DEVELOPMENT OF NOVEL EFFECTIVE CAR-T CELL THERAPY FOR HEPATOCELLULAR CARCINOMA

Abstract

Liver cancer, mainly hepatocellular carcinoma (HCC), is the sixth most common cancer worldwide and the fourth most common cause of cancer death. Treatments for HCC are limited and mostly ineffective. Over the last few years, chimeric antigen receptor (CART) immunotherapy, a therapy that modifies patients own T cells with synthetic receptors that target tumor associated antigens, has demonstrated impressive efficacy in hematological tumors, but not in solid tumors yet. Multiple immunotherapeutic approaches are being studied for HCC, including chimeric antigen receptor T cell (CART) therapy. Glypican 3 (GPC3) is expressed on the surface of about 70% of HCC tumor cells, but not in normal hepatocytes. Several anti-GPC3 CARTs have been studied, but clinical outcomes have been published only for CARTs developed from the GC33 antibody. Most GPC3 CARTs are constructed from high-affinity single chain variable fragments (scFvs) because high affinity CARs might generate better anti-tumor effects. However, they may also lead to on-target/off-tumor toxicity due to low expression of target antigens in normal cells; they may incur in higher T cell exhaustion and low persistence. Some studies have shown that the anti-tumor effect of low-affinity CARs can be potentiated by the addition of appropriate co-stimulatory domains. Furthermore, for this study, we aimed to develop novel and effective low-affinity GPC3-CARTs. We sought to identify the best costimulatory domains that would give our low-affinity CARTs enhanced persistence and expansion in GPC3+ tumors while generating potent anti-tumor effects. To attain this goal, three novel mAbs, (6G11, 8F8 and 12D7) specific for HCC, were used to generate second-generation CARTs bearing the 4-1BB co-stimulatory domain. 8F8 CARTs showed the strongest in vitro and in vivo anti-tumor effects in subcutaneous tumors. Stimulated 8F8 CARTs also had higher production of IL-2 and IFN-gamma, and maintained their function after repetitive antigen stimulation. 8F8 binds an epitope close to that of GC33, but with 17X lower affinity, offering us the opportunity to comparatively study the effects of high- versus low- avidity CARTs. To further investigate whether replacing the co-stimulatory domain would enhance the effector functions and persistence of our novel CARTs, we built second-generation CARs using the co-stimulatory domains ICOS, CD28 and 4-1BB, known for their impact on CART persistence, effector function, and memory differentiation, respectively. We found that, in vitro, 8F8 and GC33 CARTs have similar function in the three different CAR designs that were tested (ICOSZ, CD28Z, and 4-1BBZ). 8F8 and GC33 had similar function, however, 8F8 BBZ CARTs exhibited a less apoptotic and less exhausted phenotype than GC33 CARTs. Furthermore, 8F8 CARTs expanded and persisted longer than GC33 when treating HepG2 tumors. Importantly, the tumor-infiltrating 8F8 CARTs maintained a better effector function for a longer time compared to high-avidity GC33 CARTs. 8F8 CARTs generated durable anti-tumor effects. Furthermore, our data suggest that our novel 8F8 CARTs might have the potential to generate potent anti-tumor effects in high-antigen clinical settings.