Subtype-Dependent Transcriptional Coupling Between DNMT1 and E2F-Linked Proliferative Programs in Breast Cancer: A...

DNMT1 is a maintenance DNA methyltransferase involved in replication-coupled epigenetic regulation and is frequently expressed in highly proliferative tumors. Although DNMT1 has been linked to RB-E2F-related transcriptional programs, the reproducibility and subtype dependence of this association in breast cancer remain incompletely defined.

We performed a dual-cohort integrative transcriptomic analysis using TCGA-BRCA as the discovery cohort and METABRIC as the external validation cohort. DNMT1 expression was evaluated across tumor status and subtype annotations, with PAM50 used for all core multivariable and interaction analyses and METABRIC CLAUDIN_SUBTYPE used only for descriptive stratification. DNMT1-associated transcriptional programs were characterized in TCGA-BRCA using differential expression analysis and pathway enrichment. E2F-related transcriptional activity was summarized using an E2F metagene score derived from 122 available genes in the MSigDB Hallmark E2F Targets gene set. Cross-cohort Spearman correlation analyses, multivariable linear models, MKI67-adjusted sensitivity analyses, and likelihood-ratio tests for E2F2 × PAM50 interaction were performed.

DNMT1 expression was elevated in breast tumors and was highest in basal-like tumors. In TCGA-BRCA, high DNMT1 expression was associated with proliferation- and replication-related transcriptional programs, including cell-cycle progression, DNA replication, homologous recombination, and replication stress-related pathways. Across TCGA-BRCA and METABRIC, DNMT1 expression was positively associated with the E2F metagene score, E2F2, E2F3, and MKI67. In multivariable models adjusted for PAM50 subtype and age, the E2F metagene score remained associated with DNMT1 expression in both TCGA-BRCA (standardized β = 0.889, 95% CI 0.837-0.941) and METABRIC (β = 0.578, 95% CI 0.517-0.638). In joint E2F2/E2F3 models, cross-cohort evidence was strongest for E2F2, whereas the E2F3 association was more cohort-dependent and was attenuated in METABRIC. MKI67 adjustment attenuated but did not abolish the E2F2 association in both cohorts. Formal interaction testing supported subtype-dependent heterogeneity in the E2F2-DNMT1 association in both TCGA-BRCA and METABRIC.

DNMT1 shows reproducible, subtype-dependent transcriptional coupling with E2F-linked proliferative programs in breast cancer, with the most consistent cross-cohort evidence observed for the E2F metagene score and E2F2. These findings should be interpreted as hypothesis-generating transcriptomic associations rather than evidence of direct mechanistic causality.