In vivo CRISPR screening identifies Ptpn2 as a target for cancer immunotherapy
Robert T Manguso1,3, Hans W Pope1,3, Margaret D Zimmer1,3, Flavian D Brown1, Kathleen B Yates1,3, Brian C Miller1, Natalie B Collins1, Kevin Bi1, Martin W LaFleur1, Vikram R Juneja2, Sarah A Weiss1, David E Root3, Arlene H Sharpe2, John G Doench3, & W. Nicholas Haining1,3*
1Department of Pediatric Oncology, Dana-Farber Cancer Institute, Boston, MA
2Department of Microbiology and Immunology, Harvard Medical School, Boston, MA
3Broad Institute of Harvard and MIT, Cambridge, MA
Despite the dramatic clinical success of cancer immunotherapy with PD-1 checkpoint blockade, most patients don’t experience sustained clinical benefit, suggesting that additional therapeutic strategies are needed. Functional genomic screens in cancer cells to discover new therapeutic targets are usually carried out in vitro where interaction with the immune system is absent. Here we report a pooled, loss-of-function genetic screening approach using CRISPR/Cas9 genome editing that is conducted in vivo in mouse transplantable tumors treated with vaccination and PD-1 checkpoint blockade. We tested 2,400 genes expressed by melanoma cells for those that synergize with or cause resistance to checkpoint blockade, and recovered the known immune evasion molecules, PD-L1 and CD47. Loss of function of multiple genes required to sense interferon-y caused resistance to immunotherapy. Deletion of Ptpn2, a pleotropic protein tyrosine phosphatase improved response to immunotherapy. In vivo, Ptpn2 deficient tumors showed increased infiltration of activated CD8+T cells. In vitro, Ptpn2 loss by tumor cells increased antigen presentation to T cells. Biochemical, transcriptional and genetic epistasis experiments demonstrated that loss of function of Ptpn2 sensitizes tumors to immunotherapy by enhancing interferon-y-mediated effects on the tumor cell. Thus, augmenting interferon-y signaling in tumor cells could increase the efficacy of immunotherapy. More generally, in vivo genetic screens in tumor models can identify new immunotherapy targets and rationally prioritize combination therapies.