Abstract
Treatment of solid cancers with chimeric antigen receptor (CAR) T-cells is challenging because of a lack of target antigens that are both tumor-specific and homogenously expressed. While epidermal growth factor receptor (EGFR)vIII represents a glioblastoma (GBM)-specific antigen, its expression is heterogeneous within the tumor resulting in tumor escape. In contrast, more homogenously expressed GBM-associated antigens (GAA), such as EphA2, are non-ideal because of expression in other normal organs, yielding potential cross-reactive toxicity. As a way to safely target GAAs in the tumor without attacking normal cells expressing the same GAAs outside of the brain, we adapted a novel synthetic Notch (synNotch) receptor system and established a “prime and kill” sequential two-receptor CAR circuit. A synNotch receptor recognizes a specific priming antigen; the heterogeneous GBM neoantigen EGFRvIII or a brain tissue-specific antigen to prime the local expression of a CAR that mediates cytotoxicity against a GAA (e.g. EphA2). In orthotopic GBM6 glioma model, a patient-derived xenograft (PDX) with heterogeneous expression of EGFRvIII, intravenous infusion of T-cells transduced with EGFRvIII synNotch→anti-IL-13Rα2/EphA2 tandem CAR circuit resulted in long-term (over 100 days) survival and eradication of the heterogeneous tumor in all of 12 mice in two independent experiments. In contrast, constitutive CARs targeting EGFRvIII or IL-13Rα2/EphA2 (as a tandem CAR) failed to exhibit long-term anti-tumor response. Moreover, T-cells transduced with synNotch-regulated CAR maintain a less differentiated state which is associated with higher durability compared with ones with constitutive CAR in vivo. T-cells transduced with a synNotch→CAR circuit primed by a brain-specific antigen, myelin oligodendrocyte glycoprotein (MOG), exhibited a precise and potent local control of intracranial PDX without evidence of priming in extracranial organs. These data support the utility of synNotch→CAR circuits in EGFRvIII-negative GBM cases. By integrating multiple imperfect but complementary antigens, we improve both the specificity and persistence of T-cells directed against GBM.