One-Time Optimization of Advanced T Cell Culture Media Using a Machine Learning Pipeline

Interfacing data science with cell therapy manufacturing: where we are and where we need to be

Although several cell-based therapies have received FDA approval, and others are showing promising results, scalable, and quality-driven reproducible manufacturing of therapeutic cells at a lower cost remains challenging. Challenges include starting material and patient variability, limited understanding of manufacturing process parameter effects on quality, complex supply chain logistics, and lack of predictive, well-understood product quality attributes. These issues can manifest as increased production costs, longer production times, greater batch-to-batch variability, and lower overall yield of viable, high-quality cells. The lack of data-driven insights and decision-making in cell manufacturing and delivery is an underlying commonality behind all these problems. Data collection and analytics from discovery, preclinical and clinical research, process development, and product manufacturing have not been sufficiently utilized to develop a "systems" understanding and identify actionable controls. Experience from other industries shows that data science and analytics can drive technological innovations and manufacturing optimization, leading to improved consistency, reduced risk, and lower cost. The cell therapy manufacturing industry will benefit from implementing data science tools, such as data-driven modeling, data management and mining, AI, and machine learning. The integration of data-driven predictive capabilities into cell therapy manufacturing, such as predicting product quality and clinical outcomes based on manufacturing data, or ensuring robustness and reliability using data-driven supply-chain modeling could enable more precise and efficient production processes and lead to better patient access and outcomes. In this review, we introduce some of the relevant computational and data science tools and how they are being or can be implemented in the cell therapy manufacturing workflow. We also identify areas where innovative approaches are required to address challenges and opportunities specific to the cell therapy industry. We conclude that interfacing data science throughout a cell therapy product lifecycle, developing data-driven manufacturing workflow, designing better data collection tools and algorithms, using data analytics and AI-based methods to better understand critical quality attributes and critical-process parameters, and training the appropriate workforce will be critical for overcoming current industry and regulatory barriers and accelerating clinical translation.

Explainable AI for CHO cell culture media optimization and prediction of critical quality attribute

Cell culture media play a critical role in cell growth and propagation by providing a substrate; media components can also modulate the critical quality attributes (CQAs). However, the inherent complexity of the cell culture media makes unraveling the impact of the various media components on cell growth and CQAs non-trivial. In this study, we demonstrate an end-to-end machine learning framework for media component selection and prediction of CQAs. The preliminary dataset for feature selection was generated by performing CHO-GS (-/-) cell culture in media formulations with varying metal ion concentrations. Acidic and basic charge variant composition of the innovator product (24.97 ± 0.54% acidic and 11.41 ± 1.44% basic) was chosen as the target variable to evaluate the media formulations. Pearson's correlation coefficient and random forest-based techniques were used for feature ranking and feature selection for the prediction of acidic and basic charge variants. Furthermore, a global interpretation analysis using SHapley Additive exPlanations was utilized to select optimal features by evaluating the contributions of each feature in the extracted vectors. Finally, the medium combinations were predicted by employing fifteen different regression models and utilizing a grid search and random search cross-validation for hyperparameter optimization. Experimental results demonstrate that Fe and Zn significantly impact the charge variant profile. This study aims to offer insights that are pertinent to both innovators seeking to establish a complete pipeline for media development and optimization and biosimilar-based manufacturers who strive to demonstrate the analytical and functional biosimilarity of their products to the innovator. KEY POINTS: • Developed a framework for optimizing media components and prediction of CQA. • SHAP enhances global interpretability, aiding informed decision-making. • Fifteen regression models were employed to predict medium combinations.

Challenges in developing cell culture media using machine learning

Microbial and mammalian cells are widely used in the food, pharmaceutical, and medical industries. Developing or optimizing culture media is essential to improve cell culture performance as a critical technology in cell culture engineering. Methodologies for media optimization have been developed to a great extent, such as the approaches of one-factor-at-a-time (OFAT) and response surface methodology (RSM). The present review introduces the emerging machine learning (ML) technology in cell culture engineering by combining high-throughput experimental technologies to develop highly efficient and effective culture media. The commonly used ML algorithms and the successful applications of employing ML in medium optimization are summarized. This review highlights the benefits of ML-assisted medium development and guides the selection of the media optimization method appropriate for various cell culture purposes.

Employing active learning in the optimization of culture medium for mammalian cells

Medium optimization is a crucial step during cell culture for biopharmaceutics and regenerative medicine; however, this step remains challenging, as both media and cells are highly complex systems. Here, we addressed this issue by employing active learning. Specifically, we introduced machine learning to cell culture experiments to optimize culture medium. The cell line HeLa-S3 and the gradient-boosting decision tree algorithm were used to find optimized media as pilot studies. To acquire the training data, cell culture was performed in a large variety of medium combinations. The cellular NAD(P)H abundance, represented as A450, was used to indicate the goodness of culture media. In active learning, regular and time-saving modes were developed using culture data at 168 h and 96 h, respectively. Both modes successfully fine-tuned 29 components to generate a medium for improved cell culture. Intriguingly, the two modes provided different predictions for the concentrations of vitamins and amino acids, and a significant decrease was commonly predicted for fetal bovine serum (FBS) compared to the commercial medium. In addition, active learning-assisted medium optimization significantly increased the cellular concentration of NAD(P)H, an active chemical with a constant abundance in living cells. Our study demonstrated the efficiency and practicality of active learning for medium optimization and provided valuable information for employing machine learning technology in cell biology experiments.

A review of algorithmic approaches for cell culture media optimization

Cell culture media composition and culture conditions play a crucial role in product yield, quality and cost of production. Culture media optimization is the technique of improving media composition and culture conditions to achieve desired product outcomes. To achieve this, there have been many algorithmic methods proposed and used for culture media optimization in the literature. To help readers evaluate and decide on a method that best suits their specific application, we carried out a systematic review of the different methods from an algorithmic perspective that classifies, explains and compares the available methods. We also examine the trends and new developments in the area. This review provides recommendations to researchers regarding the suitable media optimization algorithm for their applications and we hope to also promote the development of new cell culture media optimization methods that are better suited to existing and upcoming challenges in this biotechnology field, which will be essential for more efficient production of various cell culture products.

Machine learning and hypothesis driven optimization of bull semen cryopreservation media

Cryopreservation provides a critical tool for dairy herd genetics management. Due to widely varying inter- and within-bull post thaw fertility, recent research on cryoprotectant extender medium has not dramatically improved suboptimal post-thaw recovery in industry. This progress is stymied by the interactions between samples and the many components of extender media and is often compounded by industry irrelevant sample sizes. To address these challenges, here we demonstrate blank-slate optimization of bull sperm cryopreservation media by supervised machine learning. We considered two supervised learning models: artificial neural networks and Gaussian process regression (GPR). Eleven media components and initial concentrations were identified from publications in bull semen cryopreservation, and an initial 200 extender-post-thaw motility pairs were used to train and 32 extender-post-thaw motility pairs to test the machine learning algorithms. The median post-thaw motility after coupling differential evolution with GPR the increased from 52.6 ± 6.9% to 68.3 ± 6.0% at generations 7 and 17 respectively, with several media performing dramatically better than control media counterparts. This is the first study in which machine learning was used to determine the best combination of constituents to optimize bull sperm cryopreservation media, and provides a template for optimization in other cell types.

Natural killer cells in clinical development as non-engineered, engineered, and combination therapies

Natural killer (NK) cells are unique immune effectors able to kill cancer cells by direct recognition of surface ligands, without prior sensitization. Allogeneic NK transfer is a highly valuable treatment option for cancer and has recently emerged with hundreds of clinical trials paving the way to finally achieve market authorization. Advantages of NK cell therapies include the use of allogenic cell sources, off-the-shelf availability, and no risk of graft-versus-host disease (GvHD). Allogeneic NK cell therapies have reached the clinical stage as ex vivo expanded and differentiated non-engineered cells, as chimeric antigen receptor (CAR)-engineered or CD16-engineered products, or as combination therapies with antibodies, priming agents, and other drugs. This review summarizes the recent clinical status of allogeneic NK cell-based therapies for the treatment of hematological and solid tumors, discussing the main characteristics of the different cell sources used for NK product development, their use in cell manufacturing processes, the engineering methods and strategies adopted for genetically modified products, and the chosen approaches for combination therapies. A comparative analysis between NK-based non-engineered, engineered, and combination therapies is presented, examining the choices made by product developers regarding the NK cell source and the targeted tumor indications, for both solid and hematological cancers. Clinical trial outcomes are discussed and, when available, assessed in comparison with preclinical data. Regulatory challenges for product approval are reviewed, highlighting the lack of specificity of requirements and standardization between products. Additionally, the competitive landscape and business field is presented. This review offers a comprehensive overview of the effort driven by biotech and pharmaceutical companies and by academic centers to bring NK cell therapies to pivotal clinical trial stages and to market authorization.

Combined multivariate statistical and flux balance analyses uncover media bottlenecks to the growth and productivity of CHO cell cultures

Chinese hamster ovary (CHO) cells are widely used for producing recombinant proteins. To enhance their productivity and product quality, media reformulation has been a key strategy, albeit with several technical challenges, due to the myriad of complex molecular mechanisms underlying media effects on culture performance. Thus, it is imperative to characterize metabolic bottlenecks under various media conditions systematically. To do so, we combined partial least square regression (PLS-R) with the flux balance analysis of a genome-scale metabolic model to elucidate the physiological states and metabolic behaviors of human alpha-1 antitrypsin producing CHO-DG44 cells grown in one commercial and another two in-house media under development. At the onset, PLS-R was used to identify metabolite exchanges that were correlated to specific growth and productivity. Then, by comparing metabolic states described by resultant flux distributions under two of the media conditions, we found sub-optimal level of four nutrients and two metabolic wastes, which plausibly hindered cellular growth and productivity; mechanistically, lactate and ammonia recycling were modulated by glutamine and asparagine metabolisms in the media conditions, and also by hitherto unsuspected folate and choline supplements. Our work demonstrated how multivariate statistical analysis can be synergistically combined with metabolic modelling to uncover the mechanistic elements underlying differing media performance. It thus paved the way for the systematic identification of nutrient targets for medium reformulation to enhance recombinant protein production in CHO cells. This article is protected by copyright. All rights reserved.

Clever Experimental Designs: Shortcuts for Better iPSC Differentiation

For practical use of pluripotent stem cells (PSCs) for disease modelling, drug screening, and regenerative medicine, the cell differentiation process needs to be properly refined to generate end products with consistent and high quality. To construct and optimize a robust cell-induction process, a myriad of cell culture conditions should be considered. In contrast to inefficient brute-force screening, statistical design of experiments (DOE) approaches, such as factorial design, orthogonal array design, response surface methodology (RSM), definitive screening design (DSD), and mixture design, enable efficient and strategic screening of conditions in smaller experimental runs through multifactorial screening and/or quantitative modeling. Although DOE has become routinely utilized in the bioengineering and pharmaceutical fields, the imminent need of more detailed cell-lineage specification, complex organoid construction, and a stable supply of qualified cell-derived material requires expedition of DOE utilization in stem cell bioprocessing. This review summarizes DOE-based cell culture optimizations of PSCs, mesenchymal stem cells (MSCs), hematopoietic stem cells (HSCs), and Chinese hamster ovary (CHO) cells, which guide effective research and development of PSC-derived materials for academic and industrial applications.