Optimizing TNFR2 antagonism for immunotherapy with tumor microenvironment specificity.

Abstract

Most approved cancer immunotherapies lack T-regulatory (Treg) or tumor specificity. TNF receptor 2 (TNFR2) antibody antagonism is emerging as an attractive immunotherapy due to its tumor microenvironment (TME) specificity. Here we show that the human TNFR2 receptor is overexpressed on both human tumor cells and on human tumor-residing Tregs, but negligibly expressed on beneficial T effectors (Teffs). Further, we found widespread, if variable, TNFR2 expression on 788 human tumor cell lines from diverse cancer tissues. These findings provided strong rationale for developing a targeted immunotherapy using a TNFR2 antibody antagonist. We designed a novel, human-directed TNFR2 antibody antagonist and tested it for function using three cell-based TME assays. The antagonist showed TME specificity by killing of TNFR2-expressing tumor cells and Tregs, but sparing Teffs, which proliferated. However, the antagonist shuffled between five isoforms, only one of which showed the desirable function. We designed and tested several new chimeric human versions of the antagonist, finding that the IgG2 isotype functioned better than the IgG1 isotype. To further improve function, we introduced targeted mutations to its amino acid sequence to stabilize the natural variability of the IgG2 isotype's hinge. Altogether, our findings suggest that optimal TNFR2 antagonists are of the human IgG2 isotype, have hinge stabilization, and have wide separation of antibody arms to bind to newly synthesized TNFR2 on rapidly growing tumor cells. Antagonistic antibodies with these characteristics, when bound to TNFR2, can form a nonsignaling cell surface dimer that functions with high TME specificity.