An Investigation Into the Relationship Between Metabolic Responses and Energy Regulation in Antibody-Producing Cell
An Investigation Into the Relationship Between Metabolic Responses and Energy Regulation in Antibody-Producing Cell
Journal of Microbiology and Biotechnology. 2013. Nov, 23(11): 1586-1597
Copyright © 2013, The Korean Society For Microbiology And Biotechnology
  • Received : July 10, 2013
  • Accepted : August 13, 2013
  • Published : November 28, 2013
Export by style
Cited by
About the Authors
Ya-Ting, Sun
Liang, Zhao
Zhao-Yang, Ye
Li, Fan
Xu-Ping, Liu
Wen-Song, Tan

Energy-efficient metabolic responses were often noted in high-productive cultures. To better understand these metabolic responses, an investigation into the relationship between metabolic responses and energy regulation was conducted via a comparative analysis among cultures with different energy source supplies. Both glycolysis and glutaminolysis were studied through the kinetic analyses of major extracellular metabolites concerning the fast and slow cell growth stages, respectively, as well as the time-course profiles of intracellular metabolites. In three cultures showing distinct antibody productivities, the amino acid metabolism and energy state were further examined. Both the transition of lactate from production to consumption and steady intracellular pools of pyruvate and lactate were observed to be correlated with efficient energy regulation. In addition, an efficient utilization of amino acids as the replenishment for the TCA cycle was also found in the cultures with upregulated energy metabolism. It was further revealed that the inefficient energy regulation would cause low cell productivity based on the comparative analysis of cell growth and productivity in cultures having distinct energy regulation.
Therapeutic proteins, as one of the best-selling biologic drugs, are demanded in considerably large quantities [25] . Their manufacture using mammalian cells has become the largest ever division in the pharmaceutical industry [1] . In order to increase both cell mass and productivity, during the past two decades, great efforts have been paid to optimize protein-producing cell cultures [27] , and many efficient culture processes have therefore been successfully developed [6 , 8 , 9] . Along with the enhanced cell performance such as increased cell density, cell viability, and production rate, distinct metabolic responses in these high-yielding processes have also been noticed. Recently, researches have further pointed out that cell growth and productivity are strongly correlated with certain metabolic responses; for example, the lactate and amino acid metabolisms [7 , 16] . However, the recognition of such a correlation calls for a further detailed investigation into metabolic responses in both high-and low-productive processes to understand why cell performance has a close correlation with metabolic responses.
Some metabolic responses in high-productive culture processes often displayed a tendency towards more efficient energy metabolism, whereas in suboptimal processes the energy metabolism was deregulated [16 , 17] . Recently, based on a proteomic study, the upregulation of proteins involved in energy metabolism was found to be associated with the enhancement in cell mass and protein synthesis, suggesting the critical roles of energy regulation on cell performance [2 , 18] . Moreover, some direct evidences showed that the high intracellular ATP concentration was helpful to enhance the synthesis of recombinant protein [12] . These findings suggest that the energy metabolism might be a key to understand the correlation between cell performance and metabolic responses.
It is well known that protein-producing mammalian cells are usually associated with deregulated central metabolism, which often leads to disturbed energy regulation. These cells preferentially undergo aerobic glycolysis, an inefficient pathway for energy generation [21] . Compared with the tricarboxylic acid cycle (TCA cycle), which is the main mechanism for ATP production, aerobic glycolysis converts glucose mainly into lactate and produces less amount of ATP [28] . Besides this, these cells consume an excessive amount of glutamine, more than their actual needs, with a large portion metabolized into ammonium or non-essential amino acids [11] . The inefficient utilization of glutamine usually impairs its primary function as a backup for the TCA cycle in energy generation, whereas other anaplerotic pathways seem trivial owing to the low-level expression of key enzymes [22] . The deregulated energy metabolism in protein-producing cells may have a negative influence on cell performance if the aerobic glycolysis or inefficient utilization of glutamine predominates the whole cultivation. Instead, if a metabolic shift towards more efficient energy metabolism is realized, cell performance should be improved. To better understand the correlation of cell performance with metabolic responses, it would be reasonable to investigate the relationship between metabolic responses and energy regulation, and further demonstrate the influence of energy regulation on cell growth and productivity. However, few studies have tried to elucidate these issues. It remains elusive which metabolic responses are correlated with efficient energy regulation and whether efficient energy regulation has a positive influence on cell growth and productivity.
To address these issues, the present study was performed by examining cellular responses to different energy conditions. Cells were cultured under conditions designed to have varied energy sources. Three conditions included (i) deficient energy source (by supplying insufficient glucose); (ii) abundant energy source (by supplying surplus glucose); and (iii) limited energy source (by supplying insufficient glucose and other carbon sources). Cell growth and productivity were monitored and the metabolic responses of glycolysis and glutaminolysis were examined regarding both fast and slow growth stages, respectively. Then amino acid metabolism as well as intracellular energy status was further investigated. Through a comparative analysis, the relationship between metabolic responses and energy regulation as well as the influence of energy regulation on cell growth and productivity was elucidated.
Materials and Methods
- Cell Culture
A DHFR-CHO cell line expressing a chimeric anti-human CD20 monoclonal antibody was used in this study [4] . Cells were routinely maintained in a protein-free cell culture medium, which was composed of a mixture of DMEM/HamR F12 (1:1, (v/v); Gibco) supplemented with vitamins, nucleic acids, ethanolamine, sodium selenite, ferric citrate, and Pluronic F-68. This basal medium containing 7.8 mM glucose and 1 mM pyruvate was further supplemented with glutamine at the final concentration of 6 mM. This basal medium was then added with various energy sources for different conditions.
Four batch cultures were performed under different conditions: (ⅰ) deficient energy source with an insufficient low concentration of glucose (supplemented with 2.2 mM glucose) (designated as LG with final 10 mM glucose); (ⅱ) abundant energy source with a sufficient high concentration of glucose (supplemented with 32.2 mM glucose) (HG with final 40 mM glucose); (ⅲ) limited energy source with a low concentration of glucose (supplemented with 2.2 mM glucose) but further supplemented with either an monocarboxylate (pyruvate, 9 mM) (LG+Pyr with final 10 mM glucose and 10 mM pyruvate) or hexose (galactose, 25 mM) (LG+Gal with final 10 mM glucose and 25 mM galactose). In addition, a fed-batch culture was carried out on the basis of LG+Gal culture (designated as FB-Gal), wherein the only difference from the LG+Gal culture was that a feeding medium composed of concentrated amino acids and vitamins was supplemented once at the time point of 72 h. All reagents were purchased from Sigma-Aldrich.
Exponentially growing cells were inoculated at a density of 4-5 × 10 5 cells/ml. All cultures were carried out in 50 ml cell culture tubes (TPP, Switzerland) with a working volume of 20 ml and agitated at 220 rpm on a rotary shaker (Adolf Kühner, Switzerland). Cells were incubated at 37℃ with 5% CO 2 and 85% relative humidity. Daily sampling was performed for cell counting and metabolite measurements. Cell density and viability were determined with a hemacytometer using trypan blue staining. Cell size was examined through visually comparing cells against the scale on the hemacytometer and found to be approximate 14 µm in diameter, thus giving a single cell volume of 1.44 × 10 -9 ml [24] .
- Metabolite Measurements and Quantification of Antibody Concentration
Extracellular metabolites. Concentrations of glucose, glutamine, glutamate, lactate, and ammonium in culture medium were measured on a BioProfile 400 Analyzer (Nova Biomedical). The concentration of pyruvate in the culture medium was determined using commercial assay kits following the manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, China). Concentrations of amino acids were quantified using HPLC (1<525 Binary HPLC Pump; 717 plus Autosampler; 2475 Multi λ Fluorescence Detector; Waters) equipped with a reversed-phase column (Nova-Pak C184 μm, 3.9×150 mm; Waters), following the pre-column derivation using the AccQ·Fluor reagent kit (Waters).
Intracellular metabolites. Intracellular ATP was directly quantified with a bioluminescent method using an ATP assay kit (Beyotime Institute of Biotechnology, China). Intracellular NADH and NAD+ were measured using an NAD + /NADH quantification kit (BioVision,USA). These measurements were all carried out according to the manufacturers’ protocols.
Intracellular pyruvate and lactate were extracted as previously described [13] . Briefly, sampled cell preparations were immediately centrifuged and the supernatant was discarded. To lyse cells, cold perchloric acid solution was added to the cell pellets, followed by neutralization with KOH and K 2 HPO 4 solutions. The whole process was carried out at 0℃. Intracellular concentrations of pyruvate and lactate were determined after extraction using commercial assay kits following the manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, China).
Antibody concentration. The concentration of the chimeric anti-human CD20 monoclonal antibody secreted by cells in the culture medium was determined using an ELISA method.
- Calculation of Specific Consumption/Production Rates of Metabolites or the Antibody
Specific rates of consumption and/or production of metabolites or the anti-human CD20 antibody were calculated by plotting measured concentrations against the respective integral of viable cell concentration (IVCC). The respective average specific rates were equal to the slopes of the linear fitted curves.
- Calculation of Relative Standard Deviation of Each Amino Acid and Intracellular Metabolite
The relative standard deviation (RSD) of each amino acid and intracellular metabolite was evaluated during the establishment of the analytical method, which was assumed as a constant characteristic parameter of statistical distribution and calculated as follow:
PPT Slide
Lager Image
  • X : measurement value for one sample
  • : average of repeated measurements
  • n : repeated n times of measurement, n ≥ 20
Results and Discussion
- Cell Growth and Antibody Production
Cell growth and production profiles for all cultures including the fed-batch culture (FB-Gal) were obtained and are summarized in Fig. 1 . Within the initial 48 h in culture, cell growth appeared similar under all conditions ( Fig.1 A). After 48 h, the growth of cells halted both with deficient energy source (LG) and limited energy source supplemented with pyruvate (LG+Pyr), accompanied by a drastic decrease in cell viability after 72 h ( Fig. 1 B). In contrast, cells kept growing under the conditions of abundant energy source (HG) and limited energy source supplemented with galactose (including LG+Gal and FB-Gal), and cell viability was maintained above 98% during 48-120 h in these cultures ( Figs.1 A and 1 B). When plotting the logarithm of viable cell density against culture time, two distinct growth stages with apparently different growth behaviors could be discerned (as shown in Fig. 1 C and denoted as I&II, respectively). The respective specific growth rates of these two stages for all cultures are also summarized in Table 1 . Based on the specific growth rates, a fast growth stage for all cultures (0-48 h) and a slow growth stage (48-120 h) for HG, LG+Gal, and FB-Gal cultures were established. However, in LG and LG+Pyr cultures, the slow growth stage was not present since cells could not thrive at all after 48 h.
The production profile of the monoclonal antibody in all cultures is shown in Fig. 1 D and the corresponding production rates are listed in Table 1 . Within the initial 72 h, the production rates were lower in LG and LG+Pyr than those in HG, LG+Gal, and FB-Gal, which was probably due to the early cell death in the latter two cultures. However, after 72 h, the antibody production rate declined significantly by 43.6% and 47.8% in HG and LG+Gal cultures, respectively, while a constant rate was sustained in FB-Gal culture in the whole cultivation process ( Table 1 ).
PPT Slide
Lager Image
Time profiles of cell growth and productivity. (A) Viable cell density; (B) logarithmic value of viable cell density (I, II, III represents three culture span); (C) cell viability; (D) antibody concentration in culture LG (■), HG (●), LG+Pyr (□), LG+Gal (○), and FB-Gal (×). The error bar represents standard deviation of three independent experiments (same for Figs. 2 and 3).
Specific growth rates and monoclonal antibody production rates in different cultures.
PPT Slide
Lager Image
aThe concentration of energy source for specific culture refers to the final concentration in the medium.
- Metabolic Responses of Glycolysis and Glutaminolysis
As shown in Fig. 2 , the concentrations of key metabolites involved in both glycolysis and glutaminolysis were plotted against IVCC and the corresponding specific rates are summarized in Table 2 . The metabolic responses under different culture conditions were analyzed with respect to distinct growth stages.
In the fast growth stage, the metabolites in glycolysis, including glucose, pyruvate, and lactate, were measured in all cultures ( Figs. 2 A- 2 C and Table 2 .) It was found that when cells were exposed to deficient (LG) or abundant energy source (HG), the consumption of glucose was similar, showing almost identical specific rates. However, the downstream metabolites in glycolysis, including pyruvate and lactate, demonstrated differential responses under these two conditions ( Figs. 2 B and 2 C). Whereas an overflow of extracellular pyruvate at a rate of 0.13 mmol 10 9 cells -1 ·day -1 took place in HG culture, which was accompanied by a significant increase in its intracellular concentration, LG culture cells consumed pyruvate at a rate of 0.11mmol 10 9 cells -1 ·day -1 and the intracellular pyruvate concentration remained relatively constant at a lower level ( Figs. 3 A and 3 B and Table 2 ). In the meantime, both extra-and intracellular lactate were at higher levels in HG culture than those in LG culture, which was consistent with the fact that the production rate of extracellular lactate in HG culture was 50.2% greater ( Figs. 3 C and 3 D and Table 2 ). In addition, when cells were subjected to limited energy supply with different carbon sources ( i.e ., LG+Gal, FB-Gal, and LG+Pyr), the metabolic responses of glycolysis also appeared different. When galactose was used (LG+Gal and FB-Gal), the glycolysis was similar to that in LG culture. However, when cells were given pyruvate (LG+Pyr), the glucose consumption rate was 44.5% lower than that in LG+Gal (FB-Gal) culture and the consumption rate of extracellular pyruvate was increased significantly (by 8.8- folds compared with LG culture), whereas no obvious uptake was noted in LG+Gal (FB-Gal) culture ( Table 2 ). On the other hand, the metabolites in glutaminolysis, including glutamine, glutamate, and ammonium, were also monitored ( Figs. 2 E- 2 G and Table 2 ). Compared with the glycolysis, glutaminolysis did not show much difference among different culture conditions ( Table 2 ). However, both the uptake and accumulation rates of glutamine and ammonium in LG+Pyr culture were found to be the lowest among all the cultures, which implicated the slowest glutaminolysis.
PPT Slide
Lager Image
Concentrations of extracellular metabolites versus integral of viable cell concentration (IVCC). (A) Glucose; (B) pyruvate; (C) lactate; (D) glutamine; (E) glutamate; and (F) ammonium in culture LG (■), HG (●), LG+Pyr (□), LG+Gal (○), and FB-Gal (×). The solid and dotted lines represent the relevant slopes before and after 48 h, respectively.
As cells entered the slow growth stage, significant metabolic shifts were observed under the conditions of abundant (HG) and limited energy source (LG+Gal and FB-Gal). In HG culture, the uptake of glucose became slower compared with that in the fast growth stage ( Figs. 2 A and 2 D). At this stage, even though the glucose concentration was still above 10 mM, the glucose consumption rate declined by 80% in HG culture ( Fig. 2 A and Table 2 ). In LG+Gal and FB-Gal cultures, since glucose was exhausted within the first 48 h ( Fig. 2 A), cells should have turned to galactose for glycolysis. However, the consumption profile of galactose was not examined in the present study. However, according to a previous report, cells generally consume a lower quantity of galactose than glucose [3] . In consistent with the decelerated consumption of carbon sources, the consumption rate of glutamine at the slow growth stage also decreased by 55.8% in HG culture and 18.4% in both LG+Gal and FB-Gal ( Table 2 ) compared with those in the fast growth stage, respectively. Despite the consistent reduced uptake of nutrients in all these three cultures, the metabolic responses of pyruvate, lactate, and glutamate varied at the slow growth stage. Under the condition of abundant energy source (HG), cells started to take up pyruvate in HG culture after 72 h ( Fig. 3 A) and at the same time extracellular lactate ceased to accumulate ( Fig. 3 C). Meanwhile, the intracellular concentrations of pyruvate and lactate dropped dramatically after 96 h in HG culture ( Figs. 3 B and 3 D). Under the condition of limited energy source (LG+Gal and FB-Gal), pyruvate was also consumed but at a much lower rate, which was nearly one third of that in HG culture ( Figs. 2 B, 3 A, and Table 2 ). However, different from that in HG culture, a steady uptake of lactate was observed after 48 h in both LG+Gal and FB-Gal cultures ( Fig. 3 C). In addition, in comparison with that in LG+Gal culture, the consumption rate of lactate was increased by 64.5% in FB-Gal culture ( Table 2 ), suggesting that the feeding of amino acids and vitamins promoted the uptake of lactate. This faster uptake caused the exhaustion of lactate in the end of FB-Gal culture ( Fig. 2 C). Moreover, the intracellular concentrations of pyruvate and lactate remained relatively steady in the slow growth stage in both LG+Gal and FB-Gal cultures when compared with those in HG culture ( Figs. 3 B and 3 D). On the other hand, consistent with the decreased consumption of glutamine, the accumulation of ammonium halted in all three cultures at the slow growth stage ( Figs. 2 A and 2 F). In contrast, whereas the glutamate remained unchanged in FB-Gal culture, in both HG and LG+Gal cultures, glutamate was consumed from the previous accumulation in the fast growth stage ( Fig. 2 E).
Specific rates of metabolites consumption/production in 0-48 h/48-120 h (negative value represents consumption).
PPT Slide
Lager Image
aThe specific consumption rate of pyruvate in HG culture was fitted after 72 h. b“--” stands for rates with no distinct consumption or production trend. cThe specific rates of glutamate in LG+Gal and FB-Gal cultures were fitted within 0-72 h/72-120 h.
PPT Slide
Lager Image
Time profiles of extracellular and intracellular metabolites concentrations. (A) and (B) Extra-/intracellular pyruvate concentrations; (C) and (D) extra-/intracellular lactate concentrations in culture LG (■), HG (●), LG+Pyr (□), LG+Gal (○), and FB-Gal (×). The arrow indicates that the value of the open square is according to the Y-axis on the right side.
- Amino Acid Metabolism and Intracellular Energy States
The time-course profiles of amino acids in HG, LG+Gal, and FB-Gal cultures are shown in Fig. 4 . Among the major amino acids, in HG and LG+Gal cultures, similar concentrations of both arginine and threonine were found with no obvious consumption or production, and these two amino acids also maintained steady concentrations in FBGal culture after the feeding point (data not shown). In contrast, during the whole cultivation process, other amino acids, except proline, glycine, and alanine, were consumed gradually to a similar extent in the two batch cultures (HG and LG+Gal). Meanwhile, cysteine, leucine, isoleucine, valine, and histidine were found at lower levels in LG+Gal culture, suggesting more utilization of these amino acids than in HG culture. Compared with the LG+Gal culture, the one-shot feeding at 72 h in FB-Gal culture increased the concentrations of most amino acids. However, the consumptions of tyrosine, cysteine, leucine, isoleucine, valine, methionine, and aspartate were found to be increased even after 96 h in FB-Gal culture. In addition to the increased utilization of these amino acids, it was also interesting to note the metabolic shift of alanine and serine, which were produced previously in FB-Gal culture. It was found that alanine was consumed after 96 h in FB-Gal, while its concentration was kept steady in both HG and LG+Gal culture with a higher level in the former. Moreover, while serine was produced in LG culture at 96 h, considering the similar concentrations at 96 h in FB-Gal with feeding and in LG+Gal without feeding, its consumption in FB-Gal culture could be inferred before 96 h and even more pronounced after 96 h.
PPT Slide
Lager Image
Relative amino acids concentrations (to 0 h) in the early, middle, and end of cultivation. LG (square blocks); LG+Gal (shaded blocks); FB-Gal (black blocks). The arrow indicates that the value of the black blocks is according to the Y-axis on the right side.
PPT Slide
Lager Image
Time-course profiles of intracellular energetic metabolites. (A) ATP; (B) NADH; and (C) NAD+/NADH in culture LG (square blocks), LG+Gal (shaded blocks), and FB-Gal (black blocks).
In addition, the intracellular energy states were also analyzed in HG, LG+Gal, and FB-Gal cultures. Fig. 5 shows the time-course profiles of indices related to energy state, including ATP, NADH, and NAD + /NADH. It was found that intracellular ATP in HG culture decreased dramatically at 72 h and then continued to decline at a much lower rate. In contrast, in both LG-Gal and FB-Gal cultures, although ATP levels also declined at 72 h, it was followed by a gradual increase, but to a more prominent extent in FB-Gal culture than that in LG-Gal culture ( Fig. 5 A). In addition, the highest level of NADH was observed at 148 h in HG ( Fig. 5 B), which was consistent with its lowest ratio of NAD + /NADH at the same time point ( Fig. 5 C). In contrast, a gradual decrease in the level of NADH was noticed in FB-Gal culture ( Fig. 5 B), which was accompanied by the gradual increase in NAD + /NADH ( Fig. 5 C).
- Relationship Between Metabolic Responses and Energy Regulation
In the present study, three cultures (HG, LG+Gal, and FB-Gal) were found to show both fast and slow growth stages. It was revealed that the energy states were different among these cultures, especially between HG and FB-Gal cultures. Intriguingly, in HG and FB-Gal cultures, different metabolic responses of glycolysis, glutaminolysis, and amino acid metabolism were observed regarding both extra-and intracellular metabolite profiles. Based on these results, a comparative analysis of the energy states and corresponding metabolic responses in HG and FB-Gal cultures would be able to unravel the relationship between metabolic responses and energy regulation.
HG culture displayed accumulated NADH level and a low ratio of NAD + /NADH at at 148 h and a drastic decrease in intracellular ATP at 72 h ( Fig. 5 ). The accumulation of NADH and decreased ratio of NAD + /NADH indicated the cells' inability to maintain the normal energy production ( i.e., ATP) [5] . Hence, in HG culture, the downregulation of energy was coincident with the transition of cells from the fast to slow growth stage, which could be further correlated with the observed aerobic glycolysis and inefficient utilization of glutaminolysis during the whole cultivation. In the fast growth stage, aerobic glycolysis was reflected in the overflow of extracellular pyruvate and fast production of lactate. The accumulation of pyruvate is generally recognized as the consequence of a rapid glycolysis. In addition, glutaminolysis can also produce pyruvate via the malate shunt to generate NADPH in most mammalian cells [11] . Hence, the pyruvate accumulation and the resultant lactate production in HG culture indicated that the majority of consumed glucose and glutamine inefficiently contributed to ATP generation via the TCA cycle. As cells in HG culture went into the slow growth stage, a significant consumption of pyruvate occurred and the accumulation of lactate ceased, which suggested that the consumed pyruvate entered the TCA cycle directly for energy formation and aerobic glycolysis became less pronounced. However, the inefficient utilization of glutaminolysis was still present in HG culture. The dysfunction of glutamine and glutamate to replenish the TCA cycle was further evidenced by the metabolism of two amino acids, proline and alanine. Proline can be synthesized directly from glutamate and alanine is a sink for the excess nitrogen intake from glutamine. Based on the observation of the steady accumulation of proline and high concentration of alanine in HG culture ( Fig. 4 ), it could be inferred that most of the consumed glutamine and glutamate contributed to the formation of non-essential amino acids instead of going through catabolism as anaplerosis for the TCA cycle.
On the other hand, FB-Gal culture had a decreased NADH level as cells entered the slow growth stage and a gradually increasing ratio of NAD + /NADH in the whole cultivation ( Figs. 5 B and 5 C). Although the intracellular ATP level after 72 h was lower than that at 48 h, it increased gradually afterwards and the level remained much higher than that in HG culture after 72 h ( Fig. 5 A). These data implicated that cells in FB-Gal culture upregulated the energy state during the whole cultivation, which could be further correlated with alleviated aerobic glycolysis in the fast growth stage and the replenishment of lactate and other amino acids into the TCA cycle in the slow growth stage. In the fast growth stage, none overflow of pyruvate accompanied by the much lower lactate accumulation was noticed in FB-Gal culture ( Figs. 3 A and ( 3 C). As cells went into the slow growth stage, a notable uptake of lactate was observed in FB-Gal culture ( Fig. 2 C). In mammalian cells, the conversion between pyruvate and lactate is feasible [8] . Since the uptake of pyruvate was found instead of accumulation ( Fig. 3 A), it could be possible that the consumed lactate was exploited to substitute pyruvate and enter the TCA cycle for energy production. In addition, the greater consumption of some amino acids and the metabolic shift of produced serine and alanine in the slow growth stage were also found in FB-Gal culture ( Fig. 4 ). Besides its contribution in the synthesis of antibody (representing the largest molar fraction in anti-human CD20 monoclonal antibody), serine can be converted to pyruvate for the subsequent entry into the TCA cycle. The relative steady concentration of glutamate in FB-Gal culture ( Fig. 2 E) also suggested that the consumed alanine might be metabolized via the transamination pathway, in which pyruvate is generated from alanine for entry into the TCA cycle. As the basic building blocks for proteins, amino acids play key roles in the antibody production process. However, amino acids can also contribute to catabolism for energy generation via the TCA cycle. Thus, the more efficient utilization of amino acids accompanied by anaplerosis of some produced amino acids in FB-Gal culture enhanced the energy generation through the efficient replenishment to the TCA cycle.
Among all the monitored metabolic responses, some were found directly correlated with energy regulation. Specifically, in the slow growth stage, a switch from production to consumption of lactate was noted in both LG+Gal and FB-Gal cultures. Feeding of amino acids and vitamins in FB-Gal culture caused even more pronounced lactate consumption. In contrast, in HG culture, although the rapid accumulation of lactate in the fast growth stage stopped quickly as cells entered the slow growth stage, none consumption of lactate was detected. In the meantime, the ATP profile in LG, FB-Gal, and HG cultures varied along with the shift in lactate metabolism, suggesting that energy regulation could be correlated with lactate metabolism. Although the shift of lactate metabolism had been described in similar cultures, in which glucose was either delivered at a very low concentration or was substituted with other hexoses [3 , 7] , its relationship with energy regulation has not yet been established. Consistently with the lactate metabolism, the intracellular concentrations of pyruvate and lactate in LG+Gal and FB-Gal cultures were less fluctuated than those in HG culture during the whole cultivation, which suggested that more stable intracellular pools of pyruvate and lactate might also be associated with more efficient energy regulation in protein-producing cell cultures.
- Influence of Energy Regulation on Cell Growth and Productivity
Although the energy regulation was found to be different between HG and FB-Gal cultures, the only difference in these two culture conditions was their condition of energy source supply. In FB-Gal culture, glucose was supplemented at a low level 10 mM and galactose was supplemented as a second carbon source. Since the glucose transporters in mammalian cells usually have a much lower affinity for other hexoses [20] , the supply of energy source with other hexoses is not as sufficient as that with glucose. As shown in this study, energy-efficient metabolic responses leading to more efficient energy regulation were induced under this condition. In contrast, when surplus glucose was supplied, aerobic glycolysis and inefficient utilization of glutaminolysis could be observed during the whole cultivation and thereby led to downregulated energy metabolism. However, the reason why the manipulation of energy source supply could induce distinct metabolic responses and resultant energy regulation was not explored in the present study. Nevertheless, these results were consistent with other reported studies. HepG2 cells were found to increase the respiration rate to maintain ATP level in galactose media [29] . It was also reported that the defective mitochondrial function of cancer cells could be considerably enhanced by altering the substrate [23] . In the present study, we were more concerned whether the energy regulation had a direct influence on cell growth and productivity. This can be of great significance for process development and optimization to control the metabolic responses towards more efficient energy regulation via effective strategies.
Despite the distinct energy states, the profiles of cell growth in HG and FB-Gal cultures were similar in both the fast and slow growth stages, suggesting that energyregulation had less influence on cell growth. However, if glucose was restricted to a low concentration or glycolysis and glutaminolysis were inhibited, cell proliferation would be affected owing to the deficient nutrients. In the present study, prolonged growth profiles were noted in HG, LG, and FB-Gal cultures, whereas transient growth profiles were in both LG and LG+Pyr cultures. In LG culture, cells consumed even more pyruvate and glutamate than those in FB-Gal culture, both of which could directly enter the TCA cycle for efficient energy generation, suggesting that the energy metabolism in LG culture might be even more efficient in the fast growth stage than that in FB-Gal culture. However, when running out of glucose after 48 h, the only carbon source in LG culture, cells could not survive any longer with deficient energy source. In LG+Pyr culture, the cell specific growth rate was even 21.7% lower than that in LG culture in the fast growth stage (Table 2) . Consistently, the consumption rates of both glucose and glutamine were found lower than those in LG culture. Since glycolysis and glutaminolysis are the main pathways in mammalian cells for supplying intermediates such as ribose, purine, and NADPH for macromolecular synthesis [19] , the growth inhibition in LG+Pyr culture might result from the inadequate nutrient supply with dysfunction of glycolysis and glutaminolysis.
On the other hand, the cell productivity profiles in HG and FB-Gal cultures were different in the slow growth stage, which suggested that energy regulation might have a direct influence on cell productivity. When cells were under the condition of abundant energy source (HG), although the deregulation of energy metabolism seemed less pronounced in the slow growth stage, the inefficient replenishment to the TCA cycle for energy generation still caused the downregulated energy state. Instead, when cells were under the limited condition of energy source with the feeding of amino acids and vitamins (FB-Gal), the upregulated energy state was achieved through the metabolic shift of lactate and steady intracellular pools of pyruvate and lactate in the slow growth stage. In the meantime, the antibody production rate decreased with time in HG culture, whereas it remained constant in FB-Gal culture. The correlation between protein production rate and energy metabolism has been implicated in several previous studies. Kuwae et al. [14] found that the limited availability of energy coincided with the decreased production of recombinant human antithrombin by CHO cells. A positive correlation between the oxidative phosphorylation rate and antibody productivity was also found in the production of the recombinant MUC1 fusion protein [15] . In the present study, it was demonstrated that inefficient energy regulation in the slow growth stage would cause the decrease of cell roductivity. In addition to the formation of cellular proteins for normal needs, protein-producing cells bear an extra burden to synthesize a large quantity of recombinant protein products. Since protein production is an extremely energy-intensive process in which the energetic demand of cells is rather high [26] , the energy requirement for production could not be omitted although cell growth became decelerated. In the present study, the decreased uptake of glucose was noted in the slow growth stage in HG culture, which implied that the energy obtained via glucose phosphorylation would be decreased in this stage. Furthermore, the inefficient replenishment to the TCA cycle for energy generation further caused the downregulated energy states in HG culture. Therefore, the decreased productivity in HG culture might result from the failure to satisfy the energy requirements for antibody production.
However, different from that in HG culture, the intracellular ATP in LG+Gal culture increased gradually after 72 h. Nevertheless, the antibody production rate still declined after 72 as that in HG culture, which suggested an alternative mechanism responsible for the decreased cell productivity in LG+Gal culture. As a matter of fact, it was found that most amino acids, including cysteine, leucine, isoleucine, valine, histidine, and aspartate, were at lower concentrations than those in HG culture. Although the intracellular ATP was at a higher level in LG+Gal culture than in HG culture, the inadequate precursors for protein synthesis might have caused the decrease in productivity in LG+Gal culture, suggesting that insufficient nutrient supply could also result in inefficient protein production. On the other hand, in FB-Gal culture, the feeding of amino acids and vitamins at 72 h might attenuate the deficiency of certain amino acids as encountered in LG+Gal culture. Moreover, it was found that more lactate and amino acids were utilized as replenishment for the TCA cycle, which stimulated more efficient energy regulation than LG+Gal culture. The probable reason for the influence of additional feeding on cell metabolism was not explored extensively. Nevertheless, it has been reported that some fed-batch cultures with more balanced nutrients or metal ions can bring about more efficient energy regulation when cells seemed to have more utilization of amino acids as well as a phenotype of lactate consumption [16 , 17] . Considering that both energy regulation and nutrient supply have direct influence on cell productivity, it is expected that a fed-batch culture with increased productivity would be established via further optimizing the feeding medium and feeding strategy to better balance the energy and nutrient equirements.
In conclusion, comparative analyses of the metabolic responses in cultures with distinct energy states revealed that the shift from production to consumption of lactate and steady intracellular pools of pyruvate and lactate were closely correlated with efficient energy regulation. In addition, the efficient utilization of amino acids as the replenishment into the TCA cycle was also associated with upregulated energy metabolism. By comparing both the cell growth and productivity in different cultures, it was revealed that the energy regulation had less influence on cell growth. When cells stepped into the slow growth stage, the inefficient energy regulation would cause the decrease in cell productivity, indicating that it is important to meet energetic needs in the cultivation of protein-producing cells.
This work was supported by the National Natural Science Foundation of China (No. 21106045, No. 21206040), the National High Technology Research and Development Program of China (863 Program) (No. 2012AA02A303), the National Science and Technology Major Project (No. 2013ZX10004003-003-003), and Shanghai Training Program for College Young Teacher (No. hlg11011).
Aggarwal S 2011 What's fueling the biotech engine-2010 to 2011. Nat. Biotechnol. 29 1083 - 1089    DOI : 10.1038/nbt.2060
Alete DE , Racher AJ , Birch JR , Stansfield SH , James DC , Smales CM 2005 Proteomic analysis of enriched microsomal fractions from GS-NS0 murine myeloma cells with varying secreted recombinant monoclonal antibody productivities. Proteomics 5 4689 - 4704    DOI : 10.1002/pmic.200500019
Altamirano C , Illanes A , Becerra S , Cairo JJ , Godia F 2006 Considerations on the lactate consumption by CHO cells in the presence of galactose. J. Biotechnol. 125 547 - 556    DOI : 10.1016/j.jbiotec.2006.03.023
Chen F , Fan L , Wang J , Zhou Y , Ye Z , Zhao L 2012 Insight into the roles of hypoxanthine and thymidine on cultivating antibody-producing CHO cells: cell growth, antibody production and long-term stability. Appl. Microbiol. Biotechnol. 93 169 - 178    DOI : 10.1007/s00253-011-3484-z
Dawson AG 1979 Oxidation of cytosolic NADH formed during aerobic metabolism in mammalian cells. Trends Biochem. Sci. 4 171 - 176    DOI : 10.1016/0968-0004(79)90417-1
Fan L , Zhao L , Sun YT , Kou TC , Zhou Y , Tan WS 2009 A high-yielding, generic fed-batch process for recombinant antibody production of GS-engineered cell lines. J. Microbiol. Biotechnol. 19 1695 - 1702    DOI : 10.4014/jmb.0904.04054
Gagnon M , Hiller G , Luan YT , Kittredge A , DeFelice J , Drapeau D 2011 High-end pH-controlled delivery of glucose effectively suppresses lactate accumulation in CHO fed-batch cultures. Biotechnol. Bioeng. 108 1328 - 1337    DOI : 10.1002/bit.23072
Gladden LB 2004 Lactate metabolism: a new paradigm for the third millennium. J. Physiol. 558 5 - 30    DOI : 10.1113/jphysiol.2003.058701
Hermes PA , Castro CD 2010 A fully defined, fed-batch, recombinant NS0 culture process for monoclonal antibody production. Biotechnol. Prog. 26 1411 - 1416    DOI : 10.1002/btpr.455
Huang YM , Hu W , Rustandi E , Chang K , Yusuf-Makagiansar H , Ryll T 2010 Maximizing productivity of CHO cell-based fed-batch culture using chemically defined media conditions and typical manufacturing equipment. Biotechnol. Prog. 26 1400 - 1410    DOI : 10.1002/btpr.436
Hu WS , Ozturk SS 2005 Cell Culture Technology for Pharmaceutical and Cell-based Therapies Taylor & Francis, Boca Raton. 82 - 100
Kim H-J , Kwon Y , Lee S , Kim P 2012 An engineered Escherichia coli having a high intracellular level of ATP and enhanced recombinant protein production. Appl. Microbiol. Biotechnol. 94 1079 - 1086    DOI : 10.1007/s00253-011-3779-0
Kochanowski N , Blanchard F , Cacan R , Chirat F , Guedon E , Marc A 2006 Intracellular nucleotide and nucleotide sugar contents of cultured CHO cells determined by a fast, sensitive, and high-resolution ion-pair RP-HPLC. Anal. Biochem. 348 243 - 251    DOI : 10.1016/j.ab.2005.10.027
Kuwae S , Ohda T , Tamashima H , Miki H , Kobayashi K 2005 Development of a fed-batch culture process for enhanced production of recombinant human antithrombin by Chinese hamster ovary cells. J. Biosci. Bioeng. 100 502 - 510    DOI : 10.1263/jbb.100.502
Link T , Backstrom M , Graham R , Essers R , Zorner K , Gatgens J 2004 Bioprocess development for the production of a recombinant MUC1 fusion protein expressed by CHO-K1 cells in protein-free medium. J. Biotechnol. 110 51 - 62    DOI : 10.1016/j.jbiotec.2003.12.008
Luo J , Vijayasankaran N , Autsen J , Santuray R , Hudson T , Amanullah A 2012 Comparative metabolite analysis to understand lactate metabolism shift in Chinese hamster ovary cell culture process. Biotechnol. Bioeng. 109 146 - 156    DOI : 10.1002/bit.23291
Ma N , Ellet J , Okediadi C , Hermes P , McCormick E , Casnocha S 2009 A single nutrient feed supports both chemically defined NS0 and CHO fed-batch processes: Improved productivity and lactate metabolism. Biotechnol. Prog. 25 1353 - 1363    DOI : 10.1002/btpr.238
Meleady P , Doolan P , Henry M , Barron N , Keenan J , O’Sullivan F 2011 Sustained productivity in recombinant Chinese hamster ovary (CHO) cell lines: proteome analysis of the molecular basis for a process-related phenotype. BMC Biotechnol. 11 78 - 88    DOI : 10.1186/1472-6750-11-78
Morgan MJ , Faik P 1981 Carbohydrate metabolism in cultured animal cells. Biosci. Rep. 1 669 - 686    DOI : 10.1007/BF01116465
Mueckler M 1994 Facilitative glucose transporters. Eur. J. Biochem. 219 713 - 725    DOI : 10.1111/j.1432-1033.1994.tb18550.x
Mulukutla BC , Khan S , Lange A , Hu WS 2010 Glucose metabolism in mammalian cell culture: new insights for tweaking vintage pathways. Trends Biotechnol. 28 476 - 484    DOI : 10.1016/j.tibtech.2010.06.005
Neermann J , Wagner R 1996 Comparative analysis of glucose and glutamine metabolism in transformed mammalian cell lines, insect and primary liver cells. J. Cell. Physiol. 166 152 - 169
Rossignol R , Gilkerson R , Aggeler R , Yamagata K , Remington SJ , Capaldi RA 2004 Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. Cancer Res. 64 985 - 993    DOI : 10.1158/0008-5472.CAN-03-1101
Schmid G , Blanch HW 1992 Extra- and intracellular metabolite concentrations for murine hybridoma cells. Appl. Microbiol. Biotechnol. 36 621 - 625    DOI : 10.1007/BF00183239
Scolnik PA 2009 mAbs: a business perspective. MAbs 1 179 - 184    DOI : 10.4161/mabs.1.2.7736
Seth G , Hossler P , Yee J , Hu WS 2006 Engineering cells for cell culture bioprocessing - physiological fundamentals, In Hu WS (ed.). Cell Culture Engineering. Springer Berlin, Heidelberg 119 - 164
Shukla AA , Thömmes J 2010 Recent advances in largescale production of monoclonal antibodies and related proteins. Trends Biotechnol. 28 253 - 261    DOI : 10.1016/j.tibtech.2010.02.001
Warburg O 1956 On the origin of cancer cells. Science 123 309 - 314    DOI : 10.1126/science.123.3191.309
Warburg O , Geissler AW , Lorenz S 1967 On growth of cancer cells in media in which glucose is replaced by galactose. Hoppe Seylers Z. Physiol. Chem. 348 1686 - 1687    DOI : 10.1515/bchm2.1967.348.1.1686