DNA methylation is an epigenetic process linked to cell development and various critical activities. Aberrant methylation patterns have been observed in cancer cell genomes (16). It is a reversible process in which methyl groups are introduced to the fifth carbon position of the cytosine from S-adenosyl methionine. Of note, two types of methylation exist: Maintenance methylation (when CpG dinucleotides on a single strand of DNA are methylated) and de novo methylation (when CpG dinucleotides on both of the strands are unmethylated) (17). DNA methylation is a reversible process facilitated by DNA methyltransferases (DNMTs). DNMT1, DNMT2, DNMT3A and DNMT3B encode proteins with different functional specificities. As previously demonstrated, two patterns of aberrant methylation have been identified: Global hypomethylation throughout the genome and hypermethylation at CpG islands within promoter regions (18,19).

Global DNA hypomethylation, a cancer hallmark coexisting with focal CpG island hypermethylation, involves genome-wide loss of cytosine methylation (20). Mechanistically driven by an impaired DNMT1 maintenance, deregulated DNMT3A/3B activity and aberrant TET-mediated demethylation, and reduced 5-methylcytosine levels are associated with tumor progression and a poor survival (21). Functionally, hypomethylation destabilizes heterochromatin, causing chromosomal instability. It reactivates transposable elements, inducing mutagenesis, and aberrantly activates oncogenes and metastasis-promoting genes (22).

The aberrant addition of methyl groups to cytosines in CpG dinucleotides, particularly within promoter-associated CpG islands, is a hallmark epigenetic alteration in cancer, causing stable gene silencing without altering the DNA sequence. In cancer, ~5-10% of promoter CpG islands become hypermethylated, repressing tumor suppressor genes involved in cell cycle control, DNA repair and apoptosis (e.g., CDKN2A, MLH1 and BRCA1), contributing to oncogenesis (23,24). These reversible changes provide targets for epigenetic therapy, as well as biomarkers for cancer diagnosis, prognosis and therapeutic stratification (25). Both modifications have commonly been documented in BC and are greatly affected by environmental factors such as aging, stress, alcohol consumption and air pollution (26).

DNA wraps around a histone octamer in the nucleus. The histone tails are sensitive to several covalent posttranslational modifications (PTMs) that regulate chromatin state. Changes in histone PTM patterns have been widely associated with cancer (27). Key modifications include the following:

Histone phosphorylation. The addition of phosphate groups promoting chromatin relaxation, critical for DNA damage responses (notably γ-H2AX at Ser139), transcriptional regulation and mitotic chromosome condensation (28).

Histone acetylation. Acetylation by histone acetyltransferases leads to decreased histone-DNA interaction, creating euchromatin. This is strongly associated with gene transcription activation. Histone deacetylases remove acetyl groups, compacting chromatin and inhibiting transcription (29).

Histone ubiquitylation. H2A ubiquitination (Lys119) usually causes gene repression, while H2B ubiquitination (Lys120) aids transcription and is required prior to methylation (30).

Histone sumoylation. SUMO conjugation drives transcriptional repression and chromatin compaction, often antagonizing acetylation and ubiquitination (31).

Histone methylation. This occurs on lysine and arginine residues with variable outcomes. Activating markers include H3K4me3; repressive markers include H3K9me3 and H3K27me3. Histone methyltransferases and demethylases control methylation dynamics (32).

Non-coding RNAs lack protein-coding ability, and ~52% of the human genome has been encoded, although only 1.2% codify proteins. ncRNAs are categorized into ‘housekeeping ncRNAs’ (tRNAs, rRNAs, snRNAs and snoRNAs) and ‘regulatory ncRNAs’ (lncRNAs, >200 nucleotides; and short ncRNAs, <200 nucleotides) (33,34). They play a critical role in controlling signaling pathways linked to tumor development, spread and therapeutic resistance (35).

Short non-coding RNAs (sncRNAs). snRNAs are a diverse group under ~200 nucleotides, including microRNAs (miRNAs), small interfering RNAs (siRNAs) and PIWI-interacting RNAs (piRNAs), that regulate gene expression primarily at the post-transcriptional level. These molecules contribute to chromatin modulation, cell differentiation, stress responses and disease processes (36,37).

Long non-coding RNAs (lncRNAs). These RNA species are longer than 200 nt without protein-coding potential (33). The human transcriptome has ~10,000 lncRNAs. They function as oncogenes or tumor suppressors, affecting cancer cell growth, apoptosis, metabolic processes, epithelial-mesenchymal transition (EMT), metastasis and treatment resistance. lncRNAs are usually exclusive to a particular tumor subtype, and BC exhibits abnormal expression levels of several lncRNAs (34,38).

MSH6 partners with MSH2 to form the MutSα complex, which recognizes and repairs base-base mismatches during DNA replication. In BC, promoter hypermethylation silences MSH6 expression in both ductal carcinoma in situ (DCIS) and invasive carcinomas, compromising mismatch repair (MMR) fidelity and elevating mutation rates (41,42). TCGA bioinformatic analyses have demonstrated inverse correlations between promoter methylation and gene expression, establishing MSH6 loss as an early driver of genomic instability in breast tumorigenesis (42,68).

MDC1 functions as a molecular scaffold at DNA double-strand breaks (DSBs), binding phosphorylated histone H2AX (γH2AX) and orchestrating the recruitment of the MRN complex, ATM kinase, RNF8, BRCA1 and 53BP1. Promoter-associated transcriptional downregulation in BC is associated with genomic instability, radioresistance, nodal metastasis and adverse clinical outcomes. In silico network analyses position MDC1 as a central signaling hub within the ATM-BRCA1-RNF8 axis, where its functional loss accelerates DNA damage accumulation and impairs checkpoint responses (43,44).

RNF168 (3q29), an E3 ubiquitin ligase, catalyzes histone H2A ubiquitination at lysines 13 and 15, facilitating the recruitment of 53BP1, BRCA1 and PALB2 to DSB sites for homologous recombination (HR). TCGA-BRCA and METABRIC dataset analyses reveal that promoter hypermethylation suppresses RNF168 expression, resulting in HR deficiency, heightened sensitivity to DNA-damaging agents and potential endocrine therapeutic resistance (45,46). TCGA data indicate RNF168 overexpression in BRCA1-mutant BCs and in BCs with HR deficiency (HRD), driven mainly by copy-number amplification and enriched in basal-like tumors. A high expression of RNF168 is associated with HRD mutational signature 3 and a worse survival, while lower RNF168 levels are associated with improved outcomes and a reduced risk of developing BC in BRCA1 mutation carriers (69).

RAD51 (17q23) is the central recombinase in homologous recombination repair, mediating strand invasion and the resolution of HR intermediates. Integrative analyses of TCGA and DNA methylation arrays identify promoter hypermethylation-mediated RAD51 silencing predominantly in triple-negative breast cancer (TNBC), where it is associated with HR-deficiency molecular signatures and reduced transcript abundance (47,48). TCGA and drug-response datasets demonstrate that a low expression of RAD51 is associated with a poor overall survival, but increased sensitivity to DNA-damaging chemotherapy, particularly in BC, reflecting homologous recombination deficiency that can be therapeutically exploited (70).

MLH1 heterodimerizes with PMS2 to execute mismatch repair, correcting replication errors including base mismatches and insertion-deletion loops. Promoter hypermethylation silences MLH1 in ~43.5% of BCs, as demonstrated by multi-platform in silico analyses demonstrating strong inverse methylation-expression correlations (49,50). The loss of MLH1 precipitates MMR deficiency, microsatellite instability, elevated tumor mutational burden, and the progression toward advanced, genomically unstable disease states (71).

RAD50 is a core part of the ATPase-containing MRN complex, which is critical for the recognition of DSBs, the activation of ATM, telomere maintenance, and selection of DSB repair pathways between HR and non-homologous end joining (NHEJ) repair (51,52). Multi-omics analysis has demonstrated that the hypermethylation of promoters suppresses RAD50 transcription and disrupts the stability of the MRN complex, resulting in improper DSB signaling, genomic instability and an increased risk of developing BC. RAD50 suppression is particularly implicated in TNBC carcinogenesis and chemotherapeutic resistance (72).

MRE11 provides the nuclease activity within the MRN complex, initiating DNA end resection, activating ATM-dependent checkpoints, and supporting both HR and NHEJ pathways (53,54). Previously, TCGA-BRCA analyses demonstrated a reduced MRE11 expression in a subset of breast cancers, while copy-number alterations affecting MRE11 were observed at a lower frequency. A low expression of MRE11 was shown to be associated with differential gene expression and pathway enrichment indicating impaired ATM signaling and homologous recombination, consistent with genomic instability and aggressive tumor phenotypes (73).

BRCC36 (Xq28) is a K63-specific deubiquitinase within the BRCA1-A complex that removes ubiquitin chains to fine-tune BRCA1 activation and regulate 53BP1 recruitment during DSB repair pathway selection (55,56). TCGA-BRCA transcriptomic data have revealed the dysregulation/overexpression of BRCC36 in BC (74). Network and pathway analyses have linked altered BRCC36 levels to a perturbed BRCA1-A complex function and ubiquitin-dependent DSB repair, with computational models suggesting the modulation of DSB pathway choice and potential radiosensitization upon BRCC36 inhibition (74).

BRCA1 is a tumor suppressor orchestrating homologous recombination, cell cycle checkpoints, and transcriptional regulation. In sporadic BC, promoter hypermethylation constitutes a major mechanism of BRCA1 silencing, producing HR deficiency, triple-negative phenotype enrichment, and synthetic lethality with PARP inhibitors (57-59). Integrated bioinformatics analyses of BC cohorts have demonstrated that BRCA1 promoter hypermethylation is tightly linked to reduced BRCA1 expression and homologous-recombination-deficient genomic and transcriptomic signatures, closely mirroring BRCA1-mutated tumors and supporting epigenetic BRCA1 silencing as a functional driver of HR deficiency (75).

BRCA2 functions as a tumor suppressor that regulates RAD51 nucleoprotein filament assembly during HR-mediated DSB repair. Promoter hypermethylation represents an early and frequent event in breast carcinogenesis, detected even in DCIS, where it induces gene silencing, genomic instability and repair incompetence (41,60). TCGA-based methylation and integrative bioinformatic analyses link BRCA2 promoter hypermethylation in pre-invasive breast lesions to homologous recombination deficiency signatures and increased genomic instability (76).

PALB2 serves as a molecular bridge connecting BRCA1 and BRCA2, facilitating RAD51 loading and enabling efficient HR. While germline PALB2 mutations confer hereditary breast cancer susceptibility, somatic promoter hypermethylation also suppresses expression in a subset of sporadic cases, resulting in HR deficiency and disease progression (61,62). Bioinformatics methylome profiling in early-onset BC samples has revealed increased methylation at PALB2 promoter-associated CpG probes in tumor DNA compared with matched blood DNA, consistent with potential epigenetic modulation of PALB2 expression in a subset of cases (77).

ATM encodes a master checkpoint kinase that phosphorylates numerous substrates, including p53, BRCA1 and CHK2 in response to DSBs, orchestrating cell cycle arrest and repair pathway activation. Promoter hypermethylation reduces ATM expression in BC, and is associated with a larger tumor size, advanced stage and an older patient age (63-65). TCGA-based repair signature analyses have associated ATM silencing with checkpoint deficiency and impaired DSB repair capacity, supporting its utility as a prognostic and predictive biomarker in sporadic disease (78).

RNF8 (6p21.3) is an E3 ubiquitin ligase initiating the ubiquitin cascade required for BRCA1 and 53BP1 recruitment to DSBs, promoting high-fidelity HR, while suppressing mutagenic NHEJ. Paradoxically, RNF8 is frequently overexpressed in BC, where it drives EMT, metastasis and chemoresistance through the activation of the Twist and β-catenin signaling pathways (66,67). Bioinformatics analyses of BC datasets has demonstrated that RNF8 promoter methylation is rare; however, RNF8 is transcriptionally dysregulated and embedded in DDR networks linked to BRCA1-RAD51-ATM signaling, indicating an altered homologous recombination and DDR pathway balance (79).