TARGETING SHARED CANCER NEOANTIGENS WITH BISPECIFIC T CELL-RETARGETING ANTIBODIES

Identification of therapeutic targets highly specific to cancer cells is one of the greatest challenges toward developing an effective cancer therapy. Somatic mutations in cancers are ideal therapeutic targets as they are expressed only in tumor cells and not normal cells. Most of these mutant proteins reside intracellularly and thus efforts to target them have focused on small molecules. However, following protein proteolysis, a fraction of the peptides derived from these mutant proteins is loaded onto human leukocyte antigen (HLA), potentially generating a mutation-containing peptide-HLA (pHLA) complex called mutation-associated neoantigens (MANAs). pHLA complexes are the natural ligands of T cell receptors (TCRs). In the series of studies described herein, we exploited the high affinity and specificity nature of antibodies and targeted these MANAs using TCR-mimic antibodies called MANAbodies. Starting from a diverse phage display library, we identified antibodies that specifically target MANAs derived from common driver mutations in oncogene RAS and tumor suppressor gene TP53. To convert these antibodies into potent immunotherapeutic agents, we linked the antibodies with an anti-CD3 antibody, generating T cell-retargeting bispecific antibodies. These bispecific antibodies were capable of activating T cells in the presence of target cells presenting extremely low antigen levels. They killed tumor cells in vitro and in mice in a fashion exquisitely dependent on the presence of the target MANAs. To screen for cross-reactivity against other pHLA complexes, we systemically interrogated mimics of the target peptide and identified putative cross-reactive peptides. The structural basis of the specificity of MANAbody was also delineated by crystallography. This body of work demonstrated that the MANAbody platform could be a powerful approach to targeting cancers containing mutations that are difficult to target in conventional ways