DNA methylation is a critical epigenetic modification that regulates gene expression without altering the underlying DNA sequence. Unlike genetic mutations, which permanently change the genome, methylation patterns are dynamic and can be influenced by environmental factors, aging, and disease. In cancer cells, these patterns become profoundly disrupted, leading to widespread changes in gene activity that fuel tumor initiation, progression, and metastasis. Understanding how methylation patterns go awry in cancer is not only central to the biology of the disease but also opens the door to novel diagnostic tools and targeted therapies.

Fundamentals of DNA Methylation

DNA methylation typically refers to the addition of a methyl group (CH3) to the fifth carbon of a cytosine base, forming 5-methylcytosine. This reaction is catalyzed by a family of enzymes known as DNA methyltransferases (DNMTs). In mammals, the most common context for methylation is a CpG dinucleotide—a cytosine followed by a guanine. Regions rich in CpG dinucleotides, called CpG islands, are often found in the promoter regions of genes.

The Methylation Machinery

Three main DNMTs orchestrate methylation patterns: DNMT1 maintains existing methylation marks during DNA replication, while DNMT3A and DNMT3B establish new methylation patterns during development. These enzymes work together to ensure that methylation landscapes are faithfully inherited across cell divisions. In addition, the ten-eleven translocation (TET) family of enzymes can oxidize methyl groups, leading to active demethylation, providing a reversible layer of control.

CpG Islands and Gene Regulation

In healthy cells, CpG islands in promoter regions are generally unmethylated, allowing transcription factors and RNA polymerase to access the DNA. Conversely, repetitive elements and transposons are typically heavily methylated to maintain genomic stability. This binary state—methylated (silent) versus unmethylated (active)—serves as a fundamental switch for gene regulation. However, methylation can also occur in gene bodies, where it often correlates with active transcription by preventing spurious initiation from internal promoters.

Methylation Dynamics in Normal Cells

DNA methylation patterns are not static; they are established during embryogenesis and undergo precise changes throughout life. During development, a wave of global demethylation erases most parental marks, followed by lineage-specific re-methylation that defines cellular identity. In adult tissues, methylation patterns are generally stable but can be altered by aging, diet, and environmental exposures.

Tissue-Specific Methylation

Each cell type maintains a unique methylome that reflects its differentiated state. For example, genes that are essential for neuronal function are hypomethylated in neurons but hypermethylated in liver cells. This tissue-specific regulation ensures that only the appropriate set of genes is expressed, preventing the activation of ectopic programs.

Methylation and Chromatin Structure

Methylation at gene promoters recruits methyl-binding domain (MBD) proteins, which then attract histone deacetylases and other repressive complexes. This leads to a compact chromatin state that physically blocks transcription. Conversely, unmethylated CpG-rich promoters are often associated with active histone marks (e.g., H3K4me3) and open chromatin, allowing gene expression.

Epigenetic Disruption in Cancer Cells

Cancers universally exhibit aberrant DNA methylation patterns. Two opposing phenomena occur simultaneously: global hypomethylation and focal hypermethylation. This "yin-yang" of epigenetic disruption creates a landscape where tumor suppressor genes are silenced and oncogenes are activated.

Global Hypomethylation

Overall, cancer genomes are hypomethylated compared to normal tissue, particularly in repetitive sequences such as LINE-1 retrotransposons and Alu elements. Hypomethylation of these repeats can lead to genomic instability by promoting chromosomal rearrangements and reactivating transposable elements. Additionally, hypomethylation at specific loci can activate oncogenes such as R-RAS and C-MYC, driving uncontrolled proliferation.

Focal Hypermethylation of Tumor Suppressor Gene Promoters

A hallmark of many cancers is the hypermethylation of CpG islands located in the promoter regions of tumor suppressor genes. This aberrant methylation silences the second allele (the remaining functional copy) in the absence of genetic mutation, effectively inactivating key protective pathways. Classic examples include:

  • p16INK4a (CDKN2A) – a cyclin-dependent kinase inhibitor frequently hypermethylated in melanomas, lung, and pancreatic cancers.
  • BRCA1 – a DNA repair gene silenced by promoter methylation in a subset of breast and ovarian cancers.
  • MLH1 – a mismatch repair gene; hypermethylation causes microsatellite instability in colorectal and endometrial cancers.
  • GSTP1 – a detoxification enzyme; its methylation serves as a biomarker for prostate cancer.

Methylation Patterns as Cancer Subtype Signatures

Tumors of the same tissue type often share specific methylation patterns, which can classify cancers into subtypes with distinct prognoses. For instance, the CpG island methylator phenotype (CIMP) identifies a subset of colorectal cancers with extensive promoter hypermethylation and better or worse outcomes depending on the genetic background. Similarly, glioma classification now includes a methylation-based group that stratifies patients for therapy.

Mechanistic Impact on Gene Expression

The effect of DNA methylation on gene expression is mediated through several interconnected mechanisms. While direct steric hindrance of transcription factor binding is important, most silencing involves the recruitment of chromatin-modifying complexes that establish a repressive state.

Direct Blockade of Transcription Factor Binding

Many transcription factors recognize sequences that contain CpG dinucleotides. When those cytosines are methylated, the major groove of the DNA is altered, preventing factor binding. For example, methylation of the E-box motif can block the binding of the MYC transcription factor, leading to gene repression.

Methyl-Binding Proteins and Repressive Complexes

Methylated CpG dinucleotides are recognized by a family of proteins with methyl-CpG-binding domains (MBDs), including MeCP2, MBD1, MBD2, and MBD4. These proteins recruit histone deacetylases (HDACs), histone methyltransferases (e.g., G9a, SUV39H1), and other co-repressors that modify histones to produce a compact, transcriptionally inert chromatin. This crosstalk between DNA methylation and histone modifications creates a self-reinforcing loop that is difficult to reverse spontaneously.

Methylation of Enhancers and Other Regulatory Elements

Beyond promoters, methylation of enhancers, insulators, and other non-genic regions can also influence expression. For example, hypomethylation of enhancers can activate oncogenic super-enhancers, while hypermethylation of insulators can break boundary elements, leading to ectopic gene activation. The classic example is hypermethylation of the CTCF binding site at the IGF2/H19 locus, which causes loss of imprinting and contributes to cancer growth.

Methylation and Alternative Splicing

Recent studies show that intragenic methylation can affect exon recognition. Methylation of exons slows RNA polymerase II elongation, promoting inclusion of the exon in the final mRNA transcript. Thus, altered methylation patterns in cancer can lead to aberrant splicing events that generate oncogenic protein isoforms.

Clinical Implications and Therapeutic Strategies

The reversible nature of DNA methylation makes it an attractive target for cancer therapy. Drugs that inhibit DNMTs can reactivate silenced tumor suppressor genes, and methylation markers are being developed for early detection, prognosis, and monitoring.

DNA Methylation Biomarkers

Because abnormal methylation occurs early in carcinogenesis, detecting methylated DNA in body fluids offers a non-invasive screening approach. For example:

  • SEPT9 methylation in plasma is FDA-approved for colorectal cancer screening.
  • MGMT promoter methylation predicts response to temozolomide in glioblastoma. Patients with methylated MGMT have longer survival due to impaired DNA repair.
  • RASSF1A and APC methylation are being investigated as biomarkers for lung and breast cancers.

Hypomethylating Agents

Two cytosine analogs, azacitidine (Vidaza) and decitabine (Dacogen), are approved for myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). These drugs incorporate into DNA and trap DNMTs, leading to their degradation and a gradual reduction in methylation. Reactivation of silenced tumor suppressor genes, including p15INK4b and various differentiation factors, contributes to their clinical activity. However, these agents are not specific to cancer cells and can cause myelosuppression. Newer, less toxic formulations and combination regimens (e.g., with HDAC inhibitors or immune checkpoint blockers) are under investigation.

Combination Therapies and Resistance

While hypomethylating agents can be effective, resistance inevitably develops. Mechanisms include upregulation of drug efflux pumps, increased DNMT activity, and compensatory pathways such as histone methylation. Combining DNMT inhibitors with agents that target other epigenetic modifiers (e.g., EZH2 inhibitors, LSD1 inhibitors) or with conventional chemotherapy may overcome resistance. Several clinical trials are testing such combinations in solid tumors. Additionally, emerging evidence suggests that demethylation can upregulate viral defense pathways, making tumors more susceptible to immunotherapy. For example, demethylating agents can activate endogenous retroviruses, triggering an interferon response and enhancing anti-tumor immunity.

DNA Methylation and Immunotherapy Response

Recent research has linked specific methylation patterns to immune checkpoint inhibitor outcomes. Tumors with a high degree of immune evasion–related hypermethylation (e.g., silencing of PD-L1 or antigen presentation genes) may be less responsive to PD-1/PD-L1 blockade. Combining hypomethylating agents with checkpoint inhibitors is a promising strategy to restore immunogenicity, and early-phase trials are showing encouraging results in non-small cell lung cancer and Hodgkin lymphoma.

Future Directions and Unanswered Questions

Despite significant progress, many aspects of methylation in cancer remain poorly understood. The field is moving toward single-cell resolution, spatial epigenomics, and targeted epigenetic editing.

Single-Cell Methylation Analysis

Bulk methylation profiling averages signals across millions of cells, masking heterogeneity within a tumor. Single-cell bisulfite sequencing technologies now allow researchers to map methylation at the single-cell level, revealing rare subclones that drive metastasis or resistance. These techniques are beginning to uncover how methylation changes occur dynamically during tumor evolution.

Targeted Epigenome Editing

Fusion proteins combining a catalytically inactive Cas9 (dCas9) with the TET enzyme or DNMT3A enable precise demethylation or methylation of specific loci. In preclinical models, reactivating silenced tumor suppressor genes such as p16 or MLH1 through targeted demethylation has shown promise in restoring cell cycle control and DNA repair. Challenges include delivery efficiency, off-target effects, and long-term stability of the edits. Nevertheless, these tools hold the potential for “epigenetic therapy” with fewer side effects than global inhibitors.

DNA methylation at specific sites changes predictably with age, forming an “epigenetic clock.” Accelerated epigenetic aging has been linked to increased cancer risk, possibly through the accumulation of stochastic methylation errors. Understanding how lifestyle factors (diet, exercise, smoking) modulate the methylome—and whether interventions can reverse age-related methylation changes—is an active area of investigation.

Methylation and Non-Coding RNAs

Intricate crosstalk exists between DNA methylation and regulatory non-coding RNAs. Long non-coding RNAs (lncRNAs) can recruit DNMTs or TETs to specific genomic regions. Conversely, DNA methylation can control the expression of microRNAs and lncRNAs that themselves regulate oncogenes. Disentangling these multilayered circuits will be essential for a complete picture of epigenetic dysregulation in cancer.

Conclusion

Methylation patterns in DNA are a fundamental layer of gene regulation that becomes deeply corrupted in cancer. Global hypomethylation and focal hypermethylation collaborate to silence tumor suppressors, activate oncogenes, and destabilize the genome. The mechanistic understanding of how methylation affects transcription, chromatin, and splicing has already translated into clinical tools—ranging from diagnostic biomarkers like SEPT9 to therapies such as azacitidine and decitabine. As the field advances toward single-cell profiling and targeted epigenome editing, we will gain the ability to dissect the epigenome with unprecedented precision. Continued research into the dynamic interplay between DNA methylation and the cancer epigenome promises to yield more effective, personalized strategies for early detection and treatment.

For further reading, see Nature Reviews Molecular Cell Biology: DNA methylation in cancer and the NCI Fact Sheet on Epigenetics. Additional information on targeted epigenetic editing can be found in Nature Reviews Cancer: Epigenetic therapy in solid tumours.