Specifically Targeting the Interface Between HER1-HER3 Heterodimer on Breast Cancer to Limit Off-Target Effects Using Chimeric Antigen Receptor Designs with Improved T-Cell Energy Balance

Human epidermal growth factor receptor (EGFR) family consists of four members i.e. EGFR (HER1), HER2 (ErbB2), HER3 (ErbB3,) and HER4 (ErbB4). Overexpression, mutation, or catalytic activation of these proteins can lead to malignancies in breast, ovarian, colorectal, pancreatic and lung. Therapies targeting EGFR-associated proteins to disrupt signaling may fail because of crosstalk within the EGFR family or among downstream pathways. One mechanism of escape is HER3 activation and concomitant heterodimer formation with HER1 causing disease relapse and treatment failure. A bi-specific monoclonal antibody (mAb, MEHD7945A) can specifically bind an epitope shared between HER1-HER3 heterodimer thereby blocking EGFR-HER3 mediated signaling (Schaefer et al., Cancer Cell, 2011). We now report that the specificity of this mAb can be used to redirect the specificity of T cells through enforced expression of a chimeric antigen receptor (CAR) targeting the HER1-HER3 heterodimer, such as expressed on breast cancer cells. A 2^nd generation CAR targeting the HER1-HER3 heterodimer was expressed from DNA plasmid constituting scFv (designated DL11f, derived from mAb MEHD7945A) coupled to CD3-zeta fused in frame with chimeric CD28 or CD137 T-cell signaling domains on a clinical-grade Sleeping Beauty (SB) backbone. T cells were electroporated with SB system and numerically expanded on irradiated “universal” activating and propagating cells (uAaPC) (Rushworth et al., J Immunotherapy, 2014). These feeder cells are derived from K-562 cells engineered to co-express a CAR activating ligand (CAR-L, a scFV specific to CAR stalk) to sustain proliferation of genetically modified T cells. We validated CAR expression on genetically modified T cells by flow cytometry and western blot. The specificity of HER1-HER3 specific CAR T cells was confirmed in situ by a proximity ligation-based assay using breast cancer cells. The redirected killing by CAR⁺ T cells to HER1⁺HER3⁺ breast cancer cells was confirmed in vitro and its efficacy evaluated in vivo in NSG mice bearing a breast tumor xenograft. HER1-HER3 specific CAR⁺ T cells activated via CD137 signaling exhibited superior proliferation compared with T cells expressing CAR with CD28 signaling domain. This is consistent with the ability of CD3-zeta/CD137 endodmain to alter mitochondrial metabolism and to suppress apoptosis leading to proliferation after initial activation. In summary, we report a new CAR design that can interrogate the conformation between two tumor-associated antigens (TAAs). This will likely improve specificity and limit on-target off-tissue side effects compared to CARs targeting only HER-1 or HER-3. Thus, targeting an epitope derived from two TAAs may help distinguish normal cells versus malignant cells and treat HER1⁺HER3⁺ malignancies that are resistant to therapies targeting single EGFR family members. These data have immediate translation appeal for targeting solid tumors as we use the SB and AaPC platforms to manufacture CAR⁺ T cells in our clinical trials.