Background Application of chimeric antigen receptor (CAR) T cell therapies to solid cancers has been limited by poor efficacy and toxicities. Because CAR T cells cannot distinguish between healthy and malignant cells, they attack and damage non-tumor sites that express their target leading to morbidity and mortality.^1–3 Development of strategies to limit toxicity will require animal models that faithfully recapitulate CAR T cell toxicity in healthy tissues. Here we present a model focused on the tumor associated antigen CEA (i.e. CEACAM5), which is abundantly expressed by the majority of colorectal and pancreas cancers, but is also present in healthy lung and colon tissue.

Methods We utilized a model with expression of human CEACAM5 (hCEA) in a physiologic distribution in C57Bl/6 mice and generated a murine CAR against human CEACAM5. To evaluate the in vivo effects of our CAR in models of pancreatic cancer, we utilized the KPC implantable model, which we transduced to stably express human CEACAM5. KPC-hCEA cells were implanted into the pancreas or liver of hCEA transgenic or wildtype mice. Five days later mice received 1–3*10⁶ CD8⁺ CAR T cells or control transduced T cells. Outcomes included change in body mass, tumor size, distribution of CAR T cells, and histological analyses.

Results Expression of human CEACAM5 was confirmed in lung and colon of hCEA transgenic mice, but not wildtype littermates nor in tissues such as liver, pancreas, or spleen. Pancreas and metastatic liver tumors both responded to treatment with CAR T cells with significant reduction in tumor size (p<0.05 n=5–6/group). However, the treatment effect was limited due to toxicity. Dose-limiting toxicity, including lethality, was observed in the hCEA transgenic mice, but not wildtype mice. Flow cytometry revealed specific accumulation of CAR T cells in organs that expressed the CAR target, including tumor, lung, and colon. The accumulation of CAR T cells in lung and colon corresponded with histologically defined tissue destruction. Organs that did not harbor CEA expression were not damaged. We identified additional biologic features, including trogocytosis of the CAR target and a role for apical sequestration of the CAR target as a mechanism impacting toxicity.

Conclusions Our results demonstrate that the hCEA transgenic mouse coupled with KPC-hCEA tumor cells and CEA-targeting CAR T cells captures on-target, off-tumor toxicity of CAR T cells. Because this system replicates the clinical reality, it allows for development of approaches to mitigate on-target off-tumor toxicity applicable to the clinic.

References

1. Morgan RA, Yang JC, Kitano M, et al. Case report of a serious adverse event following the administration of T cells transduced with a chimeric antigen receptor recognizing ERBB 2. Mol Ther. 2010;18:843–51.

2. Thistlethwaite FC, Gilham DE, Guest RD, et al. The clinical efficacy of first-generation carcinoembryonic antigen (CEACAM5)-specific CAR T cells is limited by poor persistence and transient pre-conditioning-dependent respiratory toxicity. Cancer Immunol Immunother. 2017;66:1425–36.

3. Parkhurst MR, Yang JC, Langan RC, et al. T cells targeting carcinoembryonic antigen can mediate regression of metastatic colorectal cancer but induce severe transient colitis. Mol Ther. 2011;19:620–6.

Ethics Approval All animal experiments were approved by OHSU’s Institutional Animal Care and Use Committee on 7/21/2023; approval number TR01_IP00003155.