Bioinformatics analysis identifies ULK3 as a novel prognostic and immune-related biomarker in esophageal cancer

Esophageal cancer (EC) is a leading global malignancy, exhibiting high incidence in East Asia, especially China. Although tumor, node, metastasis (TNM) staging remains the cornerstone of clinical evaluation, molecular profiling of tumors and blood now offers complementary prognostic and therapeutic information. Yet patient survival rates remain unsatisfactory. Autophagy plays a dual role in tumor. ULK3, a kinase related to autophagy regulation, remains underexplored in EC. This study aimed to investigate the role of ULK3 in EC prognosis and tumor immunity.

ULK3 expression patterns were examined across pan-cancer and EC using The University of ALabama at Birmingham CANcer data analysis Portal (UALCAN) and Gene Expression Profiling Interactive Analysis (GEPIA2), with gene set cancer analysis (GSCA) characterizing its single nucleotide variation (SNV) and copy number variation (CNV) profiles. Gene Expression Omnibus (GEO) datasets validated these findings [GSE13898: esophageal adenocarcinoma (EAC), 15 Barrett's esophagus, 28 normal; GSE53625: 179 esophageal squamous cell carcinoma (ESCC), 179 normal; GSE77861: 7 ESCC, 7 normal; GSE92396: 12 EAC, 9 normal; GSE213565: 10 ESCC, 10 normal]. GEPIA2 and KM-Plotter assessed prognostic associations with EC and immunotherapy patients, which were subsequently validated in GSE53625, while GSCA analyzed CNV-survival correlations. The Cancer Genome Atlas (TCGA) EC cohort (96 ESCC, 88 EAC, 1 mucinous cystic neoplasm), incorporating race, tumor location, pathological grade, smoking status, alcohol consumption, gender, age, stage, molecular markers, and overall survival (OS) was used to analyze ULK3's clinical associations and prognostic value. Immune infiltration, T-cell exhaustion, and immunotherapy response were evaluated using TCGA EC, complemented by GSCA's drug sensitivity assessment. Pathway enrichment analysis was used to explore potential mechanisms, while T-cell and epithelial phenotypes across ULK3 expression groups were compared using single-cell RNA sequencing (SCRNA-seq) data (GSE196756: 3 ESCC, 3 normal). Mendelian randomization (MR) employed data from FinnGen and IEU OPEN Genome-Wide Association Study (GWAS) to investigate genetic susceptibility of ULK3 and EC.

ULK3 expression was elevated in normal EC tissues, with pan-cancer SNV mutations primarily consisting of missense mutations and EC CNV mutations biased toward deletions. Multivariate Cox regression identified ULK3 as an independent survival factor [hazard ratio (HR): 0.417, 95% confidence interval (CI): 0.218-0.798, P=0.008]. Immunologically, ULK3 was associated with reduced macrophage infiltration, improved chemotherapy and immunotherapy responses. Enrichment analysis showed ULK3 association with epidermal growth, immune regulation, and microautophagy pathways. ScRNA-seq data revealed ULK3-mediated T-cell activation and suppression of tumor phenotype. MR analysis yielded consistent risk associations [IEU OPEN GWAS, odds ratio (OR): 0.9994, 95% CI: 0.9992-0.9996, P=7.89×10^-8; FinnGen, OR: 0.9047, 95% CI: 0.8218-0.9959, P=0.04].

ULK3's mutation profile, expression patterns, prognostic associations, immune correlations and MR results suggest that its elevated expression may reduce EC risk and progression, supporting its potential as a prognostic biomarker. Single-cell and pathway analyses indicate that microautophagy-mediated regulation of tumor proliferation and immunity underlies these effects.