First-in-human CAR T for solid tumors targets the MUC1 transmembrane cleavage product

Background & Aim We developed a MUC1* targeting CAR T that is scheduled for a 1st-in-human clinical trial for metastatic breast cancers at the Fred Hutchinson Cancer Research Center in Q1 2019. huMNC2-CAR44 targets MUC1*, which is the transmembrane cleavage product of MUC1. MUC1* is a growth factor receptor that is activated when onco-embryonic growth factor NME7AB dimerizes its truncated extra cellular domain. The binding site for NME7AB is masked in full-length MUC1, as is the binding site for the huMNC2 antibody. Thus, huMNC2 does not bind to full-length MUC1, which is expressed on healthy epithelium. Previous attempts at a MUC1 targeting therapeutic all failed, presumably because they targeted the tandem repeat domain of full-length MUC1, which is shed after cleavage to MUC1*. Methods, Results & Conclusion huMNC2-CAR44 T cells are highly selective for cancerous tissues: first, because they target MUC1*, not the healthy full-length form. Secondly, the huMNC2 antibody recognizes a specific conformation of the MUC1* extra cellular domain that is created when MUC1 is cleaved by an enzyme that is overexpressed in many cancers, especially breast cancers. IHC studies of thousands of normal vs. cancerous human tissue specimens show that huMNC2-scFv almost exclusively binds to tumor tissues, hitting over 90% of breast, 83% ovarian, 78% pancreatic and 71% of lung cancers. Recognition of breast cancer specimens appears not to be limited by cancer sub-type. In vivo experiments of human tumors in NSG mice (n>300), show that huMNC2-CAR44 T cells robustly inhibited MUC1* positive solid tumors. A single injection of huMNC2-CAR44 T cells eliminated tumors in NSG mice which remained tumor free to Day 100, compared to control animals that had to be sacrificed at Day 20 due to excess tumor volume. Other animal experiments showed that huMNC2-CAR44 T cell efficacy increased with increasing antigen density. Dual tumor experiments of a high antigen density and low antigen density tumor in the same animal showed that adequate MUC1* density is required for a CAR T response, further supporting the idea that huMNC2-CAR44 T cells will selectively kill MUC1* positive tumors, while sparing normal tissues. The generation of an anti-MUC1* CAR that recognizes a conformational epitope created when MUC1 is cleaved to MUC1* by a specific cleavage enzyme, argues that CARs could be made patient specific, based on which cleavage enzymes their tumors express.