Glypican-3-specific CAR-NKT cells overexpressing BATF3 mediate potent antitumor activity against hepatocellular carcinoma.

Abstract

e14521 Background: Hepatocellular carcinoma (HCC), the fourth leading cause of cancer-related deaths worldwide, is amenable to immunotherapies that target glypican-3 (GPC3), a protein expressed by over 75% of HCCs but not by healthy tissues. The invariant natural killer T cells (NKTs) prevent HCC tumorogenesis through interactions with the gut microbiome in mice, and the presence of tumor-infiltrating NKTs is associated with improved clinical outcomes. We previously tested GPC3-specific chimeric antigen receptors (CARs) and selected a receptor with both CD28 and 41BB costimulatory endodomains based on its ability to mediate superior antitumor activity of transduced NKTs in vivo. Co-expression of interleukin-15 (IL15), a cytokine that supports NKT expansion and persistence, with the chosen GPC3-CAR further enhanced this activity. GPC3-CAR NKTs co-expressing IL15 (15.GPC3-CAR) overexpressed the Basic Leucine Zipper ATF-Like Transcription Factor 3 (BATF3), a member of the AP-1 family. Since AP-1 family proteins are master regulators of cell survival and proliferation, we hypothesized that overexpressing BATF3 in GPC3-CAR NKTs (BATF3.GPC3-CAR) will lead to superior expansion, persistence, and antitumor activity compared to 15.GPC3-CAR NKTs. Methods: We used RNAseq and flow cytometry to define the transcriptomic and phenotypic characteristics, respectively, of BATF3-overexpressing NKTs at baseline and after tumor cell killing in co-culture. We also used co-culture assays to evaluate the ability of BATF3- or IL15-expressing NKTs to proliferate, produce effector cytokines, and kill tumor cells after multiple rounds of exposure. Finally, we evaluated the ability of BATF3.GPC3-CAR and IL15.GPC3-CAR NKTs to expand, persist, and target tumors in an HCC xenograft murine model. Results: BATF3 overexpression in NKTs induced significant differences in gene expression profiles compared to a GFP control at baseline (DEGs up: 20; down: 26, p < 0.01) and after tumor cell killing (DEGs up: 19; down: 32; p < 0.01; fold change > 2 or < 0.5, respectively) corresponding to enrichment of gene sets related to metabolic fitness. Importantly, CD62L+ NKTs, which have enhanced antitumor properties, were enriched in BATF3-expressing NKTs (27.6% vs 8.2%, p < 0.05, versus GFP control). Further, BATF3 significantly enhanced NKT proliferation (area under the curve of absolute NKT cell counts: 5x1011 vs 1.1x1011, p = 0.001; fold expansion: 6.12x104 vs 1.2x104, p = 0.001, versus IL15 NKTs). Finally, BATF3.GPC3-CAR NKTs induced long-term tumor control and significantly improved survival of HCC xenograft-bearing mice compared to 15.GPC3-CAR NKTs (p = 0.0002, Mantel-Cox analysis). Conclusions: BATF3.GPC3-CAR NKTs induce robust antitumor activity in pre-clinical models of HCC. These results provide a foundation for evaluating BATF3.GPC3-CAR NKTs in patients with relapsed/refractory HCC.