RESEARCH ARTICLE – Pharmaceutical Nanotechnology

Modifying the Release Properties of Liposomes Toward Personalized

Medicine
DAVID CIPOLLA,1,2 HUIYING WU,2 IGOR GONDA,2 SIMON EASTMAN,3 TOM REDELMEIER,3 HAK-KIM CHAN1
1
Faculty of Pharmacy, The University of Sydney, Sydney, NSW 2006, Australia
2
Aradigm Corporation, Hayward, California 94545
3
Northern Lipids Inc., Burnaby, British Columbia, V5J 5J1, Canada

Received 19 February 2014; revised 17 March 2014; accepted 17 March 2014

Published online 8 April 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23969

ABSTRACT: Surfactant–liposome interactions have historically been investigated as a simplified model of solubilization and breakdown of
biological membranes by surfactants. In contrast, our goal was to utilize surfactants to modify the encapsulation and release properties of
liposomes. The ability to manufacture one liposomal formulation, which could be modified by the addition of a surfactant to support a wide
range of release profiles, would provide greater flexibility than manufacturing multiple batches of liposomes, each differing in composition
and with its own specific release profile. A liposomal ciprofloxacin formulation was modified by the addition of various surfactants.
These formulations were characterized in terms of liposome structure by cryo-TEM imaging, vesicle size by dynamic light scattering,
drug encapsulation by centrifugation–filtration, and in vitro release (IVR) performance. The addition of polysorbate 20 or polysorbate
80 to liposomal ciprofloxacin, in a hypotonic environment, resulted in a concentration-dependent loss of encapsulated drug, and above
0.4% polysorbate 20, or 0.2% polysorbate 80, a modified IVR profile as well. This study demonstrates that the encapsulation and release
properties of a liposomal formulation can be modified postmanufacture by the addition of judiciously chosen surfactants in combination
with osmotic swelling of the liposomes and may support a personalized approach to treating patients.  C 2014 Wiley Periodicals, Inc. and

the American Pharmacists Association J Pharm Sci 103:1851–1862, 2014

Keywords: liposomes; surfactants; controlled delivery; aerosols; nanoparticles; personalized medicine; drug delivery; ciprofloxacin;
in vitro release

INTRODUCTION In the present study, we were interested whether the release

properties of a liposomal ciprofloxacin for inhalation (CFI) could
There are a range of formulation technologies that can be used
be modified by the addition of a surfactant to the drug product
to modulate the release properties of pharmaceutical drugs, in-
postmanufacture. One could envision a scenario in which this
cluding liposomes.1 Liposomes are phospholipid vesicles com-
simple procedure is performed by a doctor or a patient, from
posed of one or more lipid bilayers surrounding an aqueous
a kit containing ampules with different concentrations of sur-
core and can vary in size ranging from approximately 20 nm to
factant, or different surfactants, to provide a formulation tai-
10 :m.2 Liposomes as drug delivery vehicles are now well ac-
lored to the specific needs of that patient. This could prove less
cepted with more than 10 approved liposomal products on the
costly, and less complex, than developing and manufacturing
market.3 A wide variety of drugs have been formulated into li-
multiple, unique liposomal formulations which differ in compo-
posomes including small molecules, peptides, and nucleic acids;
sition to cover a broad range of release profiles. However, to our
hydrophilic drugs are generally dissolved in the aqueous com-
knowledge, this represents a new paradigm as surfactant has
partment, whereas hydrophobic drugs are associated with the
not previously been reported to be useful in the context of mod-
lipid bilayers.1,2 The liposome composition, lamellarity, and the
ifying the release properties of a liposome formulation, once it
manufacturing process can each influence the physicochemical
has been manufactured. Of course, under this new personal-
properties of a liposomal product, including its drug release ki-
ized medicine approach, new hurdles will arise such as how to
netics. Although there are a wide number of variables in play
identify the appropriate release profile for each individual, but
to produce liposomal formulations with differing pharmacoki-
that is not the focus of this paper.
netics, it would be complex and costly to market multiple lipo-
There is an extensive history of detergents or surfactants
somal compositions to cover a broad range of release profiles. A
being used to solubilize biological membranes to allow for elu-
liposomal formulation which could be easily modified postman-
cidation of membrane structure and function.4 The ability of
ufacture to produce different release profiles has the potential
surfactants to solubilize phospholipids, specifically, has been
to provide greater flexibility in treatment.
reviewed.5 As surfactant is added to phospholipids, surfactant
initially partitions between the solution and the phospholipid
bilayers and the bilayer permeability may increase without loss
Abbreviations used: CF, carboxyfluorescein; CFI, ciprofloxacin for inhalation; of structure.5 Once the phospholipid bilayers become saturated
Cryo-TEM, cryogenic transmission electron microscopy; DLS, dynamic light
scattering; DRCFI, dual release ciprofloxacin for inhalation; IVR, in vitro with surfactant, the addition of more surfactant leads to the
release; PC, phosphatidylcholine; TEA, triethylamine. formation of mixed micelles of surfactant and phospholipid un-
Correspondence to: David Cipolla (Telephone: +510-265-8849; Fax: +510-265-
8049; E-mail: [email protected])
til all of the remaining surfactant-saturated bilayers are con-
Journal of Pharmaceutical Sciences, Vol. 103, 1851–1862 (2014)
verted to mixed micelles. Any further addition of surfactant

C 2014 Wiley Periodicals, Inc. and the American Pharmacists Association leads to a decrease in the size of the micelles as they become

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 1851

1852 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

more dilute in phospholipid content. Vesicle–surfactant sys- of the liposomes might enhance the ability of the surfactant to
tems have frequently been characterized by monitoring their intercalate or associate with the liposome membrane and thus
light scattering properties, with maximum turbidity generally alter its drug encapsulation and release properties. This paper
associated with the surfactant-saturated bilayer state6 followed describes liposomal formulations of ciprofloxacin with modified
by a rapid decline in turbidity once the bilayers are completely encapsulation and release properties because of the addition
solubilized by surfactant.7–12 of surfactant under hypotonic conditions. In line with the goal
Liposomes have also been used as a simplified model of bi- of personalized medicine, to tailor a product to an individual’s
ological membranes. Interest in the development of antimicro- needs so that it releases the “right amount of drug at the right
bials that would function by disruption of the bacterial mem- time,” this strategy may provide more flexibility than the alter-
brane spurred a better understanding of the interaction of native of manufacturing multiple batches of liposomes differing
surfactant-like molecules with liposomes.13,14 During this same in composition to cover a broad range of desired release profiles.
time period, liposomes were also being investigated as drug de-
livery vehicles in their own right, so knowledge of the factors
which affected the timing and rate of release of the encapsu- MATERIALS AND METHODS
lated drug, including interaction with biological fluids, was im- Materials
portant. After in vivo administration, liposomes come into con-
tact with many natural amphiphiles present in physiological Free ciprofloxacin (FCI), 20 mg/mL in an acetate-buffered aque-
fluids. Surfactants were used as a simplification to the complex ous formulation at pH 3.3 and CFI, 50 mg/mL in a histidine-
biological milieu, allowing for characterization of drug release buffered aqueous formulation at pH 6.0 were from Aradigm
from liposomes in the presence of surfactant.15 Corporation (Hayward, California). (The ciprofloxacin concen-
The observation that surfactants can interact with liposomes trations are expressed in terms of ciprofloxacin hydrochloride.)
to cause leakage of encapsulated drug was extensively investi- Polysorbate 20 was purchased from VWR Int. (West Chester,
gated in the 1970s and 1980s, but new insights into the mecha- New Jersey). Polysorbate 80, sorbitan monolaurate (Sp20), sor-
nism of surfactant-induced liposomal breakdown and the ki- bitan monooleate (Sp80), poloxamer L44, and poloxamer L62
netics of release of the entrapped agent [carboxyfluorescein were a gift from Croda, Inc. (Edison, New Jersey). HEPES, free
(CF)] were just recently described.6 Upon the addition of sol- acid was purchased from Avantor (Center Valley, Pennsylva-
ubilizing levels of surfactant, there is a complete breakdown nia). Sodium chloride was obtained from Amresco (Solon, Ohio).
of liposome structure because of the direct interactions of the Sodium acetate was purchased from Sigma–Aldrich (St. Louis,
surfactant molecules with the lipids, resulting in complete re- Missouri). HPLC grade methanol was purchased from Fisher
lease of encapsulated drug in less than a second.6 At lower sur- Scientific (Fair Lawn, New Jersey) and triethylamine (TEA)
factant concentrations, surfactant molecules can cooperatively was purchased from JT Baker (Center Valley, Pennsylvania).
assemble into structures which form pores in the liposomes, Donor adult bovine serum was obtained from HyClone (Logan,
again allowing for almost instantaneous release of a portion Utah). Nanosep centrifugal filtration devices, 10 K and 30 K
of the encapsulated agent, but these holes then close limit- Da molecular weight, were obtained from Pall Corporation (Ann
ing further release.6 Nagawa and Regen13 demonstrated that Arbor, Michigan). Deionized water was used for all studies.
for some combinations of liposomes and surfactant, the encap- Preparation of Liposomal Ciprofloxacin
sulated agent was released from all vesicles with increasing
surfactant concentration. For other combinations, there was a Ciprofloxacin for inhalation is an aqueous dispersion of unil-
catastrophic rupture in which a subset of the vesicles rapidly re- amellar liposomes of approximately 80 nm containing hydro-
leased their entire encapsulated payload, whereas others were genated soy phosphatidylchocline (PC) and cholesterol. The
unaffected.13 Although there are many reports of loss of drug preparation of CFI has been reported previously.18 Briefly, mul-
from liposomes in response to surfactant addition, none have tilamellar liposomes are extruded through membranes to pro-
shown that the surfactant can be used to modify the pharma- duce unilamellar liposomes which are then actively loaded with
cokinetic properties and thus the in vitro release (IVR) profile ciprofloxacin.19,20 Any unencapsulated ciprofloxacin is removed
of the liposomes.6,13–17 by diafiltration resulting in >99% encapsulated ciprofloxacin
A mechanistic understanding of the interactions of surfac- at a target concentration of 50 mg/mL. Dual-release CFI (DR-
tants with liposomes is generally well understood; however, in- CFI) is an equivolume mixture of FCI and CFI resulting in
terest in utilizing surfactants to modify the release properties approximately 70% encapsulated and approximately 30% FCI.
of liposomes for therapeutic purposes, which is the focus of this
Addition of Surfactant to CFI
paper, has not been explored. Rather than using the strongly
solubilizing surfactants, like Triton X-100, we chose surfactants In the first series of experiments, CFI was diluted from
that are common to many pharmaceutical formulations, for 50 mg/mL to 30, 20, 15, 12.5, or 10 mg/mL with histidine buffer
example, the polysorbates, to generate liposomal formulations (HB: 25 mM histidine, 145 mM sodium chloride buffer, pH 6.0)
with modified release properties. We also studied other nonionic and polysorbate 20 to achieve a final polysorbate 20 concentra-
surfactants including poloxamers and spans, which are closely tion of 0.1%, 0.2%, 0.4%, 0.8%, 1.2%, 1.6%, or 2.0% (w/v). A sec-
related structurally to the polysorbates. It was expected that ond series of experiments was identical to the first except that
these “milder” surfactants may not be sufficiently disruptive to the CFI was diluted with water, instead of HB, to determine
the liposomes at low concentrations to yield formulations with whether osmotic swelling might play a role in enhancing the
significantly modified properties. To overcome this challenge, effect of the addition of surfactant. In a third experiment, the
we evaluated the combination of addition of surfactant along order of dilution was varied to determine whether the proper-
with dilution of the liposomal formulations with water, to pro- ties of these preparations depended upon the order of addition
duce a hypotonic environment. The resulting osmotic swelling of water and surfactant to the 12.5 mg/mL CFI preparations. In

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969

 RESEARCH ARTICLE – Pharmaceutical Nanotechnology 1853

a fourth experiment, after 30 min equilibration, the 30 mg/mL drug, was quantified by HPLC after diluting the CFI sample
CFI preparations containing various levels of surfactant were 20-fold into 80% methanol to solubilize the liposomes. The per-
diluted with water to 10 mg/mL CFI to determine whether the cent encapsulation was determined by comparing the free drug
final encapsulation state was dependent upon the timing of the to the total drug in each sample.
preparation steps. In all of these experiments, the encapsula-
In Vitro Release
tion state was expressed as the mean of duplicate values.
Another series of experiments explored the use of various The IVR assay measures the release of encapsulated
surfactants, including Sp20, Sp80, poloxamer L62, poloxamer ciprofloxacin when incubated at 37◦ C in 50% (v/v) bovine
L44, and polysorbate 80, to modify the properties of liposomal serum.21 Briefly, the CFI samples were diluted to 50 :g/mL
ciprofloxacin, CFI. In these studies, the formulations contained ciprofloxacin in HEPES-buffered saline (HBS: 20 mM HEPES,
12.5 mg/mL CFI (i.e., diluted fourfold with water to produce 145 mM NaCl, pH 7.4) and mixed one-to-one with chilled (2◦ C–
a hypotonic environment) in 0.01%, 0.05%, 0.1%, 0.5%, or 1% 8◦ C) bovine serum (Hyclone) and placed in a shaking water bath
(w/v) of each surfactant. A final experiment evaluated the re- [Techne, TSBS40 (Staffordshire, UK)] at 37◦ C and 150 rpm. Du-
producibility of this effect across three different batches of CFI, plicate samples were removed after incubation for 30, 60, 120,
upon the addition of polysorbate 20 in a hypotonic environment. and 240 min, diluted 1:1 with chilled (2◦ C–8◦ C) HBS and placed
In this study, 50 mg/mL CFI was diluted fourfold with water, in an ice-water bath to terminate any further release of encap-
and 1% polysorbate 80 to a final concentration of 12.5 mg/mL sulated drug from the liposomes. The released ciprofloxacin
CFI in 0%, 0.05%, 0.1%, 0.2%, 0.3%, 0.4,% 0.5%, or 0.6% (w/v) was separated from the liposome-encapsulated ciprofloxacin by
polysorbate 80. transferring 400 :L of each chilled sample to a Nanosep cen-
In all experiments, the CFI samples containing surfactant trifugal device and centrifuging at 8100g for 18 min. The fil-
were allowed to equilibrate for at least 30 min to provide ade- trate was removed for subsequent quantitation of the released
quate time for the surfactant to associate with the liposomes. ciprofloxacin by HPLC. This value was normalized by multiply-
No further change in the properties of these formulations was ing by 1.05, to correct for a small but reproducible loss of free
observed when samples were equilibrated for longer periods drug in the filtration devices in the presence of serum.21 The
of time (data not shown). For each of these preparations, the original CFI sample was diluted into 80% methanol to dissolve
vesicle size distribution and drug encapsulation states were the liposomes and allow for quantitation of the total amount of
determined. For a number of the polysorbate 20 preparations, ciprofloxacin by HPLC. The percent release at each time point
cryogenic transmission electron microscopy (cryo-TEM) images was calculated by comparing the free drug to the total drug.
and IVR profiles were also obtained.
Cryo-TEM
Vesicle Size To obtain visual images of the liposome formulations, cryo-TEM
Each CFI or CFI plus surfactant sample was diluted with iso- analysis was performed using a JEOL 2100 (Tokyo, Japan) in-
tonic saline to a concentration of approximately 1 mg/mL CFI strument operated at 200 kV. The CFI samples were diluted to
(2 mg/mL liposomes), and 0.5 mL was transferred to a dispos- approximately 5 mg/mL (10 mg/mL liposomes) with water and
able culture tube (Kimble Glass Inc. (Vineland, NJ)) for vesi- 3 :L of each sample was applied to a glow discharge Quantifoil
cle size analysis using a Submicron Particle Sizer Autodilute carbon grid (Jena, Germany) in a chamber controlled to 22◦ C
Model 370 (Nicomp (Santa Barbara, CA)). The following in- and 100% relative humidity. Grids were blotted once with fil-
strument parameters were selected: temperature: 23◦ C; viscos- ter paper, at a blotting angle of 2 mm for 2 s, and vitrified by
ity: 0.933; refractive index: 1.333; intensity set point: 300 kHz; plunging into liquid ethane using a Vitrobot (F.E.I., Eindhoven,
channel width: 10 :s; scattering angle: 90; run time: 5 min; Netherlands). The vitrified samples were stored in liquid nitro-
mode: vesicle; Gaussian distribution. The mean and SD of the gen prior to cryo-TEM analysis.
vesicle size distribution were recorded. HPLC

Drug Encapsulation The amount of ciprofloxacin in each sample was quantified us-
ing an HPLC method as described previously.22 Briefly, HPLC
Nanosep Omega centrifugation devices (Pall Corporation) with analysis was performed using a Nucleosil C-18 column (5 :m,
modified polyethersulfone membrane filters of 10,000 or 30,000 4.6 × 150 mm2 ; Canadian Life Science (Peterborough, Ontario))
molecular weight cutoffs were used to separate free drug from protected with a Nucleosil C-18 guard column (4 × 3.0 mm2 ;
liposomal encapsulated drug. Each sample was diluted 10-fold Phenomenex (Torrance, California)) both at 35◦ C. The mobile
into acetate buffer (50 mM sodium acetate, 145 nM NaCl, pH phase was a mixture of 0.5% TEA in water, pH 3.0, and 100%
4.0) and 400 :L was transferred to the centrifugation device methanol (83:17, v/v), and the isocratic elution was performed
and centrifuged for 18 min at 8100g (10,000 rpm). A centrifu- at a flow rate of 0.9 mL/min. Ciprofloxacin was detected and
gation time of 18 min and a speed of 8100g were chosen to quantified at a wavelength of 277 nm.
ensure that there was an adequate volume in the filtrate to
transfer to HPLC vials without risk of rupture to the mem-
branes because of a too high centrifugation force. These condi- RESULTS
tions were previously shown to provide quantitative recovery.21
Liposomal Ciprofloxacin Encapsulation: Effect of Addition of
(Subsequently, it was found that centrifugation times of 10 min
Polysorbate 20
were also effective when samples were further diluted by a fac-
tor of two.) The filtrate, representing the FCI component, was Polysorbate 20 is present in a number of approved pharma-
analyzed by HPLC for ciprofloxacin content. The total amount ceutical products and so was chosen for initial evaluation be-
of ciprofloxacin, representing both the encapsulated and free cause of the general knowledge about its suitability in inhaled

DOI 10.1002/jps.23969 Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014

1854 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

formulations. When 50 mg/mL CFI was diluted with HB, and Congruous to the previous studies (Fig 1a), for any given
increasing concentrations of polysorbate 20, up to 2.0%, were concentration of polysorbate 20, there was a greater extent
added, there was a small concentration-dependent loss of en- of release of encapsulated ciprofloxacin for more dilute CFI
capsulated drug but the amount of free drug did not exceed formulations than for more concentrated CFI formulations
8% (Fig. 1a). The 30 and 10 mg/mL CFI samples decreased (Fig. 1b). In contrast to the previous studies, for which there
to approximately 95% and 92% drug encapsulation, respec- was little variation in encapsulation state across a constant ra-
tively, at the highest polysorbate 20 concentration (i.e., 2%). tio of surfactant to liposomes (dotted lines in Fig. 1a), in these
As expected, intermediate CFI concentrations, that is, 12.5, 15, studies there was a dramatic difference in encapsulation states
and 20 mg/mL, had encapsulation values which fell in between across a constant ratio of surfactant to liposomes (dotted lines
those two curves. in Fig. 1b). For the 30 mg/mL CFI formulation in 2% polysorbate
For any given surfactant concentration, there was an in- 20, the encapsulation state was approximately 95%, whereas
crease in drug release with decreasing CFI concentration, as it decreased to approximately 56% for 10 mg/mL CFI in 0.67%
would be expected because the ratio of surfactant to liposomes polysorbate 20. Similarly, for the 30 mg/mL CFI formulation
is greater. Therefore, one would expect more surfactant to be in 0.8% polysorbate 20, the encapsulation state was again ap-
associated with each individual liposome for the lower CFI con- proximately 95%, whereas it decreased to approximately 67%
centration preparations and thus to have the potential for a for 10 mg/mL CFI in 0.27% polysorbate 20. Although the ratio
greater effect on drug release. At a constant ratio of surfactant of surfactant to liposomes remained constant along the dotted
to liposomes, if no other changes were introduced, it might be lines in Figure 2B, the degree of tonicity, and hence the amount
expected that the final encapsulation states would be compa- of osmotic swelling, did not remain constant. The lower concen-
rable. Although the individual data points were not selected to tration CFI preparations were more hypotonic and thus likely
keep the ratio of surfactant to liposomes constant, a comparison experienced more osmotic swelling. This may explain why the
can be made by following the dotted black lines in Figure 1a more dilute liposome solutions of CFI were more susceptible to
for both a high and low ratio of surfactant to liposomes. In surfactant-induced release of encapsulated drug than for the
both cases, there was a slight decrease in encapsulation with more concentrated solutions of CFI, at the same molar ratio of
decreasing CFI concentration over the concentration range of surfactant to liposomes.
30–10 mg/mL CFI, but the changes were modest (i.e., 1%–2%).
Liposomal Ciprofloxacin Encapsulation: Effect of Dilution Order
(Water and Polysorbate 20)
Liposomal Ciprofloxacin Encapsulation: Combination of Osmotic In a third experiment, the order of dilution was evaluated by
Swelling and Polysorbate 20 adding the surfactant to the 50 mg/mL CFI immediately af-
ter dilution with water, as was performed in the previous ex-
In the previous experiment, we were unable to produce mean-
periments, or immediately before dilution with water. In this
ingful increases in free drug, of more than 8%, even using a rel-
experiment there was no meaningful difference in the encap-
atively large amount of polysorbate 20 (i.e., 2%). The next series
sulation state between the two preparations across the range
of experiments were conducted to determine whether liposomes
of surfactant concentrations suggesting that order of dilution
subjected to osmotic swelling would be more responsive to the
is not critical to the final encapsulation state, at least as long
addition of surfactant. In these experiments, CFI was diluted
as the dilution steps are close together temporally (Fig. 1c).
with water, instead of HB, to create a hypotonic (hypoosmotic)
state. Polysorbate 20 was added immediately after dilution with
Liposomal Ciprofloxacin Encapsulation: Effect of Equilibration
water. In contrast to the previous studies for which there was
with Polysorbate 20 prior to Dilution with Water
a minimal loss of encapsulated drug (Fig. 1a), the addition of a
surfactant in conjunction with osmotic swelling of the liposomes In a fourth experiment, the encapsulation states of 30 mg/mL
resulted in a much greater loss of encapsulated drug, especially CFI in a range of polysorbate 20 concentrations were measured
for the lower CFI concentrations (Fig. 1b). The release of en- after 30 min equilibration, and then were remeasured after di-
capsulated drug happened very rapidly, well within 30 min, lution with water to 10 mg/mL CFI (Fig. 1d). In contrast to the
and no further release was observed thereafter. The degree of previous experiment which suggested that the order of dilu-
osmotic swelling would be expected to be the greatest for the tion did not have a significant impact on encapsulation state,
lowest CFI concentration preparations (i.e., 10 mg/mL) because in this experiment, the final encapsulation state was very de-
of the greater dilution of CFI and thus the largest reduction in pendent upon the dilution procedure. When the 10 mg/mL CFI
tonicity. And indeed those samples showed the greatest loss of formulation was prepared directly from the 50 mg/mL CFI by
encapsulated drug. For the 10 mg/mL CFI preparation contain- the addition of water and a surfactant, in either order (Fig. 1c),
ing 2% polysorbate 20, the amount of released drug increased there was a much greater decrease in encapsulation state ver-
from approximately 8% to >50% when the liposomes were di- sus when the 30 mg/mL CFI was allowed to equilibrate with
luted with water versus HB suggesting that osmotic swelling surfactant prior to dilution to 10 mg/mL (Fig. 1d). In the lat-
acts synergistically with the addition of surfactant to modify the ter case, there was only a minimal change in the encapsula-
encapsulation state of the liposomes. The encapsulation state tion state upon the threefold dilution (Fig. 1d). This result
of the highest concentration CFI preparation (i.e., 30 mg/mL), suggests that the final encapsulation state may be pathway
across the range of polysorbate 20 concentrations, was no dif- dependent (hence also time dependent) and predominantly ki-
ferent when diluted with water or HB. This result suggests netically based, rather than equilibrium based. In other words,
that there may not have been sufficient osmotic swelling of the there may only be a transient time period for which the os-
30 mg/mL CFI preparations to have any effect on release of motically swollen liposomes are susceptible to surfactant mod-
encapsulated drug. ification and if the surfactant has already equilibrated with

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969

DOI 10.1002/jps.23969
Figure 1. The effect of the addition of polysorbate 20 on the encapsulation state of ciprofloxacin in CFI. Each data point represents the mean (n = 2). CFI containing 50 mg/mL
encapsulated ciprofloxacin was diluted to a final concentration of 30, 20, 15, 12.5, and 10 mg/mL ciprofloxacin with HB (a) or water (b) and an aliquot of 1% or 10% (w/v) polysorbate
RESEARCH ARTICLE – Pharmaceutical Nanotechnology

20 to achieve a final surfactant concentration of 0.1%, 0.2%, 0.4%, 0.8%, 1.2%, 1.6%, or 2.0% (w/v). The dotted black lines represent a constant ratio of surfactant to liposomes.
(c) CFI containing 50 mg/mL ciprofloxacin was diluted to a final concentration of 12.5 mg/mL ciprofloxacin with water and 1% or 10% (w/v) polysorbate 20 to achieve a final surfactant
concentration of 0.1%, 0.2%, 0.4%, 0.8%, 1.2%, 1.6%, or 2.0% (w/v). In one experiment, the water was added to the CFI before the polysorbate 20 and in the other the polysorbate 20 was
added to the CFI before the water. (d) CFI containing 30 mg/mL ciprofloxacin and 0.1%, 0.2%, 0.4%, 0.8%, 1.2%, 1.6%, or 2.0% (w/v) polysorbate 20 was diluted to a final concentration of
1855

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014

10 mg/mL ciprofloxacin with water to achieve a final concentration of 0.03%, 0.07%, 0.13%, 0.27%, 0.4%, 0.53%, and 0.67% (w/v) polysorbate 20.
1856 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

the liposomes prior to osmotic swelling, as in this experiment, encapsulation before incubation with the release agent, that is,
subsequent osmotic swelling does not appear to enable the sur- the amount of “burst.” For the control CFI vial, the “burst” value
factants to associate in the same way to produce larger losses is approximately 1%. For DRCFI, which is a mixture of approx-
of encapsulated drug. imately 70% CFI and 30% FCI solution, the “burst” value is
approximately 30%, as would be expected. For the 12.5 mg/mL
Liposomal Ciprofloxacin Encapsulation: Applicable to Other CFI containing various levels of surfactant, the amount of FCI,
Surfactants? or “burst,” increased with increasing polysorbate 20 from ap-
To determine whether the surfactant-induced release of encap- proximately 2% for 0.05% polysorbate 20 to approximately 27%
sulated ciprofloxacin in a hypoosmotic environment was specific for 0.4% polysorbate 20 in the IVR assay (Fig. 3a).
to polysorbate 20, or more broadly applicable to other closely re- The last time point in the IVR assay, at 4 h, represents the
lated, nonionic surfactants, another series of experiments was complete extent of release of ciprofloxacin from the liposome
performed with the following surfactants: polysorbate 80, Sp20, preparations. By the 4 h time point, all of the formulations
Sp80, poloxamer L44, and poloxamer L62. Upon the addition reached a plateau value in the IVR assay representing approx-
of a surfactant to CFI diluted with water to a final concentra- imately 100% release of encapsulated ciprofloxacin.
tion of 12.5 mg/mL, all surfactants caused some release of en- The IVR profiles for the CFI formulations containing 0.05,
capsulated ciprofloxacin with the amount released increasing and 0.1% polysorbate 20 were similar to that for CFI alone.
with increasing surfactant concentration (Fig. 2a). However, However, the IVR profile for the CFI formulation containing
the amount of drug that was released was very dependent on 0.4% polysorbate 20 had marked differences to that for CFI as
the choice of surfactant, with polysorbate 80 resulting in sub- well as for DRCFI. The T30 min value is an approximate rep-
stantial release of ciprofloxacin, whereas none of the other non- resentation of the “midpoint” of release for CFI.21 The T30 min
ionic surfactants that were tested produced more than 1%–2% values increased with increasing polysorbate 20, from approx-
release of encapsulated drug over the surfactant concentration imately 44% for CFI alone to approximately 66% for CFI con-
ranges that were evaluated (Fig. 2a). At the same surfactant taining 0.4% polysorbate 20. However, one cannot directly com-
concentrations, polysorbate 80 was more effective than polysor- pare the T30 min values of these formulations to assess changes
bate 20 (Fig. 2a vs. Fig. 1c) at causing greater release of encap- to the release rates because the encapsulation states were not
sulated ciprofloxacin in the 12.5 mg/mL CFI preparations. More equivalent at the initial time point (T0 min ). To address this is-
data points were not collected over the surfactant concentration sue, one can “normalize” the T30 min values by subtracting the
range studied because the focus of this experiment was to un- initial value, T0 min , and dividing by the total available range for
derstand the general behavior of the surfactant interactions release (T240 min – T0 min ) and converting to a percentage: 100 ×
with CFI, but not to fully map out the design space. (T30 min − T0 min )/(T240 min − T0 min ). The normalized T30 min val-
A follow-on experiment was conducted to more fully map out ues are 44.4% for CFI, and 43.4%, 46.8%, 47.6%, and 55.7%
the design space for polysorbate 80, and to determine how re- for CFI containing 0.05%, 0.1%, 0.2%, and 0.4% polysorbate 20,
producible this effect was for different manufactured batches respectively. The encapsulated ciprofloxacin release rates were
of CFI. The response of each of the three batches of CFI to relatively comparable, except for the CFI formulation contain-
polysorbate 80 was comparable, with a large loss of 30%–35% ing 0.4% polysorbate 20. In addition, although CFI containing
encapsulated drug for 0.1% polysorbate 80, and 40%–45% loss 0.4% polysorbate 20 had a comparable initial release value,
of encapsulated drug for 0.2% polysorbate 20 (Fig. 2b). Higher or “burst,” to that for DRCFI, 27% versus 30%, respectively,
concentrations of polysorbate 20, up to 0.6%, did not signifi- its release rate was faster than DRCFI with a T30 min value
cantly alter the plateau value of approximately 45% released (un-normalized, as the initial values are similar) of 66% ver-
drug (55% encapsulated drug). sus 57%, respectively. The normalized T30 min values of 55.7%
versus 41.6%, respectively, also demonstrate this difference in
Vesicle Size release rates.
Polysorbate 80 had an even greater effect on the IVR profile
The vesicle size of the various preparations was measured to than polysorbate 20. CFI containing 0.2% polysorbate 80 had a
determine whether the addition of surfactant results in any faster IVR profile than for CFI containing twice as much (0.4%)
measurable change in liposome size. The mean vesicle size of polysorbate 20 (Figs. 3b vs. 3a). To investigate whether the
CFI was unchanged in the presence of either poloxamer L44 surfactant has altered the membrane permeability of the lipo-
or poloxamer L62, increased by 1–3 nm for Sp20, Sp80, and somes and thus its release profile, or if the surfactant was sim-
polysorbate 20, and increased by up to 8 nm for polysorbate 80 ply operating as a release vehicle itself, control CFI was diluted
(Table 1). into the release vehicle (serum) with or without polysorbate 80
and the IVR assay was repeated (Fig. 3b). The concentration of
IVR Profile
surfactant in the IVR assay in that experiment (0.0004%) was
We have shown that addition of polysorbate 20 and polysorbate identical to that for 12.5 mg/mL CFI containing 0.2% polysor-
80 surfactants to CFI, in a hypotonic environment, leads to bate 80 after the 500-fold dilution into the IVR assay buffer.
more extensive changes to the encapsulation state of the lipo- The presence of polysorbate 80 in the release buffer had no
somes than in the absence of osmotic swelling. We were also in- effect on the IVR profile of CFI (Fig. 3b).
terested whether the release properties of the liposomes could
be modified by surfactant addition. The IVR assay measures
Cryo-TEM
the release of encapsulated ciprofloxacin from the liposomes
after incubation with bovine serum at 37◦ C.21 Consistent with To better understand the effect of the addition of polysorbate 20
the encapsulation assay data (Figs. 1b and 2a), the time zero on the structure and state of the liposomal ciprofloxacin vesi-
release values in the IVR assay (Fig. 3a) represent the state of cles, 12.5 mg/mL CFI samples (diluted with water) with and

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969

 RESEARCH ARTICLE – Pharmaceutical Nanotechnology 1857

Figure 2. (a) The effect of the addition of various surfactants on the state of ciprofloxacin encapsulation. CFI containing 50 mg/mL of encapsu-
lated ciprofloxacin was diluted to a final concentration of ∼12.5 mg/mL ciprofloxacin with water and an aliquot of 1% or 10% (w/v) surfactant to
achieve a final surfactant concentration of 0.01%, 0.05%, 0.1%, 0.5%, or 1.0% (w/v). The surfactants that were investigated included: poloxamer
L44, poloxamer L62, Sp20, Sp80, and polysorbate 80. After vortexing and allowing each sample to equilibrate for at least 30 min, the ciprofloxacin
encapsulation state was determined by centrifugal filtration in duplicate. Each data point represents the mean (n = 2). There are no data points
using 0.5% or 1% (w/v) Sp20 because of the poor miscibility of the solution at those concentrations. (b) The effect of the addition of polysorbate
80 on the state of ciprofloxacin encapsulation for three different batches of CFI. CFI containing ∼50 mg/mL of encapsulated ciprofloxacin was
diluted to a final concentration of ∼12.5 mg/mL ciprofloxacin with water and an aliquot of 1% (w/v) polysorbate 80 to achieve a final surfactant
concentration of 0.05%, 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, or 0.6% (w/v). After vortexing and allowing each sample to equilibrate for at least 30 min,
the ciprofloxacin encapsulation state was determined by centrifugal filtration.

DOI 10.1002/jps.23969 Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014

1858 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

Table 1. The Vesicle Size Distribution for 12.5 mg/mL CFI in the Presence of the Following Surfactants: Poloxamer L44, Poloxamer L62,
Sp20, Sp80, Polysorbate 20, and Polysorbate 80

Surfactant (%) SPAN 80 SPAN 20 Pluronic L62 Polysorbate 80 Pluronic L44 Polysorbate 20

0 93.5 [26.0]
0.01 93.4 [26.0] 93.7 [27.6] 93.4 [27.6] 93.8 [29.0] 94.7 [34.8] 92.0 [30.6]
0.05 93.5 [26.5] 94.2 [28.9] 93.4 [26.5] 94.0 [24.9] 94.6 [30.4] 93.4 [30.8]
0.1 93.8 [28.2] 94.7 [29.9] 93.0 [25.2] 97.9 [29.2] 94.6 [30.3] 93.9 [30.7]
0.5 94.2 [34.0] NDa ND 101.3 [33.8] 94.7 [31.4] 95.2 [31.9]
1.0 96.7 [37.4] NDa 93.8 [24.0] 101.6 [26.9] 94.3 [34.0] 94.6 [27.4]
a
Preparation did not form a miscible solution.
Vesicle size data are reported as the mean (in nm) and [SD].
ND, not done.

without various levels of polysorbate 20 were imaged by cryo- Liposomal ciprofloxacin preparations were also diluted with
TEM. The CFI formulation without surfactant was composed water to create a hypotonic environment that would be ex-
of spherical, unilamellar liposomes between approximately 50 pected to result in osmotic swelling of the liposomes.24–26 When
and 100 nm in size (Fig. 4a). The CFI liposomes do not appear to polysorbate 20 was added just prior to or just after dilution
form agglomerates in contrast to liposomes composed of choles- with water, there was a much greater loss of encapsulated drug
terol and distearoyl PC, which are nonspherical and interact suggesting a synergistic effect of osmotic swelling and surfac-
attractively to form large clusters.12,23 CFI samples containing tant on release of encapsulated drug (Fig. 1c). The greater the
0.05% polysorbate 20 (Fig. 4b), 0.2% polysorbate 20 (Fig. 4c), dilution of CFI with water, and thus the greater osmotic im-
and 0.4% polysorbate 20 (Fig. 4d) all appeared to be qualita- balance between the interior and exterior of the liposomes, the
tively similar to CFI with respect to liposome size and lamellar- larger the loss of encapsulated ciprofloxacin. The combination
ity. However, with increasing concentration of polysorbate 20, of osmotic swelling and 2% polysorbate 20 led to more than 50%
there were a greater proportion of liposomes that were lighter free drug for 10 mg/mL CFI (Fig. 1b) in contrast to only 8% free
in shading suggesting a loss of encapsulated ciprofloxacin drug in the absence of osmotic swelling (Fig. 1a).
(Fig. 4d). There were also a very small fraction of disk-like The addition of polysorbate 20 (or polysorbate 80) causes
fragments, which may represent pieces of ruptured liposomes enhanced leakage of drug in a hypotonic liposome environ-
(Fig. 4d). To determine whether the lighter density liposomes ment. But does it also alter the release rate of the drug that
are consistent with loss of encapsulated drug, empty liposomes remains encapsulated? For low concentrations of polysorbate
(Fig. 4e), and a 50:50 mixture of empty liposomes and CFI, were 20, up to 0.2%, even in concert with osmotic swelling, there did
also imaged (Fig. 4f). The empty liposomes were comparable to not appear to be a meaningful effect on the IVR release rate for
CFI in terms of size and lamellarity, but had lighter density 12.5 mg/mL CFI (Fig. 3a). This result suggests that osmotic
(Fig. 4e). The 1:1 mixture of the empty liposomes and CFI re- swelling of liposomes, when simultaneously exposed to low con-
vealed a mosaic of liposomes with both light and dark shading centrations of surfactant, transiently allows enhanced release
in approximately equal proportion (Fig. 4f). These results sug- of drug, but without long-term effect on the bilayer properties
gest that the addition of increasing amounts of polysorbate 20, that could alter drug release in the IVR assay. In contrast, for
up to 0.4%, caused leakage of ciprofloxacin from a subset of higher concentrations of surfactant, for example, 0.4% polysor-
liposomes which retained their physical integrity after loss of bate 20 or 0.2% polysorbate 20, there was a significant increase
encapsulated drug. in the IVR release rate for 12.5 mg/mL CFI (Figs. 3a and 3b).
This result is consistent with the increased bilayer permeabil-
ity model.5,11
In studies of surfactant interactions with liposomes and
DISCUSSION
phospholipid vesicles, turbidity measurements of the system
We have investigated the interaction of surfactants with a often increase with surfactant concentration as surfactant in-
liposomal ciprofloxacin formulation with the goal to develop creasingly becomes associated with the bilayers, and this has
novel formulations with modified encapsulation states and re- been attributed to an increase in vesicle size.11,27 Once the vesi-
lease properties. Our strategy was to use sublytic quantities cles become saturated with surfactant, a decrease in turbidity
of surfactant, below the level that solubilizes the liposomes, is often observed because of the partial solubilization of the bi-
so that the liposome vesicles retain their integrity, but may layers and formation of mixed micelles which scatter less light.
have altered drug release rates.15,16 The addition of polysor- Although no turbidity measurements were made in our studies,
bate 20 resulted in a small loss of encapsulated drug when and our goal was to remain below the surfactant-solubilizing
mixed with CFI using isotonic HB as the dilution vehicle; concentration, we did observe a general trend of modest in-
the amount of released drug increased with greater concen- creases of vesicle size with increasing surfactant for a number
trations of surfactant, but even with 2% polysorbate 20 more of surfactants tested, consistent with this model (Table 1). The
than 92% of the ciprofloxacin remained encapsulated (Fig. 1a). increase in vesicle size was small for most surfactants but was
These data are consistent with the established mechanism the greatest for polysorbate 80 with an increase in mean vesicle
that sublytic concentrations of surfactant form transient pores size of up to 8 nm. This increase in vesicle size for polysorbate
in liposome bilayers allowing release of small quantities of 80 is likely related to its structure and its propensity to be
the encapsulated drug before the membrane barrier is fully more disruptive to the membrane bilayer. Although there may
recovered.15,16 be other explanations for this observation, the largest increase

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969

 RESEARCH ARTICLE – Pharmaceutical Nanotechnology 1859

Figure 3. Evaluation of the effect of liposomal ciprofloxacin composition on the IVR assay. The in vitro release of 25 :g/mL ciprofloxacin in 50%
(v/v) bovine serum and 10 mM HEPES buffered saline, pH 7.4 after incubation at 37◦ C for 4 h is reported. (a) IVR profiles for Control CFI (blue
diamonds), DRCFI (orange circles), 12.5 mg/mL CFI in 0.05% (w/v) polysorbate 20 (purple crosses), 12.5 mg/mL CFI in 0.1% (w/v) polysorbate
20 (green triangles), 12.5 mg/mL CFI in 0.2% (w/v) polysorbate 20 (blue stars), and 12.5 mg/mL CFI in 0.4% (w/v) polysorbate 20 (red squares).
Each value represents the mean (n = 2). (b) IVR profiles for Control CFI (blue diamonds), DRCFI (orange circles), 12.5 mg/mL CFI in 0.2% (w/v)
polysorbate 80 (red squares), and 12.5 mg/mL CFI in serum containing 0.0004% (w/v) polysorbate 20 (green stars). Each value represents the
mean (n = 2). Reprinted with permission from Ref. 21.

DOI 10.1002/jps.23969 Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014

1860 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

bate 20 (C16 vs. C10) and has a kink in the middle, unlike
polysorbate 20 that has a saturated hydrophobic tail. Thus, the
tail of the polysorbate 80 surfactant will be more disruptive
upon insertion into the liposome bilayer than the polysorbate
20 tail, and explains the greater effect that polysorbate 80 has
on the transient release of encapsulated drug prior to adjust-
ment by the lipids and cholesterol in the membrane to a new,
more-compact state. Once the membrane has accommodated
the surfactant, there does not appear to be any further release
of encapsulated drug (data not shown). For low concentrations
of polysorbate 20, up to 0.2%, there did not appear to be much
effect on the IVR release rate of 12.5 mg/mL CFI suggesting
that the membrane permeability and packing were not too per-
turbed (Fig. 3a). However, the addition of 0.4% polysorbate 20
to 12.5 mg/mL CFI led to a more rapid release of encapsulated
drug in the IVR assay (Fig. 3a). For this formulation, the ratio
of surfactant to liposomes was 1:6.25 by weight, which suggests
that the surfactant could represent up to approximately 14%
of the liposome bilayer on average, if all of the surfactant was
incorporated into the liposomes. One possible explanation for
this transition is that once the surfactant reaches a critical
level (between 0.2% and 0.4% for polysorbate 20 and less than
0.2% for polysorbate 80 in 12.5 mg/mL CFI), the cholesterol
and lipids may be less able to accommodate the surfactant,
and while the membrane retains its barrier properties to en-
capsulated drug, it may be more easily fluidized after mixing
with serum in the IVR assay. The packing of cholesterol and
lipids around the polysorbate 80 tail will be even less effective
than for the polysorbate 20 tail, thus explaining the more pro-
nounced disruption to drug encapsulation and IVR profile at
lower surfactant concentrations for polysorbate 80.
To determine whether these observations are broadly appli-
cable to other nonionic surfactants with similar chemistries, the
encapsulation studies were repeated using poloxamer L44 and
L62, and Sp20 and Sp80 surfactants. However, none of these
surfactants resulted in meaningful changes in the amount of
encapsulated drug. The Sp20 and Sp80 surfactants have iden-
Figure 4. Cryo-TEM micrographs of various preparations of liposo- tical hydrophobic tails to polysorbate 20 and polysorbate 80, re-
mal ciprofloxacin (CFI). The scale bar in the bottom left-hand corner spectively, but a different head group. These surfactants were
of each micrograph is 100 nm for a, b, e, and f and 200 nm for b and much less disruptive than the polysorbates, indicating that the
c. All CFI samples were applied at a concentration of ∼10 mg/mL li-
specific hydrophilic head group plays a key role in this process.
posomes. (a) 12.5 mg/mL ciprofloxacin; (b) 12.5 mg/mL ciprofloxacin in
0.05% (w/v) polysorbate 20; (c) 12.5 mg/mL ciprofloxacin in 0.2% (w/v)
For successful insertion of the surfactant tail into the bilayer,
polysorbate 20; (d) 12.5 mg/mL ciprofloxacin in 0.4% (w/v) polysorbate the surfactant head group apparently must successfully as-
20; (e) empty liposomes; (f) 1:1 mixture of empty liposomes and CFI. sociate with the hydrophilic head group of the phospholipid.
Both poloxamer surfactants, composed of a hydrophobic poly-
oxypropylene segment, flanked by two hydrophilic chains of
in vesicle size was also associated with the greatest loss in polyoxyethylene, had very little effect on drug encapsulation
encapsulated drug (Table 1) and the greatest effect on IVR re- (Fig. 2a), suggesting that they were also not effectively able to
lease profile (Fig. 3b), which may reflect upon the extent of the insert into the liposome bilayer.
disruption and resulting size increase of the liposomes due to Cryo-TEM studies were conducted to shed light on the con-
the association with surfactant. It is noteworthy that in the sequences of the surfactant interactions with the liposomes.
dynamic light scattering (DLS) studies, we did not observe a That so much more drug is released under hypotonic condi-
second population of smaller particles, which would have been tions raises the question of whether osmotic swelling is simply
consistent with the formation of surfactant micelles. It is pos- enhancing the transient release of encapsulated drug after ex-
sible that some micellar structures may have formed in these posure to surfactant, or if a large population of vesicles is being
studies, as DLS may have been insensitive to the presence of a completely disrupted. In other words, are a large fraction of li-
small percentage of micelles. posomes releasing all of their drug payload, or are all liposomes
Polysorbate 80 had a more pronounced effect than polysor- releasing some fraction of their payload, or is it a combina-
bate 20, both in terms of loss of drug encapsulation (Fig. 2a) tion of the two. The cryo-TEM micrographs of mixtures of CFI
and increase in IVR rate (Fig. 3b). Although the head groups containing low levels of polysorbate 20 are practically indistin-
are similar for both polysorbate surfactants, the hydrophobic guishable from CFI alone (Figs. 4a–4c). All three micrographs
tail for polysorbate 80 is almost twice as long as for polysor- show predominantly spherical, unilamellar liposomes around

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969

 RESEARCH ARTICLE – Pharmaceutical Nanotechnology 1861

50–100 nm in diameter. For these low levels of polysorbate 20 patient. The answer clearly would depend upon the specific dis-
added to CFI in a hypotonic environment, there does not appear ease being addressed, the properties of the therapeutic being
to be meaningful changes in liposome size, shape, or integrity. delivered, and the characteristics of the patient that demanded
Micellar structures were not observed in any of the liposomal a personalized regimen. Finally, how would a clinical trial be
samples with added surfactant. designed to demonstrate safety and efficacy when patients may
When 0.4% polysorbate 20 was added to 12.5 mg/mL CFI in be receiving formulations with different release profiles? This
a hypotonic environment, there was approximately 30% loss of may not be a totally foreign concept. The active drug is the
encapsulated drug. By cryo-TEM imaging, the majority of li- same for each patient so it may be possible to apply a strat-
posomes appeared unaltered in size and shape and there was egy similar to those for drugs that are dosed on a mg/kg basis.
minimal liposomal debris which might be indicative of vesicle In those studies, the total amount of drug is individualized to
rupture (Fig. 4d). However, a second population of liposomes the patient, based on their body mass, so why not modify the
emerged, which were comparable in size, shape, and lamellar- treatment based on other characteristics of a patient, like their
ity to the unmodified CFI liposomes, but had lighter shading need for a shorter or longer release profile? In the case of the
(Figs. 4d vs. 4a). These vesicles may have lost a substantial liposomal ciprofloxacin formulation in this study, it is being de-
fraction of their encapsulated drug, or all of it, as the density veloped as an aerosol therapy and has successfully completed
of their interior space is not dissimilar to that for empty lipo- Phase 2 trials in noncystic fibrosis bronchiectasis patients colo-
somes (Figs. 4d–4f). This result suggests that a second release nized with Pseudomonas aeruginosa.32 One might imagine that
mechanism may be occurring in the presence of higher concen- if a sputum sample from a patient indicated colonization with
trations of surfactant. Rather than simple leakage of a small a bacterial strain with a higher minimum inhibitory concen-
fraction of the drug content from some of the liposomes, a subset tration, then the most effective treatment might be a faster
of liposomes appear to have released all of their drug content release profile resulting in higher ciprofloxacin concentrations
and yet retained their vesicular shape. Very little liposomal de- in the lung to better kill the more resistant pathogen. The clin-
bris or deformed liposomes were observed that might be indica- ical trial design and its safety and efficacy endpoints may not
tive of liposome rupture. This is consistent with a mechanism be much changed from the standard situation when the formu-
that has been proposed that above a certain minimum surfac- lation is not individualized to the patient.
tant concentration, but below the solubilizing concentration,
the surfactant may cause an increase in the permeability of the
membranes without having any effect on their structure.5,11 CONCLUSIONS
However, that model does not explain why the majority of lipo-
We have demonstrated that the encapsulation state and IVR
somes appeared unaltered (Fig. 4d).
properties of a liposomal ciprofloxacin formulation (CFI) could
Although the use of surfactants to produce therapeutic nio-
be modified by the addition of polysorbate surfactant to the
somes has received much attention,28–30 there are few exam-
formed liposomes. Moreover, the effect of surfactant was signif-
ples in the literature where surfactant was added to phospho-
icantly enhanced if the liposomes were exposed to a hypotonic
lipid vesicles or liposomes to intentionally modify their physic-
environment, to induce osmotic swelling, at the time of surfac-
ochemical properties or modulate drug release. In one study,
tant addition. Osmotic swelling alone, in the absence of sur-
polysorbate 20, 60, or 80 was added to soy PC to produce unil-
factant addition, did not cause significant loss of encapsulated
amellar vesicles or multilamellar aggregates containing caf-
drug or a change in the IVR properties (data not shown). In the
feine as the model drug.31 There was no sustained release of
case for 12.5 mg/mL CFI containing 0.4% (w/v) polysorbate 20,
caffeine from the unilamellar liposomes, whether or not surfac-
or 0.2% (w/v) polysorbate 80, the encapsulation state was re-
tant was present, as the release profiles were identical to that
duced from >99% to approximately 70%. The 12.5 mg/mL CFI
for caffeine in solution.31 In contrast, in our study using unil-
preparation containing 0.4% (w/v) polysorbate 20, or 0.2% (w/v)
amellar liposomes, although the presence of small amounts of
polysorbate 80, also had altered IVR properties, with faster
polysorbate 20 or 80 had a minimal effect on the release rate,
rates of drug release, suggesting that the permeability of the
higher polysorbate concentrations, in combination with osmotic
bilayer was also affected. These studies suggest that it may one
swelling, caused a faster release profile in the IVR assay. An-
day be possible to tailor therapy to an individual, by dialing
other key difference is that the surfactant was added to intact
in the desired encapsulation state and release rate of a liposo-
liposomes in our study, whereas in the previous studies the
mal formulation by simple addition of an aliquot of a selected
surfactant was added prior to liposome manufacture.
surfactant, at a specific surfactant concentration, to the liposo-
Although these studies on modifying the release rate of a
mal drug product. These studies may form the basis for a new
liposomal formulation speak to the promise of tailoring ther-
paradigm of modifying the release profiles of a liposome formu-
apy to an individual’s needs, there are many challenges that
lation by the end user, to better meet their specific treatment
would need to be addressed before this approach could be real-
needs.
ized in practice. The regulatory pathway for this new treatment
paradigm is unclear and so would require discussion and agree-
ment by the innovator with the Regulatory Agencies. As the
ACKNOWLEDGMENTS
personalization involves relatively minor qualitative changes
in the formulation, and its primary purpose is to modulate the The authors acknowledge the facilities and the scientific and
PK profile with the view to optimize the treatment efficacy and technical assistance of the Australian Microscopy & Micro-
safety for each patient, this approach seems much easier to analysis Research Facility at the Electron Microscope Unit,
regulate than introduction of products with fundamentally dif- The University of New South Wales, and in particular, Judy
ferent compositions or manufacturing processes. Another hur- Loo (Ching -Yee Loo) and Delfine Cheng who provided assis-
dle would be how to determine the ideal release profile for each tance with the cryo-TEM analysis of the samples. The authors

DOI 10.1002/jps.23969 Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014

1862 RESEARCH ARTICLE – Pharmaceutical Nanotechnology

recognize Paul Young and Daniela Traini of the Woolcock 16. Memoli A, Annesini MC, Petralito S. 1999. Surfactant-induced leak-
Institute of Medical Research, and Francis Dayton, Sujata age from liposomes: A comparison among different lecithin vesicles. Int
Mudumba and Jim Blanchard of Aradigm Corporation for help- J Pharm 184(2):227–235.
ful discussions. 17. Kraft JC, Freeling JP, Wang Z, Ho RJY. 2014. Emerging research
and clinical development trends of liposome and lipid nanoparticle drug
Conflict of interest: David Cipolla, Huiying Wu, and Igor
delivery systems. J Pharm Sci 103(1):29–52.
Gonda are currently employed by Aradigm Corporation. Simon
18. Ong HX, Traini D, Cipolla D, Gonda I, Bebawy M, Agus H, Young
Eastman is currently employed by Northern Lipids, Inc. Hak- PY. 2012. Liposomal nanoparticles control the uptake of ciprofloxacin
Kim Chan declared that no conflict of interest exists. across respiratory epithelia. Pharm Res 29(12):3335–3346.
19. Yim D, Blanchard JD, Mudumba S, Eastman S, Manda K, Re-
delmeier T, Farr S. 2006. The development of inhaled liposome-
REFERENCES
encapsulated ciprofloxacin to treat cystic fibrosis. In Respiratory drug
1. Drummond DC, Noble CO, Hayes ME, Park JW, Kirpotin DB. 2008. delivery 2006; Dalby RN, Byron PR, Peart J, Suman JD, Farr SJ, Eds.
Pharmacokinetics and in vivo drug release rates in liposomal nanocar- River Grove, Illinois: Davis Healthcare Int’l Publishing, pp 425–428.
rier development. J Pharm Sci 97:4696–4740. 20. Webb MS, Boman NL, Wiseman DJ, Saxon D, Sutton K, Wong
2. Cipolla D, Gonda I, Chan HK. 2013. Liposomal formulations for in- KF, Logan P, Hope MJ. 1998. Antibacterial efficacy against an in
halation. Ther Deliv 4(8):1047–1072. vivo Salmonella typhimurium infection model and pharmacokinetics of
3. Cipolla D, Redelmeier T, Eastman S, Bruinenberg P, Gonda I. 2011. a liposomal ciprofloxacin formulation. Antimicrob Agents Chemother
Liposomes, niosomes and proniosomes—A critical update of their (com- 42(1):45–52.
mercial) development as inhaled products. In Respiratory drug delivery 21. Cipolla D, Wu H, Eastman S, Redelmeier T, Gonda I, Chan H-K.
Europe 2011; Dalby RN, Byron PR, Peart J, Suman JD, Farr SJ, Young 2014. Development and characterization of an in vitro release assay for
PM, Eds. River Grove, Illinois: Davis Healthcare Int’l Publishing, pp liposomal ciprofloxacin for inhalation. J Pharm Sci 103(1):314–327.
41–54. 22. Cipolla D, Dayton F, Fulzele S, Gabatan E, Mudumba S, Yim D, Wu
4. Helenius A, Simons K. 1975. Solubilization of membranes by deter- H, Zwolinski R. 2010. Inhaled liposomal ciprofloxacin: In vitro proper-
gents. Biochim Biophys Acta 415(1):29–79. ties and aerosol performance. In Respiratory drug delivery 2010; Dalby
5. Lichtenberg D, Robson RJ, Dennis EA. 1983. Solubilization of phos- RN, Byron PR, Peart J, Suman JD, Farr SJ, Young PM, Eds. River
pholipids by detergents. Structural and kinetic aspects. Biochim Bio- Grove, Illinois: Davis Healthcare Int’l Publishing, pp 409–414.
phys Acta 737(2):285–304. 23. Edwards K, Johnsson M, Karlsson G, Silvinder M. 1997. Effect of
6. Velluto D, Gasbarri C, Angelini G, Fontana A. 2011. Use of sim- polyethyleneglycol-phospholipids on aggregate structure in prepara-
ple kinetic and reaction-order measurements for the evaluation of tions of small unilamellar liposomes. Biophys J 73(1):258–266.
the mechanism of surfactant–liposome interactions. J Phys Chem B 24. Ghosh P, Singh UN. 1992. Liposome shrinkage and swelling under
115(25):8130–8137. osmotic-diffusional stress: Evaluation of kinetic parameters from spec-
7. Cho W, Wu SK, Yoon E, Lichtenbergova L. 1999. Fluorometric phos- trophotometric measurements. Biochim Biophys Acta 1110(1):88–96.
pholipase assays based on polymerized liposome substrates. Methods 25. De Gier J, Mandersloot JG, Van Deenen LLM. 1968. Lipid composi-
Mol Biol 109:7–17. tion and permeability of liposomes. Biochim Biophys Acta 150(4):666–
8. Paternostre MT, Roux M, Rigaud JL. 1988. Mechanisms of mem- 675.
brane protein insertion into liposomes during reconstitution procedures 26. Borochov A, Borochov H. 1979. Increase in membrane fluidity in li-
involving the use of detergents. 1. Solubilization of large unilamellar li- posomes and plant protoplasts upon osmotic swelling. Biochim Biophys
posomes (prepared by reverse-phase evaporation) by triton X-100, octyl Acta 550(3):546–549.
glucoside, and sodium cholate. Biochemistry 27(8):2668–2677. 27. Thoren PE, Soderman O, Engstron S, Von Corswant C. 2007. Inter-
9. Ribosa I, Garcia MT, Parra JL, Maza A, Sanchez-Leal J, Trullas C, actions of novel, nonhemolytic surfactants with phospholipid vesicles.
Tsi A, Balaguer F, Pelejero C. 1992. Physico-chemical modifications of Langmuir 23(13):6956–65.
liposome structures through interaction with surfactantsy. Int J Cos- 28. Terzano C, Allegra L, Alhaique F, Marianecci C, Carafa M. 2005.
met Sci 14(3):131–149. Non-phospholipid vesicles for pulmonary glucocorticoid delivery. Eur J
10. Lasch J. 1995. Interaction of detergents with lipid vesicles. Biochim Pharm Biopharm 59(1):57–62.
Biophys Acta 1241(2):269–292. 29. Desai TR, Finlay WH. 2002. Nebulization of niosomal all-trans-
11. Deo N, Somasundaran P. 2003. Effects of sodium dodecyl sulfate on retinoic acid: An inexpensive alternative to conventional liposomes. Int
mixed liposome solubilization. Langmuir 19(18):7271–7275. J Pharm 241(2):311–317.
12. Almgren M, Edwards K, Karlsson G. 2000. Cryo transmission elec- 30. Elhissi A, Hidayat K, Phoenix DA, Mwesigwa E, Crean S, Ahmed
tron microscopy of liposomes and related structures. Colloids Surf A W, Faheem A, Taylor KMG. 2013. Air-jet and vibrating-mesh nebu-
Physicochem Eng Aspects 174(1–2):3–21. lization of niosomes generated using a particulate-based proniosome
13. Nagawa Y, Regen SL. 1992. Surfactant-induced release from phos- technology. Int J Pharm 444(1–2):193–199.
phatidylcholine vesicles. Regulation of rupture and leakage pathways 31. Fadda AM, Baroli BM, Maccioni AM, Siniso C, Valenti D, Alhaique
by membrane packing. J Am Chem Soc 114(5):1668–1672. F. 1998. Phospholipid-detergent systems: Effects of polysorbates on the
14. Liu Y, Regen SL. 1993. Control over vesicle rupture and leakage release of liposomal caffeine. Il Farmaco 53(10–11):650–654.
by membrane packing and by the aggregation state of an attacking 32. Serisier DJ, Bilton D, De Soyza A, Thompson PJ, Kolbe J, Greville
surfactant. J Am Chem Soc 115(2):708–713. HW. Cipolla D Bruinenberg P Gonda I. 2013. Inhaled, dual release
15. Ruiz J, Goni FM, Alonso A. 1988. Surfactant-induced release of liposomal ciprofloxacin in non-cystic fibrosis bronchiectasis (ORBIT-2):
liposomal contents. A survey of methods and results. Biochim Biophys A randomised, double-blind, placebo-controlled trial. Thorax 68(9):812–
Acta 937(1):127–134. 817.

Cipolla et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:1851–1862, 2014 DOI 10.1002/jps.23969