Potency assays are required for cellular and gene therapy (CGT) final product (FP) lot release, stability testing, and manufacturing updates, as per the FDA draft guidance on Potency Tests for CGT Products (2011), and Manufacturing Changes and Comparability for Human CGT Products (2023). These assays by definition measure the ability of a product to effect a desired result, and the FDA allows for “considerable flexibility” in determining the appropriate potency assay. However, standard in vitro potency assays are limited in their capacity to meaningfully assess manufacturing comparability and consistency, or product quality to provide predictive value for clinical performance. For chimeric antigen receptor (CAR)-T cell products, certificates of analysis typically report key FP attributes such as CAR expression, cell concentration and viability, impurities and target-specific in vitro potency, among others. However, clinical effectiveness of CAR-T is dependent on complex pharmacokinetics (PK)/ pharmacodynamics (PD) including in vivo engraftment, trafficking, drug exposure kinetics, tumor antigen-driven T cell expansion in tumor microenvironment, survival and persistence, and other factors that are not easily recapitulated in short-term in vitro assays. We developed a bioassay for assessing CAR-T FP potency using a tumor-bearing murine NSG xenograft model to evaluate in vivo performance of GMP-manufactured autologous (Auto) or allogeneic (Allo) CAR-T FPs. We hypothesized that this complex in vivo system may better model PK/PD and provide a more meaningful way to assess CAR-T FP potency. Preclinical dose-finding data from both RUO and GMP anti-BCMA CAR-T lots were used to establish a potency specification for RPMI-8226 tumor control at a “stress dose” of 5e6 CAR-T cells at <60% of the tumor volume compared to the untreated control group at 21 days after CAR-T cell infusion. Using these specifications, we evaluated the in vivo potency performance of GMP-manufactured Auto FP samples (P-BCMA-101; NCT03288493) generated from multiple myeloma (MM) patients and administered in the clinic with known responses ranging from poor (no response or stable disease) to significant (very good partial response [VGPR] or stringent complete response [sCR]). In addition, GMP-manufactured Allo FP samples generated from healthy donors (P-BMCA-ALLO1; NCT04960579) were also tested. We report that half of the P-BCMA-101 products (50%) previously demonstrating favorable clinical outcomes in MM patients were also able to control RPMI-8226 tumor in vivo, while none of the P-BCMA-101 products (0%) that resulted in poor clinical responses were capable of tumor control. These data show translational relevance considering the known clinical outcomes in MM patients, albeit at a high level of stringency at current specifications. Interestingly, all the Allo products (100%) performed at least as well as or better than the Auto products capable of controlling tumor, demonstrating a greater potency of Allo CAR-T cells manufactured from healthy donors. There is great demand for potency assays that can reliably assess manufacturing consistency and product quality, and that might provide some level of predictive value for clinical performance, especially considering the current limitations of standard in vitro potency assays. In vivo potency bioassays have been proposed or utilized for other products with complex biological mechanisms of action including vaccines, gene therapies, and other cell therapy products. The potency bioassay described herein demonstrated a level of positive predictive value for clinical performance using Auto FPs of known clinical outcome in MM patients. In the Auto CAR-T setting, this assay could be utilized for correlative studies including possible biomarker identification. In the Allo CAR-T setting, this potency bioassay may inform FP manufacturing consistency and quality for lot release to ensure patients receive only the highest quality products.
Read full abstract