Optimizing the Manufacture of CAR-T Cells for Clinical Applications

Abstract 348

Chimeric antigen receptors (CARs) are artificial molecules that can be used to redirect T cell immune response against antigens expressed on the surface of tumor cells. Recent encouraging clinical data from our group and others has shown that T cells engineered with these molecules can effectively traffic to distant tumor sites, penetrate even bulky disease, and eradicate disseminated tumors. Although promising, most current protocols expand engineered T cells non-specifically using IL2 and OKT3, which often results in a decrease in the frequency of transgenic populations over time. Additionally, cell expansion using conventional cultureware is complicated and labor intensive, which limits the broader application of this therapy. With the purpose of optimizing and streamlining CAR-T cell manufacture, we assessed whether cell expansion could be improved by: (i) supplementing non-specific stimuli (IL2) with an artificial antigen presenting cell (a-APC) engineered to express cognate antigen and co-stimulatory molecules, and (ii) efficiently and rapidly expanding cells in a simple and scalable gas permeable culture device (G-Rex), developed by Wilson Wolf Manufacturing for expanding suspension cells. As a proof of principle, we sought to expand T cells engineered with a CAR targeting the prostate cancer antigen, PSCA. We first generated an antigen-expressing a-APC cell line by modifying K562 cells, which already expressed a range of co-stimulatory molecules including CD80, CD86, and 41BBL, with a retroviral vector encoding the PSCA antigen. After the co-culture of CAR-PSCA T cells with the irradiated a-APC, we found that a-APCs co-expressing PSCA antigen, CD80, and 41BBL were the most effective in inducing T cell expansion, with a 1.9 fold increase in total cell numbers when compared with CAR T cells expanded in the presence of IL2 alone. We also saw an increase in the frequency of transgenic CAR-modified T cells in cultures expanded in the presence of a-APCs co-expressing PSCA antigen, CD80, and 41BBL, which increased from 36.5% CAR-modified cells to 88.1% after 10 days of culture. In contrast, the percentage of transgenic T cells was sustained when culture in the presence of IL2 (36.5% on day 0 and 37.2% on day 10). Thus, culture of CAR-T cells with antigen-expressing a-APCs not only improves total cell output, but also enriches for transgene-expressing.

Next, to assess whether we could scale up cell production for clinical application we transferred the engineered a-APCs and CAR-PSCA modified T cells (at a 2:1 ratio) into a static GMP-compliant G-Rex with a surface area of 100cm². In these G-Rex devices, O₂ and CO₂ are exchanged across a silicone membrane at the base, which allows for the addition of an increased depth of medium above the cells, providing more nutrients while the waste products are diluted. These culture conditions have been shown to increase cell output when compared with conventional commercial products such as bags, flasks, and 24-well tissue culture plates, without increasing the number of cell doublings. From an initial seeding density of 25E+06 CAR-modified T cells (0.25E+06 cells per cm²), we obtained a total of 2200–2500E+06 cells (22-25E+06 T cells per cm²) within 10 days of culture. Thus, without any intervention we obtained a 93 fold increase in cell numbers using only 1 liter of T cell culture media. As expected, the co-culture of antigen-expressing a-APCs with CAR-T cells also resulted in an enrichment of transgenic T cells (from 33.2% to 81.7% after 10 days of culture). Thus, we achieved a 2.4±1.2 fold increase in the frequency of transgenic T cells. Taken together the total T cell fold expansion (93) and the enrichment for the transgene (2.4±1.2), we calculate a 223.5±111.6 fold expansion of CAR T cells with 10 days of culture. Importantly we demonstrated the robustness of this manufacture process by successfully extending this approach to other CAR T cell products.