Molecular and immune landscape of melanoma: a risk stratification model for precision oncology

Abstract

Background: Melanoma is a highly aggressive skin cancer with significant heterogeneity in immune infiltration and clinical outcomes. Accurate risk stratification is essential for improving personalized treatment strategies.

Methods: This study utilized data from The Cancer Genome Atlas (TCGA) to explore immune-related gene expression in melanoma. Single-sample gene set enrichment analysis (ssGSEA) was employed to classify patients into high and low immune groups. Tumor microenvironment (TME) characteristics, including immune cell infiltration, HLA gene expression, and TME scores, were analyzed. Prognostic genes were identified using univariate and multivariate Cox regression analyses. A risk score model and nomogram were constructed, and gene set enrichment analysis (GSEA) and gene set variation analysis (GSVA) were conducted to explore relevant signaling pathways.

Results: ssGSEA-based classification revealed significant differences between high and low immune groups in terms of immune infiltration and HLA gene expression. The risk model incorporated immune-related genes such as GBP2, SEMA4D, and KIR2DL4, which demonstrated distinct tumor expression profiles and strong prognostic value. GSEA and GSVA analyses uncovered critical immune-related and oncogenic pathways linked to risk stratification. A predictive nomogram integrating molecular risk scores and clinical variables improved prognostic accuracy. Computational immune deconvolution highlighted CD8+ T cell infiltration as a key prognostic factor. To validate the functional role of candidate genes, KIR2DL4 was silenced in A375 melanoma cells using shRNA. Knockdown efficiency was confirmed by qRT-PCR. Functional assays revealed that KIR2DL4 silencing significantly reduced cell proliferation, as assessed by MTT assay, and impaired migratory capacity, as demonstrated by wound healing assay. These in vitro findings support the computational predictions and suggest that KIR2DL4 may play a tumor-promoting role in melanoma.

Conclusion: This study provides a robust immune-related prognostic model for melanoma. It underscores the value of immune gene expression and T cell infiltration, particularly CD8+ T cells, in predicting patient outcomes. The model facilitates personalized treatment decisions and advances precision oncology approaches in melanoma. The integration of transcriptomic analysis with experimental validation confirms the tumor-promoting role of KIR2DL4 and enhances the translational value of the model in guiding precision immunotherapy.

Keywords: Gene expression profiling; Immune infiltration; Melanoma stratification; Prognostic biomarkers; TCGA data analysis.

Fig. 1

Immune profile classification and tumor microenvironment analysis in melanoma. A t-SNE plot demonstrating the segregation of melanoma samples into high (Immunity_H) and low (Immunity_L) immune response groups based on ssGSEA scores. B Heatmap showing the expression patterns of all genes and immune-related genes across the two immune groups. C Box plots comparing tumor microenvironment (TME) scores—including stromal score, immune score, and ESTIMATE score—between Immunity_H and Immunity_L groups. D Box plots showing differential expression of HLA-related genes across immune subgroups. E Box plots illustrating the relative abundance of various immune cell types estimated by deconvolution algorithms across the two immune groups. Statistical significance was determined using the Wilcoxon rank-sum test

Fig. 2

Identification of differentially expressed immune-related genes and development of a prognostic model. A Volcano plot displaying differentially expressed genes (DEGs) between high and low immune groups (adjusted p-value < 0.05, |log2FC|> 1). B Venn diagram showing the intersection between DEGs and immune-related gene sets. C–E LASSO regression analysis for feature selection of prognostic immune-related genes. F Nomogram integrating the immune-based risk score and clinical variables to predict 1-, 3-, and 5-year overall survival in melanoma patients

Fig. 3

Prognostic validation of the risk model and its association with clinical features. A Forest plot of univariate Cox regression analysis for candidate genes. B Forest plot of multivariate Cox regression analysis showing independent prognostic genes incorporated into the model. C Time-dependent ROC curves evaluating the predictive performance of the risk model at 1-, 3-, and 5-year overall survival. D–G Bar plots illustrating the distribution of risk scores across clinicopathological subgroups, including age, gender, AJCC stage, and tumor subtype

Fig. 4

Correlation between immune cell infiltration and risk score. A Bubble plot showing the correlation between estimated immune cell populations and risk scores using multiple immune deconvolution algorithms (e.g., TIMER, CIBERSORT, xCell). B–G Scatter plots presenting the correlation between specific immune cell types (including CD8⁺ T cells, macrophages, and dendritic cells) and the calculated risk score. Spearman correlation coefficients and p-values are indicated

Fig. 5

Pathway enrichment analysis associated with risk stratification. A–C GSEA enrichment plots of representative immune- and tumor-related pathways in high-risk versus low-risk groups. D, E Heatmaps displaying GSVA scores of selected pathways across samples, highlighting distinct pathway activity between risk subgroups

Fig. 6

Gene expression and survival analysis of key prognostic biomarkers. A Violin plots comparing the expression levels of GBP2, SEMA4D, and KIR2DL4 in normal versus melanoma tissues. B–D Kaplan–Meier survival curves showing the association between high versus low expression of each gene and patient overall survival. E–G Violin plots illustrating the relationship between gene expression levels and clinical TNM staging (T, N, and M), revealing progressive upregulation with advanced pathological stages

Fig. 7

Knockdown of KIR2DL4 inhibits proliferation and migration of A375 melanoma cells. A Quantitative real-time PCR analysis confirmed the efficient knockdown of KIR2DL4 mRNA in A375 cells transduced with shKIR2DL4 compared to negative control (shNC). p < 0.01. B Cell proliferation was assessed using the MTT assay at 0, 24, 48, and 72 h post-transduction. Knockdown of KIR2DL4 significantly reduced the proliferative capacity of A375 cells compared to shNC. p < 0.01. C Wound healing assay showing cell migration at 0 and 24 h after scratching. The shKIR2DL4 group exhibited impaired wound closure compared to the shNC group, indicating reduced migratory ability