绝经年龄、表观遗传衰老与癌症风险：PLCO试验的二次分析

Background : Our previous studies have linked early menopause (early-M, <45 years) with increased risks of lung-related morbidities and mortalities. However, its relationship with other cancer types and the underlying biological and causal mechanisms remains unclear. Aim : To conduct a secondary analysis evaluating the associations between early-M and blood-based epigenetic aging biomarkers, and cancer risks and mortalities using the prospective Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening trial. Methods : Genome-wide DNA methylation (DNAm) profiles were available from baseline blood samples in 1,517 PLCO participants who subsequently developed breast cancer or who remained cancer-free. Epigenetic age acceleration was estimated using four established DNAm clocks, HorvathAge, HannumAge, PhenoAge, and GrimAge, as well as DNAm-based telomere length. An epigenome-wide association study (EWAS) was performed followed by pathway enrichment analysis and exploratory analyses integrating immune marker data. For cancer risk and mortality analyses, we included all postmenopausal women of European ancestry with natural menopause from the full PLCO cohort with genotype and phenotype data available (n = 31,022). A polygenic risk score (PRS) for age at natural menopause was constructed using 154 genetic variants identified from the NHGRI-EBI GWAS Catalog, after quality control based on imputation quality, minor allele frequency, Hardy-Weinberg equilibrium, linkage disequilibrium, and ambiguity filtering. The associations between this PRS and cancer incidences and mortalities were modeled using Cox proportional hazards regression. Results : Phenotypic early-M was associated with higher GrimAge acceleration (0.57 years, 95%CI=0.04, 1.10) in the 1,517 PLCO participants. EWAS and pathway analyses identified CpG sites enriched in estrogen response and immune regulation pathways, consistent with immune marker profiling that revealed upregulation of immune-related proteins among women with early-M. In the full trial, genetically predicted younger age at natural menopause (per 1 SD change) was associated with lower risks of breast cancer incidence (HR=0.14, 95%CI=0,12, 0.16) and longer survival (HR=0.27, 95%CI=0.19. 0.40), and lower risk for incidence of ovarian (HR=0.07, 95%CI=0.04, 0.11) and lung (HR=0.30, 95%CI=0.17, 0.50) cancers. In contrast, a higher incidence of bladder cancer (HR=3.34, 95%CI=1.05, 10.61) was observed. No significant associations were found for colon, melanoma, or hematologic malignancies. Conclusion : Early-M is associated with accelerated biological aging and distinct immune and hormonal regulatory patterns, with the potential to contribute to heterogeneous cancer risk profiles across cancer types.