Integrative multi-omics and Mendelian randomization analysis reveals immune-epigenetic crosstalk potentially involved in cervical carcinogenesis.

BACKGROUND: Although immunotherapy is recommended for recurrent or metastatic cervical cancer (CC), therapeutic response rates remain suboptimal. Emerging evidence suggests that metabolite-driven post-translational modifications (PTMs) within the tumor immune microenvironment may contribute to immune evasion. We aimed to evaluate the causal roles of immune-related and histone PTM-associated genes in CC using integrative multi-omics Mendelian randomization (MR). METHODS: A total of 1,576 immune-related and histone phosphorylation-, acetylation-, and lactylation-associated genes were curated. Two-sample MR analyses were performed using cis-eQTL (eQTLGen) and cis-pQTL (deCODE) datasets as exposures and three independent CC GWAS datasets as outcomes. Sensitivity analyses, summary-data-based MR (SMR), colocalization, phenome-wide association studies, and multivariable MR were conducted to assess robustness. Subsequent exploratory bioinformatics analysis included differential expression analysis, immune infiltration profiling (ssGSEA), mediation analysis, and molecular docking. RESULTS: MR identified four genes (CASP8, MAPK3, CBS, and LTA) and three proteins (EIF2AK2, SPAG1, and LGALS3) significantly associated with CC risk (FDR < 0.05). LTA and LGALS3 increased risk, whereas CASP8, MAPK3, CBS, EIF2AK2, and SPAG1 were protective. SMR and colocalization analyses provided supportive genetic evidence for LTA, EIF2AK2, and SPAG1 as prioritized candidate genes/proteins. Multivariable MR demonstrated independence from autoimmune confounders. Key genes were differentially expressed, showed moderate diagnostic performance (AUC 0.7-0.9), and correlated with 19 immune cell subsets, particularly activated CD8+ T cells. Molecular docking suggested weak-to-moderate predicted binding affinity for LTA-acetaminophen and SPAG1-valproic acid. CONCLUSIONS: LTA, EIF2AK2, and SPAG1 emerged as genetically prioritized candidates with potential biological relevance to cervical carcinogenesis, offering hypothesis-generating evidence for future mechanistic.