DNA Methylation, Histone Modification & Gene Silencing
Your genome contains roughly 20,000 genes, yet a liver cell and a neuron — carrying identical DNA — behave entirely differently. Epigenetics explains how. This guide covers the molecular machinery that reads, writes, and erases epigenetic marks: DNA methyltransferases, histone-modifying enzymes, Polycomb and Trithorax complexes, chromatin remodelers, and non-coding RNAs — and why their dysregulation drives cancer, developmental disorders, and complex disease.
What Epigenetics Controls — and Why the Sequence Alone Doesn’t Explain Biology
When the Human Genome Project concluded in 2003, many assumed that sequencing the genome would explain human biology. It explained far less than expected. Identical twins share virtually 100% of their DNA sequence, yet one can develop schizophrenia, type 1 diabetes, or breast cancer while the other does not. A fertilised egg contains a single genome, yet the 200-plus differentiated cell types of the adult body — each with radically different morphology, function, and gene expression profile — carry that same sequence. The genome alone is not enough. What determines which of the 20,000 protein-coding genes are active, at what level, in which tissue, and at which developmental stage is an additional layer of information: the epigenome.
Epigenetics — from the Greek epi (above or on top of) and genetics — refers to heritable changes in gene expression and chromatin state that occur without alteration of the underlying DNA nucleotide sequence. The definition, first formally articulated by Conrad Waddington in 1942 to describe gene-environment interactions during development, now encompasses three interconnected molecular systems: DNA methylation (covalent modification of cytosine bases), histone post-translational modifications (chemical marks on the N-terminal tails of histone proteins that package DNA), and non-coding RNA regulation (lncRNAs, miRNAs, and piRNAs that guide epigenetic complexes or degrade transcripts). According to the National Human Genome Research Institute, epigenetic research is now central to understanding development, ageing, cancer, and the mechanisms by which environment shapes biology across a lifetime and potentially across generations.
Understanding epigenetics requires understanding chromatin — the complex of DNA and histone proteins in which the genome is packaged inside the nucleus. The fundamental repeating unit of chromatin is the nucleosome: 147 base pairs of DNA wound ~1.65 times around an octamer of histone proteins (two copies each of H2A, H2B, H3, and H4). The ~10 nm “beads-on-a-string” nucleosome array is compacted into higher-order structures through linker histone H1 interactions and histone tail modifications, ultimately forming the 30 nm fibre and chromosome-scale loops visible during mitosis. Gene regulation depends critically on chromatin accessibility — whether transcription factors, RNA polymerase II, and other machinery can physically reach the DNA. Epigenetic modifications regulate this accessibility: open, transcriptionally active euchromatin versus compacted, transcriptionally silent heterochromatin, with these states maintained heritably through successive cell divisions.
Epigenetics Knowledge Graph — Entity Attributes and Related Concepts
The table below maps the primary entity (Epigenetics) to its core attributes, directly connected molecular entities, and supporting facts — structuring the conceptual framework for all subsequent sections.
| Category | Entity / Attribute | Key Detail |
|---|---|---|
| Core Mechanism 1 | DNA Methylation | Addition of -CH₃ to cytosine-C5 at CpG dinucleotides by DNMT1 (maintenance), DNMT3A/DNMT3B (de novo); read by MBD proteins; erased by TET enzymes (active demethylation via 5-hmC, 5-fC, 5-caC) |
| Core Mechanism 2 | Histone Modification | Acetylation (HATs → active; HDACs → remove), methylation (HMTs — H3K4me3 active, H3K27me3/H3K9me3 repressive; KDMs demethylate), phosphorylation (H3S10p in mitosis), ubiquitination (H2AK119ub1 — PRC1, repressive; H2BK120ub1 — active) |
| Core Mechanism 3 | Gene Silencing (Chromatin) | Transcriptional silencing via promoter methylation + MBD/HDAC recruitment → heterochromatin; Polycomb repressive complexes (PRC1, PRC2); position effect variegation (heterochromatin spreading) |
| Core Mechanism 4 | Gene Silencing (RNA-based) | miRNA → RISC-mediated translational repression/mRNA cleavage; siRNA → post-transcriptional degradation; piRNA → transposon silencing in germline; lncRNA → chromatin targeting of PRC2 (e.g., XIST, HOTAIR) |
| Writer Enzymes | DNMT1, DNMT3A, DNMT3B, DNMT3L, EZH2 (PRC2), G9a/GLP, SUV39H1/2, PRMT enzymes, HATs (p300/CBP, PCAF, MYST family) | Add epigenetic marks; mutations in DNMT3A and EZH2 are recurrent oncogenic drivers in haematological malignancies |
| Reader Proteins | MBD1/2/4, MeCP2, HP1 (H3K9me3), PHD fingers, chromo domains, bromodomains (BET proteins — BRD4) | Recognise specific marks and recruit co-regulatory complexes; BET bromodomain inhibitors (JQ1, OTX015) are in clinical trials for cancer |
| Eraser Enzymes | TET1/2/3 (5mC → 5-hmC → 5-fC → 5-caC → unmethylated C), KDMs (histone demethylases — LSD1/KDM1A, KDM6A/B — targets H3K27me3), HDACs (classes I–IV) | TET2 is among the most commonly mutated genes in clonal haematopoiesis; KDM6A (UTX) is a tumour suppressor frequently deleted in cancer |
| Chromosomal Epigenetics | X-chromosome inactivation (XCI), Genomic Imprinting | XCI: XIST lncRNA coats inactive X → PRC2 recruitment → H3K27me3 → heterochromatin (Barr body); Imprinting: ~100 loci with parent-of-origin-specific expression maintained by ICR methylation |
| Disease Relevance | Cancer, imprinting disorders, neurodevelopmental disorders, ageing | Global hypomethylation (genomic instability) + focal hypermethylation (tumour suppressor silencing) in cancer; Rett syndrome (MeCP2 mutations); Prader-Willi/Angelman; ICF syndrome (DNMT3B mutations) |
| Clinical/Therapeutic | DNMT inhibitors, HDAC inhibitors, EZH2 inhibitors, BET inhibitors, LSD1 inhibitors | 5-azacytidine, decitabine (DNMTi — AML/MDS); vorinostat, romidepsin, panobinostat (HDACi — lymphoma, myeloma); tazemetostat (EZH2i — follicular lymphoma, epithelioid sarcoma) |
| Research Technologies | Bisulfite sequencing (WGBS, RRBS), ChIP-seq, ATAC-seq, CUT&RUN, Hi-C, single-cell multi-omics | ENCODE and Roadmap Epigenomics projects have mapped histone marks and DNA methylation across hundreds of human cell types and tissues |
DNA Methylation — the Foundational Epigenetic Mark
5-Methylcytosine: the “Fifth Base”
DNA methylation in mammals occurs predominantly at cytosine residues within CpG dinucleotides (cytosine-phosphate-guanine), where a methyl group is added to the 5-carbon position of the cytosine ring to generate 5-methylcytosine (5mC) — sometimes called the “fifth base” of the mammalian genome, alongside the canonical A, T, G, and C. In human somatic cells, approximately 70–80% of all CpG sites are methylated, giving rise to an estimated 28 million methylcytosine residues per haploid genome. Non-CpG methylation (predominantly CpA contexts) occurs at substantial levels in neurons and embryonic stem cells, where it has been linked to gene body methylation and neuronal gene regulation — but CpG methylation remains the dominant and best-characterised epigenetic DNA modification.
The distribution of DNA methylation across the genome is highly non-uniform. Most of the genome is CpG-depleted (CpG dinucleotides are underrepresented because 5mC deaminates spontaneously to thymine at a rate approximately five times higher than the spontaneous deamination of cytosine to uracil — if a G is present on the complementary strand, this produces a C→T transition mutation that, over evolutionary time, has depleted CpG dinucleotides throughout the genome). In contrast, CpG islands — dense CpG-rich regions typically 500–2,000 bp in length located at or near gene promoters — are largely protected from methylation in normal somatic cells. About 70% of human gene promoters contain a CpG island; in normal tissues, these remain unmethylated and associated with open chromatin. The critical functional principle is therefore one of contrast: the default methylated background versus the unmethylated promoter CpG islands that permit transcription factor binding and gene expression.
DNMT Enzymes — Writers, Maintainers, and the Role of TET-Mediated Demethylation
DNMT3A — De Novo Methyltransferase for Developmental Programming
DNMT3A (along with its partner DNMT3B) establishes new methylation patterns de novo during gametogenesis, early embryonic development, and cell differentiation. It operates on unmethylated CpG dinucleotides in double-stranded DNA, adding methyl groups from S-adenosylmethionine (SAM) as the methyl donor. DNMT3A contains a PWWP domain (reads H3K36me2/3 at gene bodies, directing methylation away from active genes), a PHD domain (recognises unmethylated H3K4, preventing methylation at active promoters), and the catalytic C-terminal methyltransferase domain. Somatic mutations in DNMT3A — particularly R882H, the most common point mutation — are found in ~20% of de novo acute myeloid leukaemia (AML) and are a defining feature of clonal haematopoiesis of indeterminate potential (CHIP), the age-related pre-malignant clonal expansion that confers cardiovascular as well as leukaemia risk.
DNMT3B — De Novo Methylation of Satellite Repeats and Gene Body Methylation
DNMT3B is required particularly for methylation of pericentromeric satellite repeats (satellite 2 and satellite alpha), gene bodies of highly expressed genes, and subtelomeric regions. Biallelic loss-of-function mutations in DNMT3B cause ICF syndrome (Immunodeficiency, Centromeric instability, and Facial anomalies) — a rare autosomal recessive disorder characterised by pericentromeric hypomethylation of chromosomes 1, 9, and 16, chromosomal instability, and severe combined immunodeficiency. DNMT3B is also the primary methyltransferase for inactivating retrotransposons in the genome, preventing their reactivation — a critical genome maintenance function. DNMT3L (DNMT3-Like) has no catalytic activity but acts as a regulatory co-factor for DNMT3A and DNMT3B, stimulating their activity and guiding them to unmethylated H3K4 regions via its PHD domain.
DNMT1 — Maintenance Methylation at the Replication Fork
DNMT1 is the “maintenance” methyltransferase — it copies pre-existing methylation patterns onto newly synthesised DNA strands during DNA replication. After the replication fork passes, hemimethylated DNA (methylated parental strand, unmethylated daughter strand) is recognised by UHRF1 (Ubiquitin-like with PHD and Ring Finger domains 1), which recruits DNMT1 to the replication fork. DNMT1 then methylates the daughter strand CpG dinucleotide complementary to the parental methylcytosine — faithfully copying the methylation pattern into both daughter cells. This semi-conservative methylation maintenance is the molecular basis of the heritability of DNA methylation patterns through mitosis. DNMT1 also has a preference for hemimethylated substrates (~10-fold over unmethylated) and possesses allosteric regulation: its N-terminal CXXC domain binds unmethylated CpGs and autoinhibits the catalytic domain on unmethylated substrates, preventing aberrant de novo methylation at active gene promoters.
TET1/2/3 — Active DNA Demethylation via Oxidative Intermediates
DNA demethylation was long thought to occur only passively (through replication dilution in the absence of DNMT1 activity). The discovery of TET (Ten-Eleven Translocation) dioxygenases resolved the long-standing mystery of active demethylation. TET enzymes — TET1, TET2, and TET3 — sequentially oxidise 5mC to 5-hydroxymethylcytosine (5-hmC), then to 5-formylcytosine (5-fC), then to 5-carboxylcytosine (5-caC) using molecular oxygen, Fe²⁺, and α-ketoglutarate as co-substrates. TDG (thymine-DNA glycosylase) then excises 5-fC and 5-caC via base excision repair, and the gap is filled with unmodified cytosine — achieving active demethylation. 5-hmC is itself a stable mark found at high levels in neurons and embryonic stem cells, and serves as both an intermediate and a functional modification recognised by distinct reader proteins. TET2 is among the most frequently mutated genes in clonal haematopoiesis and myeloid malignancies; IDH1/IDH2 mutations in AML produce 2-hydroxyglutarate (2-HG), a competitive inhibitor of TET enzymes (and also of histone KDMs), explaining the hypermethylator phenotype of IDH-mutant cancers.
Methylation (Writing): Cytosine (C) ──[DNMT3A/3B de novo]──► 5-methylcytosine (5mC) [Gene silencing at promoters] Hemimethylated CpG ──[DNMT1 + UHRF1]──► Fully methylated CpG [Inherited through replication] Methyl donor: S-adenosylmethionine (SAM) → S-adenosylhomocysteine (SAH) Active Demethylation (Erasing): 5mC ──[TET1/2/3 + O₂ + Fe²⁺ + α-KG]──► 5-hydroxymethylcytosine (5-hmC) [stable — enriched in neurons, ESCs] ──[TET]──► 5-formylcytosine (5-fC) ──[TET]──► 5-carboxylcytosine (5-caC) ──[TDG + BER]──► Cytosine (C) [Demethylation complete] Key inhibitors in disease: IDH1/IDH2 mutations → produce 2-HG (2-hydroxyglutarate) → competitively inhibits TET enzymes → blocks active demethylation → CpG hypermethylator phenotype (CIMP) TET2 loss-of-function → impaired 5-hmC generation → clonal haematopoiesis, AML, MDS, CMML Passive demethylation: Occurs during rapid cell division when DNMT1 is absent or limiting → Methylation diluted 50% each round of replication → unmethylated after ~4 generations Critical in pre-implantation embryo (paternal pronucleus: active; maternal: passive)
Histone Modification — Reading the Histone Code
In 2000, Strahl and Allis proposed the “histone code” hypothesis: that specific combinations of post-translational modifications (PTMs) on histone tails constitute a code that is read by downstream effector proteins to determine transcriptional output. The concept, though later nuanced to acknowledge the context-dependence and combinatorial nature of histone marks, remains a powerful organising framework. The N-terminal tails of histones H3 and H4 (and to a lesser extent H2A and H2B) protrude from the nucleosome core and are subject to over a hundred distinct types of covalent modification on more than 60 residues — including acetylation, methylation, phosphorylation, ubiquitination, SUMOylation, ADP-ribosylation, propionylation, butyrylation, and crotonylation. Each modification is deposited by “writer” enzymes, removed by “eraser” enzymes, and recognised by “reader” proteins that translate the mark into downstream chromatin and transcriptional effects.
Active Promoter Mark
Trimethylation of H3 lysine 4, deposited by COMPASS complexes (SET1A/B, MLL1–4). Marks active gene promoters and is recognised by PHD fingers of transcription co-activators (TAF3). Associates with paused RNA Pol II and H3K27ac.
Active Enhancer Mark
Acetylation of H3 lysine 27, deposited by p300/CBP acetyltransferases. Distinguishes active enhancers (H3K27ac+) from poised enhancers (H3K4me1+, H3K27ac–). Read by BET bromodomain proteins (BRD4), which drive transcriptional elongation at super-enhancers.
Polycomb Repressive Mark
Trimethylation of H3 lysine 27, deposited by EZH2 (the catalytic subunit of PRC2). The cardinal repressive mark of Polycomb-silenced developmental genes. Recognised by PRC1’s chromobox (CBX) proteins. EZH2 is amplified or gain-of-function mutated in follicular lymphoma.
Constitutive Heterochromatin
Deposited by SUV39H1/2 at pericentromeric and telomeric regions, transposons, and silenced proviral sequences. Read by HP1α/β/γ (heterochromatin protein 1) via their chromodomain — HP1 binding induces chromatin compaction and spreading of the repressive mark.
Transcribed Gene Body
Deposited by SETD2 within actively transcribed gene bodies by coupling with RNA Pol II CTD phosphorylation at Ser2. Prevents aberrant splicing and cryptic transcription initiation within gene bodies. SETD2 is a recurrent tumour suppressor (clear cell RCC, gliomas, leukaemia).
PRC1 Repressive Mark
Monoubiquitination of H2A lysine 119 by RING1A/B (catalytic subunits of PRC1). Directly compacts chromatin and inhibits RNA Pol II elongation. Works with H3K27me3 to maintain Polycomb-repressed chromatin domains. BAP1 (deubiquitinase) removes this mark; BAP1 is a tumour suppressor in mesothelioma and uveal melanoma.
Histone Acetyltransferases and Deacetylases — the Chromatin Accessibility Switch
Histone acetylation is the modification most directly and reliably linked to transcriptional activation. The addition of an acetyl group (from acetyl-CoA, the metabolic acetyl donor) to the ε-amino group of a lysine residue neutralises its positive charge, reducing the electrostatic attraction between the histone octamer and the negatively charged DNA phosphate backbone. The result is a local loosening of nucleosome-DNA contacts and increased chromatin accessibility — creating regions known as euchromatin where the transcriptional machinery can bind. Conversely, histone deacetylation removes acetyl groups, restoring lysine positive charge, tightening nucleosome compaction, and reducing transcription factor access.
Histone Methylation — the Same Mark, Opposing Functions
Unlike acetylation, which is uniformly activating, histone methylation is a context-dependent modification whose functional consequence depends entirely on which residue is methylated and to what degree (mono-, di-, or trimethylation). This context-dependence arises because methyl groups do not alter the charge of the lysine or arginine residue — they primarily function as binding platforms for reader proteins, and different readers are recruited by different methyl marks on different residues. The resulting diversity of downstream effects makes histone methylation the most informationally rich layer of the histone code.
Gene Silencing — Mechanisms from Chromatin to RNA Degradation
Gene silencing operates at two fundamentally distinct levels in the cell: at the chromatin level (transcriptional gene silencing, TGS — preventing RNA polymerase from transcribing a gene) and at the post-transcriptional level (post-transcriptional gene silencing, PTGS — allowing transcription but degrading or blocking translation of the resulting mRNA). Epigenetic mechanisms primarily drive transcriptional silencing, but non-coding RNAs bridge both levels, and the two systems communicate — particularly in RNA-directed DNA methylation pathways.
Promoter Hypermethylation → TGS
Methylation of CpG islands at gene promoters → MBD/MeCP2 binding → HDAC/HMT co-repressor recruitment → deacetylation → H3K9me3 deposition → HP1 binding → heterochromatin compaction. The process is self-reinforcing: H3K9me3 recruits DNMT3A/3B, adding further methylation, while DNMT3A/3B’s PWWP domain avoids H3K4me3. Result: stable, mitotically heritable, transcriptionally silent locus.
Polycomb Silencing → Reversible TGS
PRC2 (EZH2+EED+SUZ12) deposits H3K27me3 → PRC1 (RING1B+CBX+BMI1) adds H2AK119ub1 and compacts chromatin. Unlike DNA methylation-mediated silencing, Polycomb silencing is dynamically reversible — KDM6A/B can remove H3K27me3, allowing gene reactivation during differentiation. Bivalent domains in ESCs (H3K27me3 + H3K4me3) keep developmental genes poised for rapid activation upon differentiation signals.
RNAi and miRNA → PTGS
Double-stranded RNA triggers Dicer cleavage → short dsRNA (siRNA, ~21 nt) → loaded into RISC (RNA-Induced Silencing Complex) with Argonaute (AGO2) → sequence-specific mRNA cleavage (siRNA, perfect complement) or translational repression + deadenylation (miRNA, imperfect complement). miRNAs control ~60% of human protein-coding gene expression; single miRNAs can target hundreds of mRNAs, operating as network regulators rather than simple on/off switches.
Polycomb and Trithorax — Epigenetic Memory of Cell Identity
The Polycomb group (PcG) and Trithorax group (TrxG) protein complexes were first identified in Drosophila as systems maintaining the segment-specific expression patterns of homeotic (Hox) genes long after the transient transcription factors that initially activated or repressed them had disappeared. They are, in essence, the molecular memory of cell identity — maintaining the transcriptional states established during early development throughout the thousands of cell divisions that follow. In humans, their targets extend far beyond Hox genes to encompass thousands of developmental regulators, and their mutations are among the most common events in human cancer.
Polycomb Repressive Complex 2 (PRC2) — the H3K27me3 Writer
PRC2’s catalytic subunit is EZH2 (Enhancer of Zeste Homologue 2), which carries a SET domain responsible for mono-, di-, and trimethylation of H3K27. EED (Embryonic Ectoderm Development) reads H3K27me3 via its WD40 domain and allosterically stimulates EZH2 catalytic activity — creating a positive feedback loop that propagates H3K27me3 marks along chromatin. SUZ12 and RbAp46/48 complete the canonical PRC2 core. Accessory subunits (JARID2, AEBP2, PCLs) modulate PRC2 recruitment to chromatin — JARID2 is particularly important for recruiting PRC2 to unmethylated CpG islands and to RNA-dependent chromatin targets. EZH2 gain-of-function mutations (Y641, A677, A687) in follicular lymphoma and diffuse large B-cell lymphoma alter substrate preference, driving preferential H3K27 trimethylation. EZH2 is also frequently amplified in solid tumours (prostate, breast cancer). Conversely, EZH2 loss-of-function mutations occur in T-cell ALL, myelodysplastic syndrome, and juvenile myelomonocytic leukaemia.
Polycomb Repressive Complex 1 (PRC1) — Chromatin Compaction
PRC1 is a heterogeneous family of complexes sharing RING1A or RING1B (the E3 ubiquitin ligases that monoubiquitinate H2AK119) and one of six PCGF (Polycomb Group RING Finger) subunits, which define canonical (CBX-containing) and non-canonical (RYBP/YAF2-containing) PRC1 variants. Canonical PRC1 is recruited to H3K27me3 via CBX chromodomains — establishing a hierarchical PRC2→PRC1 cascade. Non-canonical PRC1 can act independently of H3K27me3 and may pioneer Polycomb domain formation before PRC2 arrival. PRC1 compacts chromatin directly (in a modification-independent manner) and through H2AK119ub1, which inhibits RNA Pol II elongation but also helps recruit PRC2, creating a self-reinforcing Polycomb loop. BAP1 (BRCA1-Associated Protein 1) — the major H2AK119 deubiquitinase — is a nuclear tumour suppressor frequently mutated in mesothelioma, uveal melanoma, and cholangiocarcinoma; its loss leads to sustained H2AK119ub1 and aberrant Polycomb silencing of tumour suppressive targets.
Chromatin Remodeling Complexes — Moving the Packaging
While histone modifications and DNA methylation alter the chemical properties of chromatin, a second class of epigenetic regulators — ATP-dependent chromatin remodeling complexes — physically reposition, eject, or restructure nucleosomes using the energy of ATP hydrolysis. These complexes do not modify histones chemically; instead, they alter chromatin accessibility mechanically by sliding nucleosomes along DNA, evicting histones, or swapping canonical histones for specialised variants. The four major families of chromatin remodelers — SWI/SNF, ISWI, CHD (chromodomain helicase DNA-binding), and INO80 — each use a Snf2-family ATPase domain but differ in their recruitment mechanisms, substrates, and functional outputs.
SWI/SNF (BAF/PBAF) — Nucleosome Ejector
The BAF (mammalian SWI/SNF) complex evicts nucleosomes or repositions them to open chromatin at enhancers and promoters. Subunits: SMARCA4/BRG1 (ATPase), SMARCB1/SNF5, ARID1A, and 10+ others. The most frequently mutated chromatin remodeler in cancer — ARID1A lost in ~50% of ovarian clear cell carcinoma; SMARCB1 in malignant rhabdoid tumour; SMARCA4 in small cell carcinoma of ovary.
NuRD — Histone Deacetylation + Remodeling
Unique dual-function complex: contains both a CHD3/4 ATPase (remodeling) and HDAC1/2 (deacetylation). NuRD is recruited to unmethylated CpG islands and to MBD2/3 target sites. Removes activating H3 and H4 acetylation while repositioning nucleosomes — coordinating deacetylation and remodeling for gene repression. Involved in developmental gene silencing and stem cell pluripotency maintenance.
ISWI (NURF/CHRAC/RSF) — Nucleosome Spacing
ISWI complexes primarily space nucleosomes at regular intervals, using SMARCA5/SNF2H or SMARCA1/SNF2L as ATPase subunits. Regular spacing promotes higher-order chromatin compaction and transcriptional repression. NURF (containing BPTF) is required for H3K4me3-directed transcription factor binding at active genes by remodeling nucleosomes adjacent to active TSS regions.
INO80 — Histone Variant Exchange
INO80 and SWR1-related complexes exchange canonical H2A/H2B dimers for histone variant H2A.Z at gene promoters and enhancers. H2A.Z-containing nucleosomes are less stable and flank the nucleosome-depleted region at active TSSs, facilitating transcription factor binding. H2A.Z incorporation marks active and poised regulatory elements; its genome-wide distribution is mapped by ChIP-seq as part of ENCODE reference epigenomes.
X-Chromosome Inactivation — Epigenetics at the Chromosomal Scale
X-chromosome inactivation (XCI) is the single most dramatic example of epigenetic gene silencing in mammals — and one of the clearest demonstrations that identical DNA sequences can be maintained in radically different transcriptional states within the same nucleus. In female cells (XX), one X chromosome is transcriptionally silenced in early embryogenesis, reducing gene dosage to equivalence with male (XY) cells. The choice of which X is inactivated is random in most embryonic lineages (though imprinted — always paternal — in extraembryonic tissues), and once established in a given cell, the inactive X (Xi) state is faithfully maintained through all subsequent mitotic divisions. The result is somatic mosaicism: a female is a patchwork of cells expressing either the maternal or paternal X, visible in calico cats (where orange vs. black fur patches reflect X-linked pigmentation gene activity from different Xs).
XIST lncRNA Expression — Initiating Silencing
X-inactivation is initiated by monoallelic upregulation of XIST (X-Inactive Specific Transcript), a ~17 kb long non-coding RNA produced exclusively from the future inactive X chromosome. XIST RNA spreads in cis (remaining associated with the chromosome from which it was transcribed) and coats the entire X chromosome from which it is expressed, making physical contact with chromatin at thousands of sites. XIST is the primary effector of silencing and is essential: cells lacking XIST cannot silence their X chromosomes. On the active X, XIST is repressed by TSIX (an antisense lncRNA), by PRC2 complex-mediated H3K27me3 at the XIST promoter, and by X-activating elements; the asymmetric regulation of XIST is itself controlled by X-encoded counting mechanisms that sense the number of X chromosomes relative to autosomes (the X-to-autosome ratio).
PRC1 and PRC2 Recruitment — Establishing H3K27me3
XIST RNA recruits PRC2 to the Xi through direct RNA-protein interactions (mediated by XIST’s Repeat A domain interacting with SHARP/SPEN, which recruits HDAC3, and Repeat B–F interacting with PRC1/PRC2 components). PRC2 deposits H3K27me3 across the entire inactive X, and PRC1 deposits H2AK119ub1. This early phase is rapid (within hours of XIST upregulation in differentiating ESCs) and produces a chromosome-scale Polycomb domain — the largest Polycomb-repressed domain in the mammalian genome. The Xi also undergoes dramatic histone modification changes: loss of active marks (H3K4me2/3, H3K9ac, H3K36me3) and gain of repressive marks (H3K9me2, H4K20me1, H3K27me3).
DNA Methylation — Locking in Silencing
In a later phase (days to weeks in differentiating cells), CpG islands of Xi-silenced genes become hypermethylated by DNMT3A/3B, converting the initially reversible Polycomb-based silencing into a stable, DNA methylation-reinforced state. This transition is critical for the maintenance of Xi silencing in fully differentiated somatic cells: withdrawal of XIST in differentiated cells is no longer sufficient to reactivate the Xi (as it is in early embryonic cells) precisely because DNA methylation and H3K9me3 at gene promoters have established a self-sustaining repressed state independent of XIST. The Xi is ultimately organised into a heterochromatic body (Barr body) at the nuclear periphery, incorporated into the lamina-associated domains (LADs), and retains the macro-H2A histone variant throughout the cell cycle.
Escape from X-Inactivation — the 15% Exception
Not all X-linked genes are silenced. Approximately 15–25% of human X-linked genes “escape” XCI and are expressed from both Xi and Xa. Escape genes are enriched on the short arm (Xp) and pseudoautosomal regions (PARs), which have homologous regions on the Y chromosome — their biallelic expression in females compensates for their expression from both sex chromosomes in males. Escape from XCI is achieved by the presence of insulators (CTCF binding sites) that block the spread of XIST RNA and heterochromatin into these loci. In cancer, aberrant silencing or reactivation of XCI escape genes contributes to tumour biology. ATRX (an escape gene encoding a chromatin remodeler) is mutated in ~90% of paediatric gliomas and ~65% of alternative lengthening of telomeres (ALT)-positive tumours.
Genomic Imprinting — When Your Parents’ Genes Compete
Genomic imprinting is an evolutionary conflict inscribed in the epigenome. The kinship theory of imprinting (proposed by Haig and colleagues) predicts that paternally expressed genes maximise resource extraction from the mother (favouring offspring growth), while maternally expressed genes limit resource extraction to preserve maternal fitness for future offspring. This conflict is reflected in the biology of imprinted gene function: the imprinted IGF2 gene (paternally expressed, stimulates growth) and its repressor CDKN1C (maternally expressed, restrains growth) represent opposing evolutionary strategies encoded in opposite parental alleles at the 11p15 locus. The approximately 100 imprinted genes in the human genome cluster in ~20 domains, each controlled by an imprinting control region (ICR) — a differentially methylated element that carries parent-of-origin-specific methylation established during gametogenesis and maintained throughout development.
Methylation-Sensitive Insulator — IGF2/H19 Domain (11p15.5)
The ICR between IGF2 and H19 contains CTCF binding sites. When the maternal ICR is unmethylated, CTCF binds and acts as an enhancer insulator — blocking downstream enhancers from reaching the IGF2 promoter, while allowing the same enhancers to drive H19 lncRNA expression. On the paternal chromosome, ICR methylation blocks CTCF binding → enhancers activate IGF2 → paternal IGF2 promotes foetal growth. H19 lncRNA is maternally expressed and acts as a reservoir for miR-675, which represses IGF1R. Beckwith-Wiedemann syndrome (BWS) arises from loss of maternal imprinting at this locus (gain of IGF2 from both alleles or loss of H19 expression) → overgrowth, macroglossia, hemihypertrophy, and increased Wilms tumour risk.
Methylation-Sensitive lncRNA — Prader-Willi/Angelman (15q11-q13)
The PWS/AS locus contains a paternally expressed gene cluster (SNRPN, SNORD115/116 snoRNAs, NDN, MAGEL2) and the maternally expressed UBE3A (E6-AP ubiquitin ligase). In neurons, the paternal SNRPN/UBE3A-ATS lncRNA is expressed and silences UBE3A in cis on the paternal allele — leaving only maternal UBE3A active. Loss of paternal SNRPN cluster (deletion, maternal uniparental disomy, or IC defect) → Prader-Willi syndrome (hypotonia, hyperphagia, cognitive impairment). Loss of maternal UBE3A (deletion, paternal UPD, UBE3A mutation, or IC defect silencing maternal allele) → Angelman syndrome (severe intellectual disability, absent speech, seizures, happy affect). Both are caused by failures of imprinting at the same locus, affecting opposite parental alleles.
Loss of Imprinting (LOI) — Biallelic Expression as Oncogenic Event
Loss of imprinting (LOI) — the reactivation of the normally silenced allele, typically due to ICR hypomethylation — converts a monoallelically expressed gene into a biallelically expressed one, doubling expression of growth-promoting imprinted genes. IGF2 LOI (biallelic expression) is one of the most common epigenetic events in human cancer, found in ~30% of colorectal cancers, as well as in Wilms tumour and many other solid tumours. IGF2 LOI can be detected in normal colorectal mucosa of individuals at increased colorectal cancer risk, making it a potential predictive biomarker. Paternal IGF2 is already normally expressed at high levels; LOI adds a second high-expression allele — creating a constitutive growth stimulus analogous to a gain-of-function oncogenic mutation but without any sequence change.
Gametogenesis — When Imprints Are Written and Erased
Imprints must be erased in the germline (to allow each generation to establish sex-specific marks appropriate to the parent of origin) and re-established anew in oocytes and sperm. Erasure occurs in primordial germ cells (PGCs) by a combination of TET-mediated active demethylation and passive replication-coupled dilution. Re-establishment occurs in a sex-specific manner during gametogenesis: maternal imprints are written in growing oocytes by DNMT3A + DNMT3L (which reads unmethylated H3K4 via PHD domain); paternal imprints are written in prospermatogonia prenatally. The sex-specific imprinting machinery — particularly DNMT3L’s role in guiding DNMT3A — explains why maternal and paternal imprints differ despite the cells using the same DNMT3A enzyme: DNMT3L binds unmethylated H3K4 and the chromatin environment of oocytes versus sperm (reflecting sex-specific histone marks, transcription factor binding, and lncRNA expression) differs profoundly.
Non-Coding RNAs — Guides, Scaffolds, and Silencers
The majority of the human genome is transcribed but does not encode proteins. This non-coding transcriptome — comprising long non-coding RNAs (lncRNAs, >200 nt), small non-coding RNAs (miRNAs, siRNAs, piRNAs, snoRNAs, snRNAs), and circular RNAs — represents a regulatory layer of previously unappreciated depth. Non-coding RNAs operate as epigenetic regulators by guiding chromatin-modifying complexes to specific genomic loci, modulating chromatin accessibility, acting as molecular decoys for RNA-binding proteins, and directing post-transcriptional regulation through complementary base-pairing with target mRNAs. Their discovery transformed epigenetics from a largely protein-centric field to one encompassing the RNA–chromatin interface.
lncRNAs — Chromatin-Targeting Scaffolds and Architectural RNA
Long non-coding RNAs (lncRNAs) act as epigenetic regulators through multiple, sometimes overlapping mechanisms. As scaffolds, they simultaneously bind chromatin-modifying complexes and genomic DNA/RNA, bringing writers and erasers to specific loci: HOTAIR (HOX Antisense Intergenic RNA) is transcribed from the HOXC locus and recruits PRC2 (via its 5′ domain) and the LSD1/CoREST complex (via its 3′ domain) to silence HOXD genes on a different chromosome — one of the first demonstrations of lncRNA-mediated trans epigenetic regulation. XIST, described above, is the most extensively studied scaffold lncRNA. As decoys, lncRNAs sequester transcription factors or miRNAs away from their targets — the ceRNA (competing endogenous RNA) hypothesis proposes that lncRNAs sponge miRNAs, protecting target mRNAs from degradation. As architectural components, lncRNAs participate in nuclear body formation (paraspeckles, NEAT1 lncRNA) and phase-separated chromatin compartments (FIRRE, MALAT1). The ENCODE project catalogued >50,000 human lncRNA genes, though the functional characterisation of the majority remains incomplete.
MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1) is one of the most abundant and ubiquitously expressed lncRNAs, localised to nuclear speckles where it regulates alternative splicing by sequestering serine/arginine (SR) splicing factors. It is upregulated in multiple cancers and promotes metastasis, though whether its primary oncogenic function is epigenetic, splicing-related, or structural remains debated. MEG3 is a maternally expressed imprinted lncRNA with tumour suppressive function; its promoter is hypermethylated and silenced in many cancers, including meningiomas and pituitary adenomas.
miRNAs, piRNAs, and RNA-Directed Chromatin Regulation
- miRNAs (microRNAs, ~22 nt): Processed from primary transcripts (pri-miRNA → pre-miRNA by Drosha/DGCR8 in nucleus → mature miRNA by Dicer in cytoplasm) and loaded into RISC with AGO2. The miRNA seed region (nt 2–8) guides RISC to complementary sequences in the 3’UTR of target mRNAs → translational repression and/or mRNA deadenylation and decay. A single miRNA can target hundreds of mRNAs; ~2,000 human miRNA genes regulate the majority of protein-coding gene expression. miR-21 is the most consistently overexpressed miRNA in cancer (targets PTEN, PDCD4, SPRY2 — tumour suppressor network). miR-34a is p53-induced and pro-apoptotic, targeting BCL2, CDK6, and MYC; its epigenetic silencing by promoter methylation occurs in many cancers, making it an epigenetically regulated tumour suppressive miRNA — a second-order epigenetic circuit.
- piRNAs (PIWI-interacting RNAs, 24–32 nt): Germline-specific small RNAs produced by PIWI-clade Argonaute proteins (PIWIL1/2/3/4). Their primary function is transcriptional and post-transcriptional silencing of transposable elements (LINE-1, Alu, ERVs) in the germline — a critical genome defence mechanism. piRNAs direct DNMT3L-dependent de novo DNA methylation to transposon loci in fetal male germ cells. Failure of piRNA biogenesis causes transposon reactivation, germline DNA damage, and infertility in mice. piRNAs are increasingly recognised in somatic cancer cells, where reactivated LINE-1 elements (due to global hypomethylation) produce piRNAs and contribute to genomic instability.
- siRNAs (small interfering RNAs) and RNA-directed DNA Methylation (RdDM): Well-established in plants, where siRNA guides Argonaute 4 (AGO4) to direct DNMT3 homologue DRM2 to specific genomic loci, establishing de novo DNA methylation at repeats and transposons. Evidence for RdDM in mammals is more limited but exists: endogenous dsRNA and Dicer-processed small RNAs can direct DNMT3A/DNMT3B to specific loci in mammalian cells, contributing to repeat silencing and imprinting establishment. This RNA–DNA methylation axis represents a cross-talk between the PTGS and TGS silencing pathways.
Epigenetic Inheritance — Beyond DNA Sequence Transmission
The concept that environmentally induced epigenetic changes can be transmitted across generations — influencing the biology of descendants who never experienced the original exposure — challenges the Weismann barrier (the strict separation between soma and germline) and the central dogma of information flow (DNA→RNA→protein, unidirectional). Evidence for transgenerational epigenetic effects in humans comes primarily from epidemiological cohorts studying population-level health outcomes following historical exposure events. The mechanisms, however, are still being resolved at the molecular level.
The Dutch Hunger Winter of 1944–45 provides the most extensively characterised human cohort for transgenerational epigenetic effects. Individuals whose mothers were pregnant during the famine showed altered IGF2 DMR methylation, reduced birth weight, increased adult obesity, and higher rates of schizophrenia — effects detectable across multiple decades and, in some analyses, in the grandchildren of exposed mothers.
Heijmans BT et al. (2008), Tobi EW et al. (multiple cohort studies 2009–2018) — DNA methylation changes in individuals periconceptionally exposed to famine, published in PNAS and International Journal of Epidemiology
In C. elegans, histone H3K9me3 marks at germline-expressed genes have been shown to be transmitted through at least three generations via a mechanism involving small RNAs and the nuclear RNAi pathway — independent of DNA methylation, which C. elegans largely lacks. This demonstrates that chromatin-based inheritance through germ cells is mechanistically feasible without DNA methylation as the carrier.
Gaydos LJ et al. (2014), Science; Rechavi O and Lev I (2017), Cell — inheritance of piRNA-directed chromatin silencing in the nematode germline
The mechanistic vehicles proposed for transgenerational epigenetic inheritance in mammals include: (1) DNA methylation at imprinting-resistant loci — specific genomic regions that incompletely erase DNA methylation during the germline reprogramming waves (at PGC specification and after fertilisation); (2) small RNAs in sperm — sperm carry specific pools of miRNAs and tRNA-derived fragments (tRFs) whose composition is altered by paternal diet, stress, and toxin exposure in mouse studies, and which, when injected into fertilised eggs, can phenocopy the paternal exposure in offspring; (3) histone marks in sperm — approximately 4% of sperm DNA remains wrapped in histones (the rest is packaged in protamines), and these retained nucleosomes are enriched at developmental gene promoters bearing H3K4me3 and H3K27me3 bivalent marks; (4) RNA in oocytes — the large cytoplasmic RNA pool of oocytes, including lncRNAs and ribosomal RNA-derived small RNAs, may carry epigenetic information that affects early embryonic gene expression.
Epigenetics in Embryonic Development and Cell Fate
Development is, fundamentally, an epigenetic problem. A single fertilised egg with one genome must generate over 200 distinct, specialised, self-maintaining cell types — each committed to a specific lineage while retaining the capacity to replicate itself indefinitely. The epigenome must be radically reprogrammed twice in the early embryo — a process of genome-wide methylation erasure and re-establishment — and then progressively restricted through development as cells commit to lineages and silence the regulatory networks of alternative fates.
Global DNA methylation dynamics across key stages of human development
The bivalent chromatin state in embryonic stem cells — described above in the context of Polycomb biology — is particularly important for understanding cell fate commitment. In ESCs, thousands of developmental transcription factor genes carry both activating H3K4me3 and repressive H3K27me3 marks at their promoters. These genes are transcribed at very low levels (if at all) in ESCs. Upon differentiation signals, one of two outcomes occurs: if the gene is needed in the lineage, PRC2 is evicted, KDM6A/B demethylates H3K27me3, and the gene becomes fully activated (H3K4me3 only, high expression); if the gene is not needed, TrxG complexes are evicted, KDM5 demethylates H3K4me3, DNMT3A/3B methylate the promoter CpG island, and the gene becomes stably silenced. The transition from bivalency to monovalent active or silenced states is thus a cell fate decision encoded in the epigenome — Waddington’s “epigenetic landscape” made molecular.
Epigenetics Assignments — Expert Academic Support
Working through chromatin immunoprecipitation (ChIP) data interpretation, writing a dissertation on epigenetic reprogramming, or preparing a case study on Angelman syndrome? Our specialist molecular biology and genomics writing team supports every aspect of epigenetics coursework and research writing at every academic level — from first-year essays to PhD thesis chapters.
Cancer Epigenetics — When the Epigenome Becomes Oncogenic
Cancer is both a genetic and an epigenetic disease. The two are not independent: oncogenic mutations frequently target epigenetic regulators directly (DNMT3A, TET2, EZH2, ARID1A, KMT2D, SETD2, IDH1/2, and others are among the most commonly mutated genes across cancer types), while environmental and cell-intrinsic epigenetic changes cooperate with genetic mutations to drive tumour initiation, progression, and drug resistance. The epigenetic hallmarks of cancer — first systematically described by Jones and Baylin — include global genomic hypomethylation, focal promoter CpG island hypermethylation of tumour suppressor genes, altered histone modification landscapes, and reprogrammed non-coding RNA expression.
Proportion of human cancers showing promoter CpG island hypermethylation of at least one tumour suppressor gene — making epigenetic silencing as pervasive as somatic mutation in cancer pathogenesis
According to the National Cancer Institute’s epigenetics fact sheet, aberrant DNA methylation in cancer affects genes controlling cell cycle arrest (CDKN2A/p16), DNA repair (MLH1, BRCA1), apoptosis (DAPK, RASSF1A), differentiation (CDH1/E-cadherin), and angiogenesis regulation — contributing to virtually every hallmark of cancer. The reversibility of these epigenetic changes, unlike somatic mutations, makes them attractive targets for pharmacological intervention and has driven the development of the first approved epigenetic cancer drugs.
Global Hypomethylation → Genomic Instability
Satellite repeats, LINE-1 elements, and Alu sequences — normally silenced by dense methylation — become hypomethylated in cancer cells. Demethylated LINE-1 elements can retrotranspose into new genomic locations, generating insertional mutations. Pericentromeric hypomethylation destabilises chromosome cohesion and contributes to chromosomal instability (CIN). Global hypomethylation of oncogenes (CDH3, MAGE, GAGE, NY-ESO-1 cancer-testis antigens) permits their expression in somatic tissues where they are normally silenced.
Focal Hypermethylation → TSG Silencing
Promoter CpG island hypermethylation silences tumour suppressors including CDKN2A (p16 — cell cycle; silenced in ~40% of all cancers), MLH1 (mismatch repair — causes microsatellite instability in Lynch-like CRC), BRCA1 (DNA repair — sporadic breast/ovarian cancer), VHL (HIF regulation — clear cell RCC), CDH1 (E-cadherin — invasion suppressor; lobular breast cancer, diffuse gastric cancer), RASSF1A, DAPK1, APC. Often provides the “second hit” in Knudson’s model alongside allelic deletion or somatic mutation.
Chromatin Regulator Mutations → Epigenome Reprogramming
Mutations in epigenetic writers, readers, and erasers reprogram the cancer epigenome: IDH1/IDH2 GOF → 2-HG → TET/KDM inhibition → hypermethylator phenotype (glioma, AML, cholangiocarcinoma); EZH2 GOF → H3K27me3 gain → TSG silencing (follicular lymphoma); ARID1A LOF → SWI/SNF dysfunction → altered enhancer accessibility (ovarian CCa, gastric, hepatocellular); H3K27M mutation (paediatric DIPG gliomas) — dominant-negative inhibitor of EZH2 → global H3K27me3 loss → transcriptional de-repression of developmental genes.
Some cancer subtypes are characterised by simultaneous hypermethylation of multiple CpG islands — the CpG Island Methylator Phenotype (CIMP). CIMP-high colorectal cancers (with MLH1 methylation-driven microsatellite instability), CIMP gliomas (driven by IDH1/2 mutations), and CIMP gastric cancers represent distinct molecular subtypes with specific prognosis and drug sensitivities. The mechanistic basis of CIMP often involves IDH mutation (2-HG inhibiting TET), EZH2 overexpression (H3K27me3 pre-marking loci for subsequent methylation), or DNMT3A/3B upregulation by oncogenic transcription factors. Identifying CIMP status through methylation array profiling has become clinically important for tumour subtyping — glioma classification (IDH-mutant vs. IDH-wildtype, 1p/19q codeletion) now requires combined genetic and epigenetic characterisation under WHO 2021 brain tumour classification criteria.
For support with biology research papers covering cancer epigenetics, epigenomics data interpretation, or oncology coursework, our specialist science writing team provides expert custom science writing services across all areas of molecular oncology.
Environmental Epigenetics — How Experience Modifies the Epigenome
The epigenome serves as the biological interface between the genome and the environment — translating chemical, nutritional, psychosocial, and toxicological exposures into heritable changes in gene expression. This interface is why identical twins diverge phenotypically over time and why chronic disease risk is shaped by developmental environment as well as genetic inheritance. Environmental epigenomics studies how exposures alter the methylome, histone modification landscape, and non-coding RNA profile, with particular focus on sensitive developmental windows (periconception, embryogenesis, early childhood, puberty) where epigenetic programming is most plastic and consequences most enduring.
Epigenetic Drugs — Writing Pharmacology Onto the Genome
The reversibility of epigenetic marks — unlike irreversible genetic mutations — provides a pharmacological opportunity unique in cancer biology: epigenetic states can be pharmacologically reset towards normal. The first epigenetic drugs to reach clinical use were DNA demethylating agents (azanucleosides) and HDAC inhibitors, both of which have transformed the treatment of haematological malignancies and expanded the epigenetic drug landscape. The field has since broadened to include inhibitors of EZH2, LSD1/KDM1A, DOT1L, BET bromodomain proteins, and PRMT5 — with dozens more in clinical trials. According to the NCBI Bookshelf’s reference on epigenetic mechanisms, the epigenetic drug pipeline now represents one of the most active areas of oncology drug development, with epigenetic combination strategies — pairing DNMTi with HDACi, or epigenetic drugs with immune checkpoint inhibitors — emerging as the most promising clinical approaches.
5-Azacytidine (Azacitidine) and 5-Aza-2′-deoxycytidine (Decitabine)
Nucleoside analogues that are incorporated into DNA (decitabine) or RNA and DNA (azacitidine) during replication. Once in DNA, they form irreversible covalent adducts with DNMT1 at CpG sites — trapping DNMT1 and preventing it from completing the maintenance methylation cycle. This leads to passive, replication-dependent demethylation of the genome over successive cell divisions, reactivating silenced tumour suppressor genes (particularly p15/CDKN2B and SOCS1 in haematological malignancies). Approved for myelodysplastic syndrome (MDS), acute myeloid leukaemia (AML) ineligible for intensive chemotherapy (azacitidine — VIALE-A trial, improved OS over best supportive care), and chronic myelomonocytic leukaemia (CMML). At low “epigenetic” doses, DNMTi also reactivates endogenous retroviruses (ERVs), producing dsRNA that triggers viral mimicry — innate immune interferon signalling — enhancing tumour immunogenicity and synergising with immune checkpoint inhibitors in preclinical models and early clinical trials.
Vorinostat, Romidepsin, Panobinostat, Belinostat, Entinostat
HDAC inhibitors act by chelating the active-site zinc ion in HDAC enzymes, preventing substrate lysine binding. Pan-HDAC inhibitors block all zinc-dependent HDACs; isoform-selective inhibitors target specific classes. Re-acetylation of histones at tumour suppressor gene promoters re-opens chromatin and restores expression of pro-apoptotic (BIM, NOXA, PUMA), anti-proliferative (p21, p27), and differentiation genes — while simultaneously acetylating non-histone targets (p53, Hsp90, α-tubulin). Approved agents: vorinostat/SAHA (cutaneous T-cell lymphoma, CTCL), romidepsin (CTCL, peripheral T-cell lymphoma), panobinostat (multiple myeloma — with bortezomib + dexamethasone in PANORAMA trials), belinostat (PTCL). Entinostat (HDAC1/3-selective) plus exemestane improved OS in ER+ breast cancer in phase III (E2112 trial — results mixed). HDAC inhibitor toxicity is class-related: fatigue, thrombocytopenia, nausea, QTc prolongation, and (for romidepsin) cardiac arrhythmia in patients with prior cardiac disease.
Tazemetostat (EPZ-6438)
Tazemetostat is a selective, competitive inhibitor of EZH2 (the PRC2 catalytic subunit), blocking S-adenosylmethionine binding and preventing H3K27 methylation. It is FDA-approved for EZH2-mutant (Y641, A686, A687) relapsed/refractory follicular lymphoma (based on the E7438-NH-101 trial) and for metastatic or locally advanced epithelioid sarcoma (where SMARCB1 loss creates EZH2 dependency — SWI/SNF loss sensitises tumour cells to EZH2 inhibition through synthetic lethality). Tazemetostat is the first approved inhibitor of a histone methyltransferase. Resistance mechanisms include EZH1 compensation (EZH1 can substitute for EZH2 in PRC2 and has reduced tazemetostat sensitivity), and co-occurring ARID1A mutations may modulate response. MAZe Therapeutics’ MTMT-001 and Constellation Pharmaceuticals’ CPI-0209 are next-generation EZH2 inhibitors in clinical trials with improved potency or combined EZH1/EZH2 inhibition profiles.
JQ1 (Research), OTX015, BMS-986158 (Clinical)
BET (Bromodomain and Extra-Terminal) proteins — BRD2, BRD3, BRD4, and BRDT — are epigenetic “readers” that bind acetylated histones via their bromodomains and recruit transcriptional elongation machinery (P-TEFb) to enhancers and super-enhancers driving expression of oncogenes including MYC, BCL2, and BCL6. BET inhibitors (JQ1 was the founding research compound; clinical candidates include OTX015/birabresib, BMS-986158, ZEN-3694) competitively displace BET proteins from acetylated chromatin, selectively suppressing transcription at super-enhancers — which are disproportionately sensitive to BET inhibition compared to typical enhancers, due to their requirement for high BRD4 occupancy. BET inhibitors suppress MYC expression in multiple cancer types and show activity in NUT carcinoma (BRD4-NUT fusion driven), AML, and some lymphomas. Clinical development has been complicated by on-target toxicity (anaemia, thrombocytopenia, GI toxicity) and rapid acquired resistance through BRD4 amplification or bromodomain mutations.
Iadademstat, Seclidemstat, INCB059872
LSD1 (Lysine-specific demethylase 1, KDM1A) is a FAD-dependent amine oxidase that demethylates H3K4me1/2 (at enhancers, causing their decommissioning) and H3K9me1/2 (at certain targets, affecting heterochromatin). LSD1 is overexpressed in AML, small cell lung cancer (SCLC), and many solid tumours, where it maintains stem cell gene expression programmes by demethylating and decommissioning enhancers of differentiation genes. LSD1 inhibitors include irreversible (tranylcypromine-based, iadademstat) and reversible (seclidemstat) mechanisms. In AML, LSD1 inhibition in combination with ATRA (all-trans retinoic acid) restores differentiation — a strategy analogous to the AML-M3 ATRA+arsenic paradigm. In SCLC, LSD1 inhibition disrupts the neuroendocrine transcription factor network that maintains the SCLC phenotype, inducing partial neuronal differentiation and sensitisation to other therapies.
Ivosidenib (IDH1), Enasidenib (IDH2), Olutasidenib
IDH1/2 mutations in cancer produce the oncometabolite 2-hydroxyglutarate (2-HG), which competitively inhibits α-ketoglutarate-dependent dioxygenases — including TET enzymes (causing DNA hypermethylation) and KDM histone demethylases (causing histone hypermethylation). Selective inhibitors of mutant IDH1 (ivosidenib, olutasidenib) and IDH2 (enasidenib) reverse 2-HG production, allowing TET and KDM re-activation and partial epigenome reprogramming — restoring the ability of AML blasts to differentiate. These are approved for IDH-mutant AML and cholangiocarcinoma. The clinical lesson: targeting an upstream oncometabolite reverses downstream epigenetic dysfunction — demonstrating that the metabolic and epigenetic landscapes of cancer are intimately interconnected through the shared currency of one-carbon and TCA cycle intermediates.
Epigenomics — Mapping the Full Epigenetic Landscape
Just as the genome is the complete set of DNA sequences in a cell, the epigenome is the complete set of epigenetic modifications across the genome — DNA methylation, histone modifications, chromatin accessibility, and three-dimensional chromatin organisation — and it differs between cell types even when the genome is identical. Epigenomics is the field that maps and integrates these layers at genome-wide scale, typically using next-generation sequencing technologies coupled to chromatin biochemistry. The ENCODE (Encyclopedia of DNA Elements) Project and Roadmap Epigenomics Mapping Consortium have produced reference epigenomes across hundreds of human cell types, while the International Human Epigenome Consortium (IHEC) has standardised reference epigenomes for 50 cell types globally.
ChIP-seq — Mapping Histone Marks and TF Binding
Chromatin immunoprecipitation followed by sequencing (ChIP-seq) uses antibodies to immunoprecipitate chromatin associated with a specific histone mark or protein. The pulled-down DNA is sequenced and mapped to the genome, revealing the genome-wide distribution of that mark or factor. ChIP-seq for H3K4me3 identifies active promoters; H3K27ac identifies active enhancers; H3K27me3 identifies Polycomb domains. ATAC-seq (Assay for Transposase-Accessible Chromatin using sequencing) maps open chromatin genome-wide using Tn5 transposase, which preferentially inserts into accessible nucleosome-depleted regions. Single-cell variants (scChIP-seq, scATAC-seq) resolve cell-to-cell epigenomic heterogeneity within tumours and developmental systems.
Bisulfite Sequencing — DNA Methylation at Base Resolution
Sodium bisulfite converts unmethylated cytosines to uracil (read as thymine after PCR) while leaving 5-methylcytosines unchanged — enabling distinction of methylated from unmethylated CpGs at single-base resolution. Whole Genome Bisulfite Sequencing (WGBS) covers all ~28 million CpGs genome-wide. Reduced Representation Bisulfite Sequencing (RRBS) enriches CpG-dense regions (promoters, CpG islands) at lower cost. MBD-seq and RRBS are used clinically for methylation biomarker detection. Oxford Nanopore long-read sequencing can directly detect 5mC without bisulfite conversion — enabling phased methylation analysis that links methylation state to specific alleles and haplotypes, crucial for imprinting and X-inactivation studies.
Hi-C and 3D Epigenomics
Hi-C (genome-wide chromosome conformation capture) maps three-dimensional chromatin organisation by crosslinking, fragmenting, ligating, and sequencing chromatin contacts. The genome is organised into topologically associating domains (TADs) — self-interacting chromatin compartments of 200 kb to 1 Mb that insulate regulatory interactions (enhancers act predominantly within their TAD). TAD boundaries are maintained by CTCF and cohesin occupancy. In cancer, CTCF binding site deletions can dissolve TAD boundaries, allowing inappropriate enhancer-promoter contacts that activate proto-oncogenes — as in T-cell ALL (TAL1 and LMO2 activation by ectopic enhancer contacts) and medulloblastoma (GFI1/GFI1B activation). Hi-C integration with ChIP-seq and ATAC-seq data is now standard in comprehensive tumour epigenome analysis.
Need specialist help with epigenetics assignments, data interpretation, literature reviews, or research papers? Explore our dedicated support: biology assignments · biology research papers · literature reviews · science writing · data analysis · dissertations and theses · complex technical assignments · biostatistics · research paper writing
