REBEL and CGT: Process analysis is key to optimizing cell therapy processes and ensuring successful production.

For many indications, cell therapy is becoming an increasingly viable treatment option—particularly for patients who have exhausted conventional treatment options. The substantial clinical activity and the approval of several autologous, patient-specific therapies have driven increased industry demand and highlighted bottlenecks in the manufacturing workflow. These bottlenecks must be addressed to enhance production efficiency, ensure safety, and enable timely delivery to patients.

In the autologous setting, each patient’s health status and demographic data translate into a unique cellular population, introducing variability into the subsequent cell therapy manufacturing process. Given this variability, it underscores the critical importance of process control in ensuring drug consistency.

One key area of focus is Process Analytical Technology (PAT), which is essential for providing critical quality attribute (CQA) information throughout the entire production process and during QC testing. Real-time process feedback is crucial for enhancing process monitoring and ensuring production success; however, critical process feedback is often delayed due to excessively long detection times.

At the recent online meeting of the American Society for Gene & Cell Therapy (ASGCT), Rich Rogers from Bristol Myers Squibb outlined a process analytical strategy based on mass spectrometry to support the development and optimization of cell therapy processes. Rogers discussed the analytical requirements and instrumentation needs for PAT, as well as how MS-based PAT holds promise as a powerful technology capable of overcoming many of the challenges currently faced by existing analytical tools.

Attribute Monitoring of CAR-T Therapy

Rogers first provided an overview of BMS’s platform for developing its CAR-T platform based on autologous lentiviral vectors. In the production of autologous CAR-T therapies, the process begins with the collection of a patient’s cells using leukemia cells. Once mononuclear cells have been harvested, numerous steps must be taken—such as cell separation, activation, modification, and expansion—to prepare the T cells for reinfusion into the patient. Currently, BMS uses PAT to monitor cell proliferation (i.e., viability and cell density) and employs flow cytometry at each stage to analyze T cells and impurities (i.e., the presence of contaminating cell types). Although the data generated by these tools are sufficient, there remains a need for more in-depth monitoring of cell health and metabolic status, as this could significantly impact the efficacy of the final product.

The possibilities for implementing PAT are limitless and applicable to every aspect of the manufacturing process. PAT can be used to identify process leverage points, ensuring product consistency and enhancing the success rate of both upstream processes (such as temperature, pH, glucose, amino acids, cell viability, and metabolites) and downstream processes (such as process-related impurities). Additionally, PAT enables the quantification of Critical Quality Attributes (CQAs) in the final product—namely, purity, potency, safety, and CAR frequency.

Rogers outlined several key features that make the instrument suitable for process PAT:

• Instrument footprint: Production cleanrooms typically have limited space, so instrument size needs to be taken into account.

• Capital investment: Cost considerations distinct from those for R&D equipment

• Number of biometric measurements: Instruments capable of performing multi-attribute testing are preferred.

• Low-volume sample requirements: Especially during the production process, it is often impossible to collect large-volume samples for in-process testing.

• Time for operational analysis: The time from sampling to obtaining data should be very short; the detection time is critical for providing relevant process feedback.

• Low technical expertise required: The training needed for operating the instrument should be minimal, and no expert involvement is necessary.

• Data speed: As short as possible—or near real-time—to maximize benefits for the production process.

Mass Spectrometry-Based PAT Strategy

Rogers provided a detailed overview of the MS-based non-targeted and targeted PAT strategies developed by BMS for process analytics. MS is a well-established technique for generating robust and reproducible data in large-molecule therapies such as monoclonal antibodies. However, applying this strategy to cell therapies involving thousands of proteins presents several challenges that need to be addressed. One such challenge is the complexity of the cells underlying CAR-T therapies. While large-molecule therapies typically require only the purification of monoclonal antibodies, cell therapies may involve thousands of proteins that need to be considered. Consequently, traditional large-molecule MAM (Multiple Attribute Methodology) approaches cannot be directly applied to cell therapies.

PAT based on proteomics

Rogers also described a MS-based cell surface proteomics approach used by BMS, which performs non-targeted analysis of T-cell surface proteins throughout the entire manufacturing process. Cell surface proteins are biotin-labeled, followed by cell lysis, enrichment using labeled streptavidin, enzymatic digestion, and MS analysis.

This non-targeted approach is advantageous because it does not require reagents that specifically target cell surface proteins (unlike flow cytometry-based methods), thereby providing an unbiased, global view of the T-cell surface protein repertoire.

Monitoring of process residues and secreted proteins represents another area where multi-attribute targeted proteomics approaches are being developed to replace ELISA, which is constrained by the need to develop target-specific detection reagents. This MS-based approach is highly sensitive, enabling the simultaneous multiplex quantification of hundreds of protein targets. By selecting targeted peptides from chromatographic separation, fragmentation, and Orbitrap MS analysis, it provides quantitative detection of process residues and T-cell-secreted proteins.

In summarizing these proteomics approaches, Rogers commented that although they are encouraged by the accuracy of the results, this method still requires further effort to meet the PAT requirements he outlined earlier in his talk. MS instruments occupy a large footprint, are costly, and demand extensive personnel training. Moreover, there is a need to improve data transmission speeds so that experimental staff can receive timely information in real time during production runs.

PAT based on targeted metabolomics

In the latter half of the presentation, Rogers used the REBEL analyzer (908 Devices) to focus on targeted PAT analysis based on metabolomics. This benchtop device performs online metabolic analysis of T cells by sampling the culture medium from bioreactors. Analysis of cell-secreted and metabolized products can provide valuable insights into cell health and efficacy. BMS plans to leverage the REBEL analyzer at every stage of process development to conduct metabolomic studies. This includes both the construction of lentiviral vectors and the T-cell proliferation process itself, during which the REBEL analyzer has already been employed to deliver crucial information about T-cell health and metabolic status.

The REBEL analyzer meets all the requirements of PAT:

• Small footprint (microwave-sized), reasonable cost

• Simultaneously monitors more than 30 analytes

• Requires only a 10 μl sample volume

• Pipetting skills are the only technical requirement.

• Minimal training required

• Fast detection time—approximately ~7 minutes per sample

• Quick data acquisition: The integrated software performs analysis and automatically generates reports, which can be exported as CSV or PDF files compatible with LIMS and PIMS.

During the method and instrument evaluation process, five replicates of the same bioreactor run were performed using REBEL on three separate samples to demonstrate high reproducibility, as shown in Figure 1 below.

At the conclusion of his talk, Rogers emphasized the importance of PAT in enhancing our understanding of cell therapy processes and ensuring successful manufacturing. Investments in PAT and innovative technologies are pushing the boundaries of throughput and accuracy in current approaches, Rogers noted. He believes that MS-based methods hold great potential for both targeted and non-targeted proteomic and metabolomic characterization, enabling a comprehensive, global assessment of T cells throughout the entire cell therapy manufacturing process.

In the next presentation at the seminar, Kerin Gregory, Product Manager at 908 Devices, shared how to use the REBEL analyzer to perform rapid compositional analysis of cell culture media—specifically, amino acids, vitamins, and biogenic amines—to achieve timely and meaningful process analytics. Cell culture media provide essential nutrients for cells grown in vitro and directly influence cell growth, productivity, and the functionality of subsequent therapeutic applications. In-depth analysis of culture media can help gain deeper insights into nutrient consumption, thereby facilitating optimization efforts and ultimately improving CQAs.

Nutritional Composition Analysis and Medium Optimization

In particular, for the production of viral vectors, meeting the demands of both clinical and commercial objectives requires increasing viral titers. Since clonal variations can affect the metabolism of host cells (i.e., HEK293), analyzing the composition of culture media/reactive solutions can provide a better understanding of the host cell’s nutritional needs and offer opportunities to optimize the culture medium for enhanced viral titers. Although numerous commercial culture medium formulations are available on the market, their relative concentrations and compositional ingredients can vary significantly—as illustrated in Figure 2 below. These different medium choices can have a substantial impact on both the titer and integrity of viral vectors. REBEL can also be used to monitor batch-to-batch variations within selected formulations, ensuring consistency and reproducibility.

Moreover, studies have shown that T-cell culture media require several essential amino acids to regulate ex vivo activation and expansion, which play a role in the metabolic readiness of CAR-T cells and their antitumor efficacy after being introduced into the tumor microenvironment in vivo. This underscores the profound impact of cell culture medium components on cell growth, metabolism, and productivity.

Serum exchange

Gregory added that the cell therapy industry has shifted from serum-containing culture media formulations to serum-free ones for two main reasons: 1) Serum exhibits batch-to-batch variability; 2) In addition to the immunogenicity risks associated with clinical applications, serum also carries the potential risk of contamination by extraneous components. The development of chemically defined media (CDM) represents an area in which REBEL can be used to characterize ingredients and support formulation efforts for culture media. REBEL has a dynamic range of 5–100 μM, enabling accurate detection of low concentrations of amino acids and other nutrients. Unlike other quantitative methods, this analysis is not affected by the presence of serum proteins.