Gene Signature Reveals Decreased SOX10 Activity in Malignant Cells from Immunotherapy-Resistant Cuta

Timothy J. Purwin1, 2, Claudia Capparelli1, Ahmet Sacan2 and Andrew E. Aplin1, 3 1 Department of Pharmacology, Physiology, and Cancer Biology, 3 Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA 19107, USA. 2 School of Biomedical Engineering, Science and Health Systems, Drexel University, Philadelphia, PA 19104, USA.

Immunotherapies and targeted therapies have advanced the treatment options for cutaneous melanoma but resistance and relapse still occur. Evidence is mounting that second-line therapy is less effective compared to the same therapy as first-line in treatment naïve tumors; however, details of cross-resistance mechanisms are limited. Here, we used computational methods to determine the extent that malignant cells lacking SOX10 activity, a known resistance mechanism to MAPK pathway inhibitors, are associated with immunotherapy-resistance. Using publicly available datasets, we integrated gene and transcript annotation, SOX10 ChIP-seq, CRISPR knockout RNA-seq, and knockdown ATAC-seq data from melanoma cell models to develop a robust SOX10 gene signature. We used gene signature scoring and enrichment analysis methods to validate the ability of the signature to determine SOX10 activity in multiple independent single cell and bulk RNA-seq SOX10 knockdown, melanoma cell line panel, and BRAFi -/+ MEKi drug-resistant datasets. Evaluation of patient scRNA-seq data revealed that immunotherapy-resistant patient tumors have lower SOX10 activity than treatment-naïve tumors. SOX10 IHC staining confirmed a higher prevalence of SOX10-deficient melanoma cells in tumor samples following immunotherapy. These data suggest that SOX10-deficient melanoma cells are associated with cross-resistance and highlight the need to identify therapeutic strategies that target this subpopulation.