B7H3 (CD276) is a promising target for overcoming the limited efficacy of CAR-T therapies in solid tumors. Overexpressed in tumor cells and vasculature, it suggests potential for dual targeting, making it crucial for enhancing the therapeutic window.

In this study, we developed B7H3-targeting CAR-T cells using novel antibodies that are cross-reactive with human and mouse B7H3. Conventional studies previously performed on human CAR-T cells often relied on human B7H3-specific antibodies, which precluded the assessment of on-target, off-tumor (OTOT) toxicities in host tissues. We aimed to simultaneously evaluate both efficacy and toxicity in a murine syngeneic model using our cross-reactive CAR-T cells. Subsequently, we sought to confirm whether human B7H3 CAR-T cells could effectively treat human tumors.

Our proprietary C12 platform—comprising a CTLA4-CD28 chimera (CTC28) and inducible IL-12—previously demonstrated that omitting lymphodepletion was essential for ensuring safety, under which condition potent anti-tumor efficacy was achieved. Consistently, mouse B7H3-C12 CAR-T cells successfully treated tumors without toxicity in non-lymphodepleted hosts. In contrast, lymphodepletion caused lethal toxicity by B7H3 CAR-T cells that was reversed in B7H3-KO mice, indicating that lymphodepletion precipitates OTOT toxicity. Similarly, human B7H3-C12 CAR-T cells induced lethal toxicity in innately lymphopenic NSG mice. Nonetheless, these human CAR-T cells efficiently eliminated human tumors in B7H3-KO NSG mice, confirming their efficacy when host-driven OTOT toxicity was excluded.

In conclusion, B7H3-C12 CAR-T cells exhibit safe and potent anti-tumor activity without lymphodepletion. This supports the clinical rationale of omitting pre-conditioning lymphodepletion to prevent severe OTOT toxicity in patients with solid tumors.