A Single Nutrient Feed Supports Both Chemically Defined NS0 and CHO

Fed-Batch Processes: Improved Productivity and Lactate Metabolism

Ningning Ma, JoAnn Ellet, Centy Okediadi, Paul Hermes, Ellen McCormick, and Susan Casnocha
Bioprocess R&D, Global Biologics, Pfizer Inc, Chesterfield, MO 63017

DOI 10.1002/btpr.238
Published online July 27, 2009 in Wiley InterScience (www.interscience.wiley.com).

A chemically defined nutrient feed (CDF) coupled with basal medium preloading was
developed to replace a hydrolysate-containing feed (HCF) for a fed-batch NS0 process. The
CDF not only enabled a completely chemically defined process but also increased recombi-
nant monoclonal antibody titer by 115%. Subsequent tests of CDF in a CHO process indi-
cated that it could also replace the hydrolysate-containing nutrient feed in this expression
system as well as providing an 80% increase in product titer. In both CDF NS0 and CHO
processes, the peak lactate concentrations were lower and, more interestingly, lactate metab-
olism shifted markedly from net production to net consumption when cells transitioned from
exponential to stationary growth phase. Subsequent investigations of the lactate metabolic
shift in the CHO CDF process were carried out to identify the cause(s) of the metabolic
shift. These investigations revealed several metabolic features of the CHO cell line that we
studied. First, glucose consumption and lactate consumption are strictly complementary to
each other. The combined cell specific glucose and lactate consumption rate was a constant
across exponential and stationary growth phases. Second, Lactate dehydrogenase (LDH) ac-
tivity fluctuated during the fed-batch process. LDH activity was at the lowest when lactate
concentration started to decrease. Third, a steep cross plasma membrane glucose gradient
exists. Intracellular glucose concentration was more than two orders of magnitude lower
than that in the medium. Fourth, a large quantity of citrate was diverted out of mitochondria
to the medium, suggesting a partially truncated tricarboxylic acid (TCA) cycle in CHO cells.
Finally, other intermediates in or linked to the glycolytic pathway and the TCA cycle, which
include alanine, citrate, isocitrate, and succinate, demonstrated a metabolic shift similar to
that of lactate. Interestingly, all these metabolites are either in or linked to the pathway
downstream of pyruvate, but upstream of fumarate in glucose metabolism. Although the spe-
cific mechanisms for the metabolic shift of lactate and other metabolites remain to be eluci-
dated, the increased understanding of the metabolism of CHO cultures could lead to future
improvements in medium and process development. V C 2009 American Institute of Chemical

Engineers Biotechnol. Prog., 25: 1353–1363, 2009

Keywords: CHO, NS0, chemically defined process, lactate, metabolic shift, metabonomics,
metabolomics

Introduction
Hydrolysates, especially those of nonanimal origin, are
widely used as serum replacements in therapeutic protein
production processes using mammalian cells. Hydrolysates
can improve cell growth and protein yield in serum-free
processes (Burteau et al., 2003; Heidemann et al., 2000;
Sung et al., 2004; Xie et al., 1997). However, there are a
few limitations that complicate hydrolysates’ utilization. First
of all, lot-to-lot variation exists for hydrolysates (Luo and
Chen, 2007; Zhang et al., 2003). Prescreening using small-
scale performance testing is typically required to eliminate
underperforming lots. Second, the nutrient composition in
hydrolysates is not balanced for cell consumption. Various
components, such as ash and salts, are either not required by
Correspondence concerning this article should be addressed to N. Ma
at ningning.ma@pfizer.com.
the cells or are required at much lower levels than supplied.
Third, the chemically undefined nature of hydrolysates hin-
ders the scientists’ understanding process of cell metabolism.
For instance, it is not well understood how mammalian cells
metabolize amino acids in peptides, so that amino acid con-
sumption analysis based on free amino acids measurements
may or may not be accurate (Nyberg et al., 1999).
Development of various chemically defined media for
NS0 cells and CHO cells has been reported (Chen et al.,
2000; Hata et al., 1992; Huang et al., 2007; Zhang and Rob-
inson, 2005), and commercial chemically defined media are
readily available from all major mammalian cell culture me-
dium manufacturers. However, the development of chemi-
cally defined fed-batch processes that support rapid cell
proliferation, high cell density, and high production rate has
been challenging. Chemically defined fed-batch processes
have been reported recently. Gong et al. (2006) developed a
chemically defined fed-batch process for hybridoma cells.

C 2009 American Institute of Chemical Engineers

V 1353
1354
Peak viable cell density reached about 8
106 cells/mL, but
viable cells dropped sharply after a 5-day culture. Burky
et al. (2006) developed a generic chemically defined fed-
batch process for NS0 cells. The process supported peak via-
ble cell density around 1
107 cells/mL and product titer of
2.64 g/L on Day 13.
Lactic acid, or lactate, has long been recognized as a
major by-product in cell culture processes (Hu et al., 1987;
Miller et al., 1989; Ozturk and Palsson, 1992). Base addition
to neutralize lactic acid elevates medium osmolarity, which,
in turn, inhibits cell growth and causes viability to drop.
Reducing lactic acid production had been shown to increase
peak cell density, extend process duration, and increase
recombinant protein yield. Four strategies had been reported
in the literature to reduce lactate production. The first one is
to reduce medium glucose concentration to limit glucose
supply (Cruz et al., 1999; Zhou et al., 1995, 1997b). Using
an online dynamic feeding protocol, Zhou et al. (1995,
1997b) successfully maintained glucose at a low level of
about 0.5 mM. As a result, lactate production was reduced
and higher peak cell densities and higher viability were
achieved. A second strategy is to use alternative, slow-con-
suming carbon sources (Altamirano et al., 2004). Altamirano
et al. (2004) used a feeding strategy in which glucose and
galactose feeds were alternated. This strategy forced CHO
cells to consume more residual lactate than the culture fed
with glucose only. As a result, harvest viability was mod-
estly improved. The third strategy is to downregulate lactate
dehydrogenase-A (LDH-A) (the LDH subunit that effectively
catalyzes pyruvate to lactate conversion) using a wide vari-
ety of techniques including homologous recombination
(Chen et al., 2001), antisense mRNA (Jeong et al., 2001;
Paredes et al., 1999), and small interfering RNA (Kim and
Lee, 2007a). Chen et al. (2001) compared a LDH-A downre-
gulated hybridoma clone to a control clone. In the culture
utilizing the LDH-A downregulated cell line, lactate concen-
tration was about 35% lower and monoclonal antibody titer
was about 200% higher. However, although Kim and Lee
(2007a) reduced lactate production rate even further (by 45–
79%), the recombinant thrombopoietin product rates of three
LDH-A downregulated clones were not statistically better
than the control clone. A fourth strategy is to enhance the
flow of glucose carbon into tricarboxylic acid (TCA) cycle.
A focus area has been the overexpression of pyruvate car-
boxylase of yeast (Irani et al., 1999, 2002) or human origin
(Kim and Lee, 2007b). In a perfusion culture, Irani et al.
(2002) demonstrated that pyruvate carboxylase-transfected
BHK-21 cells produced half of the lactate when compared
with the control cells. Cell density was about 50% higher
and recombinant erythropoietin production rate was about
100% higher. More recently, it was reported that the expres-
sion of antiapoptosis genes, E1B-19K and Aven (EA167),
could reduce lactate production of CHO cells (Dorai et al.,
in press). In batch cultures with or without glucose limita-
tion, a CHO clone transfected with antiapoptotic genes gave
lower lactate concentration compared with the control cell
line at corresponding conditions.
In this study, we developed a chemically defined nutrient
feed (CDF) with companion production medium preloading
to replace a hydrolysate-containing feed (HCF) for a fed-
batch NS0 process. The CDF was later investigated in a
CHO process. In both NS0 and CHO processes, CDF
resulted in higher product titer and lower peak lactate con-
centration. The CDF processes also exhibited a dramatic lac-
Biotechnol. Prog., 2009, Vol. 25, No. 5
tate metabolic shift from net production to net consumption.
At the end of the process, lactate concentration in the CDF
processes was almost undetectable. Subsequent investigations
were carried out to better understand this metabolic shift in
the CDF CHO process.
Materials and Methods
Cells
NS0 and CHO cells producing test recombinant monoclo-
nal antibodies were used in this study. Both cell types used
glutamine synthatase as the transfection selection marker.
Process
NS0 and CHO cells were routinely subcultured in 250 mL
shake flasks (Corning, Acton, MA) in chemically defined hy-
bridoma and CHO media, respectively. Both hybridoma and
CHO media are commercially available. The CHO medium was
supplemented with 25 mM methionine sulfoximine (MSX).
Two-liter Applikon bioreactors (Applikon, Foster city,
CA) with a 1-L working volume were used in this study. In
these bioreactors, two pitched blade impellers provided mixing
and gas bubble dispersion. Temperature was controlled at
36.5 C using a heating blanket. pH was controlled at 7.0 by
carbon dioxide sparging through a ring sparger and by adding
either 0.5 N sodium hydroxide or 7.5% sodium bicarbonate.
Dissolved oxygen (DO) was controlled at 30% air saturation
using air overlay and oxygen sparging. The fed-batch process
for NS0 and CHO cells was very similar. Cells are seeded at
2.5
105 viable cells/mL in corresponding NS0 or CHO pro-
duction media. Starting from Day 3, a nutrient feed was added
into the bioreactors at 28 mL/L/day and 16 mL/L/day for NS0
and CHO cells, respectively, till the end of the process. Glucose
was fed separately to control glucose concentration around 2 g/
L. Bioreactors were sampled daily for offline measurements.
Routine bioreactors offline measurements
Cell density and viability were measured using a Cedex
cell counter (Innovatis, Bielefeld, Germany). Offline DO,
pH, glucose concentration, and lactate concentration were
measured with a BioProfile chemical analyzer (Nova Bio-
medical, Waltham, MA). A portion of the broth from each
daily sampling was filtered through a 0.2 lm polyethersul-
fone filter (Pall, Ease Hills, NY) for product titer and amino
acid measurements.
Monoclonal antibody titer was measured using an Agilent
HPLC (Hewlett Packard, Waldbronn, Germany) with a
POROS protein-A affinity column (Applied Biosystems, Fos-
ter City, CA). The column was calibrated using correspond-
ing reference standards before titer quantification.
Intracellular LDH activity measurement
About 2–4 mL of broth was taken out of bioreactors and
transferred to a centrifuge tube chilled on ice. The cells were
washed twice with cold phosphate buffered saline (PBS). In
each wash, the cells were pelletted in a refrigerated centri-
fuge at 150g for 5 min, the supernatant was discarded, and
the cell pellet was gently resuspended in 4 mL of cold PBS.
After the second wash, cells were diluted to 5
105 viable
cells/mL and lysed using freeze-thaw. In the freeze-thaw
procedure, cells were first flash frozen in liquid nitrogen and
Biotechnol. Prog., 2009, Vol. 25, No. 5 1355

then thawed overnight at 2–8 C. LDH activity was quantified

using a colorimetric LDH assay kit (CytoTox 96 nonradioac-
tive cytotoxicity assay, Promega, Madison, WI) and a spec-
trophotometer (Molecular Devices, Sunnyvale, CA). Sample
preparation and measurement followed manufacturers’
instructions. The LDH assay kit quantifies LDH activity
based on the conversion rate of lactate and NADþ to pyru-
vate and NADH.

Sample collection for metabonomics assay

Both cell and spent medium samples were collected for
metabonomics analysis. To collect cell samples, cell culture
broth containing 2
107 viable cells was removed from the
bioreactor and immediately chilled in an ice bucket. The
cells were pelleted at 1,000g for 3 min in a refrigerated cen-
trifuge (Bechman Coulter, Fullerton, CA) and then rinsed
twice using cold PBS. After two washes, the cell pellet was
flash frozen in liquid nitrogen and stored at 80 C. A 2 mL
spent medium sample was collected where cell broth was
first centrifuged at 1,000g for 3 min and then the supernatant
was filtered through a 0.2 lm polyethersulfone filter. Spent
medium samples were also stored at 80 C. After all sam-
ples were collected, they were sent on dry ice to the analysis
provider.

Sample preparation and metabonomics assay

Sample preparation and metabonomics analysis were con-
ducted at Metabolon (Durham, NC). The metabonomics
methodology is detailed elsewhere (Lawton et al., 2008) but
briefly, the samples were thawed, small molecules were
extracted, and the reconstituted extracts were split for analy-
sis by GC and LC mass spectrometry. Chromatographic sep-
aration of all ions in each sample was followed by library
matching of these ions to Metabolon’s reference library of
standards. The identity of metabolites was determined by
matching the combination of chromatographic retention
index and mass spectra signatures compared with the refer-
ence library entries.
Results and Discussion
Development of a chemically defined nutrient
feed for NS0 process
A CDF was developed to replace a hydrolysate-containing
nutrient feed (HCF) that was previously used for a fed-batch
NS0 process. In the hydrolysate-containing fed-batch pro-
cess, the basal medium is chemically defined, but the nutri-
ent feed contains a nonanimal-derived hydrolysate. The
limitations of hydrolysates, such as lot-to-lot variation,
prompted the development of the CDF to achieve a fully
chemically defined process.
The development took a top-down approach, as illustrated
in Figure 1. There were three major development steps: (1)
develop an over-rich nutrient feed, (2) simplify formula by
removing unnecessary components and reduce the concentra-
tion of overfeed components, and (3) polishing and formula-
tion. The plan for the over-rich solution was that it should
contain all nutrients that NS0 cells require and all required
nutrients should be at sufficiently high concentrations. To
ensure that all necessary components were included in the
over-rich solution, over 90 components were incorporated.
These components came from the formula of a Pfizer propri-
Figure 1. The top-down approach used in the chemically
defined nutrient feed development.
The development included three major steps: (I) obtain an over
rich nutrient feed, (II) formula simplification, and (III) polish-
ing and formulation.
etary chemically defined NS0 medium and a previously
developed CDF. To ensure that all necessary components
were in ample supply, most components were added at a
high concentration. In the next step, nutrients in the over-
rich solution were divided into eight rough groups to balance
their ratio. Grouping was based on nutritional category (trace
elements, nucleotides, etc.) or by solubility (low pH, ethanol,
etc.). In the subsequent formula simplification steps, compo-
nents were screened in groups for two cycles, first in rough
groups and then in refined groups. In the first cycle, two of
eight rough groups were found unnecessary where their
exclusion rendered no detectable difference in process per-
formance (cell density, viability, and product titer). Compo-
nents in these two subgroups were subsequently removed
from the formula. The remaining components were divided
into 13 refined groups again based on nutritional category
and solubility. Most groups contained less than five compo-
nents. In the second cycle of screening, five refined groups
1356
Figure 2. Comparison of the CDF (solid lines) and HCF (dot-
ted lines) NS0 processes.
(a) Viable cell density and viability. (b) Product titer and lac-
tate concentration. The results shown are the average of three
replicate batches. Error bar indicates 1 SD.
were found to offer no detectable beneficial (or detrimental)
effect on the process performance and were removed. In the
final step, the components in the simplified formula, with
less than 40 components, were balanced for optimization.
Amino acids were optimized based on spent medium analy-
sis, whereas the other components were optimized relied on
refined group titration study. Finally, all components were
formulated into a single pH neutral nutrient feed. For com-
ponents with low solubility in the feed, the portion over the
solubility limit was preloaded into the basal medium. All de-
velopment studies were conducted in fed-batch shake flasks.
At each milestone, the performance was confirmed in 2 L
bench-scale bioreactors. The whole development took 9
months, with each of three major steps for about 3 months.
A comparison of the CDF process and HCF process in
2 L bioreactors is presented in Figure 2, where results are
the average of three bioreactor runs and error bars indicate
1 SD. Cell growth phase was extended in the CDF process
where the peak viable cell density reached 15
106 cells/
mL, 50% higher than that in the HCF process. Viability
was modestly improved in the CDF process, the harvest via-
bility was about 10% higher in the CDF process. Product ti-
ter was much higher in the CDF process after Day 5. At
harvest on Day 11, product titer in the CDF process was
115% higher than that in the HCF process. This 115%
increase was from a 33% increase in the integral viable cell
Biotechnol. Prog., 2009, Vol. 25, No. 5
concentration and a 62% increase in cell specific productiv-
ity (Qp). These results demonstrate that CDF not only allows
for removing hydrolysate from the process but also delivers
better cell growth and productivity. Another benefit of the
CDF process is the lower lactate concentration. As shown in
Figure 2b, lactate in the CDF process was lower throughout
the whole process, and the peak concentration was only half
of that in the HCF process. In both processes, cells switched
from net lactate production to net consumption around Day
5. In the CDF process, residual lactate was consumed in 2
days and remained lower than the detection limit until har-
vest. In the HCF process, the lactate concentration was gen-
erally higher and net lactate production resumed, although at
a very slow rate, after Day 8.
The benefit of the top-down approach to nutrient feed de-
velopment is that it eliminates nutrient limitation up front,
thereby allowing the development to focus on nutrient bal-
ancing and simplification. The risk of this strategy is the
inclusion of unnecessary components and overfeeding, which
could be toxic or inhibitory to cell growth or production.
The successful development of a CDF demonstrates that the
risk is manageable. During the development, many compo-
nents, such as vitamins and minerals, were found to be non-
toxic even at a concentration one or two orders of magnitude
higher than what the cells needed. This observation agrees
with Xie and Wang (1994), who found that a high concentra-
tion of vitamins was not harmful to animal cells. Amino
acids have the potential to be inhibitory to cell growth at
high concentrations. Fortunately, amino acid overfeeding can
be avoided using spent medium analysis.
Evaluation of the CDF in CHO processes
The CDF developed for NS0 cells was later tested on
CHO cells and compared with the CHO HCF process. The
medium used in both processes is a commercially available
chemically defined CHO medium. As shown in Figure 3, the
CDF process offered clear advantages over the HCF process.
Results shown in Figure 3 were the average of two duplicate
bioreactors. Peak viable cell density reached 18
106
cells/mL in both processes. However, viability was main-
tained better in the CDF process. When the processes were
terminated on Day 15, viability for the CDF process was
around 80%, whereas that for the HCF process had declined
to 50%. Product titer was similar in both processes till Day
10, after which titer started to level off in the HCF process,
whereas that in the CDF process kept rising linearly till har-
vest. By Day 15, product titer in the CDF process was 82%
higher than that in the HCF process. Lactate concentration
was much lower in the CDF process after Day 7. By Day
10, lactate concentration in the CDF process became close to
undetectable, whereas that in the HCF process stayed at 1.8–
2.6 g/L.
Lactate metabolic shift in the CDF processes
A major metabolic feature of the CDF processes is the
dramatic shift of lactate metabolism from net production to
net consumption. On one hand, the metabolic shift is benefi-
cial for the process as it significantly reduced the concentra-
tion of a major cellular metabolic by-product. At the end of
the process, lactate concentration was almost undetectable.
On the other hand, the shift could be due to a glucose carbon
limitation, which may have a detrimental effect on the cell
Biotechnol. Prog., 2009, Vol. 25, No. 5
Figure 3. Comparison of the CDF (solid lines) and HCF (dot-
ted lines) CHO processes.
(a) Viable cell density and viability. (b) Product titer and lac-
tate concentration. The results shown are the average of two
duplicate batches.
growth and productivity. Although glucose supply was not
limited in both CDF NS0 and CDF CHO processes (glucose
concentration [1 g/L), possible limitations could still exist
in the glucose transportation step or the flux through the gly-
colytic pathway. To better understand and to identify the
possible mechanism(s) for the lactate metabolic shift, we
examined carbon metabolisms in the CDF CHO process. The
lactate metabolic shift in fed-batch processes has been previ-
ously reported with different cell expression systems, includ-
ing DHFR-CHO (Tsao et al., 2005), GS-NS0 (Zhou et al.,
1997a), and a non-GS-NS0 (Burkey et al., 2006). However,
no thorough investigation of this phenomenon has been pub-
lished. We focused our investigation on CHO cells, where
the cell line used was a subclone of the CHO cell line used
in the initial CDF evaluation (Figure 3). The CDF fed-batch
process was kept the same.
As indicated in Figure 4a, the lactate metabolic shift took
place in the same period when cells were transitioning from
exponential to stationary growth phase. The same timing cor-
relation between the lactate metabolic shift and cell growth
phase transition exists in previous publications using differ-
ent mammalian cell types (Burkey et al., 2006; Tsao et al.,
2005; Zhou et al., 1997a). Whether there is a cause-effect
relationship between lactate metabolic shift and cell growth
slowdown is unknown. One possibility is that the metabolic
shift results from the cell growth phase transition. High lac-
1357
Figure 4. Cell growth and glucose and lactate metabolism of
CHO cells in the CDF process in two duplicate
bioreactors.
Bioreactor 1, solid symbols; Bioreactor 2, hollow symbols. (a)
Cell growth and lactate profile. The arrow indicates the correla-
tion between shift of lactate metabolism and cell growth phase
transition. (b) Weight cumulative glucose consumption, lactate
production, and net glucose consumption. The net glucose con-
sumption is defined as glucose consumption minus lactate pro-
duction. The glucose and lactate metabolism can be divided
into three phases: (I) high glucose consumption and net lactate
production, (II) low glucose consumption and net lactate
consumption, and (III) high glucose consumption and very low
residual lactate concentration.
tate production is a hallmark of transformed cells, a phenom-
enon established more than 70 years ago (Warburg, 1930).
High lactate production has been attributed in the literature
to the upregulation of various enzymes in the glycolytic
pathway (Board et al., 1990; Dang and Semenza, 1999;
Kondoh et al., 2005). As cells exit the rapid proliferating
phase, they may concurrently cease high lactate production.
Alternatively, the lactate metabolism shift could be a result
of glucose carbon limitation because of restricted flux
through the glycolytic pathway.
Analysis of glucose consumption revealed a strong com-
plementary relationship between glucose and lactate con-
sumption. Cumulative glucose utilization, lactate production,
and net glucose consumption (defined as combined glucose
and lactate consumption) from the two duplicate bioreactors
referred to in Figure 4a are illustrated in Figure 4b. The
slope of the curves in Figure 4b represents either cell-spe-
cific substrate consumption (glucose or net glucose) or pro-
duction (lactate) rate. The glucose consumption and lactate
production rates fluctuated during the process. However, net
glucose consumption rate was steady throughout the whole
process. This observation inferred that (1) cell-specific
1358 Biotechnol. Prog., 2009, Vol. 25, No. 5

Figure 5. Major NAD1 and NADH conversion pathways.

pyruvate consumption rate in TCA cycle was a constant

throughout the whole process and (2) glucose and lactate
consumption rate was tightly controlled so that the flux from
glucose to pyruvate and the flux from lactate to pyruvate are
strictly complementary to each other.
NADH homeostasis, as previously being suggested (e.g.,
Zhou et al., 1997a), could be the cause of the tight control
of glucose and lactate consumption. The three major path-
ways related to NADH homeostasis are illustrated in Figure
5. The conversion of pyruvate to lactate recycles NADH pro-
duced in glycolysis back to NADþ. When the conversion is
reversed, it competes with glycolysis for NADþ. The
reduced availability of NADþ could slow down the glycoly-
sis rate. NADH is also recycled back to NADþ inside mito-
chondria through the aspartate–malate shuttle (Figure 5) or
similar shuttle systems. We need to assume that the shuttle Figure 6. Variation of LDH activity in the CHO CDF process.
rate was a constant for NADH homeostasis to exert the com- Results were from two similar bioreactor runs using the same
clone. Bioreactor 1, solid symbols; Bioreactor 2, hollow
plementation effect on glucose and lactate consumption. It symbols.
should be noted that in Stage II and III of Figure 4b, all
NADH needed to be recycled in mitochondria as no NADH
was recycled through pyruvate to lactate conversion.
to lactate production. More studies are needed to fully under-
stand the role of LDH activity variation in lactate metabolic
LDH activity changes during the process shift, such as the ratio between the two major LDH subunits,
Cell-specific LDH activity was monitored in two bioreactor LDH-A and LDH-B, which favor the pyruvate–lactate con-
runs using the same CHO subclone and CDF process. As version in opposite directions, and the ratio between the other
shown in Figure 6, cell-specific LDH activity oscillated in two substrates in the pyruvate–lactate conversion reaction,
both runs. During the exponential growth phase, LDH activ- NADþ and NADH, where a high NADþ/NADH ratio favors
ity first increased and then decreased. It further decreased lactate to pyruvate conversion.
and reached the lowest level during the growth phase transi-
tion from exponential to stationary phase. With the decrease
of cell-specific LDH activity, lactate metabolism shifted from Shift of amino acid metabolism
net production to net consumption. After the lactate meta- In addition to lactate metabolic shift, several amino acids
bolic shift, LDH activity gradually increased and reached its also demonstrated shifts of metabolism during the fed-batch
maximum at the end of the process. Cell size was consistent process, as shown in Figure 7. The results were from a sin-
during the fed-batch process (data not shown), so that the gle bioreactor run. The alanine shift is of particular interest
LDH activity variation was not due to a cell volume change. as similar to lactate, alanine can be derived from pyruvate.
Although there seems to be a linkage between LDH activity Figure 7 showed that the pattern of alanine metabolic shift
and lactate metabolism, the results were hard to interpret. was similar to that of lactate, although the alanine production
First, in the LDH assay that we used, LDH activity is quanti- and consumption rate was much lower than that of lactate.
fied by the rate of lactate to pyruvate conversion, hence it is Asparagine net consumption rate was initially very high, but
counterintuitive to find that the lowest activity measured off- decreased rapidly midway through the process. The decrease
line correlated to the highest lactate to pyruvate conversion was due to asparagine supply limitation. As most asparagine
rate in the process. Second, although decreased LDH activity is likely utilized in the TCA cycle through aspartate and
was correlated with net lactate consumption, subsequent then oxaloacetate, under asparagine limitation, net aspartate
increase in LDH activity did not reverse the metabolism back consumption rate would increased, as observed in Figure 7.
Biotechnol. Prog., 2009, Vol. 25, No. 5
Figure 7. Cumulative consumption of amino acids whose me-
tabolism shifted in the CHO CDF process.
Other amino acids, except for glutamine, showed constant con-
sumption rates. Glutamine consumption was not monitored as
no glutamine was supplied in the process.
It should be noted that asparagine limitation was not the
cause of the lactate metabolic shift. Increasing asparagine
supply did not alter either cell growth or lactate metabolism
(data not shown). For the other two amino acids, glycine
shifted from net production to net consumption and serine
shifted from high to low net consumption during the process.
Glycine can be produced from 3-phosphoglycerate, a glyco-
lytic pathway intermediate, through serine. But whether this
link is related to the metabolic shift of glycine and serine is
unknown.
Glycolytic pathway and TCA cycle metabolite profiling
A more broad investigation of metabolites in the glyco-
lytic pathway and TCA cycle was conducted as part of a
metabonomics study. In this study, we monitored the change
of intracellular and extracellular metabolite pools in the CDF
CHO fed-batch process in one bioreactor run. One other spe-
cific purpose of the metabonomics study was to verify
whether there were any flux limitation steps in the glycolytic
pathway that limit glucose utilization and force cells to use
lactate.
The profiling of glycolytic pathway intermediates and the
TCA cycle intermediates are shown in Figures 8 and 9,
respectively. It should be noted that the results shown in Fig-
ures 8 and 9 are relative ion count results from mass spec-
trometry measurements. As different metabolites have
different ionization potential, data shown in Figures 8 and 9
cannot be used to compare relative concentration among dif-
ferent metabolites. However, for a specific metabolite, ion
counts are linearly proportional to its concentrations. Hence,
the ion count data of one specific metabolite can be used to
study its concentration change during the process and to
compare its abundance between extracellular and intracellu-
lar samples. Because the intracellular samples were more
diluted than the medium (extracellular) samples, the intracel-
lular results were adjusted based on the difference of sample
dilution factors between intracellular and extracellular sam-
ples, as shown in Eq. 1, so that the intracellular and extracel-
lular results could be directly comparable. IC indicates ion
counts in Eq. 1.
Intracelluar sample dilution factor
ICAdjusted ¼ ICOriginal (1)
Extracelluar sample dilution factor
1359
Figure 8. Concentration of glycolytic pathway metabolites in
the CHO CDF process.
(a) Intracellular concentration. (b) Extracellular concentration.
Only metabolites that were detected on consecutive days are
linked with lines. The sporadically detected metabolites are
shown as individual data points. Missing data points were
below detection limit.
Figure 8a indicates that the intracellular glucose ion count
was very low, being detected only sporadically in the first 5
days. In contrast, the extracellular glucose ion count was
much more abundant (Figure 8b). The difference between
the intracellular and extracellular glucose concentration is
more than two orders of magnitude, indicating a steep con-
centration gradient across the plasma membrane. The intra-
cellular glucose concentration is regulated by two factors,
glucose transportation and its subsequent phosphorylation.
Glucose transportation is accomplished through a family of
13 facilitative glucose transporter (GLUT) proteins (Macheda
et al., 2005). GLUT1 (Harrison et al., 1991) and GLUT4
(Bogan et al., 2001) are likely expressed on CHO cells. Pho-
phorylation of glucose by hexokinase II (HK II) is the first
reaction in the glycolysis pathway. In transformed cells, HK
II activity was found being upregulated by 100 times
(Peterson et al., 2002). This upregulation was achieved by
two means: overexpression of HK II enzyme and docking of
HK II to porins on the mitochondrial membrane. If glucose
transportation was the rate limiting step regulating the intra-
cellular glucose concentration, the immediate metabolite of
glucose, glucose-6-phosphate, should be depleted because of
glucose limitation. However, intracellular glucose-6-phos-
phate ion count was at a relative high level during most of
the process, and its extracellular concentration increased by
a factor of 10 during the process. These data suggest that the
glycolytic pathway had an ample supply of glucose-6-
1360
Figure 9. Concentration of TCA cycle metabolites in the CHO
CDF process.
(a) Intracellular concentration. (b) Extracellular concentration.
Only metabolites that were detected on consecutive days are
linked with lines. The sporadically detected metabolites are
shown as individual data points. Missing data points were
below detection limit.
phosphate. Therefore, the low intracellular glucose concen-
tration is more likely due to high HK II activity, instead of
low glucose transportation rate.
The extracellular lactate concentration profile matches that
measured using a conventional chemistry analyzer (BioPro-
file of Nova Biomedical). As shown in Figure 8b, the lactate
concentration increased 10-fold from Day 2 to 6, after
which the concentration dropped sharply by 30-fold and
stayed at a low level until the end of the process. It is inter-
esting to observe that on Day 9, when the extracellular lac-
tate concentration decreased to a very low level, pyruvate in
the broth became undetectable (Figure 8b). At the same
time, the two glycolytic metabolites preceeding pyruvate for-
mation, 3-phosphoglycerate and phosphoenolpyruvate, had
accumulated. These observations seem to suggest that the
flow from phosphoenolpyruvate to pyruvate might be par-
tially blocked after Day 9, which resulted in the buildup of
metabolites in front of the blockage, and depletion of the
metabolites downstream of it. However, this hypothesis of a
partial glycolytic pathway blockage conflicts with the glu-
cose consumption rate change illustrated in Figure 4b. Figure
4b showed that after all lactate was consumed, glucose con-
sumption rate went up (stage II to stage III transition). Alter-
natively, the accumulation of 3-phosphoglycerate and
phosphoenolpyruvate could be due to a slow down of cell
proliferation. As both 3-phosphoglycerate and phosphoenol-
Biotechnol. Prog., 2009, Vol. 25, No. 5
pyruvate are intermediates for biomass formation, slow down
of cell proliferation would reduce their consumption rates.
Intracellular and extracellular TCA cycle intermediates are
shown in Figures 9a,b respectively. One important metabo-
lite, oxaloacetate, was not measured as its volatility makes it
hard to preserve during sample preparation. All other
detected intermediates were maintained at a steady concen-
tration in the cells during the fed-batch process (Figure 9a).
Isocitrate detection was sporadic, because of its concentra-
tion being close to the detection limit. Considerably, more
variation was observed with the extracellular metabolites
(Figure 9b). All metabolite levels changed during the pro-
cess. Citrate, isocitrate, and succinate showed a similar meta-
bolic shift as lactate and alanine. Their concentration
increased till Day 9 or 10 and then decreased. As none of
these intermediates was supplemented during the process,
any extracellular accumulation was due to an efflux of
metabolites from mitochondria to the cytosol and then into
the broth.
Citrate concentration was much higher in the broth. Citrate
must be transported out of the cells against a steep concentra-
tion gradient. Previous studies had revealed that various can-
cer and transformed cells have a truncated TCA cycle where
flux from citrate to a-ketoglutarate was downregulated (Bag-
getto, 1992). One proposed mechanism for the downregula-
tion is a strong efflux of citrate out of mitochondria to the
cytosol. Once in the cytosol, citrate can be used to synthesize
cholesterol and fatty acids, critical building blocks for mem-
brane production in rapidly growing cancer or transformed
cells (Hatzivassiliou et al., 2005). In this study, it was clearly
shown that citrate was not only transported out of mitochon-
dria but also out of cells. The high citrate accumulation rate
in the first 9 days suggests that the TCA cycle in CHO cells
might be partially truncated in these days. After Day 9, net
citrate accumulation was switched to net consumption.
Because the absolute citrate concentration was not known,
we cannot estimate the percentage of pyruvate carbon that
was diverted out of TCA cycle as citrate. However, consider-
ing the vast volume difference between broth and cells, the
efflux of citrate out cells must be a considerable drain for the
TCA cycle. The extracellular isocitrate concentration was
also higher than that in the cells, indicating a transportation
of isocitrate out of cells as well. It should be noted that the
intracellular concentration shown in Figure 9a is neither the
cytosolic concentration nor the mitochondrial concentration,
but an average based on cellular volume. However, the use of
intracellular average concentration does not change the citrate
and isocitrate cellular efflux conclusion.
The extracellular malate and fumarate concentrations
increased during the process. Their concentrations were
much lower initially, but reached a similar level to their cor-
responding intracellular concentrations at the end of the pro-
cess. Malate can be a TCA cycle overflow gate. When extra
malate is produced, it is converted to pyruvate by malate
enzyme (Godia and Cairo, 2006). Results shown in this
study indicate that the conversion of malate to pyruvate
might be limited as malate concentration continued increas-
ing in the medium while pyruvate became undetectable after
Day 8. The concentration of a-ketoglutarate varied during
the process without a clear trend. Similar metabolic shift
was observed in a NS0 metabonomics study where lactate,
alanine, citrate, and succinate shifted from net production to
net consumption (data not shown). Isocitrate was not
detected in the NS0 study.
Biotechnol. Prog., 2009, Vol. 25, No. 5
Figure 10. Glycolytic pathway and TCA cycle.
Metabolites that demonstrated a shift from net accumulation to net consumption during CHO CDF process are in bold.
Metabolic shift of glycolytic pathway and TCA
cycle intermediates
Metabolites discussed in the previous sections are sum-
marized in Figure 10. The intermediates showing a transition
from net accumulation to net consumption are in bold, and
the intermediates that were undetectable or not monitored
are labeled with an asterisk. It is interesting to note that the
shifted metabolites are clustered together, downstream of py-
ruvate and upstream of fumarate with citrate considered the
start of the TCA cycle. Of all the intermediates monitored in
this segment, a-ketoglutarate is the only one showing a dif-
ferent metabolic profile from lactate. The similar metabolism
of a group of closely related metabolites suggests that the
metabolic shift was controlled by flux or by a group of
highly coordinated enzymes. A possible flux controlling
point is the synthesis of pyruvate from phosphoenolpyruvate.
Again, the flux limitation hypothesis does not agree with the
increased glucose consumption rate (Figure 4b from Stage II
to Stage III). More studies are needed to elucidate the true
mechanism of the metabolic shift. First, in this study, we
monitored only a single enzyme (LDH). A more broad pro-
teomics study is needed to assist interpreting the metabolo-
mics results. Second, the metabolic analysis was largely
based on trend analysis. A quantitative flux analysis could
increase the quality of the analysis and might lead to new
findings. Third, it might be necessary to consider other path-
ways, such as urea cycle and lipid synthesis pathways that
1361
link closely to the central metabolic pathway, to understand
the mechanism(s).
Conclusions
A CDF with companion medium preloading for a fed-
batch NS0 process was successfully developed to replace
a HCF. The new feed not only allowed us to achieve a
fully chemically defined process but also offered higher
peak cell density and higher product titer. When the CDF
with its companion medium preloading was tested in a
CHO process, it again demonstrated multiple significant
benefits. In addition to allowing the removal of hydroly-
sates, both viability and product titer were substantially
improved. A lower lactate concentration is an additional
benefit of the CDF processes in both NS0 and CHO cul-
tures. In the CDF processes, there was a rapid shift in lac-
tate metabolism that resulted in a nearly complete
depletion of lactate in the second half of the process. It
was also observed that the lactate metabolic shift occurred
concurrently with the transition of cell growth from expo-
nential to stationary phase.
Analysis of net glucose consumption by CHO cells indi-
cated that the glycolytic pathway and pyruvate to lactate
conversion are closely coordinated. Their flows are strictly
complementary to each other so that the combined cell spe-
cific glucose and lactate consumption rate is a constant. A
1362
plausible factor that could exert this tight control is cellular
maintenance of the NADþ/NADH balance.
LDH activity oscillated during the process. The lowest ac-
tivity measured was during the lactate metabolic shift from
net accumulation to net consumption. However, the results
seem to be counterintuitive so that a more in-depth study of
the LDH subunits and the other pair of substrates in the lac-
tate–pyruvate reaction, NADþ and NADH, is needed to
determine the link between the LDH activity change and the
lactate metabolic shift.
Additional studies with the CHO process indicated that
several amino acids demonstrated metabolic shifts as well.
Alanine, another byproduct of pyruvate, showed a similar
net accumulation to net consumption shift as lactate. This
demonstrates that the metabolic shift is not limited to pyru-
vate–lactate conversion only.
Finally, profiling of most glycolytic and TCA cycle inter-
mediates in the CHO CDF process was conducted. In this
study, both intracellular and extracellular metabolite levels
were measured. A steep glucose concentration gradient was
observed across the plasma membrane, where intracellular
glucose concentration was more than two orders of magni-
tude lower than the extracellular concentration. The low glu-
cose concentration is likely due to the high phosphorylation
rate of glucose in the cytosol, rather than a glucose transpor-
tation limitation.
When lactate was nearly depleted, the medium pyruvate
became undetectable in the CHO CDF process. Meanwhile,
two glycolytic intermediates upstream of pyruvate formation, 3-
phosphoglycerate and phosphoenolpyruvate, accumulated in the
medium. The opposite profile between phosphoenolpyruvate and
pyruvate suggests that the flux from the former to the latter
might be partially blocked around the time at which the lactate
metabolic shift occurred. However, this hypothesis is not sup-
ported by the glucose consumption data, where glucose con-
sumption rate increased after the lactate metabolic shift.
A considerable amount of citrate was exported out of mi-
tochondria into the medium in the first 9-day culture, after
which citrate metabolism converted from net accumulation
to net consumption in the CHO CDF process. The extracel-
lular citrate concentration increased up to 10-fold higher
than the intracellular average concentration. This observation
suggests that the TCA cycle in CHO cells might be partially
truncated. Besides citrate, isocitrate and succinate also dem-
onstrated similar metabolic shift as lactate and alanine. More
interestingly, all the metabolites showing a metabolic shift
are clustered together, generally downstream of pyruvate and
upstream of fumarate.
Although the CDF fed-batch process demonstrated signifi-
cant empirical benefits over the previous HCF process and
provides a platform for the study of cellular metabolism, the
cause(s) of the metabolic shifts in the CDF CHO process is
still unknown. It could be a flux restriction, where a possible
restriction point is the conversion of phosphoenolpyruvate to
pyruvate or a systematic enzymatic shift during cell growth
phase transition. Follow-up studies, especially a companion
proteomics study, will help elucidate the mechanism(s).
Acknowledgments
The authors appreciate Dr. Kurt Droms and Dr. Amit Bane-
rjee for their critical review of the manuscript and Process De-
velopment Analytical group for the amino acid analyses.
Biotechnol. Prog., 2009, Vol. 25, No. 5
Literature Cited
Altamirano C, Paredes C, Illanes A, Cairo JJ, Godia F. Strategies
for fed-batch cultivation of t-PA producing CHO cells: substitu-
tion of glucose and glutamine and rational design of culture me-
dium. J Biotechnol. 2004;110:171–179.
Baggetto LG. Deviant energetic metabolism of glycolytic cancer
cells. Biochimie. 1992;74:959–974.
Board M, Hummt S, Newsholme EA. Maximum activities of key
enzymes of glycolysis, glutaminolysis, pentose phosphate pathway
and tricarboxylic acid cycle in normal, neoplastic and suppressed
cells. Biochem J. 1990;265:503–509.
Bogan JS, McKee AE, Lodish HF. Insulin-responsive compartments
containing GLUT4 in 3T3-L1 and CHO cells: regulation by
amino acid concentrations. Mol Cell Biol. 2001;21:4785–4806.
Burky JE, Wesson MC, Young A, Farnsworth S, Dionne B, Zhu Y,
Hartman TE, Qu L, Zhou W, Sauer PW. Protein-free fed-batch
culture of non-GS NS0 cell lines for production of recombinant
antibodies. Biotechnol Bioeng. 2006;96:281–293.
Burteau CC, Verhoeye FR, Mols J, Ballez JS, Agathos SN, Schnei-
der YJ. Fortification of a protein-free cell culture medium with
plant peptones improves cultivation and productivity of an inter-
feron-c -producing CHO cell line. In Vitro Cell Dev Biol Anim.
2003;39:291–296.
Chen K, Liu Q, Xie L, Sharp PA, Wang DIC. Engineering of a mam-
malian cell line for reduction of lactate formation and high mono-
clonal antibody production. Biotechnol and Bioeng. 2001;72:55–61.
Chen Z, Iding K, Lukemeyer D, Lehmann. A low-cost chemically
defined medium for a recombinant CHO cell line producing pro-
thrombin. Biotechnol Lett. 2000;22:837–841.
Cruz HJ, Moreira JL, Carrondo MJT. Metabolic shifts by nutrient
manipulation in continuous cultures of BHK cells. Biotechnol and
Bioeng. 1999;66:104–113.
Dang CV, Semenza GL. Oncogenic alterations of metabolism.
Trends Biochem Sci. 1999;24:68–72.
Dorai H, Kyung YS, Ellis D, Kinny C, Lin C, Jan D, Moore G,
Betenbaugh MJ. Expression of anti-apoptotic genes alters lactate
metabolism of Chinese hamster ovary cells in culture. Biotechnol
Bioeng. In press.
Godia F, Cairo JJ. Cell metabolism. In:Ozturk SS,Hu W-S, editor-
s.Cell Culture Technology for Pharmaceutical and Cell-Based
Therapies. New York: Taylor and Francis; 2006. pp 81–112.
Gong X, Li D, Li X, Fang Q, Han X, Wu Y, Yang S, Shen BQ.
Fed-batch culture optimization of a growth-associated hybridoma
cell line in chemically defined protein-free media. Cytotechnol-
ogy. 2006;52:25–38.
Harrison SA, Buxton JM, Czech MP. Suppressed intrinsic catalytic
activity of GLUT1 glucose transporters in insulin-sensitive 3T3-
L1 adipocytes. Proc Natl Acad Sci USA. 1991;88:7839–7843.
Hata J, Tamura T, Yokoshima S, Yamashita S, Kabeno S, Matsu-
moto K, Onadera K. Chemically defined medium for the
production of biologically active substances of CHO cells. Cyto-
technology. 1992;10:9–14.
Hatzivassiliou G, Zhao F, Bauer DE, Andreadis C, Shaw AN, Dhanak D,
Hingorani SR, Tuveson DA, Thompson CB. ATP citrate lyase inhibi-
tion can suppress tumor cell growth. Cancer Cell. 2005;8:311–321.
Heidemann R, Zhang C, Qi H, Rule JL, Rozales C, Park S, Chuppa
S, Ray M, Michaels J, Konstantinov K, Naveh D. The use of pep-
tones as medium additives for the production of a recombinant
therapeutic protein in high density perfusion cultures of mamma-
lian cells. Cytotechnology. 2000;32:157–167.
Hu WS, Dodge TC, Frame KK, Himes VB. Effect of glucose on
the cultivation of mammalian cells. Developments in biological
standardization. 1987;66:279–290.
Huang EP, Marquis CP, Gray PP. Development of super-CHO pro-
tein-free medium based on a statistical design. J Chem Technol
Biotechnol. 2007;82:431–441.
Irani N, Beccaria AJ, Wagner R. Expression of recombinant cyto-
plasmic yeast pyruvate carboxylase for the improvement of the
production of human erythropoietin by recombinant BHK-21
cells. J Biotechnol. 2002;93:269–282.
Irani N, Wirth M, van Den Heuvel J, Wagner R. Improvement of
the primary metabolism of cell cultures by introducing a new
cytoplasmic pyruvate carboxylase reaction. Biotechnol and Bio-
eng. 1999;66:238–246.
Biotechnol. Prog., 2009, Vol. 25, No. 5
Jeong DW, Kim TS, Lee JW, Kim KT, Kim HJ, Kim IH, Kim IY.
Blocking of acidosis-mediated apoptosis by a reduction of lactate
dehydrogenase activity through antisense mRNA expression. Bio-
chem Biophys Res Commun. 2001;289:1141–1149.
Kim SH, Lee GM. Down-regulation of lactate dehydrogenase-A by
siRNAs for reduced lactic acid formation of Chinese hamster
ovary cells producing thrombopoietin. Appl Microbiol Biotechnol.
2007a;74:152–159.
Kim SH, Lee GM. Functional expression of human pyruvate carbox-
ylase for reduced lactic acid formation of Chinese hamster ovary
cells (DG44). Appl Microbiol Biotechnol. 2007b;76:659–665.
Kondoh H, Lleonart ME, Gil J, Wang J, Degan P, Peters G, Marti-
nez D, Carnero A, Beach D. Glycolytic enzymes can modulate
cellular life span. Cancer Res. 2005;65:177–185.
Lawton KA, Berger A, Mitchell M, Milgram KE, Evans AM, Guo
L, Hanson RW, Kalhan SC, Ryals JA, Milburn MV. Pharmacoge-
nomics. 2008;9:383–397.
Luo Y, Chen G. Combined approach of NMR and chemometrics for
screening peptones used in the cell culture medium for the pro-
duction of a recombinant therapeutic protein. Biotechnol Bioeng.
2007;97:1654–1659.
Macheda ML, Rogers S, Best JD. Molecular and cellular regulation
of glucose transporter (GLUT) proteins in cancer. J Cell Physiol.
2005;202:654–662.
Miller WM, Wilke CR, Blanch HW. Transient responses of hy-
bridoma cells to nutrient additions in continuous culture: I. Glu-
cose pulse and step changes. Biotechnol and Bioeng. 1989;33:
477–486.
Nyberg GB, Balcarcel RR, Follstad BD, Stephanopoulos G, Wang
DIC. Metabolism of peptide amino acids by Chinese hamster
ovary cells grown in a complex medium. Biotechnol Bioeng.
1999;62:324–335.
Ozturk SS, Riley MR, Palsson BO. Effects of ammonia and lactate
on hybridoma growth, metabolism, and antibody production. Bio-
technol and Bioeng. 1992;39:418–431.
Paredes C, Prats E, Cairos JJ, Azorin F, Cornudella Ll, Godia F.
Modification of glucose and glutamate metabolism in hybridoma
cells through metabolic engineering. Cytotechnology. 1999;30:
85–93.
1363
Peterson PL, Mathupala S, Rempel A, Geschwind JF, Ko YH. Mito-
chondrial bound type II kinase: a key player in the growth and
survival of many cancers and an ideal prospect for therapeutic
intervention. Biochim Biophys Acta. 2002;1555:14–20.
Sung YH, Lim SW, Chung JY, Lee GM. Yeast hydrolysate as a
low-cost additive to serum-free medium for the production of
human thrombopoietin in suspension cultures of Chinese hamster
ovary cells. Appl Microbiol Biotechnol. 2004;63:527–536.
Tsao YS, Cardoso AG, Condon RGG, Voloch M, Lio P, Lagos JC,
Kearns BG, Liu Z. Monitoring Chinese hamster ovary cell culture
by the analysis of glucose and lactate metabolism. J Biotechnol.
2005;118:316–327.
Warburg O.The Metabolism of Tumors. London: Arnold Constable;
1930.
Xie L, Nyberg G, Gu X, Li H, Möllborn F, Wang DIC. Gammain-
terferon production and quality in stoichiometric fed-batch cul-
tures of Chinese hamster ovary (CHO) cells under serum-free
conditions. Biotechnol Bioeng. 1997;56:577–582.
\Xie L, Wang DIC. Stoichiometric analysis of animal cell growth and its
application in medium design. Biotechnol Bioeng. 1994;43:1164–1174.
Zhang J, Reddy J, Buckland B, Greasham R. Toward consistent and
productive complex media for industrial fermentations: studies on
yeast extract for a recombinant yeast fermentation process. Bio-
technol Bioeng. 2003;82:640–652.
Zhang J, Robinson D. Development of animal-free, protein-free, and
chemically-defined media for NS0 cell culture. Cytotechnology.
2005;48:59–74.
Zhou W, Chen CC, Buckland B, Aunins J. Fed-batch culture of
Recombinant NS0 myeloma cells with high monoclonal antibody
production. Biotechnol Bioeng. 1997a;55:783–792.
Zhou W, Rehm J, Hu WS. High viable cell concentration fed-batch
cultures of hybridoma cells through online nutrient feeding. Bio-
technol and Bioeng. 1995;46:579–587.
Zhou W, Rehm J, Europa A, Hu WS. Alteration of mammalian cell
metabolism by dynamic nutrient feeding. Cytotechnology. 1997b;
24:99–108.
Manuscript received Sept. 15, 2008, and revision received Feb. 28, 2009.