How To Grow CHO Cells Without Cysteine, Tyrosine Feeds

A conversation with Bhanu Chandra Mulukutla, Pfizer

Cysteine and tyrosine may be relatively cheap amino acids, but their poor solubility makes them costly headaches in cell culture media development. To get around the issue, most manufacturers rely on dipeptide versions, an approach that improves solubility but can more than double medium costs and complicate production logistics. These amino acids also carry implications for product quality, from trisulfide formation to disulfide scrambling.

Bhanu Chandra Mulukutla, a Research Fellow at Pfizer, and his colleagues have been exploring another way forward, engineering CHO cells to bypass the need for supplemental cysteine and tyrosine altogether. He recently agreed to answer questions about his team’s work in building prototrophic CHO cell lines, the potential to eliminate perfusion, and what this could mean for the future of commercial bioproduction. Here’s what he told us.

To set the stage, how much do cysteine and tyrosine supplementation contribute to media costs — and how do they influence critical quality attributes during production?

Mulukutla: Cysteine and tyrosine are relatively inexpensive amino acids; however, their native forms exhibit limited solubility. As a result, there is an increasing interest in utilizing dipeptide forms of these amino acids, which offer significantly improved solubility. Nonetheless, adopting these alternatives can increase medium costs by at least 100%. Both cysteine and tyrosine play important roles in determining product quality. In intensified bioprocesses, depletion of these amino acids due to inadequate supplementation may lead to protein misincorporations that adversely impact protein structure and function. Elevated cysteine concentrations can promote trisulfide formation in proteins containing disulfide bonds, potentially compromising protein quality. Additionally, cysteine may cap thiol groups, resulting in disulfide scrambling and consequently altering protein structure and function.

How did you solve the problem? What genetic or pathway modifications were introduced to kick-start cysteine and/or tyrosine prototrophy, and how did you choose those targets?

Mulukutla: A potential solution to this issue involves engineering CHO cells to biosynthesize tyrosine and cysteine autonomously. Although tyrosine and cysteine are classified as non-essential amino acids in mammals due to existing biosynthetic pathways using phenylalanine and methionine as a source, respectively, CHO cells lack complete expression of these pathways, rendering the amino acid supplementation essential for their growth. This determination was made through comprehensive multi-omics analysis including transcriptomics and metabolomics, which evaluated gene expression of the enzymes involved in these biosynthetic routes along with the levels of associated metabolite intermediates.^1,2 The study also identified promising genetic engineering targets to impart prototrophy (meaning, ability to biosynthesize) for tyrosine and cysteine. For tyrosine prototrophy, ectopic overexpression of phenylalanine hydroxylase (PAH) and pterin-4 alpha-carbinolamine dehydratase 1 (PCBD1) genes, in the phenylalanine-tyrosine pathway, proved both necessary and sufficient.¹ Similarly, establishing cysteine prototrophy required overexpressing genes encoding glycine N-methyltransferase (GNMT), cystathionine β-synthase (CBS), and cystathionine γ-lyase (CTH) enzymes across methionine catabolism and transulfuration pathways.²

You noted that this technology obviates perfusion. Can you explain how?

Mulukutla: Amino acid catabolic pathways in CHO cells exhibit variable activity. For instance, certain pathways such as phenylalanine-tyrosine and methionine-cysteine are either inactive or only partially active, while others, including the branched-chain amino acid (BCAA) catabolic pathway, display overflow metabolism. As a result of these metabolic processes, intracellular accumulation and subsequent extracellular secretion of metabolite intermediates and their byproducts — produced by non-specific enzymatic activity — occurs.³ Some of these metabolites exert a growth inhibitory effect, leading to cell growth cessation in fed-batch cultures when they reach growth inhibitory concentrations.³

Our research has demonstrated that CHO cells degrade phenylalanine, methionine, and BCAAs into 3-phenyllactate, homocysteine, and short-chain fatty acids (SCFAs; specifically, isovalerate, isobutyrate, and 2-methylbutyrate), respectively.^1,3 These compounds inhibit CHO cell growth at much lower concentrations than traditional inhibitors such as lactate and ammonia. The use of perfusion enables the removal of these inhibitory byproducts from the culture environment resulting in higher growth. Alternatively, metabolic engineering strategies can be implemented to reduce or eliminate byproduct production, thereby decreasing their accumulation during fed-batch processes and potentially obviating the need for perfusion operations.

For phenylalanine-tyrosine and methionine-cysteine pathways, overexpression of missing enzymes redirects the metabolic flux away from byproduct formation and toward tyrosine¹ and cysteine² biosynthesis, respectively. This reduces byproduct accumulation, resulting in better growth in fed-batch cultures. Similarly, in the BCAA catabolic pathway, knockout of the gene encoding branched-chain amino acid transaminase 1 (BCAT1) — enzyme catalyzing the pathway’s first reaction step — abolishes SCFA production.¹ BCAT1 KO CHO cells consumed 30% fewer BCAAs and generated negligible levels of these byproducts, which resulted in higher peak cell densities and viabilities in fed-batch cultures. Enhanced viability enabled extended fed-batch culture durations, facilitating significantly increased volumetric productivity and eliminating the need for perfusion cultures.

What trade-offs, if any, did you observe in terms of growth rate or protein quality?

Mulukutla: Notably, minimal trade-offs have been observed with these metabolic engineering strategies. For instance, in the case of BCAT1 knockout, cell growth rates remained unchanged in passage cultures; however, in fed-batch cultures, cells achieved similar growth rates but attained significantly higher densities due to the absence of growth-inhibitory metabolite accumulation. Importantly, product quality was unaffected. Likewise, tyrosine and cysteine prototrophic cells exhibited comparable initial growth rates and achieved either similar or higher maximum cell densities in fed-batch conditions. Although tyrosine biosynthesis from phenylalanine is energetically costly, reduced byproduct formation compensates for this, resulting in enhanced peak cell densities. For cysteine prototrophy, endogenous cysteine synthesis may be more energetically favorable than uptake of cystine from the medium, as exogenous cystine needs to undergo energy-intensive reduction to cysteine intracellularly before cells can use the amino acid. Additionally, cysteine biosynthesis yields 2-ketobutyrate, which can enter the TCA cycle to further support cellular energy production. No adverse effects on the quality of proteins produced by either tyrosine or cysteine prototrophic cells were observed.^1,2

From a media development perspective, how much cost savings could this approach realistically achieve, particularly at commercial scale?

Mulukutla: Savings can be achieved through both direct and indirect cost reductions. Direct costs pertain to the use of alternatives for cysteine and tyrosine — such as dipeptides — particularly in intensified processes where the limited solubility of these amino acids poses challenges. It is estimated that employing dipeptides for cysteine and tyrosine results in at least a 100% increase in cost. Indirect costs encompass those related to preparation, storage, and transportation of media. The removal of cysteine and tyrosine streamlines medium preparation, eliminating the need for additional supplemental feeds if all amino acids cannot be included in a single nutrient feed medium. Furthermore, excluding cysteine and tyrosine may improve medium stability, potentially extending shelf life and expiry at GMP scale. Reducing or removing cysteine, tyrosine, and branched-chain amino acids in the nutrient feed medium also increases available osmolality space, facilitating the formulation of concentrated media. Enhanced stability combined with concentrated media minimizes preparation time and labor at GMP scale, while also simplifying transportation across various sites within the manufacturing network.

What do you see as the biggest technical hurdle to implementing engineered CHO cells in a GMP setting? Stability, consistency, or regulatory acceptance seem like they could be problematic.

Mulukutla: For any engineered host, it is essential to demonstrate that there are no unintended effects on protein production. Comprehensive evaluations are conducted to verify the capability of the new host to produce not only standard monoclonal antibodies but also other complex proteins. Additionally, the host must exhibit a stable phenotype and consistently yield high-quality protein across multiple generations. We have successfully established these criteria for a CHO host cell line harboring a BCAT1 KO event. Regarding cysteine and tyrosine prototrophic cells, we are currently assessing both their stability and their ability to produce a range of protein types. I do not anticipate significant challenges related to phenotype stability, as the constant metabolic selection pressure maintains the prototrophic characteristics of the cells for either cysteine or tyrosine.

What’s next? Are other amino acids or nutrients on your radar for engineering? Are we moving toward fully self-sufficient CHO lines?

Mulukutla: The approach we have taken involves examining the amino acid metabolism of CHO cells to identify optimization opportunities that benefit the cell. By eliminating BCAA catabolic byproduct production and redirecting catabolic byproducts from phenylalanine and methionine towards tyrosine and cysteine, respectively, approximately 75%-80% of the dominant metabolic byproducts produced by CHO cells in culture have been addressed. There are a few additional byproducts produced at higher levels that are currently under investigation, with efforts focused on either eliminating them or converting them into useful compounds. However, these efforts are approaching diminishing returns, as achieving fully self-sufficient cells may not be necessary; synthesizing amino acids is energetically expensive and could impact the energy available for biomass and protein of interest (POI) production. Except for BCAAs, cysteine, and tyrosine, the uptake rates for most other essential amino acids are sufficiently low such that their less soluble native forms may be used in culture at minimal expense, reducing the need for more soluble analogues such as dipeptides or engineered cells that biosynthesize those amino acids.

References:

1. Mulukutla, B. C., Mitchell, J., Geoffroy, P., Harrington, C., Krishnan, M., Kalomeris, T., Morris, C., Zhang, L., Pegman, P., & Hiller, G. W. (2019). Metabolic engineering of Chinese hamster ovary cells towards reduced biosynthesis and accumulation of novel growth inhibitors in fed-batch cultures. Metab Eng, 54, 54-68. https://doi.org/10.1016/j.ymben.2019.03.001
2. Greenfield, L., Brantley, M., Geoffroy, P., Mitchell, J., DeWitt, D., Zhang, F., & Mulukutla, B. C. (2024). Metabolic engineering of CHO cells towards cysteine prototrophy and systems analysis of the ensuing phenotype. Metab Eng, 84, 128-144. https://doi.org/10.1016/j.ymben.2024.06.003
3. Mulukutla, B. C., Kale, J., Kalomeris, T., Jacobs, M., & Hiller, G. W. (2017). Identification and control of novel growth inhibitors in fed-batch cultures of Chinese hamster ovary cells. Biotechnol Bioeng, 114(8), 1779-1790. https://doi.org/10.1002/bit.26313

About The Expert:

Bhanu Chandra Mulukutla, Ph.D., is a Research Fellow and group leader at Pfizer, where he began in 2012 as a Senior Scientist. His group works on developing cell culture processes for vaccines and recombinant proteins and building novel technologies that enhance cell culture process productivities by optimizing cellular metabolism. He received his Ph.D. in Chemical Engineering from the University of Minnesota.