Gastric cancer (GC) and pancreatic adenocarcinoma (PAAD) are highly lethal digestive cancers with poor prognoses. Current treatments often face challenges with drug resistance, recurrence, and metastasis. Adoptive cell transfer (ACT) using chimeric antigen receptor (CAR)-modified T cells (CAR-T cells) offers a potential new approach, particularly given its success in treating blood cancers. Claudin 18.2 (CLDN18.2), a stomach-specific isoform of claudin-18, is a promising target for GC and PAAD due to its prevalent expression in these cancers and limited expression in normal tissues. The clinical success of the CLDN18.2-targeting monoclonal antibody zolbetuximab further validates this target. CLDN18.2-targeting CAR-T cells are under development, but current limitations include short-term response durations. Armoring strategies, incorporating cytokines or chemokines, are being explored to enhance CAR-T cell efficacy in solid tumors. IL-15, a pleiotropic cytokine boosting T and NK cell activity, has shown promise in improving CAR-T cell antitumor activity across various tumor models. This study aimed to develop and characterize CLDN18.2-targeting mouse CAR-T cells with and without IL-15 armoring, evaluating their in vitro and in vivo antitumor activities in immunocompetent mouse models.

The literature review highlights the unmet need for effective treatments for GC and PAAD due to the challenges posed by drug resistance and metastasis. The review establishes CLDN18.2 as a promising therapeutic target, citing the success of CLDN18.2-targeting monoclonal antibodies in clinical trials. Existing CLDN18.2-targeting CAR-T products, while showing promise, suffer from limited duration of response. Previous studies have demonstrated the potential of cytokine armoring strategies, particularly using IL-7, CCL21, and IL-15, to enhance CAR-T cell efficacy. The use of IL-15 armoring has shown efficacy in various cancer models, improving both expansion and antitumor activity of CAR-T cells. This work builds on this existing literature by specifically investigating the effects of IL-15 armoring on CLDN18.2-targeting CAR-T cells within an immunocompetent mouse model.

The study involved the generation of second-generation mouse anti-CLDN18.2 CAR-T cells with and without IL-15 expression (H9 CAR and H9 CAR-IL15, respectively). Murine splenic CD3+ T cells were isolated and activated before transduction with retroviruses encoding the CAR constructs. In vitro characterization included assessment of CAR expression, fold expansion, viability, cytokine production (IFN-γ, TNF-α, IL-2), and cytotoxicity against CLDN18.2-expressing and CLDN18.1-expressing tumor cells (Panc02 and B16F10 cell lines). Cytotoxicity was measured using luciferase-based assays and a real-time impedance-based xCELLigence system. In vivo studies used subcutaneous Panc02-claudin 18.2 pancreatic tumor and B16F10-claudin 18.2 melanoma models in immunocompetent C57BL/6 mice. Mice received CAR-T cell infusions with or without cyclophosphamide (CPA) lymphodepletion pretreatment. Tumor growth was monitored, and tumor growth inhibition (TGI) calculated. In vivo CAR-T cell expansion and distribution were assessed via flow cytometry of peripheral blood and spleen samples. Immunohistochemical (IHC) analysis of tumor tissues assessed CD3 (T-cell infiltration), CD31 (angiogenesis), and claudin 18 expression. Statistical analysis utilized t-tests and ANOVA with Bonferroni's correction where appropriate.

H9 CAR-IL15 T cells demonstrated significantly greater in vitro fold expansion and viability compared to H9 CAR T cells. This was associated with a less differentiated phenotype, characterized by higher percentages of central memory T cells (TCM) and lower expression of exhaustion markers (TIM-3, TIGIT). Both H9 CAR and H9 CAR-IL15 T cells showed antigen-specific cytokine production and cytotoxicity in vitro, with only a slight advantage for H9 CAR-IL15 cells in a serial dynamic killing assay. In vivo, H9 CAR-IL15 T cells showed superior antitumor efficacy compared to H9 CAR T cells in both pancreatic and melanoma models, particularly with CPA lymphodepleting preconditioning. Even without CPA, H9 CAR-IL15 T cells showed noticeable tumor suppression. Improved antitumor activity was linked to greater in vivo expansion of CAR-T cells, increased T-cell infiltration into tumors, and surprisingly, increased angiogenesis (CD31+ cells) in Panc02 tumors. Interestingly, tumors treated with H9 CAR-IL15 T cells showed a loss of CLDN18.2 expression, while those treated with H9 CAR T cells retained target antigen expression. No significant systemic toxicity was observed in mice treated with H9 CAR-IL15 T cells.

The findings demonstrate that IL-15 armoring significantly enhances the antitumor activity of CLDN18.2-targeting CAR-T cells in immunocompetent mouse models. The improved efficacy is likely multifactorial, involving enhanced in vivo expansion and improved tumor infiltration. The observed increase in tumor angiogenesis in the early stages of treatment with H9 CAR-IL15 T cells is noteworthy and warrants further investigation. While the mechanism for this remains unclear, it may be beneficial for CAR-T cell trafficking and efficacy. The observation of tumor recurrence in some mice treated with H9 CAR-IL15 T cells, despite the initial response, suggests the potential for antigen-loss escape mechanisms. The lack of a significant host antitumor immune response independent of the targeted antigen indicates that further optimization of the treatment regimens or incorporation of additional modifications might be necessary to achieve long-lasting antitumor efficacy. The absence of significant toxicity in the mouse models supports the potential clinical translation of this approach. However, caution is advised due to previous reports of cytokine release syndrome in humans treated with similar constructs.

IL-15 armoring significantly improves the antitumor efficacy of CLDN18.2-targeting CAR-T cells in immunocompetent mouse models, primarily through enhanced in vivo expansion and tumor infiltration. Increased angiogenesis was observed, though its role requires further study. Future work should investigate the detailed mechanisms underlying the enhanced efficacy and the potential for long-term protection. Clinical translation of this approach warrants careful consideration of potential toxicities.

The study used syngeneic mouse models, which may not fully replicate the complexity of human cancers and the human immune response. The observed increase in tumor angiogenesis, while potentially beneficial, requires further investigation to clarify its role in the observed therapeutic effect. The study's duration may not capture long-term outcomes, and larger, longer-term studies may be necessary. The use of CPA lymphodepletion pretreatment in some experiments limits the generalizability of results to settings without such preconditioning.