Combining single-cell and bulk RNA sequencing to identify CAF-related signature for prognostic prediction and treatment response in patients with melanoma

Abstract Cancer-associated fibroblasts (CAFs) play complex roles in the tumor microenvironment (TME) of melanoma. However, their impact on prognosis and treatment response in melanoma remains unclear. In this study, ScRNA-seq data (GSE115978) were utilized to characterize CAF heterogeneity and identify marker genes in melanoma. Prognostic CAF genes were identified from the TCGA dataset and employed to construct a risk signature, which was subsequently validated in an independent cohort (GSE65904). Mutation, copy number variation (CNV), pathway enrichment, immune infiltration, and drug sensitivity were analyzed to determine the signature’s clinical relevance. Immunohistochemistry (IHC), immunofluorescence (IF), and qPCR were performed to validate the expression of CAF signatures on clinical melanoma samples. We identified CAFs in patients with melanoma through single-cell RNA sequencing data. A 28-gene CAF signature was constructed using the Least Absolute Shrinkage and Selection Operator (LASSO) regression based on 271 prognostic CAF genes. This signature demonstrated excellent prediction accuracy for survival, with area under the curve (AUC) values of 0.737, 0.737, and 0.779 for 1-year, 3-year, and 5-year survival, respectively. The signature was an independent prognostic factor and was correlated with CNVs, and immunosuppressive TME features (reduced CD8 + T cells, M1 macrophages). Additionally, our CAF signature could predict the efficacy of multiple chemotherapy drugs and serve as a potential prognostic marker for immunotherapy. Experimental validation confirmed the expression of CAF signature genes in melanoma tissue. Our model may help predict the prognosis and response to chemotherapy and immunotherapy in patients diagnosed with melanoma.