Histone lysine methylation patterns in prostate cancer microenvironment infiltration: Integrated bioinformatic analysis and histological validation.

Background: Epigenetic reprogramming through dysregulated histone lysine methylation (HLM) plays a crucial role in prostate cancer (PCa) progression. This study aimed to comprehensively evaluate HLM modification patterns in PCa microenvironment infiltration. Materials and methods: Ninety-one HLM regulators in The Cancer Genome Atlas (TCGA) dataset were analyzed using bioinformatics. Differentially expressed genes (DEGs) and survival analyses were performed using TCGA-PRAD clinicopathologic and follow-up information. Consensus clustering analysis divided patients into subgroups. Gene ontology (GO) function and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed on the DEGs. Tumor mutation burden (TMB) and tumor microenvironment (TME) cell infiltration were evaluated in different HLM clusters. Quantitative real-time PCR (qPCR) analysis assessed HLM regulators in clinical PCa tissues. Results: The tumor vs. normal (TN), Gleason score (GS) > 7 vs. GS < 7, pathological T stage (pT) = 2 vs. pT = 3, and TP53 mutation vs. wild-type comparisons using TCGA-PRAD dataset revealed 3 intersecting HLM regulators (EZH2, NSD2, and KMT5C) that were consistently upregulated in advanced PCa (GS > 7, pT3, HR > 1, and TP53 mutation) (P < 0.05) and verified in clinical PCa tissues. Consensus clustering analysis revealed three distinct HLM modification patterns (HLMclusters). However, no significant differences in recurrence-free survival (RFS) rates were found among the groups (P > 0.05). We screened 189 HLM phenotype-related genes that overlapped in the pairwise comparisons of HLMclusters and P < 0.01 in the Cox regression analysis. Three distinct subgroups (geneClusters) were revealed based on the 189 genes, in which cluster A involved the most advanced PCa (PSA > 10, T3-4, GS8-10, and biochemical recurrence) and the poorest RFS. The HLM score (HLMscore) was calculated by principal component analysis (PCA) of HLM phenotype-related genes that have positive predictive value for RFS (P < 0.001) and immune therapy responses (in the CTLA4-positive and -negative responses accompanied by a PD1-negative response). Conclusion: We comprehensively evaluated HLM regulators in the PCa microenvironment using TCGA-PRAD, revealing a nonnegligible role of HLM patterns in PCa complexity and heterogeneity. Elucidating the effects of HLM regulators in PCa may enhance prognostics, aggressiveness assessments, and immunotherapy strategies.