PC-5 En masse discovery of human anti-cancer antibodies

Abstract

Generation of specific antibodies for cancer therapy is a major endeavor. As a radical departure from conventional approach, we hereby describe rapid and potentially en masse identification of cancer-specific antibodies directly from human cancer tissues. A computational framework was developed and successfully tested for antibody discovery by mining TCGA RNA-seq data from 1945 cases of solid tumor samples. Surprisingly, synthetic antibodies based on high-abundance CDR3 sequences from lung adenocarcinoma (LUAD) patients bound all lung cancer samples tested and cross-reactive to other cancer types but rarely to normal tissues. Targeted DNA sequencing of the B cell receptor from 5 lung tumortissues also allowed us to identify variable region sequences with somaticmutations. Antibodies based on predominant variable region sequences showed specific binding to autologous lung cancer samples. Our platform dramatically reduces barrier in developing human anti-cancer antibodies and paved the way for cancer treatment using patient-derived tumor-reactive monoclonal antibodies.