The role of the CD58:CD2 axis in cancer immune evasion and resistance to immunotherapy

PROJECT SUMMARY Immune checkpoint inhibitors (ICI), such as anti-PD-1 antibodies (aPD1), enable CD8+ T cell-mediated anti- tumor immunity and produce durable responses in a portion of patients with metastatic melanoma. However, most patients do not benefit from ICI and the underlying mechanisms of resistance remain elusive. The overall goal of this work is to elucidate mechanisms of resistance to ICI mediated by cancer cells, and develop novel molecular or immune-based therapies to overcome resistance. Prior work demonstrated that expression of co- stimulatory protein CD28 on CD8+ T cells is required for efficacy of aPD1, however, we show that expression of CD28 is frequently absent in tumor-infiltrating CD8+ T lymphocytes (TILs), including those with preserved progenitor function. In contrast, all TILs maintain expression of CD2, the most potent alternative co-stimulatory protein. Indeed, we find that patient-derived CD28- TILs had preserved capacity for proliferation, clonal expansion, cytokine production and lysis of autologous cancer cells. The ligand of CD2 is CD58, and we discovered that CD58 was expressed on melanoma cells. The role of the CD58:CD2 axis in anti-tumor immunity, and response and resistance to ICI is unexplored. In preliminary studies, we established patient-derived pairs of melanoma cells and autologous TILs. We developed Perturb-CITE-seq, a novel method enabling pooled CRISPR-Cas9 screens with single-cell RNA-seq and protein readouts. We performed a Perturb-CITE-seq screen in patient models, perturbing hundreds of genes with putative roles in ICI resistance, followed by co-culture with TILs and profiled survivor cells that escaped T cell killing. We recovered known mechanisms of cancer immune evasion and novel ones, including loss/downregulation of CD58. In humanized mouse models, loss of CD58 in tumor cells resulted in impaired T cell infiltration and resistance to adoptive cell transfer (ACT). Additionally, we observed compensatory PD-L1 upregulation in CD58 loss. Using genome-scale CRISPR-Cas9 screens and co- immunoprecipitation mass spectrometry screens, we identified a putative regulator of this interaction, CMTM6. Rescue of CD58 re-sensitized cancer cells to T cell killing, increased T cell activation, enhanced T cell infiltration in vivo, and abolished compensatory PD-L1 expression. Based on these findings, we will dissect the role of the CD58:CD2 axis in anti-tumor immunity, cancer immune evasion and resistance ICI resistance in three aims: i) to determine the mechanisms by which CD58 loss confers cancer immune evasion, ii) to determine the role of CMTM6 in regulating compensatory PD-L1 upregulation in CD58 loss, and iii) to determine whether aPD1 resistance is associated with an impaired CD58:CD2 axis in patients. We propose to systematically address these using genetic, biochemical, patient and in vivo models, and multi-modal single-cell genomics analyzed with novel machine-learning tools. We further test two therapies that overcome resistance to ICI due to defects in the CD58:CD2 axis. Upon completion, this work will shed light on the role of the CD58:CD2 pathway in cancer and cancer immunotherapy, and may provide novel avenues for therapeutic development.