Prediction of survival and immunotherapy response by the combined classifier of G protein-coupled receptors and tumor microenvironment in melanoma

Abstract

Background Melanoma is the deadliest form of skin tumor, and G protein-coupled receptors (GPCRs) play crucial roles in its carcinogenesis. Furthermore, the tumor microenvironment (TME) affects the overall survival (OS) and the response to immunotherapy. The combination of GPCRs and TME from a multi-omics perspective may help to predict the survival of the melanoma patients and their response to immunotherapy. Methods Bulk-seq, single-cell RNA sequencing (scRNA-seq), gene mutations, immunotherapy responses, and clinicopathologic feature data were downloaded from public databases, and prognostic GPCRs and immune cells were screened using multiple machine learning algorithms. The expression levels of GPCRs were detected using real-time quantitative polymerase chain reaction (qPCR) in A375 and HaCaT cell lines. The GPCR–TME classifier was constructed and verified using different cohorts and multi-omics. Gene set enrichment analysis (GSEA), weighted gene co-expression network analysis (WGCNA), and tracking tumor immunophenotype (TIP) were used to identify the key biological pathways among the GPCR–TME subgroups. Then, tumor mutational burden (TMB), vital mutant genes, antigen presentation genes, and immune checkpoints were compared among the subgroups. Finally, the differences in immunotherapy response rates among the GPCR–TME subgroups were investigated. Results A total of 12 GPCRs and five immune cell types were screened to establish the GPCR–TME classifier. No significant differences in the expression levels of the 12 GPCRs were found in the two cell lines. Patients with high GPCR score or low TME score had a poor OS; thus, the GPCRlow/TMEhigh subgroup had the most favorable OS. The scRNA-seq result revealed that immune cells had a higher GPCR score than tumor and stromal cells. The GPCR–TME classifier acted as an independent prognostic factor for melanoma. GSEA, WGCNA, and TIP demonstrated that the GPCRlow/TMEhigh subgroup was related to the activation and recruitment of anti-tumor immune cells and the positive regulation of the immune response. From a genomic perspective, the GPCRlow/TMEhigh subgroup had higher TMB, and different mutant genes. Ultimately, higher expression levels of antigen presentation genes and immune checkpoints were observed in the GPCRlow/TMEhigh subgroup, and the melanoma immunotherapy cohorts confirmed that the response rate was highest in the GPCRlow/TMEhigh cohort. Conclusions We have developed a GPCR–TME classifier that could predict the OS and immunotherapy response of patients with melanoma highly effectively based on multi-omics analysis.

Comprehensive single-cell transcriptomic profiling reveals molecular subtypes and prognostic biomarkers with implications for targeted therapy in esophageal squamous cell carcinoma

“…In summation, the culmination of our endeavors casts forth a compelling proposition: PRKCB might indeed play a pivotal role in expediting ESCC progression by potentiating the migratory capabilities of these cells. In addition to enhancing ESCC cell migration, PRKCB may also modulate DNA replication dynamics and replication stress responses based on its characterized activities in other cancer types [ 17 , 32 ]. Further studies are warranted to investigate the mechanistic links between PRKCB expression, DNA replication regulation, and ESCC progression.…”

“…The indicative genes for B cells include BANK1, CD79A, TCL1A, MS4A1, and CD79B; for T cells, CD3D, CD3G, and CD2; for Macrophage cells, C1QA, CD14, CD68, and CSF1R; for Fibroblast cells, BGN, AEBP1, and MMP2; for Smooth Muscle cells, TAGLN, and CNN1; for Mast cells, ADIRF, CD9, CLU, LMNA, DDX1, and RAB34; for Epithelial cells, KRT18, KRT8, and WFDC2; and for Endothelial cells, IFIT1. In reference to the meticulously annotated eight distinct cellular subgroups, we employed the "FindAllMarkers" function to compute differentially expressed genes across all cellular clusters [17] . We employed a stringent criterion, with |log2FoldChange| > 0.25 and P.Value < 0.05, to filter genes for subsequent investigation as single-cell differentially expressed genes (scDEGs).…”

Decoding the metastatic potential and optimal postoperative adjuvant therapy of melanoma based on metastasis score

“…Univariate Cox regression analysis with a bootstrap algorithm, least absolute shrinkage and selection operator (LASSO) regression analysis, and multivariate Cox regression analysis with a bootstrap algorithm were performed to screen the potential metastatic-related genes that were highly correlated with OS. The bootstrap coefficient of the included potential metastatic-related gene was defined using the formula: bootstrap coefficient = [ 66 ]. The MET score was obtained using the following formula: MET score = …”

Computational Characterization of Membrane Proteins as Anticancer Targets: Current Challenges and Opportunities

“…Beyond cancer types, GPCR expression signatures extracted with ML models have also been shown to allow head and neck cancer patient stratification into subtypes leading to differential sensitivity to immunotherapy [88]. A similar approach enabled the classification of melanoma patients based on survival and response to immunotherapy based on combined GPCR-TME multi-omics data [89]. These applications have the high potential to define GPCRs as immune biomarkers to help in cancer treatment and patient stratification.…”