HDACs are enzymes found in virtually all eukaryotic cells. Their principal job is to remove acetyl chemical groups from histones, the spool-like proteins around which DNA is wound. When histones are heavily acetylated, chromatin is loose and genes can be read; when HDACs strip those acetyl groups away, chromatin condenses and gene transcription is silenced. The enzymes also modify non-histone proteins, influencing protein stability and activity directly. Based on their similarity to yeast proteins, HDACs are sorted into four classes: Classes I, II, and IV are zinc-dependent enzymes, while Class III, the sirtuins, depend on the metabolic cofactor NAD+. Class I members HDAC1, 2, 3, and 8 operate within multiprotein complexes and have been linked to cancer, inflammation, and infection. Class II enzymes shuttle between nucleus and cytoplasm, with HDAC6 in particular targeting non-histone substrates such as alpha-tubulin and molecular chaperones, thereby regulating autophagy and cellular architecture.

What the new review makes clear is that HDAC family members are not uniformly disrupted in endometriosis; instead, each subtype follows its own pattern, and those patterns appear to map onto different stages and features of disease. HDAC1 is broadly upregulated. Studies of endometriotic stromal cell lines show elevated HDAC1 mRNA and protein compared with normal cells, and immunohistochemistry of human lesions reveals strong staining in ovarian, skin, and gastrointestinal implants, with ovarian lesions showing the highest levels. Experiments with exosomes derived from ovarian endometrioma tissue found that the long non-coding RNA HOTAIR upregulates HDAC1 by suppressing microRNA-761, and that these exosomes promote ectopic lesion growth in mice. Genetic work has even linked HDAC1 single nucleotide polymorphisms to endometriosis risk in South Indian women, and in mouse models HDAC1 staining intensity rises as lesions progress. In deep infiltrating endometriosis, the most aggressive form, HDAC1 levels correlate positively with the degree of fibrosis.

HDAC2 tells a more complicated story. Its expression is elevated in some cell models and in skin lesions and eutopic endometrial stroma, yet lower in ovarian and peritoneal implants, and one study found HDAC2 staining actually declines as mouse lesions progress while being significantly lower in deep infiltrating disease than in ovarian endometrioma. Functionally, however, silencing HDAC2 in endometriotic cells activates the HNF4A/ARID1A axis, curbing proliferation and invasion while promoting apoptosis, and reduces lesion size in animals. HDAC3, by contrast, is downregulated, particularly in women with infertility. In baboon models of the disease, HDAC3 expression falls as disease advances, and mice lacking HDAC3 specifically in the uterus show implantation failure and defective decidualization, tied to de-repression of collagen genes and impaired progesterone signaling. HDAC8, meanwhile, is upregulated and strongly linked to fibrosis: a specific HDAC8 activator accelerated lesion growth and scarring in mice, while the inhibitor PCI-34,051 dose-dependently suppressed lesions, eased fibrosis, and reduced pain.

Among the Class II and III enzymes, HDAC6 stands out for its dependence on lesion subtype and cellular context. A tissue microarray analysis of 168 endometriotic lesions confirmed that HDAC6 is significantly elevated in both epithelium and stroma of deep infiltrating disease and positively correlated with fibrosis, but unchanged in ovarian endometrioma. In cases lacking ARID1A expression, epithelial HDAC6 was higher still, suggesting a possible route toward malignant transformation. On the sirtuin side, SIRT1 is persistently expressed throughout the menstrual cycle in patients, whereas in healthy endometrium it appears only during menstruation. Mice engineered to overexpress Sirt1 in the uterus show implantation failure and downregulation of progesterone target genes, and the SIRT1/BCL6 axis has been implicated in the progesterone resistance that characterizes the disease. SIRT3, which governs mitochondrial energy metabolism and antioxidant defense, is reduced in ovarian endometrioma tissue, while SIRT7 is upregulated in ectopic lesions and drives a metabolic shift from mitochondrial respiration toward glycolysis through the TUFM/SIRT7 axis, promoting proliferation and lesion growth in mice.

These expression anomalies feed into five interconnected pathological pathways. HDACs promote transcription of pro-inflammatory cytokines such as TNF-alpha, IL-1 beta, and IL-6, and weaken immune surveillance that would otherwise clear ectopic cells; exosomal miR-22-3p from peritoneal macrophages, for example, suppresses SIRT1 and activates the pro-invasive NF-kappaB pathway. HDAC1 and HDAC2 silence cyclin-dependent kinase inhibitors such as p21 and p57, pushing cells through the G1/S checkpoint, while SIRT1 upregulation is linked to reduced p53-mediated apoptosis. Invasion and migration are enhanced through epithelial-mesenchymal transition: HDAC1 suppresses RHOB, HDAC2 modulates the HNF4A/ARID1A axis, and HDAC6 remodels the cytoskeleton via alpha-tubulin deacetylation. HDAC1 also activates STAT3 to promote angiogenesis, and under hypoxia HDACs bind HIF-1 to drive adhesion, proliferation, and vessel-forming genes. Finally, HDACs regulate aromatase and estrogen receptor alpha, boosting local estrogen, and mediate epigenetic silencing of progesterone receptor B, the molecular basis of progesterone resistance.

The therapeutic logic follows directly. Histone deacetylase inhibitors, already established as anti-cancer drugs, can restore acetylation and reactivate silenced genes, and because epigenetic modifications are reversible, a single inhibitor can theoretically correct multiple pathological programs at once. Preclinical evidence is accumulating rapidly. Vorinostat, or SAHA, upregulated thioredoxin-binding protein-2 in endometrial stromal cells and promoted apoptosis through the TRX/TBP-2 system; in a mouse model of endometriosis-associated ovarian cancer it reduced tumor volume by inhibiting M2 macrophage polarization and lowering IL-10. Romidepsin, a Class I-specific inhibitor, inhibited HDAC activity in endometriotic epithelial cells with an IC50 of 6.5 nanomolar, induced histone acetylation, upregulated p21, and triggered caspase-dependent apoptosis. Trichostatin A reduced lesion area by 54.3 percent and relieved hyperalgesia in mice, suppressed NF-kappaB activation in endometriotic cells, which proved more than ten times more sensitive than normal cells, and induced apoptosis via NAG-1 in primary stromal cells.

Perhaps the most unexpected player is butyrate, a short-chain fatty acid produced by gut bacteria. Mouse studies showed endometriosis depletes butyrate-producing microbiota and colonic butyrate levels; fecal microbiota transfer confirmed that stool from diseased mice promotes lesion growth while normal stool inhibits it. Mechanistically, butyrate suppresses endometrial cell proliferation by activating GPR43 and GPR109A, inhibiting HDAC activity, and upregulating RAP1GAP, and it enhances ferroptosis sensitivity of ectopic cells through the FFAR2/PPAR-gamma/PINK1/Parkin axis. Serum butyrate in patients is reduced and inversely correlated with disease stage. Valproic acid, a widely used antiepileptic, also inhibits HDACs: in rats it dose-dependently shrank lesions and eased pain, and in human cells it silenced the CYP19 aromatase promoter, cutting estrogen synthesis, while reactivating the tumor suppressor C/EBPalpha. Selective HDAC8 inhibition with PCI-34,051 in a deep infiltrating disease model cut lesion weight by nearly two-thirds, reduced fibrosis, and relieved pain, hinting that subtype-selective drugs could avoid the off-target toxicity of pan-inhibitors.

Formidable obstacles remain before any of this reaches the clinic. No HDAC inhibitor is approved for endometriosis, and the known safety profile from oncology, including myelosuppression, thrombocytopenia, gastrointestinal symptoms, fatigue, and QT prolongation, is a serious concern for long-term use in benign disease. Reproductive safety is the thorniest issue: HDACs regulate gametogenesis, follicular development, implantation, and endometrial receptivity. Valproic acid carries a well-documented teratogenic risk of roughly 10 percent congenital malformation rate with pregnancy exposure, and embryonic exposure to butyrate or valproate can impair primordial germ cell development in mice. The review’s authors therefore call for subtype-selective inhibitors guided by single-cell expression maps, lesion-targeted delivery systems such as nanoparticles or exosomes to limit systemic exposure, combination strategies with hormonal, anti-inflammatory, or anti-angiogenic drugs, and biomarkers such as histone H3 acetylation states and circulating exosomal non-coding RNAs to stratify patients. If those pieces come together, epigenetic therapy could open the first genuinely new front against endometriosis in decades.

Subject of Research: The role of histone deacetylases and their inhibitors in the pathogenesis and treatment of endometriosis

Article Title: Research Progress of Histone Deacetylase and Its Inhibitors in Endometriosis

Article References: Xiong, S., Xu, J., Pu, J., Zhao, S., & Jiang, L. (2026). Research Progress of Histone Deacetylase and Its Inhibitors in Endometriosis. Reproductive Sciences. https://doi.org/10.1007/s43032-026-02184-7

Image Credits: AI Generated

DOI: 10.1007/s43032-026-02184-7

Keywords: endometriosis, histone deacetylase, HDAC inhibitors, epigenetics, vorinostat, romidepsin, trichostatin A, butyrate, valproic acid, fibrosis, progesterone resistance, reproductive health

One of the most striking conceptual advances came from Paola Scaffidi of the European Institute of Oncology in Italy, who introduced a systems-level framework for understanding epigenetic regulation in cancer. She described the epigenetic regulatory network, or ERN, as the collective assembly of epigenetic modifications that drive cellular states, with individual regulators ranging from dispensable to absolutely essential for cell survival. In healthy cells, this network is built with substantial functional redundancy, so the loss of one or a few components is generally tolerated through compensatory mechanisms. But when oncogenic drivers are added to the mix, the network becomes fragile. Scaffidi presented evidence that cancer cells globally lose roughly 30 percent of their epigenetic regulators, producing aberrant transcriptional responses to stress and an enhanced adaptive capacity compared with normal cells. Many of these losses are subclonal and take on stage-specific roles as tumours evolve, a finding that reframes epigenetic instability not simply as collateral damage but as a potential vulnerability that therapies could exploit.

Chromatin dynamics occupied centre stage throughout the meeting. Karl Ekwall of the Karolinska Institute in Sweden explored how chromatin-modifying enzyme complexes govern cell cycle kinetics in yeast and human fibroblast models, identifying these complexes as essential regulators of quiescence induction when cells face nutrient deprivation. Notably, in vitro analyses showed that targeting these complexes with ouabain, an inhibitor of the LEO1 component, altered LEO1 expression in tumour models but not in normal human fibroblasts, hinting at a therapeutic window for oncology. Geneviève Almouzni of the Curie Institute in France closed the conference with a keynote on chromatin integrity and histone variants, explaining how the replicative histone H3.1, enriched at late-replicating regions, and the variant H3.3, enriched at actively transcribed early-replicating regions, demarcate replication initiation zones with remarkable precision. Her presentation detailed how aberrant histone H3 function, whether through mutation of the protein itself or of chaperone proteins such as DAXX/ATRX, ASF1, CENP-A, HJURP and HIRA, contributes to cancer and other diseases, including evidence that restoring the chaperone HIRA can rescue defective H3.3 deposition patterns.

Cancer epigenetics featured prominently, with complementary sessions probing histone mutations, DNA methylation, transcription factors, chromatin topology and even the tumour microbiome. Jonathan Licht of the University of Florida described the contribution of histone fold mutations, which a pan-cancer analysis found in approximately 7 percent of patients, particularly in bladder, oesophageal, head and neck, and skin cancers. The most common of these, H2B E76K, destabilises the H2B/H4 interface, accelerating chromatin degradation into oligonucleosomes and freeing H2A-H2B dimers from the histone octamer. In lung epithelial cell models, this opens up chromatin and ramps up signalling through polycomb-repressed regions, epithelial-mesenchymal transition pathways, and AKT and c-Jun cascades, all of which can promote migration, proliferation, cytokine secretion and resistance to apoptosis.

DNA methylation emerged as a versatile player in cancer progression and diagnosis. Aniruddha Chatterjee of the University of Otago in New Zealand presented work on colorectal cancer showing that conserved aberrations in the DNA methylome discriminate powerfully between matched primary tumours, lymph node metastases and liver metastases. This matters because metastasis drives 90 percent of cancer deaths, yet no purely genetic drivers can fully explain successful metastatic colonisation; epigenetic plasticity appears to supply the adaptability tumours need. His team is using CRISPR-dCas9-based epigenetic editing at key regulatory loci to test whether methylation changes are cause or consequence. On the translational front, early findings suggest DNA methylome profiling of cell-free DNA, combined with artificial intelligence models, could underpin liquid biopsies for colorectal cancer screening. Susan Clark of the Garvan Institute of Medical Research in Australia widened the lens to the three-dimensional genome, describing how mutations at ‘persistent’ CTCF binding sites, which anchor loops and topologically associated domain borders, recur in prostate and breast cancers and may disrupt higher-order chromatin architecture.

In one of the more unexpected twists of the meeting, Maria Rescigno of Humanitas University in Italy connected the microbiome to epigenetics in cancer. Progressive microbial deregulation accompanies tumour development, and in mouse models the loss of the bacterium Faecalibaculum rodentium was reversed by restoring the organism, which reduced tumour growth and number. Intriguingly, F. rodentium produces butyrate as it proliferates, acting as a histone deacetylase inhibitor that epigenetically modulates apoptosis. A phylogenetically similar strain is underrepresented in human advanced adenomas, and in colorectal cancer, intra-tumoural bacteria appear to modulate treatment response through soluble metabolites, or ‘postbiotics’. In the context of anti-PD-1 checkpoint blockade, different bacterial strains produce postbiotics with different effects on HLA class I expression, suggesting postbiotic therapy could help overcome treatment resistance.

Beyond cancer, the meeting showcased epigenetics in ageing and non-malignant disease. Steve Horvath of Altos Labs in the United Kingdom traced the evolution of epigenetic clocks, machine learning algorithms built on DNA methylation patterns at CpG dinucleotides that estimate chronological or biological age. First-generation clocks estimate chronological age, second-generation clocks predict clinical phenotypes and mortality risk, and third-generation clocks aim for multi-species utility. GrimAge, designed to predict time to death, showed predictive value for chronic obstructive pulmonary disease, type 2 diabetes and ischaemic heart disease over 13 years of follow-up. Andrea Fuso of Sapienza University of Rome demonstrated cross-talk between DNA methylation and non-coding RNAs in Alzheimer’s disease, where methylation modulates PSEN1 and miR-29a, which in turn targets BACE1 and the demethylase gene TET1, feeding into amyloid plaque deposition and linking B vitamin status and one-carbon metabolism to neurodegeneration.

Neurodegeneration and cardiometabolic disease drew further epigenetic links. Eran Meshorer of The Hebrew University of Jerusalem showed that organoid models of early Huntington’s disease display global hypomethylation, altered DNMT3A and DNMT3B signatures and methylation ages older than controls. Counterintuitively, the polyglutamine inclusion bodies characteristic of the disease proved protective, with aggregate-forming cells dying at one third the rate of others; the transcription factor ATF3 was implicated, since its knockout prevented inclusion body formation. Melanie Waldenberger of Helmholtz Munich and Marie Loh of Nanyang Technological University in Singapore presented epigenome-wide studies of dyslipidaemia and cardiovascular risk, respectively, both emphasising ethnic diversity. In the Singapore-based HELIOS study of roughly 50,000 adults, 1,926 CpG sites were associated with carotid intima-media thickness in Asian individuals, 91 percent of them irrelevant in European populations, with two sentinel sites showing possible causal roles, underscoring the danger of extrapolating epigenomic findings across ancestries.

The therapeutic frontier was arguably the most electrifying territory of all. José Sardina of the Josep Carreras Leukaemia Research Institute in Spain used targeted DNA methylation to hypermethylate the IL1RN promoter, generating macrophages with aberrant inflammatory responses and transcriptional profiles resembling tumour-resident macrophages. Angelo Lombardo of the San Raffaele-Telethon Institute for Gene Therapy in Italy described an all-in-one zinc finger epigenetic editor delivered by lipid nanoparticles that silenced PCSK9 in vivo for nearly a year, persisting even through liver regeneration, outperforming CRISPR-dCas9 constructs with minimal off-target effects. Related platforms are being developed for durable silencing of hepatitis B virus and, prospectively, CAG length-specific silencing in Huntington’s disease. A round table debated the roadmap to the clinic, highlighting needs for standardised delivery, early regulatory engagement, quality control and stakeholder inclusion, including patients and low- and middle-income countries.

Clinical trials and drug discovery sessions reinforced the momentum. Christophe Le Tourneau of the Curie Institute presented the PEVO basket trial of the HDAC inhibitor vorinostat combined with immunotherapy in recurrent or metastatic squamous cell carcinomas, reporting a median response rate of 26 percent, highest in anal and cervical tumours, with better responses linked to microsatellite instability, HPV positivity and high PD-L1 scores. Emily Dykhuizen of Purdue University described SWI/SNF subcomplex targeting, including BRD9 degraders showing anti-tumour activity through macrophage activation and ARID1A inhibition flipping cold tumours toward immune-infiltrated hot states. Cheryl Arrowsmith of the University of Toronto outlined chemical probe screening revealing context-dependent EZH2 vulnerabilities, while Christoph Bock of CeMM in Austria closed with a vision of epigenetic priming, in which ‘alertness drugs’ could pre-emptively harden innate immunity against infection, and of AI tools such as CellWhisperer that let researchers interrogate sequencing data in plain English. Together, the meeting made clear that epigenetics, once a purely academic curiosity, is now positioned to reshape diagnostics and therapy across medicine.

Subject of Research: Recent advances and clinical opportunities in epigenetics research presented at the 4th Clinical Epigenetics International Conference

Article Title: Clinical epigenetics: recent advances and opportunities

Article References: Smith, J., Chatterjee, A., & Rodger, E. J. (2025). Clinical epigenetics: recent advances and opportunities. Epigenetics Communications, 5(1), Article 7. https://doi.org/10.1186/s43682-025-00038-y

Image Credits: AI Generated

DOI: 10.1186/s43682-025-00038-y

Keywords: clinical epigenetics, epigenetic editing, DNA methylation, chromatin, cancer epigenetics, epigenetic clocks, epigenetic therapeutics, HDAC inhibitors, CRISPR, microbiome, Alzheimer's disease, Huntington's disease

The underlying idea, known as transcriptomic reversal, is deceptively simple. If a disease leaves a characteristic fingerprint in gene expression, then a drug whose own expression perturbation profile opposes that fingerprint might counteract the disease state. The approach has powered countless drug-repurposing screens since the advent of the Library of Integrated Network-Based Cellular Signatures (LINCS), but the authors argue that the field’s reliability hinges on details that are too often glossed over: rigorous control of molecular identity, evaluation schemes that prevent information leakage between training and test data, chemical-space assessment, and a clear separation between predicted signatures and actually measured ones.

To address these concerns, the team trained two computational models on 55,695 quality-controlled LINCS L1000 Level 5 signatures. The first was a dual-stream architecture that processed atom-level tokens and molecular fingerprints in parallel, reflecting the current enthusiasm for richer chemical representations in machine learning. The second was a deliberately conventional comparator: a multilayer perceptron operating on standard molecular fingerprints. The comparison turned out to be one of the study’s most sobering findings. Across drug-cell pair, leave-drug-out, leave-cell-line-out, and scaffold-based evaluation settings, the fingerprint MLP was on average slightly better than the dual-stream model, and the fancier architecture delivered no measurable performance gain.

The evaluation design deserves particular attention because it embodies the leakage-aware philosophy at the heart of the paper. Rather than relying on a single random split, the researchers repeated performance assessments across multiple random seeds and multiple holdout strategies, including a corrected annotation-defined HDAC holdout comprising 1,856 profiles from 30 structures never seen during training. In that challenging setting, mean Pearson correlations between predicted and observed perturbation profiles were 0.378 for the dual-stream model and 0.379 for the fingerprint MLP, while mean Spearman correlations were 0.345 and 0.344, respectively. Strict candidate-level leave-drug evaluation, available for the compounds Mocetinostat and PCI-24781, again did not favor the more complex model. The authors conclude that the benchmarks do not justify the additional complexity and computational cost of the dual-stream representation, a result with real practical implications for groups deciding how to allocate modeling resources in perturbation biology.

With the models validated, the team turned to the actual repurposing screen, deploying the chemical-structure-only models against disease signatures from 22 TCGA cancer types to rank 28,477 compounds. Candidate stability, measured across 48 combinations of split, model, seed, and metric, was treated as the primary criterion, with a legacy metric extending the analysis to 72 configurations purely as a sensitivity check. This emphasis on stability rather than any single ranking reflects a growing recognition that repurposing pipelines can be exquisitely sensitive to arbitrary analytical choices.

The headline result concerns class I HDAC inhibitors. Using a signed weighted transcriptomic reversal score (wTRS), the analysis enriched the explicitly annotated class I HDAC subset at fixed revision cutoffs of the top 0.5, 1, 5, and 10 percent of the compound library, with fold enrichments of 30.6, 19.2, 8.46, and 4.62 respectively, all with false discovery rates below ten to the minus four. Notably, the co-primary Spearman reversal metric did not reproduce this enrichment. The authors are candid about why: signed wTRS is sensitive to perturbational amplitude, whereas Spearman correlation is scale invariant, so the observed enrichment may partly reflect response magnitude rather than purely directional reversal. This kind of metric-level honesty, they argue, is exactly what the field needs if reversal-based prioritization is to be trusted.

To distinguish genuine signal from artifact, the researchers compared predicted reversal against measured LINCS profiles, using official perturbagen, dose, time, cell-line, and quality annotations. Across eight compounds with official high-quality measured profiles and the 22 cancer signatures, predicted and measured reversal were concordant for both models, with Spearman correlations ranging from 0.640 to 0.830 and crossed-bootstrap lower 95 percent confidence limits between 0.297 and 0.653 depending on the model and metric. This predicted-to-measured concordance layer provides independent reassurance that the models were not merely generating internally consistent but biologically empty scores.

After candidate tiers were frozen, the team conducted an extensive post-hoc audit spanning identity verification, formal HDAC enrichment testing, structural-neighbor exposure, reversal-associated networks, DepMap dependency analysis, crystallographic redocking, a zinc-chelation decoy, and receptor sensitivity assessment. Mocetinostat emerged as the core candidate, with NCH-51 as a secondary candidate and TC-H-106 flagged as exploratory. Two exclusions illustrate the value of the auditing discipline. RG2833 lacked a measured LINCS signature, preventing independent validation, and Tianeptinaline, also known as BG-1010, had an identity conflict that excluded it from primary inference altogether. In an era when compound databases routinely contain ambiguous or duplicated entries, such identity control is not pedantry; it is a prerequisite for reproducible conclusions.

The biological context layers added further nuance. Analysis of the Cancer Dependency Map (DepMap) supported HDAC3, rather than HDAC1, as the dominant pan-cancer dependency among class I HDACs, sharpening the mechanistic hypothesis that the prioritized inhibitors act on. On the structural side, zinc-aware redocking with AutoDock4Zn successfully recovered the crystallographic binding pose of the HDAC inhibitor Vorinostat, but a deliberately designed decoy demonstrated that favorable docking scores alone do not establish the zinc-chelating geometry essential to true HDAC inhibition. In other words, even the structural evidence was treated as one calibrated layer among many, never as a standalone confirmation.

The authors are careful to frame the conclusions as hypothesis-generating rather than therapeutic. The framework, they emphasize, separates prediction generalization, predicted-to-measured transcriptomic concordance, biological context, and structural sensitivity without implying metric-independent class enrichment, direct target engagement, or clinical efficacy. What the study offers instead is a template: rigorous split design, systematic leakage and structural-proximity auditing, and layered evidence that can be inspected step by step. For a field where flashy deep learning architectures often outpace validation, the message that a well-tuned conventional fingerprint model can match or beat a dual-stream neural network, while a disciplined audit pipeline does the heavy lifting for credibility, may prove to be the most transferable finding of all. As transcriptomic reversal continues to feed candidate lists into experimental labs worldwide, this work raises the bar for how such lists should be built, benchmarked, and honestly reported.

Subject of Research: A leakage-aware machine learning framework for prioritizing class I HDAC inhibitors through pan-cancer transcriptomic reversal analysis

Article Title: A leakage-aware and auditable framework prioritizes class I HDAC inhibitors for pan-cancer transcriptomic reversal

Article References: Tong, S., Zhang, W., & Ji, S. (2026). A leakage-aware and auditable framework prioritizes class I HDAC inhibitors for pan-cancer transcriptomic reversal. BMC Bioinformatics. https://doi.org/10.1186/s12859-026-06650-6

Image Credits: AI Generated

DOI: 10.1186/s12859-026-06650-6

Keywords: transcriptomic reversal, drug repurposing, LINCS L1000, TCGA, HDAC inhibitors, leakage-aware evaluation, Mocetinostat, machine learning, BMC Bioinformatics, DepMap, cancer transcriptomics, molecular fingerprints