Real-Time Analytics in Automated Cell Therapies Manufacturing

Monday, March 11, 2024

Over the past few years, cell and gene therapies (CGT) have emerged as promising solutions for previously difficult-to-treat diseases. The CGT development pipeline has been growing rapidly, fueled by scientific and medical advancements, as well as significant funding from both the public and private sectors. As of third quarter of 2024, 92 CGT products have been approved by regulatory agencies worldwide, and over 2,900 more were in various phases of development.1 Chimeric antigen receptor (CAR) T cell therapies represented 47% of the genetically modified cell therapies pipeline.1 Despite rapidly expanding development pipeline and sustained investment from public and private sectors, clinical and commercial manufacturing of CGTs still face several important challenges, limiting their availability and wide adoption.

The high cost of treatment with CAR-T cell therapies, which may exceed $1 million for some patients,2 is a major hurdle. Not all insurance plans cover the treatment, and even with coverage, significant out-of-pocket costs may still be incurred. Additionally, manufacturing capacity is severely constrained by complex logistics and scaling-up difficulties. The current manufacturing capacity for CAR-T therapies is not sufficient to meet patient demand, and some cancer centers report that about 20% of patients die while waiting for treatment.3 Reducing wait time by as little as 2 months could improve overall treatment efficacy by as much as 14% in some cases.4

Compared to those of more established biologics, such as monoclonal antibodies, CGT manufacturing processes are generally more complex, difficult to characterize, and highly labor-intensive. Protein-based biotherapeutics can be purified without having to keep the host cells alive, unlike the living-cell drugs that require retention of cell function and quality to preserve their efficacy. Dealing with the variability of starting materials, meeting strict manufacturing timeframes, and maintaining compliance with Current Good Manufacturing Practices (cGMP) make it difficult to implement a robust manufacturing process. Additionally, complex and labor-intensive manufacturing requires highly- trained personnel and large and costly cleanroom facilities. Reducing manufacturing cost, lowering wait times, and increasing supply requires transitioning away from manual manufacturing protocols and towards adaptive process strategies.

Automation, when implemented early in clinical development, can significantly diminish the need for human intervention and errors, improve process robustness, and increase throughput. Several automated, closed CGT manufacturing systems, such as CliniMACS Prodigy (Miltenyi Biotech) and Cocoon (Lonza), have been introduced over the years. These systems can perform multiple manufacturing steps and are well-suited for decentralized production of individualized therapies. Automated bioreactor systems, such as Quantum Flex Cell Expansion System (Terumo Blood and Cell Technologies) and Xuri Cell Expansion System W25 (Cytiva), utilize perfusion to achieve high cell densities and fit both scale-out and scale-up manufacturing strategies. The next generation of automated manufacturing systems, utilizing robotics, is on the horizon. Cell Shuttle (Cellares) and Multiply Labs’ robotic system aim to significantly increase manufacturing throughput and greatly reduce human intervention in cell therapy manufacturing.

Most of these systems incorporate the cell expansion step, during which cells are cultured in vitro. Cell expansion is the most time-consuming step and poses several distinct challenges related to maintaining an optimal cell culture environment. Small changes in cell culture conditions, such as shear stress, can significantly affect cell metabolism and differentiation.5 CAR-T cell therapy manufacturing failure rates have been reported as high as 14%, mostly due to suboptimal cell expansion.6 Careful monitoring and control of critical process parameters (CPPs) with minimal human intervention is crucial for process robustness and is best achieved with the help of real-time monitoring technologies integrated with the automated manufacturing systems.

In upstream processes, real-time monitoring of pH, dissolved oxygen (DO), temperature, agitation or perfusion rates, cell density, cell viability, cell identity, cytokine secretion, cellular activation, and certain key nutrients and metabolites would be valuable. In 2004, the FDA set guidelines for the biopharmaceutical industry to implement Quality by Design and PAT (process analytical technologies) to achieve higher quality and more efficiencies in the manufacturing process of these drugs, notably by using more real-time monitoring of CPPs and CQAs (critical quality attributes).21 However, at present, only a few real-time monitoring technologies have been implemented in automated CGT manufacturing systems. The limited prior knowledge and added complexity of CGT manufacturing make it difficult to determine CPPs and CQAs and which are worth monitoring in real-time. However, as the CGT field continues to mature and new PAT approaches become available, the interest in transitioning to real-time monitoring of certain CPPs and CQAs is increasing.

Historically, automated CGT systems had limited real-time monitoring capabilities, instead relying on manual sampling and analysis of CPPs and CQAs by the operators. The difficulty of integrating sensors designed for much larger bioreactors, limited standardization of hardware and software, and the lack of robust measurement technologies, are partially to blame. Without the ability to monitor critical parameters in real-time, operators must rely on periodic manual sampling and off-line analysis for tracking cell culture conditions and making process adjustments. This is costly, labor-intensive, consumes valuable material and can increase the risk of process contamination. To minimize these adverse effects, sampling frequency is usually kept to the minimum and decisions are made based on limited information.

For example, glucose and lactate concentrations have been linked to cell proliferation and cell viability.7-9 Glucose is the primary energy source and its availability is linked to cell growth and CQAs of the final product. In CAR-T cell manufacturing, isolated T cells must first be activated and expanded for 7-14 days. During the expansion process, cytokines are added to promote T cell proliferation, but this may also drive effector differentiation.10,11 Less differentiated, naïve and memory T cells, have been shown to possess increased longevity and exhibit greater antitumor activity compared to effector T cells.12 There are several methods that can be used for maintaining T cells in an undifferentiated state. One way is by optimizing cell culture conditions via modulating glucose concentration - culturing cells in high concentration of glucose before switching to glucose deprivation.13 There is also ongoing research into in vitro adaptation of CAR-T cells to low-nutrient tumor microenvironment and to prevent T cell exhaustion.

Lactate, on the other hand, is primarily viewed as a metabolic waste product and its excessive accumulation has been shown to inhibit growth and induce toxicity in cells.14,15 Additionally, excessive lactate can indicate suboptimal culture conditions or inefficient nutrient utilization.16 At the same time, other studies have demonstrated potential benefits of cultivation in high extracellular lactate environments.17 This goes to show that better understanding the role that nutrients and metabolites play in regulating intracellular metabolism is highly important for advancing the CGT field. Therefore, it’s essential that suitable PAT tools and strategies continue to be identified and evaluated. Doing so, can actualize the vision of PAT and bring about revolutionary changes.

Protocols already exist for using off-line glucose and lactate measurements to estimate cell count, predict cell expansion rate, determine optimal harvest time, as well for optimization of media usage in perfusion.18 While manual sampling may be sufficient for a well-characterized process, advancing automation in CGT necessitates integration of continuous monitoring solutions with the automated systems. This is best achieved through partnerships between providers of process analytical technology (PAT) and automated manufacturing systems. One such example is a recently announced collaboration between 908 Devices and Terumo Blood and Cell Technologies.

In 2023, 908 Devices launched a PAT device for on-line monitoring of glucose and lactate in cell culture and fermentation processes. The device utilizes selective flow diffusion technology to sample glucose and lactate molecules directly from the bioreactor media without pulling a bulk sample. The analysis is fully automated and can be set to provide measurements as often as every 2 minutes. Combining flow diffusion sampling with robust biosensor technology offers several key advantages. By varying the diffusion time, a very broad measuring range and low limit of detection can be achieved. Another major advantage is that the biosensor measures concentrations in a highly diluted buffer fluid and is never directly exposed to the bioreactor media. This eliminates the need to sterilize the sensor, which can degrade its performance, and prevents potential interferences from other chemical components in the media. Finally, the sterile barrier, provided by the diffusion membrane, allows for online monitoring without increased risk of process contamination and without loss of valuable cell culture material.

The synergy between automated cell expansion systems and real-time analyzers creates a powerful combination for monitoring and optimizing cell expansion processes. The automated hollow-fiber bioreactor’s scalable cell expansion capabilities are complemented by the ability to monitor key metabolic parameters in real-time. Some users already rely on periodically measured lactate and glucose measurements obtained from manually collected media samples. Glucose consumption and lactate production are used to estimate cell density, predict cell expansion trajectory, and to adjust perfusion rate accordingly. Lactate production trend can also help operators determine when the cells are in an expansion phase and once this begins to plateau, can inform of the optimal time for harvesting. Additionally, these measurements can help detect and mitigate process deviations. For example, different cells may adhere differently to the surface of the hollow-fiber bioreactor and glucose and lactate levels can indicate that cells may be detaching.19 The integration of real-time monitoring significantly increases the quantity of data points available to process development scientists and operators from just a 1 or 2 per day to as many as 720 per day. This enables a more dynamic and adaptive approach to cell culture management, reduces need for manual sampling and lowers cleanroom dwell time.

While the integration of real-time monitoring represents a significant advancement in cell and gene therapy manufacturing, certain challenges and areas for improvement remain. One such challenge is the need for development and standardization of control strategies based on real-time monitoring of nutrients and metabolites. Standardization would facilitate comparability between manufacturing approaches and contribute to the establishment of industry-wide best practices. Another aspect to consider is the integration of AI and machine learning algorithms. These systems could be trained on process data from past runs to develop a digital twin of a cell expansion process to predict trends and recommend or automatically implement process adjustments. One of such use cases is being investigated in the EU project AIDPATH (AI powered, Decentralized Production for Advanced Therapies in the Hospital). The use case is focused on development of a bioreactor digital twin by mechanistically modeling CAR-T cell expansion using nutrients and metabolites data. The system aims to provide real-time cell concentration information as well as a short-term forecast.20

Furthermore, ongoing R&D efforts are essential to continue refining real-time monitoring technologies suitable for integration with the current and next generation of automated CGT manufacturing systems. For example, spectroscopy-based techniques, such as Raman, could provide continuous, in-situ measurements of multiple CPPs. Traditional Raman bioprocess analyzers have not been widely adopted due to reliance on empirical calibration models. The inherent variability of personalized therapies makes development and validation of these models difficult for such processes. A recent innovation in Raman-based monitoring, introduced by 908 Devices, brings this promising PAT technology one step closer to being implemented in CGT manufacturing. Unlike traditional Raman analyzers, this solution doesn’t rely on empirical calibration models. The concentrations of critical parameters are determined using a first principle approach, enabling measurements of multiple process parameters right after a simple two-point calibration of the optical probe.

In conclusion, the integration of automated CGT manufacturing systems with real-time monitoring technologies represents a transformative approach to cell and gene therapy manufacturing. This combination addresses critical challenges in cell expansion, offering improved scalability, reproducibility, and tighter process control. As the field of cell and gene therapy evolves, these innovations will play a pivotal role in shaping the future of therapeutic development and manufacturing, enhancing the efficiency of current processes and paving the way for next-generation therapies available to a broader population.

REFERENCES

  1. American Society of Gene + Cell Therapy. Gene, Cell,+ RNA Therapy Landscape Report. Q3 2024 Quarterly Data Report. https://asgct.org/global/documents/asgct-citeline-q3-2024-report. aspx. Published October 2024.
  2. Sahli B, Eckwright D, Darling E, Gleason PP, Leach JW. Chimeric antigen receptor T-cell therapy real-world assessment of total cost of care and clinical events for the treatment of relapsed or refractory lymphoma. J Clin Oncol. 2021;39(15_suppl):e19500 Events for the Treatment of Relapsed or Refractory Lymphoma among 15 Million Commercially Insured Members
  3. Chen A. “How do you decide?”: Cancer treatment’s CAR-T crisis has patients dying on a waitlist. STAT. Published June 2, 2022. Accessed January 4, 2024. https://www.statnews. com/2022/06/02/car-t-crisis-cancer-patients-die-waiting/
  4. Chen AJ, Zhang J, Agarwal A, Lakdawalla DN. Value of Reducing Wait Times for Chimeric Antigen Receptor T-Cell Treatment: Evidence From Randomized Controlled Trial Data on Tisagenlecleucel for Diffuse Large B-Cell Lymphoma. Value in Health. 2022;25(8):1344-1351.
  5. Yourek G, McCormick SM, Mao JJ, Reilly GC. Shear stress induces osteogenic differentiation of human mesenchymal stem cells. Regenerative Medicine. 2010;5(5):713-724.
  6. Lamture G, Baer A, Fischer JW, Colon-Moran W, Bhattarai N. TCR-independent Activation in Presence of a Src-family Kinase Inhibitor Improves CAR-T Cell Product Attributes. Journal of Immunotherapy. 2021;Publish Ahead of Print..
  7. Fischer K, Hoffmann P, Voelkl S, et al. Inhibitory effect of tumor cell–derived lactic acid on human T cells. Blood. 2007;109(9):3812-3819.
  8. Patel SD, Papoutsakis ET, Winter JN, Miller WM. The Lactate Issue Revisited: Novel Feeding Protocols To Examine Inhibition of Cell Proliferation and Glucose Metabolism in Hematopoietic Cell Cultures. Biotechnology Progress. 2000;16(5):885-892.
  9. Horiguchi I, Urabe Y, Kimura K, Sakai Y. Effects of glucose, lactate and basic FGF as limiting factors on the expansion of human induced pluripotent stem cells. Journal of Bioscience and Bioengineering. 2018;125(1):111-115.
  10. Crompton JG, Sukumar M, Restifo NP. Uncoupling T-cell expansion from effector differentiation in cell-based immunotherapy. Immunological Reviews. 2013;257(1):264-276.
  11. Fearon DT. The Expansion and Maintenance of Antigen‐Selected CD8+ T Cell Clones. Advances in Immunology. Published online January 1, 2007:103-139.
  12. Hosein Rostamian, Keyvan Fallah-Mehrjardi, Khakpoor-Koosheh M, et al. A metabolic switch to memory CAR T cells: Implications for cancer treatment. Cancer Letters. 2021;500:107-118.
  13. Amini A, Veraitch F. Glucose deprivation enriches for central memory T cells during chimeric antigen receptor-T cell expansion. Cytotherapy. 2019;21(5):S30-S31.
  14. Patel SD, Papoutsakis ET, Winter JN, Miller WM. The Lactate Issue Revisited: Novel Feeding Protocols To Examine Inhibition of Cell Proliferation and Glucose Metabolism in Hematopoietic Cell Cultures. Biotechnology Progress. 2000;16(5):885-892.
  15. Fischer K, Hoffmann P, Voelkl S, et al. Inhibitory effect of tumor cell–derived lactic acid on human T cells. Blood. 2007;109(9):3812-3819.
  16. Van Beylen K, Youssef A, Peña Fernández A, Lambrechts T, Papantoniou I, Aerts JM. Lactate-Based Model Predictive Control Strategy of Cell Growth for Cell Therapy Applications. Bioengineering. 2020;7(3):78.
  17. Odenwelder DC, Lu X, Harcum SW. Induced pluripotent stem cells can utilize lactate as a metabolic substrate to support proliferation. Biotechnology Progress. 2020;37(2).
  18. Martin-Manso G, Hanley PJ. Using the Quantum Cell Expansion System for the Automated Expansion of Clinical-Grade Bone Marrow-Derived Human Mesenchymal Stromal Cells. Methods in molecular biology. Published online January 1, 2015:53-63
  19. Frank ND, Miller MM, Sethi D. An optimized HEK293T cell expansion protocol using a hollow fiber bioreactor system. Biology Methods and Protocols. 2024;8(1).
  20. Niklas Bäckel, Hort S, Kis T, et al. Elaborating the potential of Artificial Intelligence in automated CAR-T cell manufacturing. Frontiers in Molecular Medicine. 2024;3. 21. PAT — A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance Guidance for Industry. OCTOBER 2004.

About the Author

Boris Aleynik is a Director of Product Marketing at 908 Devices (Boston, MA) where he leads development and implementation of marketing strategies for Life Science products. Prior to this, Boris was a Director of Strategic Product Planning for Industry 4.0 Technologies at Pall Life Sciences. Boris received his MBA from Northwestern University in 2015 and BS in Mechanical Engineering from Iowa State University in 2007.

Mindy Miller Ph.D, Senior Scientist/Manager at Terumo Blood and Cell Technologies (Lakewood, CO) where she leads a team in the development of innovative solutions for cell therapies. Mindy holds a doctorate in Immunology from University of Missouri and conducted post-doctoral studies at University of Missouri, as well as National Jewish Health in Denver, CO. She has authored multiple scientific publications and is the recipient of the Young Investigator Award from the American Association of Immunologists.

Publication Detail

This article appeared in American Pharmaceutical Review - Innovations at Interphex 2024 Supplement 
Pages: 12-15

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special
offers from American Pharmaceutical Review delivered to your inbox!
Sign up now!

  • <<
  • >>

Join the Discussion