Characterization of glutamine metabolism in two related murine hybridomas

An integrated modular framework for modeling the effect of ammonium on the sialylation process of monoclonal antibodies produced by CHO cells

BACKGROUND

Monoclonal antibodies (mABs) have emerged as one of the most important therapeutic recombinant proteins in the pharmaceutical industry. Their immunogenicity and therapeutic efficacy are influenced by post-translational modifications, specifically the glycosylation process. Bioprocess conditions can influence the intracellular process of glycosylation. Among all the process conditions that have been recognized to affect the mAB glycoforms, the detailed mechanism underlying how ammonium could perturb glycosylation remains to be fully understood. It was shown that ammonium induces heterogeneity in protein glycosylation by altering the sialic acid content of glycoproteins. Hence, understanding this mechanism would aid pharmaceutical manufacturers to ensure consistent protein glycosylation.

METHODS

Three different mechanisms have been proposed to explain how ammonium influences the sialylation process. In the first, the inhibition of CMP-sialic acid transporter, which transports CMP-sialic acid (sialylation substrate) into the Golgi, by an increase in UDP-GlcNAc content that is brought about by the augmented incorporation of ammonium into glucosamine formation. In the second, ammonia diffuses into the Golgi and raises its pH, thereby decreasing the sialyltransferase enzyme activity. In the third, the reduction of sialyltransferase enzyme expression level in the presence of ammonium. We employed these mechanisms in a novel integrated modular platform to link dynamic alteration in mAB sialylation process with extracellular ammonium concentration to elucidate how ammonium alters the sialic acid content of glycoproteins.

RESULTS

Our results show that the sialylation reaction rate is insensitive to the first mechanism. At low ammonium concentration, the second mechanism is the controlling mechanism in mAB sialylation and by increasing the ammonium level (< 8 mM) the third mechanism becomes the controlling mechanism. At higher ammonium concentrations (> 8 mM) the second mechanism becomes predominant again.

CONCLUSION

The presented model in this study provides a connection between extracellular ammonium and the monoclonal antibody sialylation process. This computational tool could help scientists to develop and formulate cell culture media. The model illustrated here can assist the researchers to select culture media that ensure consistent mAB sialylation.

Transient ammonia stress on Chinese hamster ovary (CHO) cells yield alterations to alanine metabolism and IgG glycosylation profiles

BACKGROUND

Ammonia concentrations typically increase during mammalian cell cultures, mainly due to glutamine and other amino acid consumption. An early ammonia stress indicator is a metabolic shift with respect to alanine. To determine the underlying mechanisms of this metabolic shift, a Chinese hamster ovary (CHO) cell line with two distinct ages (standard and young) was cultured in parallel fed-batch bioreactors with 0 mM or 10 mM ammonia added at 12 h. Reduced viable cell densities were observed for the stressed cells, while viability was not significantly affected. The stressed cultures had higher alanine, lactate, and glutamate accumulation. Interestingly, the ammonia concentrations were similar by Day 8.5 for all cultures. We hypothesized the ammonia was converted to alanine as a coping mechanism. Interestingly, no significant differences were observed for metabolite profiles due to cell age. Glycosylation analysis showed the ammonia stress reduced galactosylation, sialylation, and fucosylation. Transcriptome analysis of the standard-aged cultures indicated the ammonia stress had a limited impact on the transcriptome, where few of the significant changes were directly related metabolite or glycosylation reactions. These results indicate that mechanisms used to alleviate ammonia stress are most likely controlled post-transcriptionally, and this is where future research should focus.

Endoplasmic reticulum protein expression in recombinant NS0 myelomas grown in batch culture

During batch growth of two recombinant NS0 myelomas, an increase in the expression of the endoplasmic reticulum (ER) proteins (GRP78/BiP, GRP94, and ERp72) was observed. A marked increase in these proteins was associated with the decline phase of growth, an increase in the production rate of chimeric antibody, and a marked slowing or halt in the uptake of glucose and glutamate. Refeeding with glucose, glutamate, or a mixture of amino acids just prior to the onset of decline phase failed to repress induction. Although refeeding with glutamate led to an increase in specific productivity, there was no significant difference in the pattern of ER protein induction. These results indicate that an increase in ER protein expression is not solely related to productivity but also to certain changes that occur during the course of batch growth.

Characterisation of G418-induced metabolic load in recombinant CHO and BHK cells: effect on the activity and expression of central metabolic enzymes

In a previous article (Yallop and Svendsen 2001), recombinant CHO and BHK cell lines, expressing the human glucagon receptor and the gastric inhibitory peptide receptor, respectively, showed reduced growth rates and altered nutrient utilisation when grown with increasing concentrations of G418. This response was associated with an increased expression of the neo (r) protein, while expression of the recombinant membrane receptors remained unaltered. The metabolic response was characterised in both cell lines by an increase in the specific rate of glutamine utilisation and in CHO cells by a decrease in the yield of lactate from glucose, suggesting a change in the flux of glucose through central metabolism. The aim of this study was to further elucidate these metabolic changes by determining the activity and relative expression of key enzymes involved in glucose and glutamine metabolism. For both CHO and BHK cells, there was an increase in the activity of glutaminase, glutamate dehydrogenase and glutamine synthetase, suggesting an increased flux through the glutaminolysis pathway. The activity of glucose-6-phosphate dehydrogenase and pyruvate carboxylase in CHO cells was also increased whilst lactate dehydrogenase activity remained unaltered, suggesting an increased flux to the pentose phosphate pathway and TCA cycle, respectively. The activity of these enzymes in BHK cells was unchanged. Quantitative RT-PCR showed that expression levels of glutaminase and pyruvate carboxylase were the same with and without G418, indicating that the differences in activities were likely due to post-translational modifications.

Examination of primary metabolic pathways in a murine hybridoma with carbon-13 nuclear magnetic resonance spectroscopy

Primary metabolism of a murine hybridoma was probed with (13)C nuclear magnetic resonance (NMR) spectroscopy. Cells cultured in a hollow fiber bioreactor were serially infused with [1-(13)C] glucose, [2-(13)C] glucose, and [3-(13)C] glutamine. In vivo spectroscopy of the culture was used in conjunction with off-line spectroscopy of the medium to determine the intracellular concentration of several metabolic intermediates and to determine fluxes for primary metabolic pathways. Intracellular concentrations of pyruvate and alanine were very high relative to levels observed in normal quiescent mammalian cells. Estimates made from labeling patterns in lactate indicate that 76% of pyruvate is derived directly from glycolysis; some is also derived from the malate shunt, the pyruvate/melate shuttle associated with lipid synthesis and the pentose phosphate pathway. The rate of formation of pyruvate from the pentose phosphate pathway was estimated to be 4% of that from glycolysis; This value is a lower limit and the actual value may be higher. Incorporation of pyruvate into the tricarboxylic acid (TCA) cycle appears to occur through only pyruvate dehydrogenase; no pyruvate carboxylase activity was detected. The malate shunt rate was approximately equal to the rate of glutamine uptake. The rate of incorporation of glucosederived acetyl-CoA into lipids was 4% of the glucose uptake rate. The TCA cycle rate between isocitrate and alpha-ketoglutarate was 110% of the glutamine uptake rate.

Human 293 cell metabolism in low glutamine-supplied culture: interpretation of metabolic changes through metabolic flux analysis

Metabolic flux analysis is a useful tool to analyze cell metabolism. In this study, we report the use of a metabolic model with 34 fluxes to study the 293 cell, in order to improve its growth capacity in a DMEM/F12 medium. A batch, fed-batch with glutamine feeding, fed-batch with essential amino acids, and finally a fed-batch experiment with both essential and nonessential amino acids were compared. The fed-batch with glutamine led to a maximum cell density of 2.4x10(6) cells/ml compared to 1.8x10(6) cells/ml achieved in a batch mode. In this fed-batch with glutamine, it was also found that 2.5 mM ammonia was produced compared to the batch which had a final ammonia concentration of 1 mM. Ammonia was found to be growth inhibiting for this cell line at a concentration starting at 1 mM. During the fed-batch with glutamine, the flux analysis shows that a majority of amino acid fluxes and Kreb's cycle fluxes, except for glutamine flux, are decreased. This observation led to the conclusion that the main nutrient used is glutamine and that during the batch there is an overflow in the Kreb's cycle. Thus, a fed-batch with glutamine permits a better utilization of this nutrient. A fed-batch with essential amino acid without glutamine was also assayed in order to reduce ammonia production. The maximum cell density was increased further to 3x10(6) cells/ml and ammonia production was reduced below 1 mM. Flux analysis shows that the cells could adapt to a medium with low glutamine by increasing the amino acid fluxes toward the Kreb's cycle. Adding nonessential amino acids during this feeding strategy did not improve growth further and the nonessential amino acids accumulated in the medium.

Elevated glutamate dehydrogenase flux in glucose-deprived hybridoma and myeloma cells: evidence from 1H/15N NMR

The glutamine metabolism was studied in glucose-starved and glucose-sufficient hybridoma and Sp2/0-Ag14 myeloma cells. Glucose starvation was attained by cultivating the hybridoma cells with fructose instead of glucose, and the myeloma cells with a low initial glucose concentration which was rapidly exhausted. Glutamine used in the experiments was labeled with 15N, either in the amine or in the amide position. The fate of the label was monitored by 1H/15N NMR analysis of released 15NH+4 and 15N-alanine. Thus, NH+4 formed via glutaminase (GLNase) could be distinguished from NH+4 formed via glutamate dehydrogenase (GDH). In the glucose-sufficient cells a small but measurable amount of 15NH+4 released by GDH could be detected in both cell lines (0.75 and 0.31 micromole/10(6) cells for hybridoma and myeloma cells, respectively). The uptake of glutamine and the total production of NH+4 was significantly increased in both fructose-grown hybridoma and glucose-starved myeloma cells, as compared to the glucose-sufficient cells. The increased NH+4 production was due to an increased throughput via GLNase (1.6 -1.9-fold in the hybridoma, and 2.7-fold in the myeloma cell line) and an even further increased metabolism via GDH (4.8-7.9-fold in the hybridoma cells, and 3.1-fold in the myeloma cells). The data indicate that both GLNase and GDH are down-regulated when glucose is in excess, but up-regulated in glucose-starved cells. It was calculated that the maximum potential ATP production from glutamine could increase by 35-40 % in the fructose-grown hybridoma cells, mainly due to the increased metabolism via GDH.

Fluxes and enzyme activities in central metabolism of myeloma cells grown in chemostat culture

Activities of enzymes in glycolysis, the pentose phosphate pathway, the tricarboxylic acid cycle, and glutaminolysis have been determined in the mouse myeloma SP2/0.Ag14. Cells were grown on IMDM medium with 5% serum in steady-state chemostat culture at a fixed dilution rate of 0.03 h-1. Three culture conditions, which differed in supply of glucose and oxygen, were chosen so as to change catabolic fluxes in the central metabolism, while keeping anabolic fluxes constant. In the three steady-state situations, the ratio between specific rates of glucose and glutamine consumption differed by more than twentyfold. The specific rates of glucose consumption and lactate production were highest at low oxygen supply, whereas the specific rate of glutamine consumption was highest in the culture fed with low amounts of glucose. Under low oxygen conditions, the specific production of ammonia increased and the consumption pattern of amino acids showed large changes compared with the other two cultures. For the three steady states, activities of key enzymes in glycolysis, the pentose phosphate pathway, glutaminolysis, and the TCA cycle were measured. The differences in the in vivo fluxes were only partially reflected in changes in enzyme levels. The largest differences were observed in the levels of glycolytic enzymes, which were elevated under conditions of low oxygen supply. High activities of phosphoenolpyruvate carboxykinase (E.C. 4.1.1.32) in all cultures suggest an important role for this enzyme as a link between glutaminolysis and glycolysis. For all enzymes, in vitro activities were found that could accommodate the estimated maximum in vivo fluxes. These results show that the regulation of fluxes in central metabolism of mammalian cells occurs mainly through modulation of enzyme activity and, to a much lesser extent, by enzyme synthesis.

Subcellular localization of enzyme activities in chemostat-grown murine myeloma cells

As part of the development of structured models for the metabolism of myeloma cells in suspension culture, a study was made of the subcellular localization of key enzymes of glucose and glutamine metabolism. Steady state chemostat cultures of the mouse myeloma SP2/0-Ag14 were used as a reproducible source of biomass. Homogenates of the cells, obtained via mechanical disruption, were separated into a mitochondrial and a cytosolic fraction via differential centrifugation. The following conclusions are drawn: (1) approximately one fifth of the hexokinase activity of cell-free homogenates is associated with the mitochondria; (2) a malate-aspartate shuttle may operate for oxidation of cytosolic NADH, as indicated by high levels of malate dehydrogenase and aspartate aminotransferase in both particulate and soluble fractions; (3) the pentose phosphate pathway and isocitrate dehydrogenase may contribute to the provision of cytosolic NADPH; (4) phosphoenolpyruvate carboxykinase and pyruvate kinase, which are present in high activities, are exclusively cytosolic and probably play a key role in glutamine metabolism; (5) oxidation of glutamine via these enzymes leads to the formation of pyruvate that enters the same pool as pyruvate generated by glycolysis. As a result, lactate and alanine formation can occur from both glucose and glutamine.

Dichloroacetate inhibits glutamine oxidation by decreasing pyruvate availability for transamination

We have shown that dichloroacetate (DCA) inhibits growth, glutamine oxidation, and pyruvate and alanine production in a concentration-dependent manner in PQXB 1/2 hybridoma cells. The use of inhibitors indicates that glutamine oxidation proceeds by an aminooxyacetate-sensitive transamination reaction in this cell line. Addition of pyruvate to DCA-treated cells restored glutamine oxidation to control values. Our data suggest that DCA inhibits glutamine oxidation by decreasing the availability of pyruvate for transamination, which in turn results in glutamate accumulation and a consequent inhibition of glutaminase activity. Impaired glutamine catabolism in the presence of DCA has subsequent effects on overall metabolic balance and cell maintenance and growth.

Automation of selective assays for on-line bioprocess monitoring by flow-injection analysis

On-line analysis of one component in a complex media used for bioprocesses requires the application of selective tests such as enzymes assays. Because these assays are susceptible to interference by other medium components and have a limited detection range, automatic sample pretreatment is a prerequisite. The progress made with automatic sample pretreatment in flow-injection analysis makes this technique particularly suitable for on-line monitoring of bioprocesses. Moreover, newly developed software control systems may improve the necessary robustness of flow-infection analysis systems.

High-density hybridoma perfusion culture. Limitation vs inhibition

Because our earlier work indicated a strong correlation between specific antibody productivity and cell density in perfusion culture, we conducted experiments to determine the optimum means of increasing cell density while maintaining high antibody productivity. The rates of medium supply and waste removal were varied to determine whether cell density was limited or inhibited, and whether a diffusable substance could be responsible for the correlation between antibody productivity and cell density. Nutrient supply was found to be a stronger determinant of cell density than waste removal; however, the rate of waste removal had a greater effect on cell growth at lower cell densities. Even at noninhibitory levels of ammonia and lactate, cellular metabolism was regulated to minimize their concentrations at lowered rates of waste removal. Separate step changes in glucose and glutamine resulted in increased cell density and antibody concentration. Specific antibody productivity increased following the step in glutamine, but not glucose. Both steps caused changes in cellular metabolism that prevented the levels of lactate and ammonia from reaching toxic levels.

Comparative analysis of glucose and glutamine metabolism in transformed mammalian cell lines, insect and primary liver cells

Glucose and glutamine metabolism in several cultured mammalian cell lines (BHK, CHO, and hybridoma cell lines) were investigated by correlating specific utilization and formation rates with specific maximum activities of regulatory enzymes involved in glycolysis and glutaminolysis. Results were compared with data from two insect cell lines and primary liver cells. Flux distribution was measured in a representative mammalian (BHK) and an insect (Spodoptera frugiperda) cell line using radioactive substrates. A high degree of similarity in many aspects of glucose and glutamine metabolism was observed among the cultured mammalian cell lines examined. Specific glucose utilization rates were always close to specific hexokinase activities, indicating that formation of glucose-6-phosphate from glucose (catalyzed by hexokinase) is the rate limiting step of glycolysis. No activity of the key enzymes connecting glycolysis with the tricarboxylic acid cycle, such as pyruvate dehydrogenase, pyruvate carboxylase, and phosphoenolpyruvate carboxykinase, could be detected. Flux distribution in BHK cells showed glycolytic rates very similar to lactate formation rates. No glucose- or pyruvate-derived carbon entered the tricarboxylic acid cycle, indicating that glucose is mainly metabolized via glycolysis and lactate formation. About 8% of utilized glucose was metabolized via the pentose phosphate shunt, while 20 to 30% of utilized glucose followed pathways other than glycolysis, the tricarboxylic acid cycle, or the pentose phosphate shunt. About 18% of utilized glutamine was oxidized, consistent with the notion that glutamine is the major energy source for mammalian cell lines. Mammalian cells cultured in serum-free low-protein medium showed higher utilization rates, flux rates, and enzyme activities than the same cells cultured in serum-supplemented medium. Insect cells oxidized glucose and pyruvate in addition to glutamine. Furthermore, insect cells produced little or no lactate and were able to channel glycolytic intermediates into the tricarboxylic acid cycle. Metabolic profiles of the type presented here for a variety of cell lines may eventually enable one to interfere with the metabolic patterns of cells relevant to biotechnology, with the hope of improving growth rate and/or productivity.

Genetic engineering of hybridoma glutamine metabolism

The murine hybridoma PQXB1/2 cannot be adapted to grow in culture media containing < 0.5 mM glutamine. Transformants selected following electroporation of PQXB1/2 cells with vectors containing a Chinese hamster glutamine synthetase (GS) cDNA under the control of the SV40 early promoter also failed to grow in the absence of glutamine in the culture medium. PQXB1/2 cells have, however, been transformed to glutamine independence following electroporation with a vector containing this glutamine synthetase cDNA under the control of the human cytomegalovirus immediate early promoter. In these cells, sufficient active glutamine synthetase was expressed from one vector per cell to enable growth in glutamine-free media. The specific activity of glutamine synthetase in two transformed cell lines producing parental levels of antibody was increased by 128 and 152%, respectively (0.57 and 0.63 mumol min-1 per 10(6) cells in transformants compared with parental levels of 0.25 mumol min-1 per 10(6) cells). This reprogramming of glutamine synthetase expression and glutamine metabolism is important for developing strategies to deal with ammonia toxicity and the production of cell lines with improved metabolic processes.

Adaptation of mammalian cells to non-ammoniagenic media

Although glutamine is used as a major substrate for the growth of mammalian cells in culture, it suffers from some disadvantages. Glutamine is deaminated through storage or by cellular metabolism, leading to the formation of ammonia which can result in growth inhibition. Non-ammoniagenic alternatives to glutamine have been investigated in an attempt to develop strategies for obtaining improved cell yields for ammonia sensitive cell lines. Glutamate is a suitable substitute for glutamine in some culture systems. A period of adaptation to glutamate is required during which the activity of glutamine synthetase and the rate of transport of glutamate both increase. The cell yield increases when the ammonia accumulation is decreased following culture supplementation with glutamate rather than glutamine. However some cell lines fail to adapt to growth in glutamate and this may be due to a low efficiency transport system. The glutamine-based dipeptides, ala-gln and gly-gln can substitute for glutamine in cultures of antibody-secreting hybridomas. The accumulation of ammonia in these cultures is less and cell yields in dipeptide-based media may be improved compared to glutamine-based controls. In murine hybridomas, a higher concentration of gly-gln is required to obtain comparable cell growth to ala-gln or gln-based cultures. This is attributed to a requirement for dipeptide hydrolysis catalyzed by an enzyme with higher affinity for ala-gln than gly-gln.

Glucose and glutamine metabolism of a murine B-lymphocyte hybridoma grown in batch culture

The energy metabolism of a mammalian cell line grown in vitro was analyzed by substrate consumption rates and metabolic flux measurements. The data allowed the determination of the relative importance of the pathways of glucose and glutamine metabolism to the energy requirements of the cell. Changes in the substrate concentrations during culture contributed to the changing catalytic activities of key enzymes, which were determined. 1. A murine B-lymphocyte hybridoma (PQXB1/2) was grown in batch culture to a maximum cell density of 1-2 x 10(6) cells/mL in 3-4 d. The intracellular protein content showed a maximum value during the exponential growth phase of 0.55 mg/10(6) cells. Glutamine was completely depleted, but glucose only partially depleted to 50% of its original concentration when the cells reached a stationary phase following exponential growth. 2. The specific rates of glutamine and glucose utilization varied during culture and showed maximal values at the midexponential phase of 2.4 nmol/min/10(6) cells and 4.3 nmol/min/10(6) cells, respectively. 3. A high proportion of glucose (96%) was metabolized by glycolysis, but only limited amounts by the pentose phosphate pathway (3.3%) and TCA cycle (0.21%). 4. The maximum catalytic activity of hexokinase approximates to the measured flux of glycolysis and is suggested as a rate-limiting step. In the stationary phase, the hexokinase activity reduced to 11% of its original value and may explain the reduced glucose utilization at this stage. 5. The maximal activities of two TCA cycle enzymes were well above the measured metabolic flux and are unlikely to pose regulatory barriers. However, the activity of pyruvate dehydrogenase was undetectable by spectrophotometric assay and explains the low level of flux of glycolytic metabolites into the TCA cycle. 6. A significant proportion of the glutamine (36%) utilized by the cells was completely oxidized to CO2. 7. The measured rate of glutamine transport into the cells approximated to the metabolic flux and is suggested as a rate-limiting step. 8. Glutamine metabolism is likely to occur via glutaminase and amino transaminase, which have significantly higher activities than glutamate dehydrogenase. 9. The calculated potential ATP production suggests that, overall, glutamine is the major contributor of cellular energy. However, at the midexponential phase, the energy contribution from the catabolism of the two substrates was finely balanced--glutamine (55%) and glucose (45%).

Uptake of glutamate, not glutamine synthetase, regulates adaptation of mammalian cells to glutamine-free medium

Two cell lines (McCoy and MDCK) were studied in an attempt to understand the metabolic changes associated with adaptation to glutamine-free medium (GMEM + gmate). McCoy cells assumed normal growth rates after 2–3 passages in this medium whereas MDCK cells showed no growth in GMEM + gmate. The glutamine synthetase (GS) activity of both cell lines was elevated (up to × 9) as glutamine was depleted from normal media (GMEM + gmine). The high activity of GS was maintained during McCoy cell growth in GMEM + gmate. However, there was no apparent significant difference between the two cell lines in the pattern of changes of GS activity in response to glutamine. The cellular uptake rates of glutamine and glutamate from the medium differed significantly between the two cell lines. During the adaptation of McCoy cells to GMEM + gmate, the rate of glutamate uptake doubled to a value of 0.54 nmol/min per mg cell protein whereas the maximum value for MDCK cells was considerably lower (0.04 nmol/min per mg cell protein). We propose that the difference in intrinsic ability for glutamate transport accounts for the difference in growth response between the two cell lines in the glutamine-free medium.