An introduction to precision feeding for cell-based applications and bio-processes.

By Dr. Rebecca C. Vaught, with edits and commentary from Dr. Alec M. Santiago and mentors.

Cell culture processes still lack thorough understanding in reaching higher titers and robustness using technological means (Taylor et el., 2021). In this pursuit, there are often four possibilities for improving outcomes: 1) strain and cell line development, 2) infrastructure changes, 3) physiological adjustments, 4) media/feed optimization, and combinations of these.

Culture media optimization is critical but overlooked, mainly due to its complexity compared to adjusting other aspects of bio-based processes/applications and the painstaking nature of creating and implementing new formulations coupled with the reality that improvements driven via better feeding could remain modest even after considerable effort and expense (Weschselberger et al., 2012). These challenges position other aspects of the application or process as more desirable candidates for optimization instead of optimization of media formulations/feeding.

Thus it’s probably not surprising that most media formulations (traditional and modern specialized media) for both cell lines and microbes contain undefined components and are missing key vitamins, minerals, and trace elements, and include some ingredients like saccharides, salts, and certain amino acids in ratios that are completely out of sync with target cell physiology (Granat et al., 2019, Vande Voorde et al., 2019, Gardner et al., 2022).

Media component chemistry and the solubility of various components also present considerable challenges and unstable or caustic components are often included to overcome this. These shortcomings directly impact cell metabolism and physiology, negatively affecting viability and performance.

Bio-processes and in-vitro conditions induce cell stress in the forms of nutrient stress (starvation and excess), oxidative stress, shear stress, and metabolic stress and cause the build-up of toxic wastes and widespread changes to gene-expression profiles (Nicolau et al., 2010, Wellen et al., 2010, Liu and Qian, 2014, Polizzi et al., 2015), beginning the first part of a vicious cycle (O’Brien and Hu, 2020) and often resulting in widespread apoptosis. Genetically modified cells and primary cell lines may be particularly prone to such stresses. Genetic modifications often introduce their own genomic stresses that exert detrimental effects on physiology.

Further, stressful conditions can enable genetic erosion/reversion and/or rampant selective sweeps for non-focal genotypes for cells/population (Bjedov et al., 2003, Maharjan and Ferenci, 2017, O’Brien and Hu, 2020, Horwitz, 2021), adding further to declines/variance. For instance, ROS produced from oxidative stress can induce or exacerbate other physiological stress (mentioned above) and create a runaway cycle (Forrester et al., 2018). ROS produced from oxidative stress and oxygen metabolism also cause mutations (Ragu et al., 2007), and stressful physiological conditions can cause mutant clones to rise to high frequency within the batch/population for cell lines and microbes (O’Brien and Hu, 2020, Rugberg and Olsson, 2020). Thus, overall physiological stresses introduce process and product issues.

While many researchers and organizations are expending considerable resources optimizing their cell lines and strains with genetic and metabolic engineering, some experts warn against this strategy as the risks of reversion/loss at scale, under increased stress, outweigh the benefits of continued cell line development (Horwitz, 2021).

Another consequence of note regarding the stresses of laboratory and industrial environments is the production of non-optimal glycosylation patterns for heterologous proteins. Post-translational features of proteins, such as folding and their glycobiology, are critical for appropriate pharmacokinetics and pharmacodynamics of biologic therapeutics, particularly monoclonal antibodies and fusion proteins (Liu et al., 2015). Better media design can achieve optimal glycosylation patterns for biologic therapeutics (Xu et al., 2021), and improve the quality and, thus, performance of enzymes.

Optimal cell feeding can reduce common physiological stresses and consequences for the system/product. Furthermore, biological fidelity and lower cellular stress from optimal feeding can reduce process variance (Farrell et al., 2014) and improve the parameter space for further optimization potential.

For example, most working in biology can appreciate the improved growth or protein production that occurs with enhanced oxygen availability. However, this also often increases oxidative stress, and induces complications negatively affecting the cells and/or product. Similarly, physiological effects caused by nutrient excess are well documented (Eguchi et al., 1996, Wellen et al., 2010), and excess catabolic expenditures result in ROS production (Wellen et al., 2010). Nutrient starvation, however, is also detrimental and well-documented. If cells can circumvent nutrient deprivation/maintain nutrient availability, they are better positioned to respond to stressful or changing environments (Eguchi et al., 1996, Lee et al., 2009, McLeod et al., 2010, Wood et al., 2020). This balance necessitates the urgency for precision nutrient landscapes to achieve greater metabolic output with fewer negative by-products, I.e., improve process robustness with fewer negative side reactions. Thus, better feeding strategies are needed for the ultimate cellular process and application quality.

Better feeding allows for further increases in productivity/titer driven by infrastructure and physiological modifications. For instance, an optimized media with no other changes to the process could improve productivity/titer by a relative maximum of 3X, for example (Table 1). Oppositely, modifications to infrastructure (increased mixing or surface-to-volume ratio) with no media optimization could improve productivity/titer by a relative maximum of 3X, for example, but combining improvements to both combinatorially multiples the effects to 9X (Table 1), experimental evidence of exactly this prediction has been observed multiple times for Van Heron Labs’ studies.

Optimizing the macro and micronutrients within the culture media also reduces the number of undefined components often needed within the media regime and enables the transition to chemically defined conditions. Because basal media formulations are often deficient or designed for other cell types rather than the ones the media is being deployed for, undefined blood products or extracts are often used to solve deficits and solubility issues. This is common for both microbes and cell lines. Using fetal calf serum as an example, fetal calf serum will alter cell physiology, including glucose metabolism (Tildon and Stevenson, 1984), cellular senescence/senescence phenotypes (Duggal and Brinchmann, 2011), and inflammatory phenotypes (Asai et al., 2020).

For more discussion on why chemically-defined conditions are important, please see: https://www.notion.so/vanheronlabs/Ser-X 5153d8d39ebb4c609ced3a173cc6e6ff

Innovation in media and feed design is desperately needed across microbes and cell lines.

Van Heron Labs uses omics data, primarily gene-expression data, and computing (bioinformatics and AI) to create better media and feed formulations for cells to increase the productivity/titer and robustness of bio-based processes and applications. The VHL platform uses genetic information as a tool to gain insights into cellular function and determines the most efficient macro and micronutrient inputs for any biological system. The bioinformatics and AI approach VHL utilizes is proprietary, but in essence, the bioinformatics pipeline determines optimal macronutrients, and the AI platform determines micronutrients. This precision nutrition approach optimizes cellular metabolism, which, in turn, optimizes the bio-process or cell-based application of interest.