Predictive macroscopic modeling of cell growth, metabolism and monoclonal antibody production: Case study of a CHO fed-batch production.

We describe a systematic approach to establish predictive models of CHO cell growth, cell metabolism and monoclonal antibody (mAb) formation during biopharmaceutical production. The prediction is based on a combination of an empirical metabolic model connecting extracellular metabolic fluxes with cellular growth and product formation with mixed Monod-inhibition type kinetics that we generalized to every possible external metabolite. We describe the maximum specific growth rate as a function of the integral viable cell density (IVCD). Moreover, we also take into account the accumulation of metabolites in intracellular pools that can influence cell growth. This is possible even without identification and quantification of these metabolites as illustrated with fed-batch cultures of Chinese Hamster Ovary (CHO) cells producing a mAb. The impact of cysteine and tryptophan on cell growth and cell productivity was assessed, and the resulting macroscopic model was successfully used to predict the impact of new, untested feeding strategies on cell growth and mAb production. This model combining piecewise linear relationships between metabolic rates, growth rate and production rate together with Monod-inhibition type models for cell growth did well in predicting cell culture performance in fed-batch cultures even outside the range of experimental data used for establishing the model. It could therefore also successfully be applied for in silico prediction of optimal operating conditions.

Use of novel cystine analogs to decrease oxidative stress and control product quality.

Continuous improvements of cell culture media are required in order to ensure high yield and product quality. However, some components can be instable and lead to detrimental effects on bioprocess performances. l-cysteine is an essential amino acid commonly used in cell culture media. Despite its beneficial effect on recombinant protein production, in some cases, this component can be responsible for product microheterogeneity. In this context, alternative components have to be found in order to reduce product variants while maintaining high productivity. In this study, we have assessed the performance of different cysteine and cystine analogs : N-acetyl-cysteine, s-sulfocysteine, N,N'-diacetyl-l-cystine and the N,N'-diacetyl-l-cystine dimethylester (DACDM). Replacement of cysteine by cystine analogs, and especially DACDM, has shown positive impact on charge variants level and recombinant protein coloration level. Moreover, this molecule contributed to the increase of the intracellular glutathione pool, which suggests a close relationship with the oxidative stress regulation.

Characterization of glutathione proteome in CHO cells and its relationship with productivity and cholesterol synthesis.

Glutathione (GSH) plays a central role in the redox balance maintenance in mammalian cells. Previous studies of industrial Chinese hamster ovary cell lines have demonstrated a relationship between GSH metabolism and clone productivity. However, a thorough investigation is required to understand this relationship and potentially highlight new targets for cell engineering. In this study, we have modulated the GSH intracellular content of an industrial cell line under bioprocess conditions to further elucidate the role of the GSH synthesis pathway. Two strategies were used: the variation of cystine supply and the direct inhibition of the GSH synthesis using buthionine sulfoximine (BSO). Over time of the bioprocess, a correlation between intracellular GSH and product titer has been observed. Analysis of metabolites uptake/secretion rates and proteome comparison between BSO-treated cells and nontreated cells has highlighted a slowdown of the tricarboxylic acid cycle leading to a secretion of lactate and alanine in the extracellular environment. Moreover, an adaptation of the GSH-related proteome has been observed with an upregulation of the regulatory subunit of glutamate-cysteine ligase and a downregulation of a specific GSH transferase subgroup, the Mu family. Surprisingly, the main impact of BSO treatment was observed on a global downregulation of the cholesterol synthesis pathways. As cholesterol is required for protein secretion, it could be the missing piece of the puzzle to finally elucidate the link between GSH synthesis and productivity.

Oxidative stress-alleviating strategies to improve recombinant protein production in CHO cells.

Large scale biopharmaceutical production of biologics relies on the overexpression of foreign proteins by cells cultivated in stirred tank bioreactors. It is well recognized and documented fact that protein overexpression may impact host cell metabolism and that factors associated with large scale culture, such as the hydrodynamic forces and inhomogeneities within the bioreactors, may promote cellular stress. The metabolic adaptations required to support the high-level expression of recombinant proteins include increased energy production and improved secretory capacity, which, in turn, can lead to a rise of reactive oxygen species (ROS) generated through the respiration metabolism and the interaction with media components. Oxidative stress is defined as the imbalance between the production of free radicals and the antioxidant response within the cells. Accumulation of intracellular ROS can interfere with the cellular activities and exert cytotoxic effects via the alternation of cellular components. In this context, strategies aiming to alleviate oxidative stress generated during the culture have been developed to improve cell growth, productivity, and reduce product microheterogeneity. In this review, we present a summary of the different approaches used to decrease the oxidative stress in Chinese hamster ovary cells and highlight media development and cell engineering as the main pathways through which ROS levels may be kept under control.

Application of a genome-scale model in tandem with enzyme assays for identification of metabolic signatures of high and low CHO cell producers.

Biopharmaceutical industrial processes are based on high yielding stable recombinant Chinese Hamster Ovary (CHO) cells that express monoclonal antibodies. However, the process and feeding regimes need to be adapted for each new cell line, as they all have a slightly different metabolism and product performance. A main limitation for accelerating process development is that the metabolic pathways underlying this physiological variability are not yet fully understood. This study describes the evolution of intracellular fluxes during the process for 4 industrial cell lines, 2 high producers and 2 low producers (n = 3), all of them producing a different antibody. In order to understand from a metabolic point of view the phenotypic differences observed, and to find potential targets for improving specific productivity of low producers, the analysis was supported by a tailored genome-scale model and was validated with enzymatic assays performed at different days of the process. A total of 59 reactions were examined from different key pathways, namely glycolysis, pentose phosphate pathway, TCA cycle, lipid metabolism, and oxidative phosphorylation. The intracellular fluxes did not show a metabolic correlation between high producers, but the degree of similitude observed between cell lines could be confirmed with additional experimental observations. The whole analysis led to a better understanding of the metabolic requirements for all the cell lines, allowed to the identification of metabolic bottlenecks and suggested targets for further cell line engineering. This study is a successful application of a curated genome-scale model to multiple industrial cell lines, which makes the metabolic model suitable for process platform.

Application of a curated genome-scale metabolic model of CHO DG44 to an industrial fed-batch process.

CHO cells have become the favorite expression system for large scale production of complex biopharmaceuticals. However, industrial strategies for upstream process development are based on empirical results, due to a lack of fundamental understanding of intracellular activities. Genome scale models of CHO cells have been reconstructed to provide an economical way of analyzing and interpreting large-omics datasets, since they add cellular context to the data. Here the most recently available CHO-DG44 genome-scale specific model was manually curated and tailored to the metabolic profile of cell lines used for industrial protein production, by modifying 601 reactions. Generic changes were applied to simplify the model and cope with missing constraints related to regulatory effects as well as thermodynamic and osmotic forces. Cell line specific changes were related to the metabolism of high-yielding production cell lines. The model was semi-constrained with 24 metabolites measured on a daily basis in n = 4 independent industrial 2L fed batch cell culture processes for a therapeutic antibody production. This study is the first adaptation of a genome scale model for CHO cells to an industrial process, that successfully predicted cell phenotype. The tailored model predicted accurately both the exometabolomics data (r2 ≥ 0.8 for 96% of the considered metabolites) and growth rate (r2 = 0.91) of the industrial cell line. Flux distributions at different days of the process were analyzed for validation and suggestion of strategies for medium optimization. This study shows how to adapt a genome scale model to an industrial process and sheds light on the metabolic specificities of a high production process. The curated genome scale model is a great tool to gain insights into intracellular fluxes and to identify possible bottlenecks impacting cell performances during production process. The general use of genome scale models for modeling industrial recombinant cell lines is a long-term investment that will highly benefit process development and speed up time to market.

Segmented linear modeling of CHO fed-batch culture and its application to large scale production.

We describe a systematic approach to model CHO metabolism during biopharmaceutical production across a wide range of cell culture conditions. To this end, we applied the metabolic steady state concept. We analyzed and modeled the production rates of metabolites as a function of the specific growth rate. First, the total number of metabolic steady state phases and the location of the breakpoints were determined by recursive partitioning. For this, the smoothed derivative of the metabolic rates with respect to the growth rate were used followed by hierarchical clustering of the obtained partition. We then applied a piecewise regression to the metabolic rates with the previously determined number of phases. This allowed identifying the growth rates at which the cells underwent a metabolic shift. The resulting model with piecewise linear relationships between metabolic rates and the growth rate did well describe cellular metabolism in the fed-batch cultures. Using the model structure and parameter values from a small-scale cell culture (2 L) training dataset, it was possible to predict metabolic rates of new fed-batch cultures just using the experimental specific growth rates. Such prediction was successful both at the laboratory scale with 2 L bioreactors but also at the production scale of 2000 L. This type of modeling provides a flexible framework to set a solid foundation for metabolic flux analysis and mechanistic type of modeling. Biotechnol. Bioeng. 2017;114: 785-797. © 2016 The Authors. Biotechnology and Bioengineering Published by Wiley Periodicals, Inc.

Macroscopic modeling of mammalian cell growth and metabolism.

We review major modeling strategies and methods to understand and simulate the macroscopic behavior of mammalian cells. These strategies comprise two important steps: the first step is to identify stoichiometric relationships for the cultured cells connecting the extracellular inputs and outputs. In a second step, macroscopic kinetic models are introduced. These relationships together with bioreactor and metabolite balances provide a complete description of a system in the form of a set of differential equations. These can be used for the simulation of cell culture performance and further for optimization of production.

Highly efficient deletion of FUT8 in CHO cell lines using zinc-finger nucleases yields cells that produce completely nonfucosylated antibodies.

IgG1 antibodies produced in Chinese hamster ovary (CHO) cells are heavily alpha1,6-fucosylated, a modification that reduces antibody-dependent cellular cytotoxicity (ADCC) and can inhibit therapeutic antibody function in vivo. Addition of fucose is catalyzed by Fut8, a alpha1,6-fucosyltransferase. FUT8(-/-) CHO cell lines produce completely nonfucosylated antibodies, but the difficulty of recapitulating the knockout in protein-production cell lines has prevented the widespread adoption of FUT8(-/-) cells as hosts for antibody production. We have created zinc-finger nucleases (ZFNs) that cleave the FUT8 gene in a region encoding the catalytic core of the enzyme, allowing the functional disruption of FUT8 in any CHO cell line. These reagents produce FUT8(-/-) CHO cells in 3 weeks at a frequency of 5% in the absence of any selection. Alternately, populations of ZFN-treated cells can be directly selected to give FUT8(-/-) cell pools in as few as 3 days. To demonstrate the utility of this method in bioprocess, FUT8 was disrupted in a CHO cell line used for stable protein production. ZFN-derived FUT8(-/-) cell lines were as transfectable as wild-type, had similar or better growth profiles, and produced equivalent amounts of antibody during transient transfection. Antibodies made in these lines completely lacked core fucosylation but had an otherwise normal glycosylation pattern. Cell lines stably expressing a model antibody were made from wild-type and ZFN-generated FUT8(-/-) cells. Clones from both lines had equivalent titer, specific productivity distributions, and integrated viable cell counts. Antibody titer in the best ZFN-generated FUT8(-/-) cell lines was fourfold higher than in the best-producing clones of FUT8(-/-) cells made by standard homologous recombination in a different CHO subtype. These data demonstrate the straightforward, ZFN-mediated transfer of the Fut8- phenotype to a production CHO cell line without adverse phenotypic effects. This process will speed the production of highly active, completely nonfucosylated therapeutic antibodies.

Impact of RNA interference on gene networks.

Small endogenous RNAs such as microRNAs (miRNAs) and small interfering RNAs (siRNAs) have been found to post-transcriptionally control cellular gene networks by targeting complementary mRNAs for translation impairment (miRNA) or destruction (siRNA). We have developed a computational model, coordinated to molecular and biochemical parameters of RNA interference pathways, to provide (semi-) quantitative insight into the molecular events managing siRNA-mediated gene expression silencing in native and synthetic gene networks. Based on mass-conservation principles and kinetic rate laws, we converted biochemical RNA interference pathways into a set of ordinary differential equations that describe the dynamics of siRNA-mediated translation-regulation in mammalian cells. Capitalizing on mechanistic details of synthetic transactivator operation, we wired this model into a transcription control circuitry in which the siRNA and its target mRNA are independently regulated at the transcriptional level. In this context, we studied the impact of siRNA transcription timing on the onset of target gene transcription and production kinetics of target mRNA-encoded proteins. We also simulated the rate of siRNA-induced mRNA depletion and demonstrated that the relative concentrations of interacting siRNAs/mRNAs and the number of siRNA-specific target sites on a transcript modulate (i) the rate of target mRNA disappearance, (ii) the steady-state mRNA levels and (iii) induction dynamics of mRNA-encoded protein production. As our model predictions are consistent with available biochemical parameters, extrapolations may improve our understanding of how complex regulatory gene networks are impacted by small endogenous RNAs.

6-hydroxy-nicotine-inducible multilevel transgene control in mammalian cells.

The precise control of transgene expression is essential for biopharmaceutical manufacturing, gene therapy and tissue engineering. We have designed a novel conditional transcription technology, which enables reversible induction, repression and adjustment of desired transgene expression using the clinically inert 6-hydroxy-nicotine (6HNic). The 6-hydroxy-nicotine oxidase (6HNO) repressor (HdnoR), which manages nicotine metabolism in Arthrobacter nicotinovorans pAO1 by binding to a specific operator of the 6-hydroxy-nicotine oxidase (O(NIC)), was fused to the Krueppel-associated box protein of the human kox-1 gene (KRAB) to create a synthetic 6HNic-dependent transsilencer (NS) that controls chimeric mammalian promoters, which are assembled by cloning tandem O(NIC) operators 3' of a constitutive promoter. In the absence of 6HNic, NS binds to O(NIC) and silences the constitutive promoter, which otherwise drives high-level transgene expression when the NS-O(NIC) interaction stops in the presence of 6HNic. Generic NICE(ON) technology was compatible with a variety of constitutive viral and mammalian housekeeping promoters, each of which enabled specific induced, repressed, adjusted and reversible transgene expression profiles in Chinese hamster ovary (CHO-K1), baby hamster kidney (BHK-21) as well as in human fibrosarcoma (HT-1080) cells. NICE(ON) also proved successful in controlling multicistronic expression units for coordinated transcription of up to three transgenes and in the fine-tuning of transcription-translation networks, in which RNA polymerase II- and III-dependent promoters, engineered for 6HNic responsiveness, drove expression of siRNAs that triggered specific transgene knockdown. NICE(ON) represents a robust and versatile technology for the precise tuning of transgene expression in mammalian cells.

Improved transgene expression fine-tuning in mammalian cells using a novel transcription-translation network.

Following the discovery of RNA interference (RNAi) and related phenomena, novel regulatory processes, attributable to small non-protein-coding RNAs, continue to emerge. Capitalizing on the ability of artificial short interfering RNAs (siRNAs) to trigger degradation of specific target transcripts, and thereby silence desired gene expression, we designed and characterized a generic transcription-translation network in which it is possible to fine-tune heterologous protein production by coordinated transcription and translation interventions using macrolide and tetracycline antibiotics. Integration of siRNA-specific target sequences (TAGs) into the 5' or 3' untranslated regions (5'UTR, 3'UTR) of a desired constitutive transcription unit rendered transgene-encoded protein (erythropoietin, EPO; human placental alkaline phosphatase, SEAP; human vascular endothelial growth factor 121, VEGF(121)) production in mammalian cells responsive to siRNA levels that can be fine-tuned by macrolide-adjustable RNA polymerase II- or III-dependent promoters. Coupling of such macrolide-responsive siRNA-triggered translation control with tetracycline-responsive transcription of tagged transgene mRNAs created an antibiotic-adjustable two-input transcription-translation network characterized by elimination of detectable leaky expression with no reduction in maximum protein production levels. This transcription-translation network revealed transgene mRNA depletion to be dependent on siRNA and mRNA levels and that translation control was able to eliminate basal expression inherent to current transcription control modalities. Coupled transcription-translation circuitries have the potential to lead the way towards composite artificial regulatory networks, to enable complex therapeutic interventions in future biopharmaceutical manufacturing, gene therapy and tissue engineering initiatives.

A novel mammalian expression system derived from components coordinating nicotine degradation in arthrobacter nicotinovorans pAO1.

We describe the design and detailed characterization of 6-hydroxy-nicotine (6HNic)-adjustable transgene expression (NICE) systems engineered for lentiviral transduction and in vivo modulation of angiogenic responses. Arthrobacter nicotinovorans pAO1 encodes a unique catabolic machinery on its plasmid pAO1, which enables this Gram-positive soil bacterium to use the tobacco alkaloid nicotine as the exclusive carbon source. The 6HNic-responsive repressor-operator (HdnoR-O(NIC)) interaction, controlling 6HNic oxidase production in A.nicotinovorans pAO1, was engineered for generic 6HNic-adjustable transgene expression in mammalian cells. HdnoR fused to different transactivation domains retained its O(NIC)-binding capacity in mammalian cells and reversibly adjusted transgene transcription from chimeric O(NIC)-containing promoters (P(NIC); O(NIC) fused to a minimal eukaryotic promoter [P(min)]) in a 6HNic-responsive manner. The combination of transactivators containing various transactivation domains with promoters differing in the number of operator modules as well as in their relative inter-O(NIC) and/or O(NIC)-P(min) spacing revealed steric constraints influencing overall NICE regulation performance in mammalian cells. Mice implanted with microencapsulated cells engineered for NICE-controlled expression of the human glycoprotein secreted placental alkaline phosphatase (SEAP) showed high SEAP serum levels in the absence of regulating 6HNic. 6HNic was unable to modulate SEAP expression, suggesting that this nicotine derivative exhibits control-incompatible pharmacokinetics in mice. However, chicken embryos transduced with HIV-1-derived self-inactivating lentiviral particles transgenic for NICE-adjustable expression of the human vascular endothelial growth factor 121 (VEGF121) showed graded 6HNic response following administration of different 6HNic concentrations. Owing to the clinically inert and highly water-soluble compound 6HNic, NICE-adjustable transgene control systems may become a welcome alternative to available drug-responsive homologs in basic research, therapeutic cell engineering and biopharmaceutical manufacturing.

Quorum-sensing-based toolbox for regulatable transgene and siRNA expression in mammalian cells.

Technologies for regulated expression of multiple transgenes in mammalian cells have gathered momentum for bioengineering, gene therapy, drug discovery, and gene-function analyses. Capitalizing on recently developed mammalian transgene modalities (QuoRex) derived from Streptomyces coelicolor, we have designed a flexible and highly compatible expression vector set that enables desired transgene/siRNA control in response to the nontoxic butyrolactone SCB1. The construction-kit-like expression portfolio includes (i) multicistronic (pTRIDENT), (ii) autoregulated, (iii) bidirectional (pBiRex), (iv) oncoretro- and lentiviral transduction, and (v) RNA polymerase II-based siRNA transcription-fine-tuning vectors for straightforward implementation of QuoRex-controlled (trans)gene modulation in mammalian cells.

Engineered Streptomyces quorum-sensing components enable inducible siRNA-mediated translation control in mammalian cells and adjustable transcription control in mice.

BACKGROUND: Recent advances in functional genomics, gene therapy, tissue engineering, drug discovery and biopharmaceuticals production have been fostered by precise small-molecule-mediated fine-tuning of desired transgenes. METHODS: Capitalizing on well-evolved quorum-sensing regulatory networks in Streptomyces coelicolor we have designed a mammalian regulation system inducible by the non-toxic butyrolactone SCB1. Fusion of the S. coelicolor SCB1 quorum-sensing receptor ScbR to the human Kox-1-derived transsilencing domain reconstituted a mammalian transsilencer (SCS) able to repress transcription from SCS-specific operator-containing promoters in a reverse SCB1-adjustable manner. RESULTS: This quorum-sensing-derived mammalian transgene control system (Q-ON) enabled precise SCB1-specific fine-tuning of (i) desired transgene transcription in a variety of mammalian/human cell lines and human primary cells, (ii) small interfering RNA-mediated posttranscriptional knockdown (siRNA) in mammalian cells, and (iii) dosing of a human glycoprotein in mice. CONCLUSIONS: As exemplified by Q-ON technology, bacterial quorum-sensing regulons may represent a near-infinite source for the design of mammalian gene control systems compatible with molecular interventions relevant to future gene therapy and tissue engineering scenarios.

Macrolide- and tetracycline-adjustable siRNA-mediated gene silencing in mammalian cells using polymerase II-dependent promoter derivatives.

RNA interference has emerged as a powerful technology for downregulation of specific genes in cells and animals. We have pioneered macrolide- and tetracycline-adjustable short interfering RNA (siRNA) expression for conditional target gene translation fine-tuning in mammalian/human cell lines based on modified RNA polymerase II promoters. Established macrolide- and tetracycline-dependent transactivators/trans-silencers bound and activated modified target promoters tailored for optimal siRNA expression in response to clinical antibiotics' dosing regimes and modulated desired target genes in Chinese hamster ovary (CHO-K1) and human fibrosarcoma (HT-1080) cells with high precision. Further optimization of adjustable RNA polymerase II-based siRNA-specific promoters as well as their combination with various transmodulators enabled near-perfect regulation configurations in specific cell types. Devoid of major genetic constraints compared to basic RNA polymerase III-based siRNA-specific promoters, we expect RNA polymerase II counterparts to significantly advance siRNA-based molecular interventions in biopharmaceutical manufacturing and gene-function analysis as well as gene therapy and tissue engineering.