Characterization of Lentiviral Vector Derived Anti-Bcma CAR T Cells Reveals Key Parameters for Robust Manufacturing of Cell-Based Gene Therapies for Multiple Myeloma

T cells engineered with chimeric antigen receptors (CAR) specific to CD19 have caused rapid and durable clinical responses in ~90% of patients with acute lymphoblastic leukemia. These data support the development of additional CAR T cell products for the treatment of other hematological malignancies. Recently, B cell maturation antigen (BCMA) expression has been proposed as a marker for identification of malignant plasma cells in patients with multiple myeloma (MM). Nearly all MM and some non-Hodgkin's lymphoma tumor cells express BCMA, while normal tissue expression is restricted to plasma cells and a subset of mature B cells. Therefore, BCMA is an attractive CAR T cell target to treat patients with MM and some B cell lymphomas. To this end, using lentiviral vector technology, we successfully generated CAR T cells specific to BCMA that exhibit potent anti-tumor activity to both multiple myeloma and Burkitt's lymphoma in animal models. Manufacture of CAR T cells for individual patient treatment requires the establishment of a robust and reproducible process - since variability in manufacturing could impact the potency of each cell product.

To begin to understand the parameters of the manufacturing process that might contribute to the activity of the final product, we first tested the impact of lentiviral vector (LVV) multiplicity of infection (MOI) on CAR T cell phenotype and function. Using a broad range of MOIs (0.625 to 40) across multiple independent PBMC donors we observed no differences in population doubling or cell size throughout the ~10 day manufacturing process, irrespective of the MOI used. As expected, the number of anti-BCMA CAR expressing cells, the level of CAR expression per cell and the average vector copy number (VCN) in the cell product increased proportionally with MOI. Similarly, T cell function, as determined by an IFNg cytokine release assay in response to BCMA-expressing K562 target cells, was also correlated with the LVV MOI. Notably, increased IFNg expression was readily observable at MOIs as low as 1.25 and reached a plateau with T cells generated using an MOI of 20 or more - highlighting the sensitivity of this functional assay. Analogous data demonstrating MOI dependent in vitro killing activity were obtained using a BCMA-expressing tumor cell cytotoxicity assay.

Varying the LVV MOI used during transduction simultaneously alters both the amount of anti-BCMA CAR molecules expressed per cell as well as the number of T cells in the cell product that express anti-BCMA CAR. To evaluate each variable in isolation we generated T cell products containing the same frequency of anti-BCMA CAR T cells (26 ± 4% CAR+ T cells) but different levels of anti-BCMA expression per cell by diluting T cell products made with MOIs from 5 to 40 with donor-matched untransduced cells. While these populations had markedly different levels of CAR surface expression per cell (based on anti-BCMA CAR MFI levels measured by flow cytometry) both low and high expressing anti-BCMA CAR T cell products exhibited identical levels of cytotoxicity against BCMA-expressing tumor cells. These data suggest it is the number of CAR expressing cells that is the critical driver of higher functional activity (perhaps due to the efficiency of LVV mediated anti-BCMA CAR expression per transduced cell). Finally, using this information the variability in manufacturing of anti-BCMA CAR T cells was evaluated across 11 independent normal PBMC donors. All 11 products demonstrated very similar properties with respect to cell growth (population doublings, cell volume), and VCN. Importantly, using our standard MOI we obtained a consistent and high level of anti-BCMA CAR expressing T cells that resulted in robust IFNg cytokine release when co-cultured with BCMA-expression cells. Together, our data highlight the frequency of anti-BCMA CAR T cells per cell product as a key parameter for anti-tumor activity in vitro. Moreover, these data suggest that our LVV driven T cell engineering process can reproducibly generate robust anti-BCMA CAR expressing T cell products in a donor independent manner. A phase I clinical trial to evaluate this technology as a cell-based gene therapy for MM is under development.