Ammonium ion and glucosamine dependent increases of oligosaccharide complexity in recombinant glycoproteins secreted from cultivated BHK-21 cells

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.

Affecting HEK293 Cell Growth and Production Performance by Modifying the Expression of Specific Genes

The HEK293 cell line has earned its place as a producer of biotherapeutics. In addition to its ease of growth in serum-free suspension culture and its amenability to transfection, this cell line’s most important attribute is its human origin, which makes it suitable to produce biologics intended for human use. At the present time, the growth and production properties of the HEK293 cell line are inferior to those of non-human cell lines, such as the Chinese hamster ovary (CHO) and the murine myeloma NSO cell lines. However, the modification of genes involved in cellular processes, such as cell proliferation, apoptosis, metabolism, glycosylation, secretion, and protein folding, in addition to bioprocess, media, and vector optimization, have greatly improved the performance of this cell line. This review provides a comprehensive summary of important achievements in HEK293 cell line engineering and on the global engineering approaches and functional genomic tools that have been employed to identify relevant genes for targeted engineering.

Supplementing glycosylation: A review of applying nucleotide-sugar precursors to growth medium to affect therapeutic recombinant protein glycoform distributions

Glycosylation is a critical quality attribute (CQA) of many therapeutic proteins, particularly monoclonal antibodies (mAbs), and is a major consideration in the approval of biosimilar biologics due to its effects to therapeutic efficacy. Glycosylation generates a distribution of glycoforms, resulting in glycoproteins with inherent molecule-to-molecule heterogeneity, capable of activating (or failing to activate) different effector functions of the immune system. Glycoforms can be affected by the supplementation of nucleotide-sugar precursors, and related components, to culture growth medium, affecting the metabolism of glycosylation. These supplementations has been demonstrated to increase nucleotide-sugar intracellular pools, and impact glycoform distributions, but with varied results. These variations can be attributed to five key factors: Differences between cell platforms (enzyme/transporter expression levels); differences between recombinant proteins produced (glycan-site accessibility); the fermentation and sampling timeline (glucose availability and exoglycosidase accumulation); glutamine levels (affecting ammonia levels, which impact Golgi pH, as well as UDP-GlcNAc pools); and finally, a lack of standardized metrics for observing shifts in glycoform distributions (glycosylation indices) across different experiments. The purpose of this review is to provide detail and clarity on the state of the art of supplementation strategies for nucleotide-sugar precursors for affecting glycosylation in cell culture processes, and to apply glycosylation indices for standardized comparisons across the field.

Combining metabolic and process engineering strategies to improve recombinant glycoprotein production and quality

Increasing recombinant protein production while ensuring a high and consistent protein quality remains a challenge in mammalian cell culture process development. In this work, we combined a nutrient substitution approach with a metabolic engineering strategy that improves glucose utilization efficiency. This combination allowed us to tackle both lactate and ammonia accumulation and investigate on potential synergistic effects on protein production and quality. To this end, HEK293 cells overexpressing the pyruvate yeast carboxylase (PYC2) and their parental cells, both stably producing the therapeutic glycoprotein interferon α2b (IFNα2b), were cultured in media deprived of glutamine but containing chosen substitutes. Among the tested substitutes, pyruvate led to the best improvement in growth (integral of viable cell density) for both cell lines in batch cultures, whereas the culture of PYC2 cells without neither glutamine nor any substitute displayed surprisingly enhanced IFNα2b production. The drastic reduction in both lactate and ammonia in the cultures translated into extended high viability conditions and an increase in recombinant protein titer by up to 47% for the parental cells and the PYC2 cells. Product characterization performed by surface plasmon resonance biosensing using Sambucus nigra (SNA) lectin revealed that the increase in yield was however accompanied by a reduction in the degree of sialylation of the product. Supplementing cultures with glycosylation precursors and a cofactor were effective at counterbalancing the lack of glutamine and allowed improvement in IFNα2b quality as evaluated by lectin affinity. Our study provides a strategy to reconcile protein productivity and quality and highlights the advantages of PYC2-overexpressing cells in glutamine-free conditions.

Altering the central carbon metabolism of HEK293 cells: Impact on recombinant glycoprotein quality

The accumulation of metabolic by-products remains a critical challenge in the development of mammalian cells culture processes as it impacts cellular growth, productivity and product quality. Although the overexpression of the PYC2 gene was shown to significantly improve the nutrient metabolism efficiency of mammalian cells, its impact on recombinant protein quality has not been investigated yet. In this study, we assess the effect of this metabolic engineering strategy on the quality of a recombinant therapeutic glycoprotein, the human interferon α2b (IFNα2b). As inferred from densitometry analysis of SDS-PAGE gels, PYC2-overexpressing cells sustained a higher percentage of intact glycosylated IFNα2b at the late stage of batch cultures, which was correlated with prolonged viability and reduced accumulation of waste metabolites. Contrarily to the IFNα2b produced by the PYC2 cells, LC-MS analysis confirmed the presence of less glycosylated IFNα2b as well as the occurrence of proteolytic cleavage in the IFNα2b produced in the parental cells. Taken together, these results indicate that PYC2-overexpression in mammalian cells leads to extended favorable conditions for glycosylation and offer an attractive approach to mass-produce high-quality recombinant proteins.

Tuning a MAb glycan profile in cell culture: Supplementing N-acetylglucosamine to favour G0 glycans without compromising productivity and cell growth

Glycosylation is a critical quality attribute of many therapeutic proteins, particularly monoclonal antibodies (MAbs). Nucleotide-sugar precursors supplemented to growth medium to affect the substrate supply chain of glycosylation has yielded promising but varied results for affecting glycosylation. Glucosamine (GlcN), a precursor for N-acetylglucosamine (GlcNAc), is a major component of mammalian glycans. The supplementation of GlcN to CHO cells stably-expressing a chimeric heavy-chain monoclonal antibody, EG2-hFc, reduces the complexity of glycans to favour G0 glycoforms, while also negatively impacting cell growth. Although several researchers have examined the supplementation of glucosamine, no clear explanation of its impact on cell growth has been forthcoming. In this work, the glucosamine metabolism is examined. We identified the acetylation of GlcN to produce GlcNAc to be the most likely cause for the negative impact on growth due to the depletion of intracellular acetyl-CoA pools in the cytosol. By supplementing GlcNAc in lieu of GlcN to CHO cells producing EG2-hFc, we achieve the same shift in glycan complexity with marginal impacts on the cell growth and protein production.

Elucidating the effects of postinduction glutamine feeding on the growth and productivity of CHO cells

Inducible mammalian expression systems are increasingly being used for the production of valuable therapeutics. In such system, maximizing the product yield is achieved by carefully balancing the biomass concentration during the production phase and the specific productivity of the cells. These two factors are largely determined by the availability of nutrients and/or the presence of toxic waste metabolites in the culture environment. Glutamine is one of the most important components of cell culture medium, since this substrate is an important building block and source of energy for biomass and recombinant protein production. Its metabolism, however, ultimately leads to the formation of ammonia, a well known inhibitor of cellular growth and productivity. In this work, we show that nutrient feeding post-induction can greatly enhance the product yield by alleviating early limitations encountered in batch. Moreover, varying the amount of glutamine in the feed yielded two distinct culture behaviors post-induction; whereas excess glutamine allowed to reach greater cell concentrations, glutamine-limited fed-batch led to increased cell specific productivity. These two conditions also showed distinctive lactate metabolism. To further assess the physiological impact of glutamine levels on the cells, a comparative (13) C-metabolic flux analysis was conducted and a number of key intracellular fluxes were found to be affected by the amount of glutamine present in the feed during the production phase. Such information may provide useful clues for the identification of physiological markers of cell growth and productivity that could further guide the optimization of inducible expression systems.

Profiling convoluted single-dimension proton NMR spectra: a Plackett-Burman approach for assessing quantification error of metabolites in complex mixtures with application to cell culture

Single-dimension hydrogen, or proton, nuclear magnetic resonance spectroscopy (1D-(1)H NMR) has become an attractive option for characterizing the full range of components in complex mixtures of small molecular weight compounds due to its relative simplicity, speed, spectral reproducibility, and noninvasive sample preparation protocols compared to alternative methods. One challenge associated with this method is the overlap of NMR resonances leading to "convoluted" spectra. While this can be mitigated through "targeted profiling", there is still the possibility of increased quantification error. This work presents the application of a Plackett-Burman experimental design for the robust estimation of precision and accuracy of 1D-(1)H NMR compound quantification in synthetic mixtures, with application to mammalian cell culture supernatant. A single, 20 sample experiment was able to provide a sufficient estimate of bias and variability at different metabolite concentrations. Two major sources of bias were identified: incorrect interpretation of singlet resonances and the quantification of resonances from protons in close proximity to labile protons. Furthermore, decreases in measurement accuracy and precision could be observed with decreasing concentration for a small fraction of the components as a result of their particular convolution patterns. Finally, the importance of a priori concentration estimates is demonstrated through the example of interpreting acetate metabolite trends from a bioreactor cultivation of Chinese hamster ovary cells expressing a recombinant antibody.

Glycosylation and post-translational modification gene expression analysis by DNA microarrays for cultured mammalian cells

DNA microarray analysis of gene expression has become a valuable tool for bioprocessing research aimed at improving therapeutic protein yields. The highly parallel nature of DNA microarray technology allows researchers to assess hundreds of gene simultaneously, essentially enabling genome-wide snapshots. The quality and amount of therapeutic proteins produced by cultured mammalian cells rely heavily on the culture environment. In order to implement beneficial changes to the culture environment, a better understanding of the relationship between the product quality and culture environment must be developed. By analyzing gene expression levels under various environmental conditions, light can be shed on the underlying mechanisms. This paper describes a method for evaluating gene expression changes for cultured NS0 cells, a mouse-derived myeloma cell line, under culture environment conditions, such as ammonia buildup, known to affect product quality. These procedures can be easily adapted to other environmental conditions and any mammalian cell lines cultured in suspension, so long as a sufficient number of gene sequences are publicly available.

Synergizing metabolic flux analysis and nucleotide sugar metabolism to understand the control of glycosylation of recombinant protein in CHO cells

BACKGROUND

The glycosylation of recombinant proteins can be altered by a range of parameters including cellular metabolism, metabolic flux and the efficiency of the glycosylation process. We present an experimental set-up that allows determination of these key processes associated with the control of N-linked glycosylation of recombinant proteins.

RESULTS

Chinese hamster ovary cells (CHO) were cultivated in shake flasks at 0 mM glutamine and displayed a reduced growth rate, glucose metabolism and a slower decrease in pH, when compared to other glutamine-supplemented cultures. The N-linked glycosylation of recombinant human chorionic gonadotrophin (HCG) was also altered under these conditions; the sialylation, fucosylation and antennarity decreased, while the proportion of neutral structures increased. A continuous culture set-up was subsequently used to understand the control of HCG glycosylation in the presence of varied glutamine concentrations; when glycolytic flux was reduced in the absence of glutamine, the glycosylation changes that were observed in shake flask culture were similarly detected. The intracellular content of UDP-GlcNAc was also reduced, which correlated with a decrease in sialylation and antennarity of the N-linked glycans attached to HCG.

CONCLUSIONS

The use of metabolic flux analysis illustrated a case of steady state multiplicity, where use of the same operating conditions at each steady state resulted in altered flux through glycolysis and the TCA cycle. This study clearly demonstrated that the control of glycoprotein microheterogeneity may be examined by use of a continuous culture system, metabolic flux analysis and assay of intracellular nucleotides. This system advances our knowledge of the relationship between metabolic flux and the glycosylation of biotherapeutics in CHO cells and will be of benefit to the bioprocessing industry.

Glycosidic intermediates identified in 1H MR spectra of intact tumour cells may contribute to the clarification of aspects of glycosylation pathways

The glycosylation process, through the addition of carbohydrates, is a major post-translational modification of proteins and glycolipids. Proteins may be glycosylated in either the secretory pathway leading to N-linked or O-linked glycoproteins or as nucleocytoplasmic glycosylation that targets only single proteins involving a single β-linked N-acetylglucosamine. In both cases, the key precursors are the uridine diphospho-N-acetylhexosamines synthesised by the hexosamine biosynthetic pathway. Furthermore, uridine diphospho-N-acetylglucosamine participates in the biosynthesis of sialic acid. In this work, we propose MRS for the detection of uridine diphospho-N-acetylhexosamines visible in high-resolution MR spectra of intact cells from different human tumours. Signals from the nucleotide and amino sugar moieties, including amide signals observed for the first time in whole cells, are assigned, also taking advantage of spectral changes that follow cell treatment with ammonium chloride. Finally, parallel changes in uridine diphospho-N-acetylhexosamines and glutamine pools, observed after pH changes induced by ammonium chloride in the different tumour cell lines, may provide more details on the glycosylation processes.

An investigation of intracellular glycosylation activities in CHO cells: effects of nucleotide sugar precursor feeding

Controlling glycosylation of recombinant proteins produced by CHO cells is highly desired as it can be directed towards maintaining or increasing product quality. To further our understanding of the different factors influencing glycosylation, a glycosylation sub-array of 79 genes and a capillary electrophoresis method which simultaneously analyzes 12 nucleotides and 7 nucleotide sugars; were used to generate intracellular N-glycosylation profiles. Specifically, the effects of nucleotide sugar precursor feeding on intracellular glycosylation activities were analyzed in CHO cells producing recombinant human interferon-gamma (IFN-gamma). Galactose (+/-uridine), glucosamine (+/-uridine), and N-acetylmannosamine (ManNAc) (+/-cytidine) feeding resulted in 12%, 28%, and 32% increase in IFN-gamma sialylation as compared to the untreated control cultures. This could be directly attributed to increases in nucleotide sugar substrates, UDP-Hex ( approximately 20-fold), UDP-HexNAc (6- to 15-fold) and CMP-sialic acid (30- to 120-fold), respectively. Up-regulation of B4gal and St3gal could also have enhanced glycan addition onto the proteins, leading to more complete glycosylation (sialylation). Combined feeding of glucosamine + uridine and ManNAc + cytidine increased UDP-HexNAc and CMP-sialic acid by another two- to fourfold as compared to feeding sugar precursors alone. However, it did not lead to a synergistic increase in IFN-gamma sialylation. Other factors such as glycosyltransferase or glycan substrate levels could have become limiting. In addition, uridine feeding increased the levels of uridine- and cytidine-activated nucleotide sugars simultaneously, which could imply that uridine is one of the limiting substrates for nucleotide sugar synthesis in the study. Hence, the characterization of intracellular glycosylation activities has increased our understanding of how nucleotide sugar precursor feeding influence glycosylation of recombinant proteins produced in CHO cells. It has also led to the optimization of more effective strategies for manipulating glycan quality.

Influence of intracellular nucleotide and nucleotide sugar contents on recombinant interferon-gamma glycosylation during batch and fed-batch cultures of CHO cells

Both the macroheterogeneity of recombinant human IFN-gamma produced by CHO cells and intracellular levels of nucleotides and sugar nucleotides, have been characterized during batch and fed-batch cultures carried out in different media. Whereas PF-BDM medium was capable to maintain a high percentage of the doubly- glycosylated glycoforms all over the process, mono-glycosylated and non-glycosylated forms increased during the batch culture using SF-RPMI medium. Intracellular level of UTP was higher in PF-BDM all over the batch culture compared to the SF-RPMI process. UDP-Gal accumulated only during the culture performed in PF-BDM medium, probably as a consequence of the reduced UDP-Glc synthesis flux in SF-RPMI medium. When the recombinant CHO cells were cultivated in fed-batch mode, the UTP level remained at a relatively high value in serum-containing RPMI and its titer increased during the fed-phase indicating an excess of biosynthesis. Besides, an accumulation of UDP-Gal occurred as well. Those results all together indicate that UTP and UDP-Glc syntheses in CHO cells cultivated in SF-RPMI medium in batch process, could be limiting during the glycosylation processes of the recombinant IFN-gamma. At last, the determination of the energetic status of the cells over the three studied processes suggested that a relationship between the adenylate energy charge and the glycosylation macroheterogeneity of the recombinant IFN-gamma may exist.

Liver-specific increase of UTP and UDP-sugar concentrations in rats induced by dietary vitamin B6-deficiency and its relation to complex N-glycan structures of liver membrane-proteins

This is the first known report on the influence of vitamin B6-deficiency on the concentration of UDP-sugars and other uracil nucleotides in rats. Animals aged 3 weeks or 2 months were fed a vitamin B6-free diet for periods varying from 3 days to 7 weeks. Nucleotides were quantified by enzymatic-photometry and by SAX-high precision liquid chromatography. In 3 week-old rats, vitamin B6-deficiency resulted in an up to 6.3-fold increase in the concentrations of UTP, UDP, UMP and UDP-sugars and less of CTP in rat liver, while no changes were observed in older rats. In young rats, the concentration of uracil nucleotides started to increase after 1 week diet, with a maximum after 2 weeks. After 5 weeks, the concentrations returned to normal values. In heart, lungs, kidney and brain, concentrations were measured after 2 weeks diet in young rats. In contrast to liver, the heart muscle uracil nucleotide concentrations were decreased by 40%. In kidney, the sum of UTP, UDP and UMP showed a decrease of 40%, whereas UDP-sugars were increased 1.4-fold. In the lungs, nucleotide concentrations were mostly unaffected by vitamin B6-deficiency, despite a 70% increase of UDP-GA. In brain, UDP-Glc, UDP-Gal and the sum of CTP and CDP showed an increase of 30-50%. We became surprised that the increased UDP-sugar concentrations did not influence the structure of liver plasma membrane-N-glycans. Despite the 4 to 6-fold increase of UTP and UDP-sugars, no changes in the complexity or sialylation of these N-glycans could be detected. This study demonstrates that, especially in liver, pyridoxal phosphate is closely involved in the control of uracil nucleotides during a defined period of development. In contrast to in vitro experiments, in vivo N-glycan biosynthesis in liver is regulated by a more complex or higher mechanism than substrate concentrations.

Related effects of cell adaptation to serum-free conditions on murine EPO production and glycosylation by CHO cells

The necessity to perform serum-free cultures to produce recombinant glycoproteins generally requires an adaptation procedure of the cell line to new environmental conditions, which may therefore induce quantitative and qualitative effects on the product, particularly on its glycosylation. In previous studies, desialylation of EPO produced by CHO cells was shown to be dependent on the presence of serum in the medium. In this paper, to discriminate between the effects of the adaptation procedure to serum-free medium and the effects of the absence of serum on EPO production and glycosylation, adapted and non-adapted CHO cells were grown in serum-free and serum-containing media. The main kinetics of CHO cells were determined over batch processes as well as the glycosylation patterns of produced EPO by HPCE-LIF. A reversible decrease in EPO production was observed when cells were adapted to SFX-CHO(TM) medium, as the same cells partially recovered their production capacity when cultivated in serum-containing medium or in the enriched SFM(TM) serum-free medium. More interestingly, EPO desialylation that was not observed in both serum-free media was restored if the serum-independent cells were recultured in presence of serum. In the same way, while the serum-independent cells did not release a sialidase activity in both serum-free media, a significant activity was recovered when serum was added. In fact, the cell adaptation process to serum-free conditions did not specifically affect the sialidase release and the cellular mechanism of protein desialylation, which appeared to be mainly related to the presence of serum for both adapted and non-adapted cells.

Intracellular nucleotide pools and ratios as tools for monitoring dedifferentiation of primary porcine hepatocytes in culture

The effect of two culture configurations (single collagen gel and double collagen gel) and of two hormones (insulin and glucagon) on the differentiated status and the intracellular nucleotide pools of primary porcine hepatocytes was investigated. The objective was to analyze and monitor the current state of differentiation supported by the two culture modes using intracellular nucleotide analysis. Specific intracellular nucleotide ratios, namely the nucleoside triphosphate (NTP) and the uridine (U) ratio were shown to consistently reflect the state of dedifferentiation status of the primary cells in culture affected by the presence of the two hormones insulin and glucagon. Continuous dedifferentiation of the cells was monitored in parallel by the reduction of the secretion of albumin, and changes in UDP-activated hexoses and UDP-glucuronate. The presence of insulin maintained the differentiated status of hepatocytes for more than 12 days when cultivated under double gel conditions whereas glucagon was less effective. In contrast, cells cultivated in a single gel matrix immediately started to dedifferentiate upon seeding. NTP and U ratios were shown to be more sensitive for monitoring dedifferentiation in culture than the albumin secretion. Their use allowed the generation of an easily applicable NTP-U plot in order to give a direct graphical representation of the current differentiation status of the cultured cells. Moreover, the transition from functional and differentiated hepatocytes to dedifferentiated fibroblasts could be determined earlier by the nucleotide ratios compared to the conventional method of monitoring the albumin secretion rate.

Considerations on the lactate consumption by CHO cells in the presence of galactose

A CHO cell line producing t-PA was cultured using glutamate and glucose or galactose to decrease the formation of metabolic end-products and therefore improving the process. In batch cultures using glutamate (6 mM) with glucose at two different levels (5 and 20 mM) or with glucose and galactose (5 and 20 mM, respectively) a remarkable difference in cell culture parameters was evidenced. For 20 mM glucose, a usual cell pattern was observed with lactate built-up in the medium. For 5 mM glucose, cell growth was arrested due to glucose depletion and only a limited use of the excreted lactate could be observed, not supporting cell growth sufficiently. However, when glucose 5 mM and galactose 20 mM were used together, cells consumed the glucose first and, interestingly, in a second phase they continued growing on galactose with the simultaneous consumption of the endogenous lactate. Under these conditions, cell growth was even improved with respect to growth on 20 mM glucose, used as a control. This metabolic behavior is further investigated by using metabolic flux analysis, suggesting that the lactate produced is not used in the oxidative metabolism through the TCA cycle. Metabolic fate of the lactate consumed is discussed.

Optimisation of the cellular metabolism of glycosylation for recombinant proteins produced by Mammalian cell systems

Many biopharmaceuticals are now produced as secreted glycoproteins from mammalian cell culture. The glycosylation profile of these proteins is essential to ensure structural stability and biological and clinical activity. However, the ability to control the glycosylation is limited by our understanding of the parameters that affect the heterogeneity of added glycan structures. It is clear that the glycosylation process is affected by a number of factors including the 3-dimensional structure of the protein, the enzyme repertoire of the host cell, the transit time in the Golgi and the availability of intracellular sugar-nucleotide donors. From a process development perspective there are many culture parameters that can be controlled to enable a consistent glycosylation profile to emerge from each batch culture. A further, but more difficult goal is to control the culture conditions to enable the enrichment of specific glycoforms identified with desirable biological activities. The purpose of this paper is to discuss the cellular metabolism associated with protein glycosylation and review the attempts to manipulate, control or engineer this metabolism to allow the expression of human glycosylation profiles in producer lines such as genetically engineered Chinese hamster ovary (CHO) cells.

Fractionation of follicle stimulating hormone charge isoforms in their native form by preparative electrophoresis technology

Complex glycoprotein biopharmaceuticals, such as follicle stimulating hormone (FSH), erythropoietin and tissue plasminogen activator consist of a range of charge isoforms due to the extent of sialic acid capping of the glycoprotein glycans. Sialic acid occupies the terminal position on the oligosaccharide chain, masking the penultimate sugar residue, galactose from recognition and uptake by the hepatocyte asialoglycoprotein receptor. It is therefore well established that the more acidic charge isoforms of glycoprotein biopharmaceuticals have higher in vivo potencies than those of less acidic isoforms due to their longer serum half-life. Current strategies for manipulating glycoprotein charge isoform profile involve cell engineering or altering bioprocesss parameters to optimise expression of more acidic or basic isoforms, rather than downstream separation of isoforms. A method for the purification of a discrete range of bioactive recombinant human FSH (rhFSH) charge isoforms based on Gradiflowtrade mark preparative electrophoresis technology is described. Gradiflowtrade mark electrophoresis is scaleable, and incorporation into glycoprotein biopharmaceutical production bioprocesses as a potential final step facilitates the production of biopharmaceutical preparations of improved in vivo potency.

Application of a reversible immortalization system for the generation of proliferation-controlled cell lines

To employ physiological mechanisms to control cell growth primary cells were reversibly immortalized using the SV40 TAg. The cells showed a fibroblast-like morphology. When the expression of the TAg was turned off, the cells arrested in the G0/G1 cell cycle phase. The cell culture could be kept for over 1 week in the proliferation-controlled state while the growth arrest remained fully reversible. The regulation was highly efficacious in that the arrested cell population did not spontaneously resume growth, suggesting that in the absence of the immortalizing gene expression endogenous growth-control mechanisms can keep these cells in a viable state for a prolonged time. Recombinant protein expression increased in growth-controlled cells when compared to conventionally cultured cells. Analysis of a secreted pharmaceutical protein revealed high product integrity without any signs of degradation. Therefore, it is feasible to apply genetic regulation of cell immortalization to obtain proliferation-controlled cell lines and this technique may be of interest to generate novel biotechnological producer cells.

Effects of amino acid additions on ammonium stressed CHO cells

Ammonium is a toxic and inhibitory byproduct of mammalian cell metabolism. At the end of a typical recombinant protein production campaign, the ammonium concentration can be as high as 10 mM, mainly due to glutamine metabolism. Intracellular pH (pH(i)) levels are sensitive to ammonium, which negatively impacts both cell growth and recombinant protein productivity. Ammonium also negatively affects the recombinant protein glycosylation profile, thus altering quality. Many strategies have been adopted to reduce ammonium accumulation, with limited results. This study investigated the addition of amino acids to the growth media for Chinese hamster ovary (CHO) cell cultures as a means of mitigating the negative effects of ammonium. Threonine, proline, and glycine additions improved CHO cell growth and recombinant protein levels. Further, the threonine, proline, and glycine additions positively impacted important metabolic parameters, including glucose consumption, lactate production, glutamine utilization, and final ammonium levels. Additionally, threonine, proline, and glycine increased the level of alpha(2,3)-linked sialic acid, galactose-beta(1,4)-N-acetylglucosamine, and alpha(2,6)-linked sialic acid residues on the recombinant tissue plasminogen activator (t-PA). Thus, threonine, proline, and glycine can be used to mitigate some of the toxic effects of ammonium on cell growth, recombinant protein productivity, and protein quality.

Impact of temperature reduction and expression of yeast pyruvate carboxylase on hGM-CSF-producing CHO cells

Recently, we demonstrated that a recombinant yeast pyruvate carboxylase expressed in the cytoplasm of BHK-21 cells was shown to partially reconstitute the missing link between glycolysis and TCA, increasing the flux of glucose into the TCA and achieving higher yields of recombinant erythropoietin. In the present study, a CHO cell line producing recombinant human granulocyte macrophage colony stimulating factor was used to evaluate the impact of PYC2 expression and reduced culture temperature. Temperature reduction from 37 to 33 degrees C revealed a reduced growth rate, a prolonged stationary phase and a 2.1-fold increase of the cell specific rhGM-CSF production rate for CHO-K1-hGM-CSF cells. The PYC2-expressing cell clones showed a decreased cell growth and a lower maximum cell concentration compared to the control expressing rhGM-CSF but no PYC2. However, only 65% lactate were produced in PYC2-expressing cells and the product yield was 200% higher compared to the control. The results obtained for CHO cells compared to BHK cells reported previously, indicated that the PYC2 expression dominantly reduced the lactate formation and increased the yield of the recombinant protein to be produced. Finally, the growth and productivity of PYC2-expressing CHO-K1-hGM-CSF cells under both temperature conditions were investigated. The average cell specific rhGM-CSF production increased by 3.2-fold under reduced temperature conditions. The results revealed that the expression of PYC2 and a reduced culture temperature have an additive effect on the cell specific productivity of CHO-K1-hGM-CSF cells.

Endoplasmic reticulum signaling as a determinant of recombinant protein expression

Generation of functional recombinant proteins requires efficient and undisturbed functioning of the ER-Golgi secretory pathway in host cells. In large-scale production, where target proteins are highly overexpressed, this pathway can be easily congested with unfolded or misfolded proteins. Accumulating evidence suggests that, in addition to responsibility for protein processing, ER is also an important signaling compartment and a sensor of cellular stress. Two ER responses have been described to arise from the overaccumulation of proteins: unfolded protein response (UPR) and ER overload response (EOR). UPR and EOR employ various mechanisms at the transcriptional and the translational levels to deal efficiently and appropriately with encountered stress. This review will outline the molecular bases of ER functioning and stress response, highlight the relevance of ER signaling to the large-scale cell culture productivity and discuss possible approaches to the improvement of the secretion capacities of recombinant cells.

Impact of cell culture process changes on endogenous retrovirus expression

Cell culture process changes (e.g., changes in scale, medium formulation, operational conditions) and cell line changes are common during the development life cycle of a therapeutic protein. To ensure that the impact of such process changes on product quality and safety is minimal, it is standard practice to compare critical product quality and safety attributes before and after the changes. One potential concern introduced by cell culture process improvements is the possibility of increased endogenous retrovirus expression to a level above the clearance capability of the subsequent purification process. To address this, retrovirus expression was measured in scaled down and full production scaled Chinese hamster ovary (CHO) cell cultures of four monoclonal antibodies and one recombinant protein before and after process changes. Two highly sensitive, quantitative (Q)-PCR-based assays were used to measure endogenous retroviruses. It is shown that cell culture process changes that primarily alter media components, nutrient feed volume, seed density, cell bank source (i.e., master cell bank vs. working cell bank), and vial size, or culture scale, singly or in combination, do not impact the rate of retrovirus expression to an extent greater than the variability of the Q-PCR assays (0.2-0.5 log(10)). Cell culture changes that significantly alter the metabolic state of the cells and/or rates of protein expression (e.g., pH and temperature shifts, NaButyrate addition) measurably impact the rate of retrovirus synthesis (up to 2 log(10)). The greatest degree of variation in endogenous retrovirus expression was observed between individual cell lines (up to 3 log(10)). These data support the practice of measuring endogenous retrovirus output for each new cell line introduced into manufacturing or after process changes that significantly increase product-specific productivity or alter the metabolic state, but suggest that reassessment of retrovirus expression after other process changes may be unnecessary.

Erythropoietin production from CHO cells grown by continuous culture in a fluidized-bed bioreactor

A Chinese hamster ovary (CHO) cell line that expresses human erythropoietin (huEPO) was in a 2-L Cytopilot fluidized-bed bioreactor with 400 mL macroporous Cytoline-1 microcarriers and a variable perfusion rate of serum-free and protein-free medium for 48 days. The cell density increased to a maximum of 23 x 10(6) cells/mL, beads on day 27. The EPO concentration increased to 600 U/mL during the early part of the culture period (on day 24) and increased further to 980 U/mL following the addition of a higher concentration of glucose and the addition of sodium butyrate. The EPO concentration was significantly higher (at least 2x than that in a controlled stirred-tank bioreactor, in a spinner flask, or in a stationary T-flask culture. The EPO accumulated to a total production of 28,000 kUnits over the whole culture period. The molecular characteristics of EPO with respect to size and pattern of glycosylation did not change with scale up. The pattern of utilization and production of 18 amino acids was similar in the Cytopilot culture to that in a stationary batch culture in a T-flask. The concentration of ammonia was maintained at a low level (< 2 mM) over the entire culture period. The specific rate of consumption of glucose, as well as the specific rates of production of lactate and ammonia, were constant throughout the culture period indicating a consistent metabolic behavior of the cells in the bioreactor. These results indicate the potential of the Cytopilot bioreactor culture system for the continuous production of a recombinant protein over several weeks.

Metabolic control of recombinant monoclonal antibody N-glycosylation in GS-NS0 cells

Variable N-glycosylation at Asn(297) in the Fc region of recombinant therapeutic immunoglobulin G (IgG) molecules, specifically terminal galactosylation and sialylation, may affect both pharmacokinetic behavior and effector functions of recombinant therapeutic antibodies. We investigated the hypothesis that IgG Fc glycosylation can be controlled by manipulation of cellular nucleotide-sugar metabolism. In control cultures, N-glycans associated with the Fc domain of a recombinant humanized IgG1 produced by GS-NS0 cells in culture were predominantly biantennary, variably beta-galactosylated (average 0.3 mol galactose complex N-glycan(-1)) structures with no bisecting N-acetylglucosamine residues, sialylation, or alpha1,3-linked galactosylation evident. However, a variable proportion (5% to 15%) of high-mannose (Man5 to Man9) oligosaccharides were present. To manipulate the cellular content of the nucleotide sugar precursor required for galactosylation, UDP-Gal, we included either 10 mM glucosamine or 10 mM galactose in the culture medium. In the case of the former, a 17-fold increase in cellular UDP-N-acetylhexosamine content was observed, with a concomitant reduction (33%) in total UDP-hexose, although the ratio of UDP-Glc:UDP-Gal (4:1) was unchanged. Associated with these alterations in cellular UDP-sugar content was a significant reduction (57%) in the galactosylation of Fc-derived oligosaccharides. The proportion of high-mannose-type N-glycans (specifically Man5, the substrate for N-acetylglucosaminyltransferase I) at Asn(297) was unaffected. In contrast, inclusion of 10 mM galactose in culture specifically stimulated UDP-Gal content almost five-fold. However, this resulted in only a minimal, insignificant increase (6%) in beta1,4-galactosylation of Fc N-glycans. Sialylation was not improved upon the addition of the CMP-sialic acid (CMP-SA) precursor N-acetylmannosamine (20 mM), even with an associated 44-fold increase in cellular CMP-SA content. Analysis of recombinant IgG1 Fc glycosylation during batch culture showed that beta1,4-linked galactosylation declined slightly during culture, although, in the latter stages of culture, the release of proteases and glycosidases by lysed cells were likely to have contributed to the more dramatic drop in galactosylation. These data demonstrate: (i) the effect of steric hindrance on Fc N-glycan processing; (ii) the extent to which alterations in cellular nucleotide-sugar content may affect Fc N-glycan processing; and (iii) the potential for direct metabolic control of Fc N-glycosylation.

Metabolic control of recombinant protein N-glycan processing in NS0 and CHO cells

Chinese hamster ovary and murine myeloma NS0 cells are currently favored host cell types for the production of therapeutic recombinant proteins. In this study, we compared N-glycan processing in GS-NS0 and GS-CHO cells producing the same model recombinant glycoprotein, tissue inhibitor of metalloproteinases 1. By manipulation of intracellular nucleotide-sugar content, we examined the feasibility of implementing metabolic control strategies aimed at reducing the occurrence of murine-specific glycan motifs on NS0-derived recombinant proteins, such as Galalpha1,3Galbeta1,4GlcNAc. Although both CHO and NS0-derived oligosaccharides were predominantly of the standard complex type with variable sialylation, 30% of N-glycan antennae associated with NS0-derived TIMP-1 terminated in alpha1,3-linked galactose residues. Furthermore, NS0 cells conferred a greater proportion of terminal N-glycolylneuraminic (sialic) acid residues as compared with the N-acetylneuraminic acid variant. Inclusion of the nucleotide-sugar precursors, glucosamine (10 mM, plus 2 mM uridine) and N-acetylmannosamine (20 mM), in culture media were shown to significantly increase the intracellular pools of UDP-N-acetylhexosamine and CMP-sialic acid, respectively, in both NS0 and CHO cells. The elevated UDP-N-acetylhexosamine content induced by the glucosamine/uridine treatment was associated with an increase in the antennarity of N-glycans associated with TIMP-1 produced in CHO cells but not N-glycans associated with TIMP-1 from NS0 cells. In addition, elevated UDP-N-acetylhexosamine content was associated with a slight decrease in sialylation in both cell lines. The elevated CMP-sialic acid content induced by N-acetylmannosamine had no effect on the overall level of sialylation of TIMP-1 produced by both CHO and NS0 cells, although the ratio of N-glycolylneuraminic acid:N-acetylneuraminic acid associated with NS0-derived TIMP-1 changed from 1:1 to 1:2. These data suggest that manipulation of nucleotide-sugar metabolism can promote changes in N-glycan processing that are either conserved between NS0 and CHO cells or specific to either NS0 cells or CHO cells.

Multiple cell culture factors can affect the glycosylation of Asn-184 in CHO-produced tissue-type plasminogen activator

Human tissue-type plasminogen activator (t-PA) contains a variably occupied glycosylation site at Asn-184 in naturally produced t-PA and in t-PA produced in recombinant Chinese hamster ovary (CHO) cells. The presence of an oligosaccharide at this site has previously been shown to reduce specific activity and fibrin binding. In this report, the site occupancy of t-PA is shown to increase gradually over the course of batch and fed-batch CHO cultures. Additional cell culture factors, including butyrate and temperature, are also shown to influence the degree of glycosylation. In each of these cases, conditions with decreased growth rate correlate with increased site occupancy. Investigations using quinidine and thymidine to manipulate the cell cycle distribution of cultures further support this correlation between site occupancy and growth state. Comparison of the cell cycle distribution across the range of cell culture factors investigated shows a consistent relationship between site occupancy and the fraction of cells in the G(0)/G(1) phase of the cell cycle. These results support a correlation between growth state and site occupancy, which fundamentally differs from site occupancy trends previously observed and illustrates the importance of the growth profile of CHO cultures in producing consistently glycosylated recombinant glycoproteins.

A systematic approach to the validation of process control parameters for monoclonal antibody production in fed-batch culture of a murine myeloma

A systematic approach to the validation of control ranges of control parameters for a cell culture process producing a monoclonal antibody is described. Specifically, the structure and functional activity of a monoclonal IgG1 antibody produced at the outer limits of numerical ranges of fed-batch culture control parameters such as pH and temperature were examined, with the aim of providing assurance that antibody produced under varying culture conditions was of consistent quality based on a carefully defined set of specifications. An experimental design was created using a half-fractional factorial design for fed-batch culture incorporating half of the thirty two possible combinations of five selected control parameters at high and low levels. Statistical analysis of all data gathered from the study allowed an assessment of the effects of the process control parameters at either high or low outer limits on fed-batch culture response variables such as growth rate and specific antibody productivity. Measured values for the responses of growth rate and specific antibody productivity throughout this study ranged from 0.22-0.44 d(-1) and 6.4-32 microg monoclonal antibody/10(6) cells/d respectively. Analytical characterisation of monoclonal antibody purified from each fed-batch culture considered the purity, structure and biological activity of the glycoprotein. All antibody preparations were identical to each other and to the current antibody reference standard or control. Glycosylation analysis of certain samples from the study demonstrated that the distribution of glycoforms of the antibody was not affected by the varying process control conditions of the fed-batch cultures.

Enhanced productivity during controlled proliferation of BHK cells in continuously perfused bioreactors

A perfused cell-culture process was developed to investigate the stability of IRF-1-mediated proliferation control in BHK cells and to evaluate the efficacy of a novel promoter in these cells. The cell density of proliferation-controlled producer cells was effectively regulated for over 7 weeks in a microcarrier-based continuously perfused bioreactor. An IRF-1-inducible promoter was employed to express a heterodimeric IgG antibody as a relevant model protein. Basal expression levels were equivalent to that of a highly active viral promoter, while productivity increased up to sixfold during growth arrest. However, no stably expressing clone was isolated in this study. Protein expression decreased gradually with time and could not be induced further in subsequent growth-repression cycles. The results demonstrate that the regulatory system is sufficiently stable to allow controlled growth in a continuous scalable reactor system and that productivity increases can be achieved in a proliferation controlled microcarrier culture.

Recombinant glycodelin carrying the same type of glycan structures as contraceptive glycodelin-A can be produced in human kidney 293 cells but not in chinese hamster ovary cells

We have produced human recombinant glycodelin in human kidney 293 cells and in Chinese hamster ovary (CHO) cells. Structural analyses by lectin immunoassays and fast atom bombardment mass spectrometry showed that recombinant human glycodelin produced in CHO cells contains only typical CHO-type glycans and is devoid of any of the N, N'-diacetyllactosediamine (lacdiNAc)-based chains previously identified in glycodelin-A (GdA). By contrast, human kidney 293 cells produced recombinant glycodelin with the same type of carbohydrate structures as GdA. The presence of a beta1-->4-N-acetylgalactosaminyltransferase functioning in the synthesis of lacdiNAc-based glycans in human kidney 293 cells is concluded to be the cause of the occurrence of lacdiNAc-based glycans on glycodelin produced in these cells. Furthermore, human kidney 293 cells were found to be particularly suited for the production of recombinant glycodelin when they were cultured in high glucose media. Lowering the glucose concentration and the addition of glucosamine resulted in higher relative amounts of oligomannosidic-type glycans and complex glycans with truncated antennae. Human glycodelin is an attractive candidate for the development of a contraceptive agent, and this study gives valuable information for selecting the proper expression system and cell culture conditions for the production of a correctly glycosylated recombinant form.

Metabolic shifts do not influence the glycosylation patterns of a recombinant fusion protein expressed in BHK cells

BHK-21 cells expressing a human IgG-IL2 fusion protein, with potential application in tumor-targeted therapy, were grown under different nutrient conditions in a continuous system for a time period of 80 days. At very low-glucose (< 0.5 mM) or glutamine (< 0. 2 mM) concentrations, a shift toward an energetically more efficient metabolism was observed. Cell-specific productivity was maintained under metabolically shifted growth conditions and at the same time an almost identical intracellular ATP content, obtained by in vivo (31)P NMR experiments, was observed. No significant differences in the oligosaccharide structures were detected from the IgG-IL2 fusion protein preparations obtained by growing cells under the different metabolic states. By using oligosaccharide mapping and MALDI/TOF-MS, only neutral diantennary oligosaccharides with or without core alpha1-6-linked fucose were detected that carried no, one or two beta1-4-linked galactose. Although the O-linked oligosaccharide structures that are present in the IL2 moiety of the protein were studied with less detail, the data obtained from the hydrazinolysis procedure point to the presence of the classical NeuAcalpha2-3Galbeta1-3GalNAc structure. Here, it is shown that under different defined cellular metabolic states, the quality of a recombinant product in terms of O- and N-linked oligosaccharides is stable, even after a prolonged cultivation period. Moreover, unaffected intracellular ATP levels under the different metabolic states were observed.