cells

Review
Metabolic Profiling of CHO Cells during the Production
of Biotherapeutics
Mathilde Coulet 1,2 , Oliver Kepp 2,3 , Guido Kroemer 2,3,4, * and Stéphane Basmaciogullari 1, *

1 Sanofi R&D, 94400 Vitry-sur-Seine, France; [email protected]

2 Metabolomics and Cell Biology Platforms, Gustave Roussy Cancer Center, 94800 Villejuif, France;
[email protected]
3 Institut Universitaire de France, Centre de Recherche des Cordeliers, Equipe Labellisée par la Ligue Contre le Cancer,
Université de Paris Cité, Sorbonne Université, Inserm U1138, 75006 Paris, France
4 Department of Biology, Institut du Cancer Paris CARPEM, Hôpital Européen Georges Pompidou, AP-HP,
75015 Paris, France
* Correspondence: [email protected] (G.K.); [email protected] (S.B.)

Abstract: As indicated by an ever-increasing number of FDA approvals, biotherapeutics constitute

powerful tools for the treatment of various diseases, with monoclonal antibodies (mAbs) accounting
for more than 50% of newly approved drugs between 2014 and 2018 (Walsh, 2018). The pharmaceutical
industry has made great progress in developing reliable and efficient bioproduction processes to
meet the demand for recombinant mAbs. Mammalian cell lines are preferred for the production
of functional, complex recombinant proteins including mAbs, with Chinese hamster ovary (CHO)
cells being used in most instances. Despite significant advances in cell growth control for biologics
manufacturing, cellular responses to environmental changes need to be understood in order to further
improve productivity. Metabolomics offers a promising approach for developing suitable strategies to
unlock the full potential of cellular production. This review summarizes key findings on catabolism
Citation: Coulet, M.; Kepp, O.; and anabolism for each phase of cell growth (exponential growth, the stationary phase and decline)
Kroemer, G.; Basmaciogullari, S. with a focus on the principal metabolic pathways (glycolysis, the pentose phosphate pathway and
Metabolic Profiling of CHO Cells the tricarboxylic acid cycle) and the families of biomolecules that impact these circuities (nucleotides,
during the Production of amino acids, lipids and energy-rich metabolites).
Biotherapeutics. Cells 2022, 11, 1929.
https://doi.org/10.3390/cells11121929 Keywords: immunotherapy; monoclonal antibodies; metabolomics; industrial production;
Academic Editors: Ludovica Cacopardo,
process optimization
Sandeep Keshavan and Bharath
Babu Nunna

Received: 4 May 2022

1. Introduction
Accepted: 13 June 2022
Published: 15 June 2022
In 1986, Muromumab became the first FDA-approved mAb for the prevention of
kidney transplant rejection. Since 1990, and especially since the approval of the first
Publisher’s Note: MDPI stays neutral fully human antibody in 2004, the number of mAbs available in the pharmacopeia has
with regard to jurisdictional claims in increased. Biotherapeutics have been used for the treatment of a variety of medical condi-
published maps and institutional affil-
tions including cancers, organ transplants and autoimmune, cardiovascular, respiratory
iations.
and neurological diseases, which is witnessed by the steady increase in the number of
FDA-approved treatments [1]. The pharmaceutical industry has made great progress in
developing reliable and efficient bioproduction processes that meet the demand for new
Copyright: © 2022 by the authors.
biotherapeutics including monoclonal antibodies (mAbs). Sixty-eight new mAbs were
Licensee MDPI, Basel, Switzerland.
approved between 2014 and 2018, thus representing more than 50% of all new biopharma-
This article is an open access article
ceutical products on the market, and mAbs represented the most lucrative product class,
distributed under the terms and with a revenue of $123 billion registered in 2017 [2].
conditions of the Creative Commons mAbs are complex structures and require proper folding, assembly and post-translational
Attribution (CC BY) license (https:// modifications, such as glycosylation, to ensure their functionality and efficacy. Unlike microbial
creativecommons.org/licenses/by/ systems that are limited in post-translational modifications, mammalian cell lines are apt to
4.0/).

Cells 2022, 11, 1929. https://doi.org/10.3390/cells11121929 https://www.mdpi.com/journal/cells

Cells 2022, 11, 1929 2 of 21

generate diverse functional recombinant glycoproteins, with Chinese Hamster Ovary (CHO)
cells being used for the production of 84% of FDA-approved biotherapeutics in 2018 [2].
The culture of CHO cells in bioreactors is a tightly regulated process that has seen
significant advances since the 90s [3]. The steps involved in the development of a cell culture
production process include the design and the selection of a stable protein-secreting cell line,
the optimization of media and culture operating conditions at a small scale using medium-
to high-throughput screening methods and finally the upscaling of the process, which must
comply with good manufacturing practices (GMPs) [4,5]. Despite major advances in cell
culture processes and the optimization of cell culture media, the cellular response to changes
in the culture environment and the subsequent impact on cell growth and productivity needs
to be better understood in order to fully optimize the production of biologics.
In this regard, multidimensional ‘omics’ approaches are powerful tools used to im-
prove our knowledge and to design suitable strategies for unlocking the full potential
of CHO cells. Genomics, proteomics, transcriptomics and metabolomics are recent and
complementary fields of study that have been instrumental in deciphering biological mech-
anisms influencing cell growth in bioreactors. Metabolomics is a promising approach in
the bioproduction field, as it detects the downstream products of the other ‘omics’ sciences
(genomics, transcriptomics and proteomics for the characterization of DNA, RNA and
proteins/enzymes, respectively) and is believed to accurately mirror the cellular pheno-
type. Metabolomics involves the intracellular and extracellular quantification of small
molecules called metabolites [6], concentrations of which strongly vary as a function of
cellular responses to environmental changes [7]. The standardized preparation of the sam-
ples, be it debris-free culture supernatants or washed cells, is the first step of metabolomic
analysis. This is followed by the metabolic profiling of the samples either by nuclear
magnetic resonance (NMR) or mass spectrometry (MS). A bioinformatic analysis then
connects metabolites to known metabolic pathways and quantifies metabolic fluxes [8,9].
Two distinct methodologies called untargeted and targeted metabolomics are widely used
for this purpose. Untargeted metabolomics is an unbiased analysis measuring all detectable
metabolites present in the sample and can enable the discovery of new molecules impacting
cell metabolism. Targeted metabolomics focuses on groups of known metabolites that are
chemically defined. Both methods are complementary and are often used in combination to
gain maximum insights [10]. Metabolomics analysis can also be coupled with the metabolic
labeling of the cell substrate by adding non-radioactive isotope tracers (usually synthetic
metabolites labelled with carbon-13, 13 C) to the culture media prior to mass spectrometry
analysis. This combined approach allows for metabolic flux analysis (MFA) and the calcula-
tion of reaction rates as well as the identification of precursor–product relationships among
metabolites [11].
The kinetic analysis of the metabolome of cells at different time points of the cell
culture is key to the improvement of processes. This is particularly true for the detection
of the appearance and/or exhaustion of potential fate-changing molecules at the switch
from the exponential growth phase to the stationary phase, which is characterized by low
proliferation, and the final decline of the culture, which is characterized by a high rate of cell
death. Many studies managed to identify such markers [11–19] and this paved the way for
an increase in production yield through the genetic engineering of cells or by changing the
medium composition and culture parameters [20–22]. However, the metabolomic studies
reported in the literature are heterogeneous in their design and technology (Supplementary
Table S1), rendering the direct comparison of results difficult.
This review aimed to discuss the results obtained by mass spectrometry metabolomics
and 13 C MFA studies of CHO cells listed in Supplementary Table S1. Key results ob-
tained for each phase of growth (exponential growth, stationary phase, decline) regarding
catabolism and anabolism are presented. Emphasis is placed on the main metabolic path-
ways of mammalian cells (glycolysis, the pentose phosphate pathway (PPP) and the tricar-
boxylic acid (TCA) cycle) and on the families of biomolecules that impact these pathways
(nucleotides, amino acids, lipids and energy-rich molecules). Key changes marking the
Cells 2022, 11, x 3 of 20

Cells 2022, 11, 1929 3 of 21

amino acids, lipids and energy-rich molecules). Key changes marking the switch from expo-
nential growth to the stationary phase and then to decline are highlighted. The results ob-
tained in
switch distinct
from experimental
exponential growthsettings are also compared
to the stationary phase and(various
then to cell lines,
decline areclones, media
highlighted.
and results
The culture obtained
strategiesin
including batch, fed-batch
distinct experimental and continuous)
settings to identify
are also compared common
(various celltrends
lines,
affecting
clones, the CHO
media and cell metabolome
culture strategiesduring culture
including in bioreactors.
batch, fed-batch and continuous) to identify
common trends affecting the CHO cell metabolome during culture in bioreactors.
2. The Exponential Growth Phase: High Nutrient Uptake Enabling Cell Growth
2. The Exponential
The exponentialGrowth Phase:
phase is High Nutrient
characterized Uptake
by a steep Enabling
increase in theCell Growth
viable cell density
(VCD)Theinexponential phase
the cell culture. is characterized
This by a steep
growth is possible increase
because in theare
nutrients viable cell density
available in the
(VCD)
culturein the cell from
medium, culture. This(1)
which growth is possible
catabolism because
generates nutrients
energy and (2)are available
various in the
substrates
culture
are usedmedium, from
for biomass which (1) Both
generation. catabolism generates
steps will energyinand
be described this(2) various
section, andsubstrates
the main
are used for biomass generation. Both steps
metabolic trends are summarized in Figure 1. will be described in this section, and the main
metabolic trends are summarized in Figure 1.

Figure 1.
Figure 1. Global view of
Global view of the
the central
central CHO
CHO cell
cell metabolism
metabolism at
at the
the exponential
exponential phase
phase of
of growth.
growth.
Cells 2022, 11, 1929 4 of 21

2.1. Carbon Source Catabolism Generates Energy for the Cells as Nutrients Are Largely Available (Step 1)
MFA studies have shown that glucose contributes to 70% of the total carbon influx,
while pyruvate from medium serves as an additional carbon source at the beginning of the
culture, with 80% of it being consumed in the first day of culture [18,19]. Similarly, another
study concluded that glucose represents 65% of the entering carbon, with 10% coming from
glutamine and 25% from other amino acids [23]. A lower contribution was also reported
(40–50%) [24,25], although this can partly be explained by the fact that glucose consumption
can vary simply due to different culture conditions, specifically, in the present case, culture
volumes [26]. Besides these subtle differences, publications overall agree on the fact that
glucose constitutes the main carbon source for growth.
Glucose taken up by cells can be phosphorylated and supplied to the glycolysis path-
way for ATP production or to the pentose phosphate pathway (PPP), which contributes to
redox homeostasis and biosynthesis. Numerous studies report that during the exponential
phase, glucose is mainly oxidized via glycolysis, which results in the formation of pyruvate.
A significant portion of the pyruvate formed is converted into lactate, which is secreted
and acidifies the medium, resulting in growth inhibition, with the rest being used to supply
the TCA cycle [11,16,19,23,24,26–37]. This metabolic state was previously recognized as the
Warburg effect and characterizes cancer cells that consume high levels of glucose. Based
on the potential energy load of glucose, it can be considered that, in the sense, the carbon
flux is “wasted” by the cells, in that it is not targeted to the TCA and yields high lactate
levels instead. This lowers the production of the C4-6 precursors necessary for biomass
generation [38,39].
It has been estimated that 75% of glucose consumption goes toward lactate production
in cultures of adherent and suspension-growing CHO cells [11,33,40]. In perfusion cultures,
however, a more balanced proportion has been measured, with 55% of the pyruvate being
estimated to generate lactate and 45% entering the TCA cycle [32]. On the contrary, MFA
studies on batch/fed-batch cultures report that 25–40% of cytosolic pyruvate is converted
into lactate by the action of lactate dehydrogenase, an effect that is observed for both CHO
WT cells and protein-producing CHO derivatives [18,23,41], while 50–60% of the pyruvate
pool is channeled into the TCA cycle. These discrepancies might be attributed to different
media compositions. For instance, it has been shown that glutamine provided in the media
can strongly influence glucose and pyruvate metabolism [42]. It was also shown that the
overexpression of bcl2, an anti-apoptotic gene, can influence the proportion of pyruvate
entering the TCA and, as a result, lactate formation [34]. A number of studies also analyzed
the differences between high and low producer clones. Certain studies revealed that
(i) high producers possess higher levels of intracellular NADH, suggesting a higher level
of glycolysis in combination with the TCA cycle and/or oxidative phosphorylation [43],
and (ii) some differences exist in the timing of glucose consumption and lactate formation
between high and low producer cell lines, with high producers consuming lactate earlier
in the culture compared to low producers [19]. Only minor differences were reported in
another study, suggesting that further investigations are needed to clarify how productivity
might be correlated to the early phases of glucose metabolism [44].
At the pyruvate branch point, it was estimated that about 75% of pyruvate comes
from glycolysis, while 25% results from the conversion of malate by the malic enzyme and
a minor fraction from the degradation of amino acids such as serine or cysteine [40–42].
A total of 90% of pyruvate flux would enter in the TCA following its conversion to acetyl-
coA, with the remaining 10% entering the TCA cycle after conversion into oxaloacetate by
pyruvate carboxylase [32,33,42].
At the late exponential phase, a decrease by one-third in terms of glucose consumption
and glycolytic activity was measured in favor of lactate uptake [18,37]. Interestingly, it was
possible to prolong the exponential phase by the adding pyruvate and amino acids to the
medium [16].
The pentose phosphate pathway (PPP) represents an alternative route of glucose oxida-
tion that regenerates the NADPH that contributes to the redox balance in the cytoplasm and
Cells 2022, 11, 1929 5 of 21

C5 sugars that are involved in biosynthetic reactions, in particular nucleotide biosynthesis.

Pentose phosphate pathway (PPP) activity has been characterized during the exponential
phase, and would appear that its contribution to the overall carbon flux depends on the ex-
perimental models. In the perfusion mode, the proportion of glucose entering the PPP was
estimated at 20–40% [32], while studies conducted in batch/fed-batch conditions reported
a very low or negligible contribution from glucose to this pathway [11,33,34,40]. This is in
contradiction with one study conducted on fed-batch cultures, which reported that 80% of
glucose was engaged in the PPP pathway, accounting for most of the cytosolic NADPH
generation during this growth phase [23]. Thus, these studies reported major differences in
terms of glucose utilization between glycolysis and PPP, perhaps reflecting variations in
the composition of media, including in terms of glutamine levels [42], medium supply in
fed-batch cultures, specific clones or yields in the production of recombinant proteins.
Glycolysis oxidative reactions fed by glucose or PPP intermediate species result in
the generation of pyruvate that can enter the TCA cycle as acetyl-CoA. As mentioned
above, TCA cycle fluxes are low during the exponential phase, with a large portion of
pyruvate being used for lactate production [11,13,17,34]. As a possibility, any overabundant
glucose available during initial cell culture might be used for energy maintenance through
lactate production rather than for the TCA cycle and oxidative phosphorylation. It was also
hypothesized that the TCA cycle might be impaired in transformed CHO cell lines due to a
reduced capacity to convert citrate into α-ketoglutarate (AKG) [16,45]. Another bottleneck
might exist between malate and oxaloacetate due to a limited enzyme capacity, which would
result in limited malate-aspartate shuttling and, hence, NAD regeneration in the cytosol. In
addition, this limitation in oxaloacetate supply prevents the incorporation of mitochondrial
pyruvate into the TCA cycle, resulting in the conversion of pyruvate conversion into lactate
in the cytoplasm [44]. Although lactate production is significant, as mentioned earlier, the
TCA is active, with the main carbon source coming from glucose oxidation to pyruvate
and the rest from glutamine and essential amino acids [23]. Interestingly, it has been
shown that the secretion of TCA intermediates (succinate, malate, citrate and fumarate)
represents approximately the same level of TCA cycle flux during the exponential growth
and stationary phases, possibly indicating that the TCA cycle is operating at close to
maximal enzymatic capacity throughout cell culture [44]. However, this conclusion was
reached upon using 1 H-NMR, which is comparatively less sensitive than MS. It has also
been suggested that the culture scale impacts the metabolism of cells. Cells growing in
production-scale bioreactors may rely more on glycolysis for energy production than cells
growing at lab scale that synthesize more TCA cycle intermediates [46].
Regarding nucleotides and nucleosides at the exponential phase, little information
is available in the literature. Up to 90% of ATP production has been suggested to result
from the TCA cycle [41], with a downward trend along the exponential phase [17]. While
ADP and AMP concentrations were found to be similar in the cytosol and mitochondria,
ATP concentration was higher in the cytosol than in the mitochondria, suggesting that the
transport of ATP out of the mitochondria might render ATP formation thermodynamically
efficient [30]. At this phase of growth, when comparing high and low producer clones,
steady-state ATP concentrations are comparable, but the production and consumption rates
correlate with productivity in Escherichia coli and CHO cells [43]. The addition of adenosine
monophosphate (AMP) or guanosine monophosphate (GMP) to the culture media led to a 3-
fold induction of caspase activity, the highest observed for all metabolites tested, contrasting
with the addition of cytidine monophosphate (CMP) and uridine monophosphate (UMP),
which had no effects [47]. This pro-apoptotic effect was also reported in other studies
on CHO cells [48] and IEC-6 intestinal epithelial cells [49]. This means that, even in the
presence of nutrients and in the absence of high levels of ammonia and lactate, toxic
metabolites can cause cell death.
The consumption and production of amino acids has been studied in CHO cells during
the exponential growth phase (Table 1), and amino acid-relevant metabolic fluxes were
described to be significantly lower than glycolytic and TCA cycle fluxes [32]. MFA studies
Cells 2022, 11, 1929 6 of 21

reported consistent percentages in terms of the contributions of amino acid to TCA cycle
replenishment, with glutamine being the main carbon source, and it is known that its
availability in the culture media, together with asparagine or serine, has a high impact on
the metabolism of CHO cells [25,50,51]. Glutamine was found to contribute to about 40%
of the total carbon supply of the TCA cycle. The mitochondrial glutamate pool originates
from both the transport of cytosolic glutamate into the mitochondria and from the transport
of glutamine into the mitochondria followed by hydrolysis to glutamate. Glutamate is
converted into α-ketoglutarate by the mitochondrial glutamate dehydrogenase, which is
fed into the TCA cycle [23]. All the other amino acids contribute about 10% and enter the
TCA cycle at either the acetyl-CoA, α-ketoglutarate or oxaloacetate branch point. Very
similar amino acid catabolic contributions were reported in induced and non-induced
cells [41]. For low and high producer clones, consistent results were also observed, with
asparagine and glutamine generating about 30% of the citrate pool and 50% of malate and
succinate labelling, suggesting their major role in ATP generation and the replenishment of
the TCA cycle at the exponential phase of growth but not biomass production [19]. The
total contribution of amino acids together with glucose to incoming carbon flux is thought
to remain constant over the entire culture (between 30% and 50% of the total carbon) with
the uptake of other amino acids increasing after glutamine is depleted [18]. Another model
estimated that 35% of the central carbon metabolism is fed by amino acid catabolism,
with glutamine representing 10% [23]. More precisely, 60% of the aspartate formed was
transaminated to yield oxaloacetate during exponential growth, decreasing to 50% in the
stationary phase. Alanine was produced for nitrogen detoxification in the exponential
phase, corresponding to 6% of pyruvate usage. On average, 25% of the essential amino acids
consumed were not supplied for biomass or IgG production during exponential growth
and instead generated TCA cycle intermediates [44]. This interpretation is supported by a
second study in which the rate of glutamine uptake during the early exponential phase
greatly exceeded the biosynthetic demand for biomass or antibody production, which
reflects its use in catabolic energy production [18].
A study comparing cultures fed with high or low concentrations of glutamine showed
the importance of this metabolite, as major differences in pathway fluxes were observed.
These variations impacted strongly on the specific productivity, which was 2-fold higher
for cultures fed with low concentrations of glutamine, and showed how sensitive CHO cell
metabolism is to glutamine levels [42].
The replenishment of the TCA by amino acids generates intermediate molecular
species. At the end of the exponential phase, when cell density is high and lactate concen-
tration is low, these intermediates tend to accumulate, and their inhibitory effect on cell
growth has been demonstrated, with indole 3-carboxylate and isovalerate being the most
potent (refer to Tables 1–3 for more details) [21].
In addition to their role in the replenishment of the TCA and protein biosynthesis,
amino acids serve as glutathione precursors. Their metabolism thus has a major impact on
cell protection against oxidative stress caused by ROS [54]. In this respect, glutathione is
one of the key metabolites to consider, as the equilibrium between its oxidized (GSSG) and
reduced (GSH) forms influences cell growth. Both forms are present at high concentrations
during the early exponential phase at a 1:5 ratio. The GSSG/GSH ratio is then balanced
during the exponential phase [31,55]. Consistently, several publications reported the accu-
mulation of GSSG in media during the exponential growth phase, which then continued
during the stationary phase. GSSG is known to be a marker of oxidative stress and to
induce apoptosis, presumably explaining its correlation with caspase activity [12,17,20,47].
An intracellular decrease in GSSG and GSH concentrations has also been reported [17,18].
This is consistent with the accumulation of GSSG observed in media as it is excreted by
cells, and extracellular GSSG might then account for cell death in prolonged cultures [38].
Additionally, it has been suggested that the decrease in GSH concentration observed during
the late exponential phase together with the decrease in NADPH concentration triggers an
increase in PPP activity before the peak of TCA cycle fluxes. In favor of this hypothesis,
Cells 2022, 11, 1929 7 of 21

a close correlation between PPP and TCA activities was observed during the culture [18].
Based on a study focused on the redox status in the cell, it has also been suggested that cell
ageing is not responsible for the increase in glutathione metabolism and ROS management,
but that high cell densities are [54].

Table 1. Amino acids and amino acid derivatives at the exponential phase.

Amino Acid Derivative Behavior Reference Comments

Intracellular production and
[11,14,16–18,32,36,37,41,42,44]
accumulation in media
Alanine
Production mainly from
[23]
cytosolic pyruvate
Intracellular consumption leading
[12,20,32,41]
to concentration decrease in media Different conclusion might be
Arginine Accumulation from due to feeding method and
late-exponential phase onwards, medium composition that is
[17] specific to [17]
indicating over-supply
in the fed-batch
Intracellular consumption leading
[17,23,25,32,36,37,41]
to concentration decrease in media
The most consumed amino acid
from media. Intracellular
[44]
deamination to aspartic acid
generates ammonia
Consumption (measures performed
on cell substrate). Represents 5% of
[18]
Asparagine incoming carbon source. Linked to
aspartic acid production
Intracellular concentration increase
in early exponential phase
In terms of uptake from the media,
the highest compared to other
growth phases. In high producer [19,31]
clone, consumption beyond
stoichiometric requirements
together with glutamine to
replenish the TCA intermediates
Production linked to asparagine
uptake (measures performed [18,31,41,42]
on cell substrate)
Aspartic acid Consumption (measures performed
[32]
on cell culture supernatant)
Consumption (measures performed
[23,44]
on cell substrate)
Consumption (measures performed
Cysteine [41]
on cell culture supernatant)
Cells 2022, 11, 1929 8 of 21

Table 1. Cont.

Amino Acid Derivative Behavior Reference Comments

Cells use more glutamine when
cultivated in a media containing [25,42,52]
more glutamine
Consumption during all culture
phases, with the highest during the
[11,19,32,36,37]
exponential phase; source of
lactate formation

Glutamine Main amino acid consumed [23,41]

Consumption linked to glutamic
[18]
acid production
Uptake from the media, the highest
compared to other growth phases.
In high producer clone,
consumption beyond stoichiometric [19]
requirements together with
asparagine to replenish the
TCA intermediates
Intracellular production leading to
[11,23,31,32,36,41,42]
increased concentration in media
Glutamic acid
Production associated with
[18,40]
glutamine uptake
Accumulation in media
[14,16,17,23,41,42]
during culture
Accumulation in media during The low uptake identified in
culture; produced [44] [32] might be due to the
Glycine by serine catabolism analysis method (NMR) used,
Low uptake from media [32] which is different from that
used in other studies focused
Its accumulation in media together
on glycine metabolism
with the other identified
[21]
metabolites leads to growth
inhibition in HIPDOG culture
Histidine Uptake from media [32,41]
Isoleucine Uptake from media [23,32,41]
Uptake from media [23,32,41] The particularity of study [21]
Its accumulation in media together is the use of the specific
Leucine with the other identified HiDPOG culture process that
[21] might explain
metabolites leads to growth
inhibition in HIPDOG culture the difference observed
Lysine Uptake from media [23,32,41]
Uptake from media [23,32,41] The particularity of study [21]
Its accumulation together with the is the use of a specific
Methionine other identified metabolites leads to HiDPOG culture process that
[21] might explain the difference
growth inhibition in
HIPDOG culture observed
Uptake from media [12,20,23,32,41] The particularity of study [21]
Its accumulation in media together is the use of a specific
Phenylalanine with the other identified HiDPOG culture process that
[21] might explain the difference
metabolites leads to growth
inhibition in HIPDOG culture observed
Cells 2022, 11, 1929 9 of 21

Table 1. Cont.

Amino Acid Derivative Behavior Reference Comments

Consumption (measures performed
Cysteine [41]
on cell culture supernatant)
Accumulation in media [32] In study [41], a medium
different from that of the two
Proline Slight increase in concentration in
[11] other studies was used,
media with time
potentially explaining
Uptake from media [41] this discrepancy
An increase in the intracellular
concentration during the early [31]
exponential phase The particularity of study [21]
is the use of a specific
Serine Uptake from media [23,32,41,44] HiDPOG culture process that
Its accumulation together with the might explain the difference
other identified metabolites leads to observed
[21]
growth inhibition in
HIPDOG culture
Threonine Uptake from media [23,32,41]
Uptake from media [32]
Depleted in media despite constant The particularity of study [21]
[12,20]
addition of feed is the use of a specific
Tryptophan HiDPOG culture process that
Its accumulation together with the
other identified metabolites leads to might explain the difference
[21] observed
growth inhibition in
HIPDOG culture
Uptake from media [23,32,41] The particularity of study [21]
Its accumulation together with the is the use of a specific
Tyrosine other identified metabolites leads to HiDPOG culture process that
[21] might explain the difference
growth inhibition in
HIPDOG culture observed
Valine Uptake from media [23,32,41]
Amino acid derivative Behavior Reference Comments
Accumulation in media; known to
Aspartylphenylalanine [12,20]
be toxic
Glutamylalanine Accumulation in media [12,20]
Accumulation in media; known to
Glutamylphenylalanine [12,20]
be detrimental to cell growth
Formylmethionine Accumulating in media [12,20]
5-L-glutamyl-L-alanine Accumulation in media [17]
Accumulation in media; associated
Dimethyl-L-arginine [17,53]
with apoptosis
N-acetyl-L-Leucine Accumulation in media [17]
N-acetyl-L- Accumulation in media; associated
[17]
phenylalanine with apoptosis
N-acetylmethionine Accumulation in media [17]
N-formyl-L-methionine Accumulation in media [17]
Cells 2022, 11, 1929 10 of 21

Table 1. Cont.

Amino Acid Derivative Behavior Reference Comments

Accumulation in media.
Metabolic source is methionine.
Its accumulation together with
L-homocysteine [21]
the other identified metabolites
leads to growth inhibition
in HIPDOG culture
Accumulation in media.
Metabolic sources are
phenylalanine and tyrosine. Its
3-(4-Hydroxyphenyl)
accumulation together with the [21]
lactate
other identified metabolites leads
to growth inhibition
in HiPDOG culture
Accumulation in media.
Metabolic source is phenylalanine.
Its accumulation together with
Phenyllactate [21]
the other identified metabolites
leads to growth inhibition in
HiPDOG culture
Accumulation in media.
Indole 3-lactate, Metabolic source is tryptophan.
indole-3-carboxylate, Its accumulation together with
[21]
2-hydroxyburtyrate, and 4- the other identified metabolites
hydroxyphenylpyruvate leads to growth inhibition
in HiPDOG culture
Accumulation in media.
Metabolic source is leucine. Its
accumulation together with the
Isovalerate [21,44]
other identified metabolites leads
to growth inhibition
in HiPDOG culture
Accumulation in media.
Metabolic sources are serine,
threonine and glycine. Its
Formate accumulation together with the [21,44]
other identified metabolites leads
to growth inhibition
in HiPDOG culture
Isobutyrate Accumulation in media [44]
Secreted in media during
exponential and transition phase,
with this being correlated with
Ammonia glutamine and asparagine [17,44]
consumption. Accumulation is
known to affect productivity and
inhibit cell growth
Intracellular and media
accumulation mainly due to
breakdown of glutamine and
Ammonium [11,31,38]
several amino acids; detrimental
effects on growth presumably due
to apoptosis induction
Cells 2022, 11, 1929 11 of 21

Table 2. Amino acids and amino acid derivatives at the stationary phase.

Amino Acid Behavior Reference Comments

Accumulation in media [16]
Constant increase in
[11] The switch observed in [44]
concentration with time
Alanine might be due to the analysis
Switch from alanine secretion method (NMR), which is
to uptake in media during different to those used in
stationary phase at the same [44] other studies
time as aspartate and
asparagine exhaustion
Intracellular level decreases
Arginine [31]
during the stationary phase
Depleted at entry into
stationary phase. Exponential
growth and antibody [14,16,36,40,44]
production continue if it is
added again in the media The asparagine concentrations
Asparagine used in these studies might be
Intracellular consumption. different, leading to its
Represents 8% of incoming [18] exhaustion at different times
carbon source of the culture
Uptake from the media, with
this being lower than at [19]
exponential phase
Below detection level once
cells enter stationary phase. The aspartic acid
Exponential growth and [14,16] concentrations used in these
Aspartic acid antibody production resumes studies might be different,
when added to media leading to its exhaustion at
Depleted in media during different times of the culture
[44]
stationary phase
Depletion at entry into
Cysteine [14]
stationary phase
Consumption [11,31]
The glutamine concentrations
Depleted in media at entry in
[23,36,40] used in these studies might be
Glutamine stationary phase
different, leading to its
Uptake from the media, with exhaustion at different times
this being lower than at [19] of the culture
exponential phase
Below detection level once
cells enter stationary phase. The glutamic acid
Exponential growth and [16] concentrations used in these
Glutamic acid studies might be different,
antibody production resumes
when added to media leading to its exhaustion at
different times of the culture
Constant detection [11]
Glycine NA
Histidine NA
Isoleucine NA
Leucine NA
Lysine NA
Methionine NA
Cells 2022, 11, 1929 12 of 21

Table 2. Cont.

Amino Acid Behavior Reference Comments

Intracellular level dropped
Phenylalanine [31]
during stationary phase
Slight increase in
Proline [11]
concentration with time
Serine Consumption from media [44]
Intracellular level decreases The particularity of study [21]
[31]
during stationary phase is the use of a specific
Threonine Accumulation in media HiDPOG culture process that
identified as growth inhibitor [21] might explain the difference
in HiPDOG culture observed

Intracellular level decreases

[31]
during stationary phase
Increased availability in
media correlates with
diminished viable cell density
Tryptophan and accumulation of an
intermediate during
[22]
tryptophan metabolism,
5-hydroxyindolacetaldehyde
(5-HIAAld), which is
suspected to be an inhibitor of
cell growth
Depletion at entry into
Tyrosine [14]
stationary phase
Intracellular level decreases in
Valine [31]
the stationary phase
Amino acid derivative Behavior Reference Comments
Accumulation in media at the
beginning of stationary phase.
Acetylphenylalanine [12,20]
Known to be detrimental to
cell growth
Accumulation in media at the
beginning of stationary phase.
Dimethylarginine By-product of protein [12,20]
degradation and can
induce apoptosis
Accumulation in media as
culture enters stationary
N-formimino-L-glutamate phase. Metabolite of the [22]
degradation of histidine
or glutamate
Ammonia Accumulation in media [44]
Intracellular and media
accumulation mainly due to
breakdown of glutamine and
Ammonium several amino acids; has [11,31]
detrimental effects on growth
presumably due to
apoptosis induction
Cells 2022, 11, 1929 13 of 21

Table 3. Amino acids and amino acids derivatives at the decline phase.

Amino Acid Behavior Reference Comments

Accumulation in extracellular
Alanine media after addition of [46]
anti-apoptotic agent
Arginine NA
Depleted from extracellular
media after addition of [46] The asparagine and aspartic
anti-apoptotic agent acid concentrations used in
these studies might be
Asparagine and aspartic acid Intracellular concentration
different, leading to their
decreases compared to other [31]
exhaustion at different times
growth phases
of the culture
Depletion during stationary
[44]
phase
Cysteine NA
Glutamine NA
Accumulation in extracellular
media after addition of [46] The glutamic acid
Glutamic acid concentration used in these
anti-apoptotic agent
studies might be different,
Decreased intracellular
leading to its exhaustion at
concentration compared to [31]
different times of the culture
other growth phases
Glycine NA
Extracellular concentration
Histidine and isoleucine decreases after addition of [46]
anti-apoptotic agent
Exhaustion at entry into
Leucine [16]
decline phase
Extracellular concentration
decreases after addition of [46]
The glutamic acid
anti-apoptotic agent.
concentration used in these
Lysine Exhaustion at entry into studies might be different,
[16]
decline phase leading to its exhaustion at
different times of the culture
Detected; indicates
[52]
over-supply
Extracellular concentration
Methionine and
decreases after addition of [46]
phenylalanine
anti-apoptotic agent
Proline NA
Extracellular concentration
decreases after addition of [46]
anti-apoptotic agent The glutamic acid
Serine concentration used in these
Exhaustion at entry into
[16] studies might be different,
decline phase
leading to its exhaustion at
Intracellular concentration different times of the culture
decreases compared to other [31]
growth phases
Extracellular concentration
Threonine, tryptophan,
decreases after addition of [46]
tyrosine, valine
anti-apoptotic agent
Cells 2022, 11, 1929 14 of 21

Table 3. Cont.

Amino Acid Behavior Reference Comments

Detected. Known to have
Ornithine [52]
apoptotic properties
Amino acid derivative Behavior Reference Comments
Extracellular concentration
Pyroglutamate increases after addition of [46]
anti-apoptotic agent
Extracellular concentration
4-hydroxyproline decreases after addition of [46]
anti-apoptotic agent
Dimethylarginine,
Caspase activity increases
flutamylphenylalanine,
after exposition of cells to [47]
glycerophosphocholine,
these metabolites
hexanoglycine

2.2. The Use of Energy Enables Biomass Production (Step 2)

To our knowledge, no metabolomics studies focused on nucleic acid and protein
synthesis. This section will focus on lipid metabolism, which is extensively covered by
metabolomics studies.
One group of studies measured an increase in intracellular concentrations of medium
and long chain fatty acids in line with the need for cell growth during the exponential
phase. In line with this observation, the contribution of fatty acid oxidation to the acetyl-
CoA mitochondrial pool was found to be negligible during the exponential phase [11],
with acetyl-CoA being mainly directed toward fatty acid synthesis at this phase of cell
growth [23]. A large fraction of citrate was found to be transported to the cytosol for its
conversion into oxaloacetate and acetyl-CoA [31,33]. An MFA study calculated that about
25% of the citrate synthase flux is channeled via citrate lyase to the biosynthesis of fatty
acids. In this study, the TCA cycle intermediates mostly fueled lipid synthesis and rendered
negligible amounts of amino acids [32]. A study performed on adherent CHO cells demon-
strated that acetyl-CoA produced from glutamine does not contribute significantly to fatty
acid biosynthesis, suggesting that other sources of carbon such as glucose do [33]. As a
component of the glycerol-based phospholipids that compose cell membranes, glycerol was
also found to accumulate extracellularly, representing 5% of the carbon engaged in glycoly-
sis [26,44]. In their review, Pereira et al. (2018) stated that glycerol released into the medium
results in a loss of carbon to the cells. For the authors, glycerol can be regarded as a storage
compound or redox sink in the regeneration of NAD pools [38,39]. Other studies have
shown that choline, choline phosphate, ethanolamine phosphate, cytidine-diphosphate-
choline (CDP) and CDP-ethanolamine are almost depleted during the exponential phase,
while other precursors of the cell membrane constituents are accumulated [12,17]. It was
suggested by Selvarasu et al. that this might be a reflection of poor membrane lipid
metabolism regulation in CHO cells and that this might trigger the transition from the
exponential to the stationary phase [17]. In support of this hypothesis, an increase in
extracellular phosphocholine and glycerol-3-phosphocholine was reported and correlated
with intracellular caspase activity and apoptosis [44,47]. A thorough investigation of the
altered metabolic flux, including in the glycerophospholipid metabolic pathway, might
help in the design of strategies to enhance cell culture viability.
Unlike the exponential growth phase, for which some carbon flux consensus can be
drawn across cell culture models, the stationary and decline phases are characterized by
metabolic changes that appear to be cell culture model-dependent. Readers are encouraged
to refer to Supplementary Table S1 and compare their results to those obtained in relevant
cell culture models.
Cells 2022, 11, 1929 15 of 21

3. The Stationary Phase: Stabilized Growth and High Recombinant Protein Production
3.1. Alternative Carbon Source Catabolism Enables Energy Production and Cell Maintenance
As a general trend, most of the carbon and nitrogen sources were found to be consumed at
lower rates during the stationary phase compared to the exponential phase [34,40,44]. Thus, a
5-fold decrease in glycolytic activity was reported to be accompanied by an equivalent reduction
in glutaminolysis and AKG production [33]. The stationary phase also witnesses a major shift
in the metabolism of lactate, which is not so much produced but consumed, enabling TCA
cycle replenishment, as the glucose concentration decreases [11,16–18,31,33,34,44,56]. Glucose
consumption continues, even though at a decreased rate, until it is fully depleted [11,27,41,44].
One report, namely [18], indicated a slight decrease in glutaminolysis, while glucose fluxes do
not drop below 50% of their initial rate. A decrease in glucose consumption and glycolysis fluxes
by one-third was observed at the late exponential phase and the stationary phase [18]. These
key characteristics were also observed for antibody-producing GS-NS0 mouse myeloma cells.
In such cells, lactate was consumed earlier when proliferation was artificially stopped, which
also correlated with increased productivity [28]. Evidence was provided that glucose is not
depleted during the stationary phase but at the entry in the decline phase [16]. The depletion of
asparagine and aspartate was reported to occur at the entry into the stationary phase, well before
the exhaustion of glucose. Moreover, pyruvate was identified as a growth-limiting metabolite,
meaning that its depletion marks the transition from the exponential to the stationary phase [44].
The depletion of pyruvate was confirmed in a more recent study [31].
PPP fluxes increase during the stationary phase compared to the exponential phase by
a factor of 5 to 6-fold, corresponding to 30% of the glucose uptake rate [11,33] or possibly
even more [13,18,34]. Sengupta et al. hypothesized that during the stationary phase,
cells might experience more oxidative stress, leading to a higher utilization of the PPP
to generate NADPH for reductive reactions. This is supported by a study on Escherichia
coli that showed that the bacteria fail to fight oxidative stress during the stationary phase
because of the downregulation of genes involved in aerobic electron transport [57]. One
recent report showed that the downregulation of the PPP, as opposed to its activation,
occurred during the stationary phase in CHO cells [31], but this relationship has not been
extensively studied in CHO cells and deserves further investigation.
Most studies agree that succinate, malate, citrate and fumarate accumulate during
the stationary phase in the culture supernatant [12,20,27,46,58]. At this stage, carbon
influx into the TCA cycle is provided by glucose (50%), glutamine (40%) and other amino
acids (10%) entering the TCA cycle in the form of acetyl-CoA or anaplerotic substrates
including malate, α-ketoglutarate or oxaloacetate [41]. More importantly, a switch in
pyruvate utilization is observed from lactate production and secretion to its translocation
into the mitochondria and use in the TCA cycle, which reaches its peak activity [13,18,34].
Such TCA cycle activation may be secondary to exacerbated oxidative stress, resulting in a
decreased NADH/NAD+ ratio and the consequent counter regulation of the regeneration
of the NADH pool.
In addition to the publications that report the increased activity of the TCA cycle in
the stationary phase, equivalent or decreased activity has also been reported. For example,
MFA performed on adherent CHO cells during the stationary phase led to the conclusion
that most of the pyruvate is diverted to the TCA cycle (instead of yielding lactate), with
no impact on the TCA cycle, which operates at similar levels as during the exponential
phase [11,33]. This was also observed in a study using 1 H-NMR, where the secretion of
succinate, malate, citrate and fumarate were found to represent the same proportion of the
TCA cycle flux as during the exponential phase. This indicates that in these models, the
TCA is operating at close-to-maximum enzymatic capacity throughout the culture time [44].
In other studies, the activities of glutamate dehydrogenase and pyruvate carboxylase, key
anaplerotic enzymes, were found to be reduced 2-fold, and the flux from α-ketoglutarate
to succinyl-CoA was also lowered to 50% [11,33]. Such decreased TCA cycle activity has
also been observed in other works [16,17] and led to the conclusion that glucose might be
redirected to other metabolic pathways to enable ATP generation and NADPH oxidation
Cells 2022, 11, 1929 16 of 21

in the stationary phase, during which proliferation is limited and productivity is increased
(see below).
Regarding amino acids, observations made during the stationary phase of growth
are reported in Table 3. According to a number of studies, anaplerosis might be strongly
decreased during the stationary phase compared to exponential phase, resulting in 3-
fold lower malic enzyme fluxes and negligible fluxes of amino acids toward the TCA
cycle [11,13]. However, another more recent study estimated that the consumption of
essential amino acids for anaplerosis increases from 25% during the exponential phase
to 35% during the stationary phase, pointing to the key role of amino acids in energy
production throughout the culture [44]. These conflicting results demonstrate that cell
metabolism across culture phases might be model and clone dependent.

3.2. The Use of Energy Enables Recombinant Protein Production

The production of recombinant proteins by CHO cells ramps up during the stationary
phase, accounting for 15% of incoming carbon fluxes, in contrast to only 3% during the
early exponential phase [18].
In contrast to the exponential growth phase, about 25% of pyruvate formation results from
lactate consumption during the stationary phase, with other minor sources being malate and
serine together with pyruvate uptake from the media [33,40,44]. The entry of pyruvate into the
TCA cycle by conversion into acetyl-CoA was measured to be similar in the exponential and
stationary phases. On the contrary, its conversion into oxaloacetate via pyruvate carboxylase
activity was found to be undetectable during the stationary phase [33].
Regarding glycerophospholipids, one study reported that the transition between
exponential growth and the stationary phase is marked by the intracellular appearance
and then accumulation in media of glycerol-3-phosphate and glycerol, the latter being
presumably synthesized from glycerol-3-phosphate [14,16]. Carinhas et al. estimated that
25% of the carbon engaged in glycolysis ends up in glycerol [44]. The surge in glycerol
concentration observable at the transition from the exponential to the stationary phase was
interpreted to reflect the lipid biosynthesis that is required for cell proliferation-associated
membrane formation, the turnover of membrane lipids as well as the secretion vesicles
used for protein excretion [14]. An increase of glycerophosphocholine, a product of the
degradation of phosphatidylcholine, which is a major component of the plasma membrane,
was also observed. This was interpreted to reflect cell membrane degradation and cell
growth limitation [17]. Surprisingly, fatty acid biosynthesis was found to be as high during
the stationary phase as during the exponential phase, suggesting another role in addition
to cell growth that was not elucidated by the authors [33]. This is consistent with the
hypothesis that lipid synthesis might reflect the need for membrane lipid turnover (plasma
membrane and intracellular membranes involved in vesicular trafficking).

4. The Decline Phase: Media Exhaustion Resulting in Cell Death

Relatively few studies have investigated CHO metabolism during the decline of cell
cultures. The final phase was found to be accompanied by the exhaustion of glucose, with
a consequent decrease in glycolytic intermediates [16]. Pyruvate is depleted, while lactate
remains relatively constant in batch culture [14]. In fed-batch [14], and HiPDOG (hi-end
pH delivery of glucose) culture conditions [18], lactate utilization was observed during the
decline phase. In this latter study, PPP flux was maintained even when cell density began
to decline.
Regarding the TCA, Matuszczyk et al. have demonstrated the depletion of cytosolic
pyruvate while mitochondrial pyruvate was available at higher concentrations, suggesting
high TCA cycle activity [30]. This is in line with another study reporting that TCA cycle
activity is even higher during the decline phase than during the early exponential phase [18].
On the contrary, a significant decrease in the metabolites of the TCA cycle compared to the
other culture phases was reported in one study, and this was interpreted to be due to less
carbon skeleton from glucose fueling the TCA during this phase [16]. This discrepancy may
Cells 2022, 11, 1929 17 of 21

be explained by the fact that the former study was focused on an early phase of decline
while the latter was focused on a terminal phase of decline and medium exhaustion.
Regarding nucleotide metabolism, ATP, ADP and AMP concentrations decreased with
the culture time until the decline phase, but their distributions across the cytoplasmic and
mitochondrial compartments differed. The ATP concentration was found to be higher in
the cytosol than in the mitochondria during this culture phase, unlike in the cases of ADP
and AMP, which were found at similar concentrations in both compartments [30].
Regarding amino acids, observations made during the decline phase are reported in
Table 3 There is no comprehensive analysis of fatty acid metabolism during the decline phase,
but a peak of medium and long chain fatty acid concentrations was reported at this phase [31].
Finally, regarding the oxidative state of the cell, an increase in oxidized glutathione was
observed, leading to an unfavorable GSSG/GSH ratio for ROS management [31].
In addition to a comprehensive understanding of cell metabolism during the decline
phase, which needs further investigation, several strategies have been suggested to extend
the productive phase of the cell suspension. For example, one study recommended the
growing of cells in galactose-containing medium, as it is metabolized when glucose is
depleted, which enables the maintenance of cell viability [59]. Another study suggested
the addition of an anti-apoptotic agent to avoid excessive cell death [46].

5. Conclusions
The comparison of all of the metabolomic studies on CHO cells reveals general trends for
each of the phases of cell cultures, as they highlight the most important shifts from expansion
to stabilization and then to final viable cell density decline. Overall, the exponential phase
of growth corresponds to a state of high consumption in terms of nutrients, in particular
glucose and glutamine. This enables high anabolic activity, a balanced redox state and the
formation of excess energy that can be utilized for biomass production, favoring intense
cell proliferation. The oxidation of glucose via glycolysis is high, enabling bioenergetic
homeostasis and energy “storage” by means of lactate production, contrasting with the
moderate pyruvate-mediated fueling of the TCA cycle, which is instead replenished by
anaplerosis due to the availability of all amino acids. A characteristic of the growth phase is
low mAb production, which makes the subsequent stationary phase the most interesting with
regard to the synthesis of recombinant proteins. This reduced proliferation enables a higher
investment in cellular energy for biosynthesis. The stationary phase is characterized by a
comparably low consumption of carbon and nitrogen sources and reduced glycolytic activity.
In contrast, TCA cycle fluxes are increased during this phase, with lactate consumption
supplying the acetyl-CoA pool. In this phase, the role of amino acids in TCA replenishment
deserves further investigation. Also, compared to the exponential phase, PPP is largely
activated and thus generates reducing equivalents such as NADPH and glutathione, which
fight oxidative stress during the stationary phase. The final decline of cultures is marked by
the exhaustion of many metabolites and the accumulation of toxic products.
Importantly, this review also highlights important discrepancies between results from
various studies. These discrepancies may be explained by the unique features of the CHO
cell lines derived from a single host, acquired by genetic drift in separate laboratories [60].
Differences in terms of the media used and the culture strategies have also shown to give
rise to genetic and epigenetic diversity, and this could explain phenotypic differences when
distinct processes are used for the production from the same clone [61,62]. This diversity is
known to have a major impact on the metabolomic profile of CHO culture.
For instance, principal component analysis performed on cell culture supernatants of
six mAbs producing cell lines derived from two distinct hosts highlighted that variations
in metabolic profiles were highly dependent on the cell line’s host lineage [22]. When com-
paring the metabolic profile of high and low producer clones, key differences were also
identified [19,43]. In the first study, Chong et al. [43] identified seven metabolites involved
in the main metabolic pathways of CHO cells (glycolysis, the PPP and the TCA cycle), as-
sociated with high productivity. It was suggested that the abundance of these metabolites
Cells 2022, 11, 1929 18 of 21

improved the control of the oxidative state, thus increasing the production of recombinant
mAb. Furthermore, Dean et al. [19] highlighted several metabolic differences between high
and low producers’ cell lines. The low producer clone consumed more glucose at the start of
the culture, while accumulating less lactate. In addition, more than 50% of the intracellular
lactate originated from extracellular pyruvate in the high producer clone, whereas it repre-
sented only 30% in the low producer clone. After four days of culture, the high producer
was able to consume the accumulated lactate, which was not the case for the low producer.
More importantly, TCA metabolites were replenished from glucose in high producer clones,
suggesting increased activity in terms of the TCA cycle, likely representing a hallmark of high
production [19,63]. It was consistently shown that the production of the recombinant protein
correlated with a more efficient utilization of glucose for the TCA cycle [41].
Specific gene silencing or over-expression was also shown to alter cell metabolism.
For example, the endogenous expression level of the bcl-2 anti-apoptotic gene [34] or the
overexpression of malate dehydrogenase II [20] resulted in diminished lactate accumulation
and decreased NADH abundancy when compared with control cells.
Finally, the diversity in terms of media formulation across the studies referenced in
this review might explain some of the discrepancies identified. For instance, the peak
lactate concentration can be decreased by the use of a chemically defined CHO medium
when compared to a chemically defined hybridoma medium [64]. It was also shown
that maintaining glucose concentrations to almost negligible levels in HiPDOG cultures
enabled the maintenance of lactate, osmolarity and ammonia below inhibitory levels [21].
Similarly, the concentration of glutamine, another major carbon source for cells, in the
culture medium was shown to have a major impact on the CHO metabolic states [42]. In
low glutamine-containing medium, glycolytic fluxes were significantly increased while
PPP was decreased 2-fold when compared to cultures growing in high glutamine medium.
Pyruvate usage by the TCA cycle instead of lactate production was more important in low
glutamine medium cultures, which was also observed by Kirsch et al. [25]. An increase
in TCA flux in such culture conditions was also reported in other studies, including in
CHO-DHFR cells [50].
Nevertheless, the comparison of results obtained by MS metabolomics and 13 C MFA
to results obtained using other methods can be an efficient strategy when used to improve
our knowledge of CHO cells metabolism and identify main trends. Future investigation
must address the cause–effect relationships between metabolic shifts and mAb production
yields. This could be achieved by connecting information obtained on experiments using
several “omics” tools beyond metabolomics such as genomics, epigenomics, transcrip-
tomics and proteomics [65]. Moreover, the increasing availability of single-cell “omics”
may increase our knowledge of individual cells, thus allowing for a refined analysis of
likely heterogeneous cell populations, eventually improving clone selection.

Supplementary Materials: The following supporting information can be downloaded at: https:
//www.mdpi.com/article/10.3390/cells11121929/s1, Table S1: Comparison of metabolomic studies
of CHO cells.
Author Contributions: Conceptualization: M.C., S.B. Review literature: M.C. Writing—original draft
preparation: M.C., S.B. Writing—review and editing: all authors. Prepared figures: M.C. Supervision:
S.B. All authors have read and agreed to the published version of the manuscript.
Funding: M.C. is supported by a grant from the French ANRT agency (CIFRE No 2020/0106). O.K.
received funding from the Institut National du Cancer (INCa) and the DIM ELICIT initiative of the
Ile de France. G.K. is supported by the Ligue contre le Cancer (équipe labellisée); Agence National de
la Recherche (ANR)—Projets blancs; AMMICa US23/CNRS UMS3655; Association pour la Recherche
sur le Cancer (ARC); Cancéropôle Ile-de-France; Fondation pour la Recherche Médicale (FRM);
a donation by Elior; Equipex Onco-Pheno-Screen; European Joint Programme on Rare Diseases
(EJPRD); Gustave Roussy Odyssea, the European Union Horizon 2020 Projects Oncobiome and
Crimson; Fondation Carrefour; Institut National du Cancer (INCa); Institut Universitaire de France;
LabEx Immuno-Oncology (ANR-18-IDEX-0001); a Cancer Research ASPIRE Award from the Mark
Cells 2022, 11, 1929 19 of 21

Foundation; the RHU Immunolife; Seerave Foundation; SIRIC Stratified Oncology Cell DNA Repair
and Tumor Immune Elimination (SOCRATE); and SIRIC Cancer Research and Personalized Medicine
(CARPEM). This study contributes to the IdEx Université de Paris ANR-18-IDEX-0001.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Acknowledgments: We acknowledge Christine DeMaria, Henry Lin and Shawn Barrett at Sanofi for
a critical review of the manuscript. The graphical abstract was created with BioRender.com, accessed
on 2 June 2022.
Conflicts of Interest: G.K. received research support by Sanofi. M.C. and S.B. are full-time employees
of Sanofi.

References
1. Singh, S.; Kumar, N.K.; Dwiwedi, P.; Charan, J.; Kaur, R.; Sidhu, P.; Chugh, V.K. Monoclonal Antibodies: A Review. Curr. Clin.
Pharmacol. 2018, 13, 85–99. [CrossRef] [PubMed]
2. Walsh, G. Biopharmaceutical benchmarks 2018. Nat. Biotechnol. 2018, 36, 1136–1145. [CrossRef] [PubMed]
3. Kunert, R.; Reinhart, D. Advances in recombinant antibody manufacturing. Appl. Microbiol. Biotechnol. 2016, 100, 3451–3461.
[CrossRef] [PubMed]
4. Li, F.; Vijayasankaran, N.; Shen, A.Y.; Kiss, R.; Amanullah, A. Cell culture processes for monoclonal antibody production. MAbs
2010, 2, 466–479. [CrossRef] [PubMed]
5. Ritacco, F.V.; Wu, Y.; Khetan, A. Cell culture media for recombinant protein expression in Chinese hamster ovary (CHO) cells:
History, key components, and optimization strategies. Biotechnol. Prog. 2018, 34, 1407–1426. [CrossRef] [PubMed]
6. Oliver, S.G.; Winson, M.K.; Kell, D.B.; Baganz, F. Systematic functional analysis of the yeast genome. Trends Biotechnol. 1998, 16,
373–378. [CrossRef]
7. Zhang, A.; Sun, H.; Xu, H.; Qiu, S.; Wang, X. Cell metabolomics. Omics A J. Integr. Biol. 2013, 17, 495–501. [CrossRef]
8. León, Z.; García-Cañaveras, J.C.; Donato, M.T.; Lahoz, A. Mammalian cell metabolomics: Experimental design and sample
preparation. Electrophoresis 2013, 34, 2762–2775. [CrossRef]
9. Cuperlović-Culf, M.; Barnett, D.A.; Culf, A.S.; Chute, I. Cell culture metabolomics: Applications and future directions. Drug
Discov. Today 2010, 15, 610–621. [CrossRef]
10. Roberts, L.D.; Souza, A.L.; Gerszten, R.E.; Clish, C.B. Targeted metabolomics. Curr. Protoc. Mol. Biol. 2012, 98, 30–32. [CrossRef]
11. Ahn, W.S.; Antoniewicz, M.R. Metabolic flux analysis of CHO cells at growth and non-growth phases using isotopic tracers and
mass spectrometry. Metab. Eng. 2011, 13, 598–609. [CrossRef] [PubMed]
12. Chong, W.P.; Goh, L.T.; Reddy, S.G.; Yusufi, F.N.; Lee, D.Y.; Wong, N.S.; Heng, C.K.; Yap, M.G.; Ho, Y.S. Metabolomics profiling
of extracellular metabolites in recombinant Chinese Hamster Ovary fed-batch culture. Rapid Commun. Mass Spectrom. 2009, 23,
3763–3771. [CrossRef] [PubMed]
13. Sengupta, N.; Rose, S.T.; Morgan, J.A. Metabolic flux analysis of CHO cell metabolism in the late non-growth phase. Biotechnol.
Bioeng. 2011, 108, 82–92. [CrossRef] [PubMed]
14. Sellick, C.A.; Croxford, A.S.; Maqsood, A.R.; Stephens, G.M.; Westerhoff, H.V.; Goodacre, R.; Dickson, A.J. Metabolite profiling of
CHO cells: Molecular reflections of bioprocessing effectiveness. Biotechnol. J. 2015, 10, 1434–1445. [CrossRef]
15. Karst, D.J.; Steinhoff, R.F.; Kopp, M.R.G.; Serra, E.; Soos, M.; Zenobi, R.; Morbidelli, M. Intracellular CHO Cell Metabolite Profiling
Reveals Steady-State Dependent Metabolic Fingerprints in Perfusion Culture. Biotechnol. Prog. 2017, 33, 879–890. [CrossRef]
16. Sellick, C.A.; Croxford, A.S.; Maqsood, A.R.; Stephens, G.; Westerhoff, H.V.; Goodacre, R.; Dickson, A.J. Metabolite profiling of
recombinant CHO cells: Designing tailored feeding regimes that enhance recombinant antibody production. Biotechnol. Bioeng.
2011, 108, 3025–3031. [CrossRef]
17. Selvarasu, S.; Ho, Y.S.; Chong, W.P.; Wong, N.S.; Yusufi, F.N.; Lee, Y.Y.; Yap, M.G.; Lee, D.Y. Combined in silico modeling and
metabolomics analysis to characterize fed-batch CHO cell culture. Biotechnol. Bioeng. 2012, 109, 1415–1429. [CrossRef]
18. Templeton, N.; Dean, J.; Reddy, P.; Young, J.D. Peak antibody production is associated with increased oxidative metabolism in an
industrially relevant fed-batch CHO cell culture. Biotechnol. Bioeng. 2013, 110, 2013–2024. [CrossRef]
19. Dean, J.; Reddy, P. Metabolic analysis of antibody producing CHO cells in fed-batch production. Biotechnol. Bioeng. 2013, 110,
1735–1747. [CrossRef]
20. Chong, W.P.; Reddy, S.G.; Yusufi, F.N.; Lee, D.Y.; Wong, N.S.; Heng, C.K.; Yap, M.G.; Ho, Y.S. Metabolomics-driven approach
for the improvement of Chinese hamster ovary cell growth: Overexpression of malate dehydrogenase II. J. Biotechnol. 2010, 147,
116–121. [CrossRef]
21. Mulukutla, B.C.; Kale, J.; Kalomeris, T.; Jacobs, M.; Hiller, G.W. Identification and control of novel growth inhibitors in fed-batch
cultures of Chinese hamster ovary cells. Biotechnol. Bioeng. 2017, 114, 1779–1790. [CrossRef] [PubMed]
Cells 2022, 11, 1929 20 of 21

22. Alden, N.; Raju, R.; McElearney, K.; Lambropoulos, J.; Kshirsagar, R.; Gilbert, A.; Lee, K. Using Metabolomics to Identify Cell
Line-Independent Indicators of Growth Inhibition for Chinese Hamster Ovary Cell-based Bioprocesses. Metabolites 2020, 10, 199.
[CrossRef] [PubMed]
23. Nicolae, A.; Wahrheit, J.; Bahnemann, J.; Zeng, A.P.; Heinzle, E. Non-stationary 13C metabolic flux analysis of Chinese hamster
ovary cells in batch culture using extracellular labeling highlights metabolic reversibility and compartmentation. BMC Syst. Biol.
2014, 8, 50. [CrossRef] [PubMed]
24. Nargund, S.; Qiu, J.; Goudar, C.T. Elucidating the role of copper in CHO cell energy metabolism using 13 C metabolic flux analysis.
Biotechnol. Prog. 2015, 31, 1179–1186. [CrossRef]
25. Kirsch, B.J.; Bennun, S.V.; Mendez, A.; Johnson, A.S.; Wang, H.; Qiu, H.; Li, N.; Lawrence, S.M.; Bak, H.; Betenbaugh, M.J.
Metabolic Analysis of the Asparagine and Glutamine Dynamics in an Industrial CHO Fed-Batch Process. Biotechnol. Bioeng. 2021,
119, 807–819. [CrossRef]
26. Torres, M.; Elvin, M.; Betts, Z.; Place, S.; Gaffney, C.; Dickson, A.J. Metabolic profiling of Chinese hamster ovary cell cultures at
different working volumes and agitation speeds using spin tube reactors. Biotechnol. Prog. 2021, 37, e3099. [CrossRef]
27. Miccheli, A.T.; Miccheli, A.; Di Clemente, R.; Valerio, M.; Coluccia, P.; Bizzarri, M.; Conti, F. NMR-based metabolic profiling
of human hepatoma cells in relation to cell growth by culture media analysis. Biochim. Biophys. Acta 2006, 1760, 1723–1731.
[CrossRef]
28. Khoo, S.H.; Al-Rubeai, M. Metabolic characterization of a hyper-productive state in an antibody producing NS0 myeloma cell
line. Metab. Eng. 2009, 11, 199–211. [CrossRef]
29. Luo, J.; Vijayasankaran, N.; Autsen, J.; Santuray, R.; Hudson, T.; Amanullah, A.; Li, F. Comparative metabolite analysis to
understand lactate metabolism shift in Chinese hamster ovary cell culture process. Biotechnol. Bioeng. 2012, 109, 146–156.
[CrossRef]
30. Matuszczyk, J.C.; Teleki, A.; Pfizenmaier, J.; Takors, R. Compartment-specific metabolomics for CHO reveals that ATP pools in
mitochondria are much lower than in cytosol. Biotechnol. J. 2015, 10, 1639–1650. [CrossRef]
31. Vodopivec, M.; Lah, L.; Narat, M.; Curk, T. Metabolomic profiling of CHO fed-batch growth phases at 10, 100, and 1,000 L.
Biotechnol. Bioeng. 2019, 116, 2720–2729. [CrossRef] [PubMed]
32. Goudar, C.; Biener, R.; Boisart, C.; Heidemann, R.; Piret, J.; de Graaf, A.; Konstantinov, K. Metabolic flux analysis of CHO cells in
perfusion culture by metabolite balancing and 2D [13 C, 1 H] COSY NMR spectroscopy. Metab. Eng. 2010, 12, 138–149. [CrossRef]
[PubMed]
33. Ahn, W.S.; Antoniewicz, M.R. Parallel labeling experiments with [1,2-13 C]glucose and [U-13 C]glutamine provide new insights
into CHO cell metabolism. Metab. Eng. 2013, 15, 34–47. [CrossRef]
34. Templeton, N.; Lewis, A.; Dorai, H.; Qian, E.A.; Campbell, M.P.; Smith, K.D.; Lang, S.E.; Betenbaugh, M.J.; Young, J.D. The impact
of anti-apoptotic gene Bcl-2∆ expression on CHO central metabolism. Metab. Eng. 2014, 25, 92–102. [CrossRef] [PubMed]
35. Slade, P.G.; Caspary, R.G.; Nargund, S.; Huang, C.J. Mannose metabolism in recombinant CHO cells and its effect on IgG
glycosylation. Biotechnol. Bioeng. 2016, 113, 1468–1480. [CrossRef] [PubMed]
36. Wijaya, A.W.; Verhagen, N.; Teleki, A.; Takors, R. Compartment-specific 13 C metabolic flux analysis reveals boosted NADPH
availability coinciding with increased cell-specific productivity for IgG1 producing CHO cells after MTA treatment. Eng. Life Sci.
2021, 21, 832–847. [CrossRef]
37. Wijaya, A.W.; Ulmer, A.; Hundsdorfer, L.; Verhagen, N.; Teleki, A.; Takors, R. Compartment-specific metabolome labeling enables
the identification of subcellular fluxes that may serve as promising metabolic engineering targets in CHO cells. Bioprocess Biosyst.
Eng. 2021, 44, 2567–2578. [CrossRef]
38. Pereira, S.; Kildegaard, H.F.; Andersen, M.R. Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview
of Toxic and Inhibiting Metabolites and Nutrients. Biotechnol. J. 2018, 13, e1700499. [CrossRef]
39. Quek, L.E.; Dietmair, S.; Krömer, J.O.; Nielsen, L.K. Metabolic flux analysis in mammalian cell culture. Metab. Eng. 2010, 12,
161–171. [CrossRef]
40. Junghans, L.; Teleki, A.; Wijaya, A.W.; Becker, M.; Schweikert, M.; Takors, R. From nutritional wealth to autophagy: In vivo
metabolic dynamics in the cytosol, mitochondrion and shuttles of IgG producing CHO cells. Metab. Eng. 2019, 54, 145–159.
[CrossRef]
41. Sheikholeslami, Z.; Jolicoeur, M.; Henry, O. Probing the metabolism of an inducible mammalian expression system using
extracellular isotopomer analysis. J. Biotechnol. 2013, 164, 469–478. [CrossRef]
42. Sheikholeslami, Z.; Jolicoeur, M.; Henry, O. Elucidating the effects of postinduction glutamine feeding on the growth and
productivity of CHO cells. Biotechnol. Prog. 2014, 30, 535–546. [CrossRef] [PubMed]
43. Chong, W.P.; Thng, S.H.; Hiu, A.P.; Lee, D.Y.; Chan, E.C.; Ho, Y.S. LC-MS-based metabolic characterization of high monoclonal
antibody-producing Chinese hamster ovary cells. Biotechnol. Bioeng. 2012, 109, 3103–3111. [CrossRef] [PubMed]
44. Carinhas, N.; Duarte, T.M.; Barreiro, L.C.; Carrondo, M.J.; Alves, P.M.; Teixeira, A.P. Metabolic signatures of GS-CHO cell clones
associated with butyrate treatment and culture phase transition. Biotechnol. Bioeng. 2013, 110, 3244–3257. [CrossRef]
45. Blondeel, E.J.M.; Ho, R.; Schulze, S.; Sokolenko, S.; Guillemette, S.R.; Slivac, I.; Durocher, Y.; Guillemette, J.G.; McConkey, B.J.;
Chang, D.; et al. An omics approach to rational feed: Enhancing growth in CHO cultures with NMR metabolomics and 2D-DIGE
proteomics. J. Biotechnol. 2016, 234, 127–138. [CrossRef] [PubMed]
Cells 2022, 11, 1929 21 of 21

46. Aranibar, N.; Borys, M.; Mackin, N.A.; Ly, V.; Abu-Absi, N.; Abu-Absi, S.; Niemitz, M.; Schilling, B.; Li, Z.J.; Brock, B.; et al.
NMR-based metabolomics of mammalian cell and tissue cultures. J. Biomol. NMR 2011, 49, 195–206. [CrossRef] [PubMed]
47. Chong, W.P.; Yusufi, F.N.; Lee, D.Y.; Reddy, S.G.; Wong, N.S.; Heng, C.K.; Yap, M.G.; Ho, Y.S. Metabolomics-based identification
of apoptosis-inducing metabolites in recombinant fed-batch CHO culture media. J. Biotechnol. 2011, 151, 218–224. [CrossRef]
[PubMed]
48. Carvalhal, A.V.; Santos, S.S.; Calado, J.; Haury, M.; Carrondo, M.J. Cell growth arrest by nucleotides, nucleosides and bases as a
tool for improved production of recombinant proteins. Biotechnol. Prog. 2003, 19, 69–83. [CrossRef]
49. Chen, X.G.; Wang, R.S.; Deng, M.X.; Ran, X.Z. Effects of exogenerous nucleotides on the apoptosis of intestinal epithelial cells
IEC-6. J. Hyg. Res. 2005, 34, 701–704.
50. Zhang, L.-X.; Zhang, W.-Y.; Wang, C.; Liu, J.-T.; Deng, X.-C.; Liu, X.-P.; Fan, L.; Tan, W.-S. Responses of CHO-DHFR cells to ratio
of asparagine to glutamine in feed media: Cell growth, antibody production, metabolic waste, glutamate, and energy metabolism.
Bioresour. Bioprocess. 2016, 3, 5. [CrossRef]
51. Duarte, T.M.; Carinhas, N.; Barreiro, L.C.; Carrondo, M.J.; Alves, P.M.; Teixeira, A.P. Metabolic responses of CHO cells to limitation
of key amino acids. Biotechnol. Bioeng. 2014, 111, 2095–2106. [CrossRef] [PubMed]
52. Mohmad-Saberi, S.E.; Hashim, Y.Z.; Mel, M.; Amid, A.; Ahmad-Raus, R.; Packeer-Mohamed, V. Metabolomics profiling of
extracellular metabolites in CHO-K1 cells cultured in different types of growth media. Cytotechnology 2013, 65, 577–586. [CrossRef]
[PubMed]
53. Jiang, D.J.; Jia, S.J.; Dai, Z.; Li, Y.J. Asymmetric dimethylarginine induces apoptosis via p38 MAPK/caspase-3-dependent signaling
pathway in endothelial cells. J. Mol. Cell. Cardiol. 2006, 40, 529–539. [CrossRef] [PubMed]
54. Zamani, L.; Lundqvist, M.; Zhang, Y.; Aberg, M.; Edfors, F.; Bidkhori, G.; Lindahl, A.; Mie, A.; Mardinoglu, A.; Field, R.; et al.
High Cell Density Perfusion Culture has a Maintained Exoproteome and Metabolome. Biotechnol. J. 2018, 13, e1800036. [CrossRef]
55. Taschwer, M.; Hackl, M.; Hernández Bort, J.A.; Leitner, C.; Kumar, N.; Puc, U.; Grass, J.; Papst, M.; Kunert, R.; Altmann, F.;
et al. Growth, productivity and protein glycosylation in a CHO EpoFc producer cell line adapted to glutamine-free growth. J.
Biotechnol. 2012, 157, 295–303. [CrossRef]
56. Sumit, M.; Dolatshahi, S.; Chu, A.A.; Cote, K.; Scarcelli, J.J.; Marshall, J.K.; Cornell, R.J.; Weiss, R.; Lauffenburger, D.A.; Mulukutla,
B.C.; et al. Dissecting N-Glycosylation Dynamics in Chinese Hamster Ovary Cells Fed-batch Cultures using Time Course Omics
Analyses. iScience 2019, 12, 102–120. [CrossRef]
57. Chang, D.E.; Smalley, D.J.; Conway, T. Gene expression profiling of Escherichia coli growth transitions: An expanded stringent
response model. Mol. Microbiol. 2002, 45, 289–306. [CrossRef]
58. Chrysanthopoulos, P.K.; Goudar, C.T.; Klapa, M.I. Metabolomics for high-resolution monitoring of the cellular physiological state
in cell culture engineering. Metab. Eng. 2010, 12, 212–222. [CrossRef]
59. Altamirano, C.; Illanes, A.; Becerra, S.; Cairó, J.J.; Gòdia, F. Considerations on the lactate consumption by CHO cells in the
presence of galactose. J. Biotechnol. 2006, 125, 547–556. [CrossRef]
60. Vcelar, S.; Jadhav, V.; Melcher, M.; Auer, N.; Hrdina, A.; Sagmeister, R.; Heffner, K.; Puklowski, A.; Betenbaugh, M.; Wenger, T.;
et al. Karyotype variation of CHO host cell lines over time in culture characterized by chromosome counting and chromosome
painting. Biotechnol. Bioeng. 2018, 115, 165–173. [CrossRef]
61. Huhn, S.; Chang, M.; Kumar, A.; Liu, R.; Jiang, B.; Betenbaugh, M.; Lin, H.; Nyberg, G.; Du, Z. Chromosomal instability drives
convergent and divergent evolution toward advantageous inherited traits in mammalian CHO bioproduction lineages. iScience
2022, 25, 104074. [CrossRef] [PubMed]
62. Weinguny, M.; Klanert, G.; Eisenhut, P.; Lee, I.; Timp, W.; Borth, N. Subcloning induces changes in the DNA-methylation pattern
of outgrowing Chinese hamster ovary cell colonies. Biotechnol. J. 2021, 16, e2000350. [CrossRef] [PubMed]
63. Templeton, N.; Smith, K.D.; McAtee-Pereira, A.G.; Dorai, H.; Betenbaugh, M.J.; Lang, S.E.; Young, J.D. Application of 13 C flux
analysis to identify high-productivity CHO metabolic phenotypes. Metab. Eng. 2017, 43, 218–225. [CrossRef] [PubMed]
64. Ma, N.; Ellet, J.; Okediadi, C.; Hermes, P.; McCormick, E.; Casnocha, S. A single nutrient feed supports both chemically defined
NS0 and CHO fed-batch processes: Improved productivity and lactate metabolism. Biotechnol. Prog. 2009, 25, 1353–1363.
[CrossRef]
65. Huang, Z.; Lee, D.Y.; Yoon, S. Quantitative intracellular flux modeling and applications in biotherapeutic development and
production using CHO cell cultures. Biotechnol. Bioeng. 2017, 114, 2717–2728. [CrossRef]