
DNA and protein sequence, 263–264 chemical degradation, 25
DOE. See Design of experiment (DOE) freeze/thaw and shear, 24
Drug candidates, developability assessment high-throughput characterization and
aggregation, 134–136 preformulation development, 23–24
biophysical tools, 132–134 HTS, 21
chemical stability, 136 preparation, 21–22
CMC, 129–130 protein concentration, 24
description, 127–128 structural comparisons, 25–26
engineering desired properties, 128–129 temperature and pH, 22
formulation stability, 140–141 UV circular dichroism (CD), 22
high-concentration liquid formulation, Energy-dispersive X-ray spectrometer (EDX/
130–132 EDS), 230
mAbA and 3M mutant, comparison, EPD. See Empirical phase diagram (EPD)
138–140 ES-DMA. See Electrospray mass spectrometer
novel targets, 128 with differential mobility analyzers
process stability, 141–142 (ES-DMA)
protein instabilities, 132 ESI. See Electrospray ionization (ESI)
solution properties, 136–138 Extrinsic fluorescence
Drug-to-antibody ratio (DAR), 196 1-anilino-phthalene-8-sulfonic acid
DSC. See Differential scanning calorimetry (1,8-ANS), 42
(DSC) excited-state reactions, 42–43
d-SEC. See Denaturing SEC (d-SEC) solvent relaxation, 43
DSF. See Differential scanning fluorometry steady-state fluorescence spectroscopy,
(DSF) 43–44
Dynamic light scattering (DLS)
autocorrelation function, 18, 64–65
biopharmaceutical products, 19–20 F
Brownian motion, 63–64 FACS. See Fluorescence-activated
diffusion coefficient, 65 cell sorting (FACS)
drug substance, 161 Fc-fusion protein
nanoparticles, 18–19 aggregates, 156–157
286 Index
Fc-fusion protein (cont.) routine product testing and structural

AUC, 155 heterogeneity, 272
AUC-SV results, 157–158 spectrum, physicochemical properties, 273
biotherapeutic, 154 Formulation screening
gravitational sweep, 159 chimeric and human IgG, 176
SEC, 154–155 DLS measurements, 178
size and molar mass, 159 DOE approach and mAbs, 178
FCS. See Fluorescence correlation protein–protein interactions, 178–179
spectroscopy (FCS) UV–Vis spectroscopy, 176
FDA. See Food and drug administration (FDA) Fourier transform infrared (FTIR)
FFF. See Field-flow fractionation (FFF) biopharmaceutical manufacturing
Field-flow fractionation (FFF) processes, 107
AF4, 199–200 proteins, 106
ES-DMA and TDA, 208 protein therapeutics, 104
hydrodynamic radius, 208 QC compare function, 105–106
molecular sizes, 199 qualification, 105
SEC, AF4 and AUC analysis, 200 thermo electron OMNIC software, 105
Flow microscopy, 86, 92–93, 231 Fragmentation, 132, 136
Fluorescence-activated cell sorting (FACS) FRET. See Fluorescence resonance energy
flow cytometry, 228 transfer (FRET)
hydrodynamic focusing techniques, FTIR. See Fourier transform infrared (FTIR)
228–229 FTIR microscopy
light scattering and/fluorescence advantages, 229
emission, 229 analytical tool, microparticle analysis, 249
Fluorescence correlation chemical mapping and imaging, 229–230
spectroscopy (FCS), 41 database and inorganic materials, 249
Fluorescence microscopy, 229 experienced vibrational spectroscopist,
Fluorescence resonance energy transfer 249–250
(FRET), 41–42 and Raman, 246, 248, 250
Fluorescence spectroscopy FTIR spectroscopy
aromatic amino acids, 13–14 advantages, 44
conformational dynamics amide I and amide II vibrational modes,
FCS, 41 46–47
fluorophores, 39–40 amide I IR spectrum, lysozyme, 47, 48
folding and unfolding proteins, 39 interference signal, 45
steady-state measurements, 39 Michelson interferometer, 44, 45
time-resolved measurements, 40 use, 48
extrinsic fluorescence (see Extrinsic VCD, 51, 52
fluorescence) vibrational modes, CO2 molecule, 45, 46
FRET, 41–42
hydrophobic dyes, 14
intrinsic protein, 38–39 G
quenching, 41 Glass gel-like particulates
SYPRO Orange, 14 electron image, gold coated filter, 255, 256
Food and drug administration (FDA) frozen placebo vials, 253
characterization, protein neutral pH and storage, placebo vials, 256
aggregates, 273–275 optical and FTIR-microscopic analyses,
functional and structural properties, 272 254–255
ICH, 270 polysorbate-20 molecules, 256
manufacturing process, 272–273 protein product and placebo vials, 256
nucleic acid coding sequence/transcription quantitative comparative analysis, 255
products, characterization, 271 root cause analysis, 256
proteins, 273 SEM/EDS analysis, 255
purity and impurities, 271–272 Glycosylation, 264
Index 287
H HOS. See Higher-order structure (HOS)

HDX-MS. See Hydrogen/deuterium exchange HPLC. See High-performance liquid
mass spectrometry (HDX-MS) chromatography (HPLC)
High concentration liquid formulation, HPSEC. See High-performance size-exclusion
130–132 chromatography (HPSEC)
Higher-order structure (HOS) HTS. See High-throughput screening (HTS)
absorption spectroscopy, 36–38 HX. See Hydrogen–deuterium isotope
AUC, 66–71 exchange (HX)
CD spectroscopy, 49–50 Hydrogen–deuterium (H/D) exchange, 211
DSC, 59–60 Hydrogen/deuterium exchange mass
fluorescence (see Fluorescence spectrometry (HDX-MS)
spectroscopy) bevacizumab, 189
FTIR spectroscopy, 44–48 description, 188
HX, NMR and isotope exchange, 58–59 IgG1 and peptide, 189
ITC, 59 Hydrogen–deuterium isotope exchange (HX),
light scattering, 60–66 58–59
proteins, 33–35
Raman spectroscopy, 53–57
rheology, 72–74 I
VOA, 50–53 IgG1. See Immunoglobulin G1 (IgG1)
X-ray crystallography, 57–58 Immunoglobulin G1 (IgG1)
High-molecular weight species (HMWS), 198 anti IgE, 179
High-performance liquid chromatography DSC analysis, 196
(HPLC), 274 humanized, 190
High-performance size-exclusion mAb, 176, 189
chromatography (HPSEC) molecules, 193–194
description, 84, 85 monomer, 192
field-flow fractionation (AF4), 88 β-sheet character, 184
glycosylation, 84 IMS. See Ion mobility spectrometry (IMS)
limitations, 84 In-situ Raman microscopy
MALS and DLS, 90 advantages and FTIR measurements, 250
mass spectrometry (MS), 93 database, organic and polymeric
self-association, 84 materials, 250
High-throughput biophysical approaches, inorganic crystal material and visible laser
therapeutic proteins beam, 250
description, 7 microparticles, 248
DOE and data analysis, 20–21 rich electron density, molecules, 250
downstream and formulation Intrinsic fluorescence, 171
development, 8–9 Ion mobility spectrometry (IMS), 217
DSC and DSF, 15–17 Isothermal titration calorimetry
early stages, product development, 7–8 (ITC), 15, 59, 171
EPD (see Empirical phase diagram (EPD)) ITC. See Isothermal titration calorimetry (ITC)
fluorescence spectroscopy, 13–14
implementation, advantages and
challenges, 26–28 L
ITC, 15 LC. See Liquid chromatography (LC)
LC, 10–11 Light absorption spectroscopy, 11–12
light absorption spectroscopy, 11–12 Light microscopy and visual assessments
light scattering, 18–20 advantages, staining protein
SPR, 9–10 aggregates, 227
vibrational spectroscopy, 13 certified analysts, 228
High-throughput screening (HTS), 21, 176 protein-containing solutions, 228
HMWS. See High-molecular weight species Light obscuration (LO)
(HMWS) advantages and disadvantages, 227
288 Index
Light obscuration (LO) (cont.) filtration and glass vials, 248

gold standard, measurement, 227 flow chart, 247
micro-flow digital imaging, 120 primary container, 247–248
particle size method, 119 Microspectroscopy
polystyrene, 120 FTIR and Raman, 246
protein samples, 118 vibrational, 249–250
SbVP, 118 Monoclonal antibodies (mAbs)
Light scattering AUC, 198
aggregation, 18 binding, affinity, 132
description, 60 biophysical properties, 178
DLS, 18–20, 63–66 chimeric and human IgG1, 176
SANS and MALS, 209 far-and near-UV CD spectra, 185
SLS, 18, 20, 61–63 formulations, 178
Stokes–Einstein equation and DLS, 208 free-circulating IgE, 179
UV optical density spectroscopy IgG1 (see Immunoglobulin G1 (IgG1))
and SLS, 209 in silico method, 175
Liquid chromatography (LC), 10–11 lyoprotectant, 187
LMWS. See Low-molecular weight and 3M mutant, comparison, 138–140
species (LMWS) novel targets, 128
LO. See Light obscuration (LO) orthogonal methods, 194
Low-molecular weight species (LMWS), 198 SC, 176
Lumry–Eyring framework, 134–135 SEC analysis, 191
Lyophilized biotherapeutics, 267 therapeutics, 190
Mooney equation, 192
MS. See Mass spectrometry (MS)
M Multiangle light scattering (MALS), 89, 90, 209
mAbs. See Monoclonal antibodies (mAbs)
MALDI-MS. See Matrix-assisted laser
desorption/ionization mass N
spectrometry (MALDI-MS) Nanoparticle tracking analysis (NTA), 90, 220
MALS. See Multiangle light scattering NMR spectroscopy. See Nuclear magnetic
(MALS) resonance (NMR) spectroscopy
Mass spectrometry (MS) Nonconformance (NC) investigation
bottom-up approach and peptide map, 210 biopharmaceutical products, 259
chromatographic separation and H/D description, 245
exchange, 211 extraneous particulate, 246
ESI-MS and MALDI-MS, 210 glass gel-like particulates, 253–257
soluble protein aggregates intrinsic particles, 245–246
amyloid fibril, 216–217 microparticles, 245
ESI, glucocebrosidase, 217, 218 particle analysis methods, 247–251
H/D exchange, 218–220 PFSs, 246
IMS and temperature increases, 217 protein particulates
TWIMS, 217–218 silicone oil, 257–258
top-down approach, 210–211 tungsten, 251–253
Matrix-assisted laser desorption/ionization and root cause analyses, 258
mass spectrometry (MALDI-MS), USP, 245
210 NTA. See Nanoparticle tracking analysis (NTA)
MFI. See Micro-flow imaging (MFI) Nuclear magnetic resonance (NMR)
Micro-flow imaging (MFI) spectroscopy, 58
LO, 99
performance, 120
Microparticle analysis O
chromatographic and spectroscopic Optical microscopy
techniques, 248 Axioskop 2 MAT polarized microscope, 249
Index 289
Carl Zeiss Stemi 2000C Protein higher-order structure and stability

stereomicroscope, 249 and DNA sequence, 263–264
description, 248–249 glycosylation, 264
polarized microscope, 249 lyophilized biotherapeutics, 267
purification process, 266
recombinant protein production
P process, 265
Particle formation storage and transportation, 266–267
IgG4 antibody, 160 Protein particle
mechanism, 163 silicone oil, 257–258
real-time monitoring, 162 tungsten, 251–253
PFS. See Prefilled syringes (PFS) Protein pharmaceuticals. See Biophysical
Physical stability analysis, protein pharmaceuticals
far-and Near-UV CD, 189 Proteins
methionine oxidation, 189 soluble aggregates, 206–219
met oxidation, 190 submicron particles, 219–225
SPR, 189–190 subvisible and visible particles, 225–235
Prefilled syringes (PFS) Protein size distribution analysis. See Size
manufacturing, 246 distribution, biopharmaceutical
protein particles, 251 proteins
silicone oil, 258 Protein stability. See Protein higher-order
tungsten residues, 253 structure and stability
Product quality guidance, 271, 272 Protein structure. See also Higher-order
Protein aggregation structure (HOS)
characterization aggregation and chemical modification, 35
heterogeneous group, 273 “backbone” interactions, 34
HPLC and extinction coefficients, 274 domains, 34–35
in-line filters, 275 higher-order structure and stability
light obscuration and microscopy (see Protein higher-order structure
method, 275 and stability)
size, 273–274 hydrophobic effect, 34
therapeutic protein injections manufacturing processes, 35
and quantification, 275 polypeptides subunits, 35
USP, 274–275 primary and secondary structure, 1–2
description, 83 thermodynamic characteristics, 33
detection content, 268 Protein therapeutics
diseases association and formation, 262 ADA, 262
DLS, 90 attributes and requirements, 3
gas-phase electrophoretic mobility, 93–84 biophysical techniques, 4
and higher-order structure (see Protein commercialization, 258
higher-order structure and stability) complex structure, 261–262
HPSEC, 84, 93 complex three-dimensional structure, 276
impurities, 271–272 description, 4–5
in-line filters, 267 development and manufacturing, 245
product quality and safety, 262–263 extrinsic factors, 265–267
review and evaluate analytical formulation development, 3–4
methods, 269 frozen placebo vials, 253
Protein conformation, 2, 4 growth factors and cytokines, 2
Protein fragments industry standards and regulatory
AF4 separation, 88 considerations, 270–275
AUC, 88–89 inflammation, 1
description, 83 intrinsic factors, 263–264
HPSEC, 84 life cycle development, 3
mass spectrometry (MS), 93 primary and secondary structure, 1–2
290 Index
Protein therapeutics (cont.) intrinsic viscosity measurements, 73–74

primary containers, 251 liquids and semisolids, 72
product quality assurance, 267–270 particle shape and size, 73
safety and efficacy, marketed products, 276 viscosity functions, 72–73
safety and efficacy, product, 2–3 rhGAA. See Recombinant human acid
Protein therapeutics manufacturing alpha-glucosidase (rhGAA)
biosynthetics, 263 ROA. See Raman optical activity (ROA)
characterization, 268–270
and clinical application, 264
formation, protein aggregates, 262 S
Protein therapeutics quality control SANS. See Small angle neutron scattering
analytical techniques and aggregate (SANS)
content, 268 SAP. See Spatial aggregation
biological materials, 267 propensity (SAP)
chromatography techniques and CD, 268 SbVP. See Subvisible particle (SbVP)
DSF and biophysical methods, 269 Scanning electron microscope/energy
fluorescence single particle tracking, 268 dispersive spectroscopy (SEM/
H/D exchange mass spectrometry, 268 EDS)
inter-operator precision, 270 cameo mapping function, 251
molecule drugs and product chemical microanalysis technique, 250
characterization, 267 electron vacancy, 250
search space, 269 identification, microparticles, 250–251
separation methods and X-ray Scanning electron microscopy (SEM), 221
crystallography, 268 Screening
tertiary and quaternary structure, 268 DSC, 15–16
DSF, 16–17
HTS, 21
Q Screening assessments, biophysical analysis
QCM. See Quartz crystal microbalance (QCM) formulation, 176–179
Quartz crystal microbalance (QCM) molecule
aqueous measurements, 222 assays and viscosity measurements, 176
chemical modification, electrode surface, 222 bar plots kD and mAb solution
mechanical oscillations and commercial viscosity, 176, 177
systems, 222 description, 175
submicron protein particulates, 221 DLS and HTS, 176
in silico method, 175–176
SAP and amino acid residues, 175
R SEC and DSC, 175
Raman optical activity (ROA), 52–53 SE-AUC. See Sedimentation equilibrium
Raman spectroscopy analytical ultracentrifugation
advantage, 53–54 (SE-AUC)
amide I and amide III bands, 56 SEC. See Size-exclusion chromatography (SEC)
components, 54–55 Sedimentation equilibrium analytical
electron cloud distortion, 55 ultracentrifugation (SE-AUC),
vs. infrared (IR), 54 70–71
molecular geometry, 55 Sedimentation velocity analytical
sulfhydryl bond (S–H) stretching vibration, ultracentrifugation (SV-AUC)
56–57 approaches, 67–68
tryptophan, 57 biopharmaceutical industry, 110
Recombinant human acid alpha-glucosidase comparability, 69–70
(rhGAA), 212, 214 description, 109
Recombinant protein production process, 265 forces, 66
Rheology formulation development, 68–69
characteristic properties, protein, 72 Lamm equation, 67
“electroviscous” effects, 74 lower and higher molecular weight, 207
Index 291
measurements, 110 Size heterogeneity

molecular weight, 67, 68 aggregation, 83
optical centrifugation cell and software conformation/shape heterogeneity, 83
data analysis packages, 207 fragmentation, 83
protein concentration, 68, 69 glycosylation, 84
protein product, 111–112 light scattering techniques, 89–90
sedimentation coefficient, 67, 68 self-association, 83
and SE-HPLC, 207–208 SLS. See Static light scattering (SLS)
stress condition, 112 Small angle neutron scattering (SANS), 209
Svedberg equation, 66–67 Soluble protein aggregation
SE-HPLC. See Size-exclusion high- analytical method
performance liquid CE-SDS, 209–210
chromatography (SE-HPLC) description, 206–207
Self-association d-SEC, 210
AUC, 88–89 FFF, 208
description, 83 light scattering, 208–209
HPSEC, 84 MS, 210–211
light scattering techniques, 89 SDS-PAGE, 209
SEM. See Scanning electron microscopy SE-HPLC, 207
(SEM) SPR, 211
SEM/EDS. See Scanning electron SV-AUC, 207–208
microscope/energy dispersive MS, 216–219
spectroscopy (SEM/EDS) physical composition and biological
Silicone oil/protein particle activity
infrared spectra, 257 biosensors (SPR) and correlations, 215
optical micrograph and FTIR spectrum, 257 Fc receptor, monomers, 215–216
SEM/EDS analyses, 258 IgG1 monoclonal antibody, 215
Size distribution, biopharmaceutical proteins size and amount
advantages and limitations, 84–87 analytical characterization, IgG1
aggregation, 83 monoclonal antibody
AUC, 88–89 solutions, 212, 213
conformation/shape heterogeneity, 83 monomers, dimers and HMW,
description, 94 analytical characterization, 214–215
ELS, EM and AFM, 91–92 quantitative analysis, aggregate
field-flow fractionation (AF4), 88 levels, 213–214
flow microscopy, 92–93 rhGAA, 212
fragmentation, 83 Spatial aggregation propensity (SAP), 175
gas-phase electrophoretic mobility, 93–94 Spectrophotometric techniques
glycosylation, 84 amide I IR spectra, rhDNase I, 187–188
HPSEC, 84 CD, 184–186
light obscuration, 90–91 far-and near-UV CD spectra, 184–185
MALS and DLS, 89–90 FTIR spectroscopy, 186–187
mass spectrometry (MS), 93 near-UV CD, human relaxin, 185–186
NTA technique, 90 protein lyophilized, 188
resonant mass measurement, 94 SPR. See Surface plasmon resonance (SPR)
self-association, 83 SRCD. See Synchrotron radiation circular
Size-exclusion chromatography (SEC), 154, dichroism (SRCD)
175, 196, 197 Stability
Size-exclusion high-performance liquid chemical, 136
chromatography (SE-HPLC) DSC, 15–16
chromatography process, 207 DSF, 16–17
disadvantages, 207 formulation, 140–141
sizing and quantifying soluble protein process, 141–142
aggregates, 207 thermal, 135–136
292 Index
Static light scattering (SLS) size and count

cumulative scattered intensity, 61–62 coulter counter, MFI and LO, 230, 231
intensity, scattered light, 61 differential particle levels, 231, 233
molecular weight measurements, 62–63 Subvisible particle (SbVP)
oscillations, electron, 61 filtration and clearance, 119
protein development, 62 LO method, 118
protein–protein interaction, 20 protein therapeutics, 118
protein’s molecular mass, 15 Surface plasmon resonance
Rayleigh ratio, 62 (SPR), 9–10, 189–190, 211
Stopped-flow fluorescence SV-AUC. See Sedimentation velocity
calorimetry, 171 analytical ultracentrifugation
molecules, 169 (SV-AUC)
Submicron protein particulates Synchrotron radiation circular dichroism
analytical methods (SRCD), 100
AFM, 221
description, 219–220
NTA, 220 T
QCM, 221–222 Tangential flow filtration (TFF), 190
SEM and TEM, 221 Taylor dispersion analysis (TDA), 208
counting and sizing TDA. See Taylor dispersion analysis (TDA)
IgG aggregation, NTA, 222, 224 T-DM1. See Trastuzumab-DM1 (T-DM1)
nanomechanical resonators, 222–223 TEM. See Transmission electron microscopy
morphology (TEM)
AFM, 223, 225 TFF. See Tangential flow filtration (TFF)
TEM, 223–226 Therapeutic development, protein. See Protein
Subvisible and visible particles therapeutics
description, 160 Therapeutic proteins. See High-throughput
DLS, 161 biophysical approaches, therapeutic
dynamic light-scattering size, 161 proteins
mechanism, 163 Thermodynamics and kinetics
real-time monitoring, 162 biphasic kinetic model, 170
SEC method, 161 bis-ANS, 168
Subvisible and visible protein particles hydrophobic patches, 169
analytical methods monoclonal antibodies, 168
coulter counter, 226 parameters, 169
description, 225 stopped-flow measurements, 169
EDX/EDS, 230 Transmission electron microscopy (TEM)
FACS, 228–229 description, 221
fluorescence microscopy, 229 morphology, submicron protein
FTIR microscopy, 229–230 particulates, 223–226
light microscopy and visual Trastuzumab-DM1 (T-DM1), 195
assessments, 227–228 Traveling voltage-wave ion mobility mass
LO and MFI, 227 spectrometry (TWIMS), 217–218
zeta potential, 228 Tungsten particle
morphology and composition biophysical characterization and
analytical characterization, oxidization, 253
isolation, 235, 236 cameo mapping, 253, 254
bright-field image, 232, 234 identification, 251
flow cytometer, 231–232 microparticles and FTIR
fluorescence microscopy permits, 232 spectrum, 251, 252
FTIR microscopy, 234–235 prefilled glass syringe and PFSs, 251, 253
Nile red staining, 232–233 root cause and proteins, 253
SYPRO Orange and extrinsic SEM micrograph and EDS spectrum,
fluorescent dye, 232 251, 252
Index 293
TWIMS. See Traveling voltage-wave ion measurements, 50

mobility mass spectrometry PEM frequency, 50–51
(TWIMS) ROA, 52–53
VCD, 51–52
Vibrational spectroscopy, 13
U Viscoelastic properties, 176
United States Pharmacopeial Convention Viscosity
(USP), 245, 274–275 bar plots and mAb solution, 177, 178
USP. See United States Pharmacopeial mAbs1–3, 191
Convention (USP) measurements, 176
solvent, 191
TFF, 190
V thermostability, 178
VCD. See Vibrational circular dichroism VOA. See Vibrational optical activity (VOA)
(VCD)
Vibrational circular dichroism (VCD), 51–52
Vibrational microspectroscopy, 249–250 Z
Vibrational optical activity (VOA) Zeta potential, 228

Biophysics For The Life Sciences: Series Editor: Norma Allewell

Uploaded by

Copyright:

Available Formats

Biophysics For The Life Sciences: Series Editor: Norma Allewell

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Biophysics For The Life Sciences: Series Editor: Norma Allewell

Uploaded by

Copyright:

Available Formats

Biophysics for the Life Sciences

Series Editor: Norma Allewell

For further volumes:

Biophysics for Therapeutic

ISBN 978-1-4614-4315-5 ISBN 978-1-4614-4316-2 (eBook)

Library of Congress Control Number: 2013931098

© Springer Science+Business Media New York 2013

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

9 Case Studies Applying Biophysical Techniques to Better

Index ................................................................................................................. 283

Fatma AlAzzam TechnoPharmaSphere LLC, Downingtown, PA, USA

Chava Kimchi-Sarfaty Division of Hematology, Office of Blood Research and

Steven J. Shire Late Stage Pharmaceutical Development, Genetech, South San

L.O. Narhi ()

L.O. Narhi (ed.), Biophysics for Therapeutic Protein Development, Biophysics 1

stability and stability to administration by the chosen delivery method. Determination

Feng He, Vladimir I. Razinkov, C. Russell Middaugh, and Gerald W. Becker

Protein therapeutic products typically experience many development cycles, in

F. He () • V.I. Razinkov • G.W. Becker

L.O. Narhi (ed.), Biophysics for Therapeutic Protein Development, Biophysics 7

2.2 High-Throughput Biophysical Techniques That Can Be

Due to increasing demands in sensitivity, diverse sample applicability and through-

2.2.1 Surface Plasmon Resonance

Characterization of target binding is an essential component during the develop-

Protein–protein interactions can be characterized with the help of surface plasmon

2.2.2 Liquid Chromatography

2.2.3 Light Absorption Spectroscopy

Protein concentration measurements are essential during purification and formulation

2.2.4 Vibrational Spectroscopy

Vibrational spectroscopy includes Raman, Fourier transform infrared (FTIR), and

2.2.5 Fluorescence Spectroscopy

The high-throughput capability of fluorescence techniques also typically relies on a

2.2.6 Isothermal Titration Calorimetry

Isothermal titration calorimetry (ITC) is another sensitive method to evaluate

2.2.7 Differential Scanning Calorimetry/Differential

Stability of protein therapeutics under thermal stress is believed to be critical and

as the VP-Capillary DSC Platform from GE Healthcare (Piscataway, NJ), the

2.2.8 Light Scattering

Aggregation is one of the major problems in protein pharmaceutical development.

2.2.9 Design of Experiment and Data Analysis

All high-throughput methodologies mentioned above share a common ability to

2.3 Empirical Phase Diagrams as Tools to Interpret Results

2.3.1 Combination of Biophysical Techniques

As discussed above and in other chapters in this text, high-throughput screening

2.3.2 High-Throughput Characterization

The EPD method provides a comprehensive overview of how a protein responds to

2.3.3 Additional Application of High-Throughput Methods

It is also possible to create EPDs based on phenomena such as aggregation.

2.4 Advantages and Challenges of Implementing

High-throughput technologies can be applied across all aspects of biopharmaceuti-

would be screened with a readout indicating a binding event or inhibition of binding.

Ahrer K, Jungbauer A (2006) Chromatographic and electrophoretic characterization of protein

James Kranz, Fatma AlAzzam, Atul Saluja, Juraj Svitel,

3.1 Introduction: Protein Structure

J. Kranz • W. Al-Azzam (*)

L.O. Narhi (ed.), Biophysics for Therapeutic Protein Development, Biophysics 33

3.2 Spectroscopic Techniques: Structural Characterization

3.2.1 Absorption Spectroscopy

3.1 Introduction: Protein Structure

3.2 Spectroscopic Techniques: Structural Characterization

3.2.1 Absorption Spectroscopy

3.2.2.1 Intrinsic Protein Fluorescence

3.2.2.2 Fluorescence and Conformational Dynamics

3.2.2.3 Fluorescence Quenching

3.2.2.4 Fluorescence Resonance Energy Transfer

3.2.2.5 Extrinsic Fluorescence

3.2.3 Fourier Transform Infrared Spectroscopy

3.2.4 Circular Dichroism

3.2.5 Vibrational Optical Activity

3.2.6 Raman Spectroscopy

3.2.7 Other Structural Characterization Methods

3.3.1 Calorimetric Methods

3.3.1.1 Isothermal Titration Calorimetry

3.3.1.2 Differential Scanning Calorimetry

3.3.2 Light Scattering

3.3.2.1 Static Light Scattering

3.3.2.2 Dynamic Light Scattering

3.3.3 Analytical Ultracentrifugation

3.3.3.1 Sedimentation Velocity

3.3.3.2 Sedimentation Equilibrium