Development and Validation of a Novel Signature to Predict the Survival and Affect the Immune Microenvironment of Esophageal Squamous Cell Carcinoma: Epigenetic-Related Genes

Yani Su¹Ming Zhang²Qiong Zhang²Pengfei Wen²Ke Xu²Jiale Xie²Xianjie Wan²Lin Liu²Peng Xu²Zhi Yang²Mingyi Yang^2*

• ¹Air Force Medical University Tangdu Hospital, Xi'an, China
• ²Xi'an Honghui Hospital, Xi'an, China

Objective: Esophageal squamous cell carcinoma (ESCC) is a malignancy with well-documented epigenetic dysregulation. This study seeks to construct a robust prognostic model based on epigenetic-related genes (ERGs), while also elucidating potential mechanisms of immune microenvironment modulation. Methods: This study utilized transcriptomic data from TCGA as the training cohort, with GSE53625 from the GEO serving as the independent validation cohort. From the EpiFactors database, we curated 796 epigenetic regulator genes (ERGs). Differential expression analysis was performed to derive differentially expressed ERGs (DE-ERGs) in ESCC. Univariate Cox regression and LASSO regression were employed to construct a prognostic risk model, which was rigorously evaluated through risk stratification curves, survival status distribution maps, risk score heatmaps, survival analysis, ROC curves and multivariate Cox regression analyses. Subsequently, the differential analysis of risk prognostic model and clinical feature analysis were conducted. To elucidate the immune microenvironment, we conducted immune cell infiltration correlation analysis, single-sample gene set enrichment analysis (ssGSEA), and immune checkpoint profiling. Additionally, drug sensitivity analysis identified potential therapeutic agents with differential efficacy across risk subgroups. Finally, RT-qPCR validation confirmed the expression patterns of key prognostic ERGs. Results: Through comprehensive differential expression analysis, we identified 345 DE-ERGs in ESCC. Subsequent univariate Cox proportional hazards regression followed by LASSO penalized regression analysis yielded a robust prognostic signature comprising 13 critical ERGs: PIWIL4, SATB1, GSE1, NCOR1, BUB1, SAP30L, CHEK1, MASTL, ATM, BMI1, DNAJC2, UBE2D1, and SSRP1. External validation using independent GEO datasets confirmed the model's prognostic efficacy through multidimensional assessments. Immunological characterization revealed significant enrichment of CD8+_T_cells, DCs, and pDCs in high-risk patients, accompanied by elevated cytolytic_activity, HLA, and MHC_class_I. Furthermore, we identified three immune checkpoint molecules (TMIGD2, IDO1, and CD44) showing differential expression between risk groups. Drug sensitivity analysis highlighted four promising therapeutic compounds (PD-0325901, Bryostatin-1, 2 ATRA, and Roscovitine) exhibiting potential clinical utility for ESCC treatment. Experimental validation via RT-qPCR demonstrated consistent overexpression of GSE1, NCOR1, BUB1, CHEK1, UBE2D1 and SSRP1 in ESCC cell lines, while PIWIL4 and ATM showed significant downregulation. Conclusion: The findings of this study provide clinically relevant insights for prognostic stratification and immune microenvironment characterization in ESCC patients.