HDAC3 prevents enhancer hyperactivation to enable developmental transitions

HDAC3 prevents enhancer hyperactivation to enable developmental transitions | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results HDAC3 prevents enhancer hyperactivation to enable developmental transitions View ORCID Profile Nikolaos Stamidis , View ORCID Profile Anne Wenzel , View ORCID Profile Konrad Kacper Uściło , View ORCID Profile Smaragda Kompocholi , View ORCID Profile Sandra Bages-Arnal , View ORCID Profile Gemma Noviello , View ORCID Profile Lea Haarup Gregersen , View ORCID Profile James Alexander Hackett , View ORCID Profile Jan Jakub Żylicz doi: https://doi.org/10.1101/2025.08.09.669452 Nikolaos Stamidis 1 Novo Nordisk Foundation Center for Stem Cell Medicine – reNEW, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Nikolaos Stamidis Anne Wenzel 1 Novo Nordisk Foundation Center for Stem Cell Medicine – reNEW, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Anne Wenzel Konrad Kacper Uściło 1 Novo Nordisk Foundation Center for Stem Cell Medicine – reNEW, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Konrad Kacper Uściło Smaragda Kompocholi 2 Center for Gene Expression (CGEN), Institute for Cellular and Molecular Medicine, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Smaragda Kompocholi Sandra Bages-Arnal 1 Novo Nordisk Foundation Center for Stem Cell Medicine – reNEW, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sandra Bages-Arnal Gemma Noviello 3 Epigenetics and Neurobiology Unit, European Molecular Biology Laboratory ( EMBL) , Rome, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Gemma Noviello Lea Haarup Gregersen 2 Center for Gene Expression (CGEN), Institute for Cellular and Molecular Medicine, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Lea Haarup Gregersen James Alexander Hackett 3 Epigenetics and Neurobiology Unit, European Molecular Biology Laboratory ( EMBL) , Rome, Italy Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for James Alexander Hackett Jan Jakub Żylicz 1 Novo Nordisk Foundation Center for Stem Cell Medicine – reNEW, Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen , Copenhagen, Denmark Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jan Jakub Żylicz For correspondence: jan.zylicz{at}sund.ku.dk Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Dynamic gene regulation requires precise cooperation between transcription factors and chromatin modifiers at regulatory elements to achieve not only activation or repression, but also appropriate transcript dosage. However, the molecular mechanisms that ensure developmental genes are transcribed at specific levels remain largely unknown. Here, we discover that the epigenetic repressor histone deacetylase 3 (HDAC3), together with co-activators, binds a subset of the most active putative enhancer elements. Using a tunable mouse embryonic stem cell degron system, we uncover that HDAC3 directly prevents enhancer—and consequently gene—overactivation, effectively establishing a molecular “speed limit” for many naïve pluripotency and housekeeping genes. Interestingly, we find that both the catalytic and non-catalytic functions of HDAC3 contribute to establishing this physiological transcript dose. Specifically during early development, HDAC3 rather than functioning as a canonical repressor, constrains the activity of highly transcribed genes of the implanting epiblast and ensures timely exit from naïve pluripotency in vitro . Altogether, this indicates that a dynamic equilibrium between activators and repressors coexists at highly active enhancer elements in pluripotent stem cells, establishing appropriate transcriptional dosage and rendering them responsive to signaling cues, thereby enabling timely and coordinated developmental progression. Introduction Maintenance of cell identity and differentiation depends on precise control of gene expression. This regulation is orchestrated at both enhancers and promoters, primarily by transcription factors (TFs), which act in concert with co-regulators 1 . Together, they facilitate the loading of RNA polymerase II (RNAPII), often through modulation of chromatin structure 2 , 3 . This is particularly relevant at enhancer elements, which not only establish the spatio-temporal pattern of gene expression but also regulate the transcript dosage through promoting RNAPII initiation 4 . Among the various enzymatic activities of the co-regulators that are active at enhancers, acetylation of proteins, primarily histones, plays a central role in transcriptional activation 1 , 5 . Notably, histone acetylation is also one of the most rapidly turned-over chromatin modifications, due to the opposing functions of histone acetyl-transferases and deacetylases 6 – 8 . Understanding how these opposing functions modulate transcription repression and dosage is particularly important for the dynamically evolving transcriptional states of the developing embryo. Based on sequence homology and mode of action, the family of histone deacetylase (HDAC) enzymes can be divided into four classes 8 . Class I HDACs have emerged as important regulators of cellular identity 9 – 22 . Indeed, their absence during embryonic development results in gastrulation failure, and early embryonic lethality 15 , 19 , 22 , 23 . Class I of HDACs comprise HDAC1, HDAC2, HDAC3 and HDAC8, and are known to act as part of large multiprotein co-repressor complexes 8 . While HDAC1 and HDAC2 participate in many such complexes and can mutually compensate for one another, HDAC3 specifically interacts with the nuclear receptor corepressor 1 (NCOR1 or NCOR) or its paralogue the silencing mediator of retinoid acid (NCOR2 or SMRT) 24 . Together, they are recruited to chromatin primarily by various TFs, such as ETS or AP-1 family members, and ligand-unbound nuclear receptors, including retinoid acid receptor (RAR), thyroid hormone receptor (TR), and PPARγ 25 . NCOR1/2 not only bind HDAC3 but are also essential for its catalytic activity 24 , 26 – 28 . Other confirmed partners of the NCOR-HDAC3 co-repressor complex are GPS2 and TBL1 (and/or its homolog TBLR1) 29 – 31 . In multiple cellular contexts, the canonical NCOR1/2-HDAC3 complex is recruited to gene regulatory elements 32 . Specifically in the case of nuclear receptors, ligand binding leads to displacement of the co-repressor complexes and their replacement by co-activator complexes, thereby activating transcription 33 . Whether this mode of action of HDAC3 also regulates early development remains unclear. In somatic cells, HDAC3 is involved in multiple physiological and developmental processes 34 , encompassing both catalytic-dependent and -independent functions. Intriguingly, while genetic ablation of HDAC3 leads to embryonic lethality by gastrulation, mice carrying mutations in the NCOR1/2 interaction domains survive to adulthood, despite lacking detectable catalytic activity 28 . Conditional knockout of HDAC3 in adult mouse liver disrupts liver homeostasis and circadian regulation, and induces liver hepatosteatosis 27 , 35 – 38 . In this system, HDAC3 is recruited to promoters via the nuclear receptor REV-ERBa in a diurnal manner, leading to reduced H3K9ac and activation of genes related to lipid metabolism 35 , 38 . However, HDAC3’s catalytic activity and histone acetylation levels are dispensable for lipid metabolism regulation, as a deacetylase catalytic mutant can rescue the phenotype 27 . In contrast, in brown adipose tissue, both the recruitment and the catalytic activity of HDAC3 are required for the regulation of lipid metabolism-associated genes 39 , 40 . Models of early mouse development have shown that the HDAC3-NCOR1/2 complex is involved in cell identity and differentiation of trophoblast stem cells 41 , endoderm specification 42 , and X-chromosome inactivation (XCI) in female mouse embryonic stem cells (ESCs) 43 , 44 . Nevertheless, the reasons why HDAC3-deficient embryos die during early development, and the molecular mechanisms by which HDAC3 regulates autosomal genes, remain unclear. Here, using Hdac3 knock out mice, tuneable acute HDAC3 degron ESCs, chromatin profiling and nascent transcriptomics, we reveal an unanticipated function for HDAC3. We find that it is recruited to the most active enhancers where it imposes a “speed limit” for their activation. Specifically in pluripotent cells, HDAC3 prevents overactivation of many housekeeping and pluripotency genes. In line with this, HDAC3 is vital for timely exit from naïve pluripotency, efficient response to signalling cues and peri-implantation development. Through tethering and complementation experiments we reveal that HDAC3 has both catalytic-dependent and - independent functions, which cooperate to allow cells to transit through differentiation. Results Histone deacetylase 3 localizes at putative active enhancer elements Our recent work uncovered that exit from naïve pluripotency is associated with increased histone acetylation turnover 45 . We reasoned that enzymes allowing for dynamic acetylation changes are likely important in the context of development and pluripotency. We focused on HDAC3, which is expressed throughout development, mediates XCI, and whose loss leads to post-implantation lethality due to unknown reasons 19 , 23 . To discern the molecular function of HDAC3 on autosomes, we first asked where in the genome is HDAC3 localized in the context of pluripotent mouse ESCs. For this, we utilized our published ChIP sequencing data for HDAC3 as well as for H3K27ac, H3K4me1, and H3K4me3 chromatin marks generated from female ESCs growing in 2iLIF conditions 44 . H3K27ac and H3K4me1 co-enrichment typically demarcates active enhancer elements 46 – 48 , while H3K27ac and H3K4me3, along with low H3K4me1 levels, mark active promoters 47 , 49 – 51 . Upon peak calling using MACS3, we identified 11,390 HDAC3, 46,052 H3K27ac, 115,531 H3K4me1, and 39,429 H3K4me3 peaks. When focusing on all HDAC3-bound loci, we found that these regions are also co-enriched for enhancer-associated H3K27ac and H3K4me1 marks ( Fig. 1a ). On the other hand, there was only moderate H3K4me3 enrichment at HDAC3-bound sites ( Fig. 1a ). Based on the overlap between the H3K27ac and H3K4me1 peaks, we identified a set of 47,507 putative active regulatory elements, which include both enhancers and promoters. When compared to these regulatory elements, HDAC3 peaks were enriched in intergenic regions and depleted from promoter elements ( Fig. 1b ). Of the intergenic peaks, the median distance from transcription start site (TSS) is approximately 18.4 kb (ranging from 1002 to 905,093 bp – Fig. 1c ). Download figure Open in new tab Figure 1. HDAC3 binds to active enhancers. a , Heatmaps of all HDAC3 peaks (±5,000 bp) showing enrichment of HDAC3-FLAG and no FLAG control as well as H3K4me1, H3K4me3, and H3K27ac. Dataset was downloaded from Zylicz et al. (GEO GSE116990) 44 and originates from female ESCs. Regions are sorted in descending order of HDAC3 signal. Color intensity indicates reads per million per 1 kb (RPKM), binned at 100 bp intervals. b , Bar plot illustrating the genomic distribution of HDAC3 peaks and putative regulatory elements. Bar lengths represent percentages. Statistical significance was assessed using the chi-square test. c , Density plot showing the distribution of distances from each intergenic HDAC3 peak to the nearest annotated transcription start site (TSS). d , Dot plot displaying top 20 enriched Gene Ontology (GO) biological function terms related to genes in proximity to HDAC3 peaks. Dot size indicates the number of genes associated with each GO term, and color intensity represents –log 10 (adjusted p- value). P-values were corrected for multiple testing using the Benjamini–Hochberg method. e , Dot plot displaying top 30 enriched TF motifs at HDAC3-bound regulatory elements identified using HOMER tool. Dot size indicates the number of regulatory elements with a TF motif, and color intensity represents –ln( p- value). f , Metaplots of putative regulatory elements with or without HDAC3 binding, showing the average ChIP signal (in RPKM) for HDAC3-FLAG, no FLAG control, H3K4me1, H3K4me3, and H3K27ac, binned at 50 bp intervals and centered on the regulatory element (±5,000 bp). Further analysis revealed that the genes in proximity to HDAC3 peaks are enriched for Gene Ontology (GO) terms related to both pluripotency (response to leukemia inhibitory factor - LIF), and lineage commitment during gastrulation, i.e. stem cell differentiation, embryonic pattern specification, anterior-posterior pattern specification, segmentation and somite development ( Fig. 1d ). Next, we classified the putative regulatory elements based on the presence or absence of HDAC3 peaks. We found that around 18.2% of the putative active regulatory elements identified in this study contain an HDAC3 peak or localize within 100 bp from one. This comparison further confirmed that HDAC3-bound regulatory regions are predominantly intergenic rather than associated with promoters ( Extended Data Fig. 1c ). This portion of regulatory elements accounts for a potentially highly active set of enhancers, since it is enriched for H3K4me1 and H3K27ac but has lower levels of H3K4me3 ( Fig. 1f , Extended Data Fig. 1a, b, d ). Further corroborating our GO term enrichment analysis, we found that HDAC3-bound elements are more likely to contain motifs of core pluripotency-related TFs, when compared to unbound putative regulatory elements ( Fig. 1e ). Altogether, these data imply that in naïve ESCs, HDAC3 predominantly resides in the most active enhancers bound by core pluripotency TFs. This is further corroborated by recent proximity ligation experiments, which profiled proteins residing in proximity to the key histone acetyltransferase P300 in mouse ESCs 52 . Upon reanalysing this proteomic study, we found that NCOR1, NCOR2 along with HDAC3 and other components of the complex (i.e. TBL1X, TBL1XR1, GPS2) are among the most enriched proteins in proximity to P300 reaching enrichment levels comparable to those of the mediator complex subunits ( Extended Data Fig. 1e ). Overall, we found that in naïve ESCs HDAC3, contrary to its canonical role as a transcriptional repressor, predominantly localizes to highly active autosomal enhancers, which are also co-bound by such co-activators as P300 and the pluripotency TFs. HDAC3 regulates highly active regulatory elements and their nearby genes Having established that HDAC3 co-occupies the most active enhancers with potent transcriptional activators, we set out to assess HDAC3’s role at such sites. To manipulate the levels of HDAC3 in naïve ESCs, we generated two independent clones of an auxin-inducible degron system for HDAC3 in male E14Tg2a ESCs ( Fig. 2a ). These HDAC3-mAID-FLAG ESC lines allow for rapid and acute depletion of HDAC3 within 3 hours of treatment with auxin (indole acetic acid - IAA; Fig. 2b , Extended Data Fig. 2a ). The use of a male cell line allowed us to focus on the role of HDAC3 in autosomal transcription regulation, and exclude phenotypes related to the XCI. We combined these acute loss-of-function models with transient transcriptome profiling using TT chem -seq in order to discern immediate transcriptional effects 3 hours post IAA treatment ( Fig. 2c ) 53 . Importantly, this method allowed us to study nascent transcription, as well as quantify unstable regulatory element-associated RNAs such as enhancer-associated RNAs (eRNAs) and promoter upstream transcripts (PROMPTs). Such transcripts, produced from regulatory elements, both aid in enhancer element identification and also correlate with their levels of activity 54 – 56 . Download figure Open in new tab Figure 2. HDAC3 prevents overactivation of highly active enhancers and their nearby genes. a , Schematic illustrating the design of the HDAC3 degron system at the endogenous Hdac3 and Igs7 (Tigre) loci. b, Immunoblot analysis of whole-cell extracts from HDAC3-mAID-FLAG degron ESCs treated with indole acetic acid (IAA) for varying durations. Blots were probed with anti-FLAG and anti-VINCULIN antibodies. c , Schematic of the experimental design for the 3-hour HDAC3 depletion. ESCs cultured in naïve 2iLIF conditions were treated with or without IAA for 3 hours, with a 15-minute 4sU nucleotide pulse prior to harvesting for TT chem -seq. d , Euler plot showing overlap of HOMER-identified transcript loci with putative H3K27ac/H3K4me1-enriched regulatory elements and HDAC3 peaks. Regulatory element associated RNAs: reRNAs. e , Volcano plot of reRNAs in IAA-treated versus untreated 2iLIF conditions with shrunk log 2 (fold change) and -log10(adjusted p- value) calculated using DESeq2. Differentially expressed reRNAs with adjusted p- value<0.05 are highlighted in color. f , Violin plot showing nascent transcription levels in the untreated 2iLIF condition of upregulated reRNAs compared to unchanged reRNAs. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Wilcoxon-–Mann–Whitney test. g, Genome browser tracks of Klf2 , Klf4 and Prdm14 loci showing HDAC3-FLAG, no FLAG control, H3K4me1 and H3K27ac ChIP signal from published datasets in female ESC 44 . Shown is also TT chem -seq signal in IAA-treated or untreated male HDAC3-mAID-FLAG ESC. In gray, we highlighted enhancer elements identified using STARR-Seq in previous study 58 . h , MA-plot showing the difference in shrunk log2(fold change) of differentially expressed genes in IAA-treated versus untreated 2iLIF conditions against nascent transcription levels in unperturbed 2iLIF degron ESCs. i , Violin plot showing nascent transcription levels in the untreated 2iLIF ESCs. Plotted are upregulated, downregulated and unchanged genes upon IAA treatment. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test for pairwise comparisons. P- values were corrected using the Benjamini–Hochberg method. j , Violin plot showing the distances from genes to their nearest differentially expressed reRNA, with genes categorized as upregulated, downregulated, or unchanged upon IAA treatment. Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test for pairwise comparisons. P -values were corrected using the Benjamini–Hochberg method. At first, we depleted HDAC3 for 3 hours in self-renewing naïve 2iLIF conditions and used HOMER to identify high confidence nascent transcription from non-genic regions. Specifically in our analysis, we included RNAs that do not overlap with the sense strand of any annotated transcript in ENSEMBLE. To exclude signal from inefficient transcription termination, we also excluded all transcripts mapping to the regions 5kb downstream of all annotated transcription termination sites. We then merged all transcripts identified within 100 bp and selected those that have substantial coverage (mean count > 7, based on DESeq2 independentFiltering), referring to them as HOMER transcripts. Upon overlapping HOMER transcripts with active regulatory elements marked by H3K27ac and H3K4me1 co-enrichment, we found 6563 putative regulatory RNA transcripts (reRNAs). Of these, 1208 originate from elements with an HDAC3 peak ( Fig. 2d ). Notably, when analysing from the perspective of putative regulatory elements, we observed that HDAC3 occupies approximately 32% of those that exhibit significant transcriptional activity ( Extended data Fig. 2b ). All in all, we successfully identified reRNAs transcribed primarily from enhancer elements, many of which are bound by HDAC3. To determine the effect of HDAC3 depletion on reRNA transcription, we performed differential expression (DE) analysis using DEseq2. We found that HDAC3 depletion results in significant upregulation of 2181 HOMER transcripts, in contrast to only 9 downregulated, thereby confirming the role of HDAC3 as a bona fide negative regulator of transcription ( Fig. 2e ). Interestingly, under untreated 2iLIF conditions, HOMER transcripts that become upregulated upon HDAC3 loss already exhibited significantly higher basal transcript levels compared to those that remain unchanged ( Fig. 2f ). Gene Set Enrichment Analysis of DE HOMER transcripts showed significant positive enrichment for HDAC3-bound reRNA loci ( Extended data Fig. 2c ). Indeed, HDAC3-bound reRNA loci were twice as likely (2.2 fold) to become upregulated upon IAA treatment compared to HDAC3-unbound sites ( Supplementary Table 1 ). This indicates that HDAC3 likely acts in cis as a “speed limit” for highly active regulatory elements, rather than repressing poised or inactive sites. Based on our finding that HDAC3 prevents enhancer overactivation, we hypothesized that HDAC3 could regulate genes that are highly expressed in 2iLIF conditions. To test this, we measured the levels of nascent transcripts produced from genic regions and performed differential gene expression analysis (DGE). Indeed, HDAC3 depletion caused upregulation of 1989 genes ( Fig. 2h ), many of which are related to pluripotency as well as housekeeping cellular functions such as ribosome biogenesis and translation, metabolism, and RNA processing ( Extended data Fig. 2d ). Indeed, such naïve pluripotency loci as Klf2, Klf4 and Prdm14 showed upregulation of transcription at both genes and their previously validated enhancer elements ( Fig. 2g ) 57 – 60 . At these well-described loci, HDAC3 is predominantly bound to proximal and distal enhancers, however some binding at promoter elements is also observed. Notably, genes that are upregulated upon HDAC3 loss already exhibit significantly higher transcriptional activity in untreated 2iLIF conditions and reside closer to DE HOMER transcript loci compared to HDAC3-unresponsive genes ( Fig. 2i-j ). In line with the possible direct in cis regulation of these genes by HDAC3, we found that upregulated genes are also in close proximity to HDAC3 peaks ( Extended data Fig. 2e ). Unlike for the HOMER transcripts, we also identified a number of downregulated genes upon IAA treatment (2351 - Fig. 2h ). However, these genes on average are transcribed significantly lower in untreated 2iLIF conditions, compared to the rest of the genes, and reside further away from DE HOMER transcript loci ( Fig. 2i-j ). Moreover, downregulated genes are also more remote from HDAC3 peaks ( Extended data Fig. 2e ). This indicates that in ESCs, HDAC3 is unlikely to function as a cis -acting transcriptional activator, rather, the observed changes are either due to indirect effects or HDAC3’s trans -acting role. In line with our TT chem -seq results, an additional EU-labelling experiment confirmed that HDAC3 does not lead to a global increase in transcription levels ( Extended Data Fig. 2f ). Overall, our results reveal that HDAC3 thresholds the maximum expression of pluripotency and house-keeping genes, thereby establishing appropriate transcript dose. The buffering role of HDAC3 relies in part on its catalytic activity and is specific to ESCs To gain further mechanistic insight into the role of HDAC3 in transcriptional regulation, we set out to complement the phenotype of HDAC3 loss. To this end, in the HDAC3-AID-FLAG degron ESC line, we used a PiggyBac-based vector to ectopically express either wild-type HDAC3 or the catalytically inactive HDAC3-Y298F mutant 27 . Following a 3-hour IAA treatment, we performed TT chem -seq to measure nascent transcription ( Fig. 3a-b ). By focusing on HDAC3-dependent genes identified in previous experiments ( Fig. 2h ), we found that the wild-type HDAC3 can fully complement IAA treatment and prevent overactivation of such genes as Nanog, Esrrb, Klf2 and Prdm14 among others ( Fig. 3d-e , Extended Data Fig. 3a ). On the other hand, the catalytically inactive HDAC3-Y298F mutant could partially complement the phenotype of endogenous HDAC3 loss ( Fig. 3d-e ). Notably, unbiased principal component analysis revealed that the overexpression of the catalytically inactive HDAC3-Y298F mutant, but not of the wild-type protein, leads to a transcriptional change ( Fig. 3c , Extended Data Fig. 3b-e ). However these alterations are less pronounced than those observed upon HDAC3 loss and do not affect the pluripotency genes. Despite this confounding effect, our results indicate that HDAC3 prevents gene overactivation through its catalytic activity but with the aid of the catalytic-independent function. Download figure Open in new tab Figure 3. Catalytic activity of HDAC3 prevents gene overactivation in ESCs but not in TSCs. a , Schematic of the complementation experiment design, involving transfection with an empty vector, HDAC3 WT , or the catalytically inactive mutant HDAC3 Y298F . b , Immunoblot analysis of the complementation experiment ( a ) with whole-cell extracts from HDAC3-mAID-FLAG degron ESCs treated with indole acetic acid (IAA) for 3 hours. Blots were probed with anti-FLAG (endogenous HDAC3), anti-HA (transfected HDAC3) and anti-VINCULIN antibodies c , Principal Component Analysis (PCA) plot of nascent transcriptomics data from the 3-hour HDAC3 depletion and complementation experiment. d , Violin plot illustrating the shrunk log 2 (fold change), in degron ESCs complemented with empty vector, HDAC3 WT , or HDAC3 Y298F upon treatment with or without IAA, for genes that were previously identified as upregulated genes upon HDAC3 loss ( Fig. 2h ). Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test. P- values were corrected using the Benjamini–Hochberg method. e , Transcription levels of representative pluripotency-associated genes in ESCs degron line complemented by empty vector, HDAC3 WT , or HDAC3 Y298F and treated with IAA for 3-hours. f , Schematic of the dCas9-GCN4-based epigenome editing system. Constitutively expressed dCas9 fused to GCN4-binding sites and TagBFP was stably transfected with doxycycline-inducible vectors expressing sfGFP-scFV, HDAC3 WT -sfGFP-scFV, or HDAC3 Y298F -sfGFP-scFV. Guide RNAs were used to target a safe-harbor locus expressing mCherry-H2B under the Ef1a promoter. g–h , Bar plots showing the percentage of sfGFP+/TagBFP+/mCherry2+ positive cells and their median fluorescence intensity (A.U.) upon doxycycline treatment. Statistical analysis was performed using ANOVA followed by Šidák’s test for multiple comparisons. i , Schematic of the protocol for derivation of trophoblast stem cells (TSCs) from ESCs. j , Immunoblot analysis of whole-cell extracts from HDAC3-mAID-FLAG degron TSCs treated with indole acetic acid (IAA) for varying durations. Blots were probed with anti-FLAG and anti-VINCULIN antibodies. k , Schematic of the experimental design for the 3-hour HDAC3 depletion in TSCs cultured in N2B27 with FGF2, Activin A, XAV939 and Y27632 (FAXY). l , Volcano plot of genes in IAA-treated versus untreated TSCs with shrunk log 2 (fold change) and -log10(adjusted p- value) calculated using DESeq2. Differentially expressed genes with adjusted p-value<0.05 are highlighted in color. m , Violin plot showing nascent transcription levels in the untreated TSCs. Plotted are upregulated, downregulated and unchanged genes upon IAA treatment. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P- values were corrected using the Benjamini–Hochberg method.. Complementation experiments, nascent transcriptomic analysis in combination with HDAC3 profiling, all indicate that HDAC3 acts in cis on chromatin to prevent gene overactivation. In order to further test this model, we asked whether de novo recruitment of HDAC3 activity is sufficient to regulate gene expression. To this end, we implemented a recently developed ESC line expressing a catalytically inactive dCas9 fused to an array of GCN4 motifs (dCas9-GCN4). This system allows tethering of target proteins, when fused to a single-chain variable fragment (scFV), to a constitutively active promoter driving the expression of mCherry-H2B ( Fig. 3f ) 61 . As a negative control, we used superfolded GFP (sfGFP) fused to scFV. Additionally, we have established lines which upon doxycycline treatment express full-length wild-type HDAC3 or the HDAC3-Y298F mutant fused to sfGFP and scFV. Upon doxycycline treatment we have observed that wild-type HDAC3, and to a much lesser extent HDAC3-Y298F mutant, is capable of reducing the expression of mCherry-H2B, as assessed both by the percentage of cells in which silencing is observed and the respective fluorescence intensity of the reporter ( Fig. 3g-h ). Together with our complementation assays, these data indicate that cis -acting HDAC3, through its catalytic activity, is sufficient to reduce the expression of active genes. Therefore, our experiments reveal that in pluripotent stem cells HDAC3, co-binds with activating complexes (e.g. P300) at the most active regulatory elements to establish a dynamic equilibrium and reduce the transcriptional dose of highly active genes. Our findings in ESCs are in contrast to previous work performed in other cellular contexts, where HDAC3 maintains gene silencing and its displacement allows for gene activation 34 . Therefore, we wanted to determine whether the role of HDAC3 in regulating highly transcribed genes is a unique feature of pluripotent stem cells. We decided to focus on the trophoblast stem cells (TSCs), which, similarly to ESCs, approximate early mouse development, but do not express the core pluripotency genes 62 , 63 . For this, we trans- differentiated HDAC3-AID-FLAG ESCs into TSCs and established two independent clones ( Fig. 3i , Extended Data Fig. 4a ) 62 – 66 . Next, we performed TT chem -seq upon depletion of HDAC3 for 3 hours ( Fig. 3j-k ). Our DGE analysis showed that HDAC3 loss in TSCs results in upregulation of 3124 genes and downregulation of an additional 1405 genes ( Fig. 3l , Extended Data Fig. 4b ). Unlike ESCs, in TSCs, the genes upregulated upon HDAC3 loss showed significantly lower basal expression in untreated TSCs compared to the HDAC3-independent genes ( Fig. 3m ). Furthermore, trophoblast-specific markers Hand1 , Elf5 , and Eomes were upregulated but not Cdx2, Krt8 or Gata3 ( Fig. 3l ). Further data integration revealed limited overlap in the genes differentially expressed in ESCs and TSCs upon HDAC3 loss ( Extended Data Fig. 4d ). More strikingly, even the genes highly expressed in both ESCs and TSCs control conditions did not respond to HDAC3 loss in the same way ( Extended Data Fig. 4e-f ). Therefore, in the context of TSCs, HDAC3 acts similarly to what was described for somatic cells, where it maintains the repression of genes, preventing their activation. All in all, the role of HDAC3 in regulating highly transcribed genes seems specific to pluripotent cells and in part relies on its catalytic activity. HDAC3 follows gene expression changes allowing for robust response to signalling cues Our work up to this point focused on the role of HDAC3 in self-renewing conditions. Next, we set out to address what the role of HDAC3 is when cells undergo cell-state transitions. In particular, we aimed to understand if there is a role for HDAC3 co-binding with transcriptional activators at naïve pluripotency genes when cells respond to signalling cues. For naïve pluripotent stem cells to become competent for differentiation, they need to undergo a process described as capacitation 67 – 70 . This is characterized by extensive chromatin and gene expression changes, shutdown of naïve pluripotency genes with maintenance of core pluripotency TFs 71 , 72 . By treating 2iLIF ESCs with FGF2 and Activin A (FA), we can induce epiblast-like cells (EpiLCs), which faithfully recapitulate epiblast capacitation of the E5.5-E6.25 67 , 68 , 70 . Initially we asked what role HDAC3 plays in mediating acute response to signalling cues. To this end, we treated HDAC3-AID-FLAG ESCs with FA for 3 hours in the presence or absence of IAA, and then performed TT chem -seq ( Fig. 4a-b ). DGE analysis comparing FA to 2iLIF conditions validated the induction process. Indeed, we found genes that are known to be immediate targets of ERK-signalling, downstream of FGF receptor, to be upregulated such as Egr1 , Arc , Dusp4/6 , Plaur and Myc 73 ( Extended Data Fig. 5a-b ). Concomitantly, there was downregulation of such naïve pluripotency markers as Nanog, Klf2/4, Esrrb and Prdm14 ( Extended Data Fig. 5b ). Having validated the induction, we subsequently focused on the phenotypes linked to HDAC3 loss. We identified 4443 upregulated genes and 2809 downregulated genes in FA conditions upon IAA addition ( Fig. 4c ). In line with the results in 2iLIF conditions, HDAC3 loss combined with FA treatment also resulted in upregulation and downregulation, of highly and lowly expressed genes respectively ( Extended Data Fig. 5c ). Moreover, the highly expressed genes that are differentially expressed associate to housekeeping biological processes such as RNA splicing, translation initiation, and ribosome biogenesis ( Extended Data Fig. 5d ). Interestingly, Gene Set Enrichment Analysis (GSEA) revealed that developmental genes downregulated upon FA induction alone, are significantly enriched in genes that become upregulated upon HDAC3 loss in FA treated cells ( Fig. 4d ). Similarly, genes upregulated upon FA induction alone are significantly enriched for genes that are downregulated in combined IAA and FA treatment. Together these data reveal that HDAC3 is crucial to efficiently respond to signalling cues and initiate the exit from naïve pluripotency. Strikingly, when we verified individual naïve pluripotency genes, which are downregulated upon the 3-hour FA induction, we observed that the levels of nascent transcription were even higher than the levels in the starting 2iLIF conditions ( Fig. 4e ). These effects demonstrate that the HDAC3-dependent hyperactivation of highly transcribed genes impacts the response to signalling not only by preventing silencing upon induction but also over-activating genes above their physiological baseline. Download figure Open in new tab Figure 4. HDAC3 loss prevents exit from naïve pluripotency in mouse embryonic stem cells. a , Schematic of the experimental workflow for ESCs exiting naïve pluripotency using EpiLC induction. Naïve ESCs in 2iLIF were switched to FGF2 and Activin A (FA) for 3 hours, with or without IAA, and samples were collected for TT chem -seq. Untreated 2iLIF cells and cells treated for 3 hours with FA, were also collected for ChIP-seq analysis. b , PCA plot showing gene transcriptomic profiles of different conditions analyzed by TT chem -seq. c , MA-plot showing shrunk log 2 (fold change) of differentially expressed genes in IAA-treated versus untreated FA conditions against nascent transcription levels in unperturbed 2iLIF degron ESCs. d , Gene Set Enrichment Analysis (GSEA) of ranked differentially expressed genes upon FA induction of IAA-untreated ESCs, for genes upregulated or downregulated upon HDAC3 loss in the FA sample. Ranking was based on DESeq2 stat values. e , Changes in nascent gene transcription upon HDAC3 loss in 2iLIF±IAA versus FA±IAA for selected naïve pluripotency markers. f , GSEA for genes near HDAC3 peaks that gained or lost ChIP signal during the 3-hour FA induction, using DE genes upon FA induction of IAA-untreated ESCs as the reference. Ranking was based on DESeq2 stat values. g , Volcano plot illustrating the distribution of HDAC3 ChIP signal changes at HDAC3 peaks during the 3-hour FA treatment. h , Box plots showing HDAC3 ChIP signal at peaks that gained, lost, or showed no change in HDAC3 signal during FA treatment. Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test for pairwise comparisons. P -values were corrected using the Benjamini–Hochberg method. i , Tracks showing average HDAC3 ChIP signal on both strands, binned at 50 bp intervals and centered on gained or lost HDAC3 peaks (±1,000 bp). Having established that HDAC3 is important for ESCs to respond to signalling cues, we next asked how HDAC3 binding responds to FA treatment. To this end we performed ChIP-seq for HDAC3 in 2iLIF conditions and at 3 hours post FA induction. We used MACS3 to identify peaks in both conditions and then merged them together. Subsequent differential peak enrichment analysis revealed that within the 3-hour FA time-window, HDAC3 signal is partially repositioned ( Fig. 4f-g ). Strikingly, peaks with decreased HDAC3 binding upon FA treatment are preferentially located near to developmental genes becoming downregulated upon 3-hour FA treatment in control conditions ( Fig. 4 f-j ). This implies that HDAC3 becomes rapidly unloaded from regulatory elements upon their silencing. On the other hand, we observed some increased binding of HDAC3 in proximity to genes that are upregulated upon FA treatment alone ( Fig. 4 f-j ). Collectively, we find that HDAC3 is pre-bound on the most active regulatory elements, where it establishes a dynamic equilibrium with transcriptional co-activators. In this context, HDAC3 prevents overactivation of the most active genes, while maintaining their responsiveness to developmentally programmed signalling cues. Additionally, in pluripotent cells HDAC3 binding follows the underlying enhancer activation status, leaving sites upon their silencing and becoming recruited to newly activated regulatory elements. Loss of HDAC3 delays exit from naïve pluripotency While HDAC3 is important for ESCs to acutely respond to FA, it is unclear if this is relevant for functional cell state transitions. To identify the long-term effects of HDAC3 loss during capacitation we performed RNA-seq in HDAC3-AID-FLAG Day 2 EpiLCs induced in the presence or absence of IAA ( Fig. 5a , Extended Data Fig. 6a ). We identified 1015 upregulated genes and 1198 downregulated genes ( Fig. 5b ). Notably, these numbers differ compared to the number DE genes that we identified at the 3-hour time point post FA treatment, implying the existence of potential compensatory mechanisms. Moreover, there is a significant but not complete overlap between the DE genes from the two transcriptomics datasets, most likely due to secondary effects of HDAC3 loss ( Extended Data Fig. 6b, Supplementary Table 3 ). Nevertheless, GSEA showed that genes upregulated upon HDAC3 loss are found significantly enriched in genes downregulated upon EpiLC induction alone. The inverse is also true, as genes downregulated upon HDAC3 loss are found enriched in genes upregulated upon EpiLC induction ( Fig. 5d-e ). These data confirmed that ESCs cannot efficiently exit naïve pluripotency in the absence of HDAC3. This phenotype is apparent both in inability to efficiently silence genes specific to naïve pluripotency, and in the failure to establish the EpiLC-specific transcriptional profile ( Fig. 5c-e ). We confirmed our results by immunofluorescence for TBX3, a naive pluripotency marker, which fails to become completely silenced in the absence of HDAC3 ( Extended Data Fig. 6c-d ). Next, we set out to test if this apparent delay in exiting naïve pluripotency leads to functional outcomes. To this end, we performed short reversion experiments, where ESCs are induced towards EpiLCs for 1 day and then returned to 2iLIF conditions for another 24 hours. Without IAA treatment, cells are not able to efficiently re-express TBX3 and form dome-shaped colonies. On the other hand, HDAC3 loss rendered cells more plastic and allowed them to readily upregulate TBX3 in response to 2iLIF and establish characteristic morphology ( Extended Data Fig. 7a-c ). All in all, HDAC3 loss prevents ESCs from efficiently exiting naïve pluripotency at both molecular and functional levels. Download figure Open in new tab Figure 5. Long term loss of HDAC3 impairs exit from naïve pluripotency and capacitation. a , Schematic overview of experimental design. Mouse ESCs were cultured in 2iLIF and induced into EpiLCs with or without IAA. RNA-seq and RT-qPCR analyses were performed 48 hours post-induction. b , Volcano plot analysis of gene expression levels in IAA-treated versus untreated EpiLCs with shrunk log 2 (fold change) and -log10(adjusted p- value) calculated using DESeq2. Differentially expressed genes with adjusted p -value<0.05 are highlighted in color. c , GSEA for genes associated with GO terms of biological processes, using ranked DE genes from EpiLC induction with or without IAA. Top enriched pathways are shown. d , Heatmap showing scaled normalized expression of differentially expressed genes with absolute shrunk log2(fold change) > log2(1.5) in IAA-treated versus untreated EpiLCs. Cells expressing only TIR1 (not HDAC3-mAID-FLAG) served as controls. e , Gene Set Enrichment Analysis (GSEA) of ranked differentially expressed genes upon 2-day induction of IAA-untreated EpiLCs, for genes upregulated or downregulated upon HDAC3 loss in the FA sample. Ranking was based on DESeq2 stat values. f, Schematic of experimental workflow as in (a), complemented with vectors encoding HDAC3 WT , HDAC3 Y298F , or empty vector. g , RT-qPCR analysis of key naïve pluripotency markers in EpiLCs as in (e). Statistical analysis was performed using two-way ANOVA followed by Tukey’s test. h , Schematic of extended reversion experiment. Naïve ESCs were induced to EpiLCs for two days, then reverted in 2iLIF for 5 days, and colonies were stained for alkaline phosphatase. i , Alkaline phosphatase staining of ESCs after EpiLC induction and reversion, demonstrating the effect of HDAC3 loss on maintenance and re-acquisition of naïve pluripotency. Statistical analysis was performed using one-way ANOVA followed by Tukey’s test. Under self-renewing conditions wild-type HDAC3, and to a lesser extent the HDAC3-Y298F mutant, can complement the loss of endogenous HDAC3. Next, we set to address, whether the catalytic activity of HDAC3 is needed for efficient EpiLC induction. For this we used the HDAC3-AID-FLAG ESCs complemented with an empty vector or vectors encoding for wild-type HDAC3 or the HDAC3-Y298F mutant. These cells were induced for 2 days toward EpiLC in the presence or absence of IAA. RT-qPCR analysis for the HDAC3-sensitive naïve pluripotency genes revealed that constitutive expression of wild-type HDAC3 complements HDAC3 loss and allows the establishment of silencing of Nanog, Tbx3 and Zfp42 ( Fig. 5f-g ). On the other hand, HDAC3-Y298F mutant can partially complement the phenotype of HDAC3 loss, suggesting both catalytic and non-catalytic functions of HDAC3 are at play. Following this, we performed a series of short reversion experiments, as described previously, this time by complementing the loss of endogenous HDAC3 with an empty vector, wild-type HDAC3 or catalytic inactive HDAC3-Y298F ( Extended Data Fig. 6e ). These experiments confirmed that the catalytic function of HDAC3 is primarily, but not solely, required for efficient exit from naïve pluripotency ( Extended Data Fig. 6f ). We further explored this phenotype by performing a set of extended reversion experiments. Briefly, we have induced the complemented ESC lines toward EpiLCs for 2 days in the presence or absence of IAA. This was followed by a reversion in 2iLIF conditions and a colony formation assay with alkaline phosphatase staining as a readout ( Fig. 5h ). While the efficiency of reversion varied across lines in untreated conditions, stable expression of wild-type HDAC3 complemented the IAA phenotype ( Fig. 5i , Extended Data Fig. 6g ). Similarly, the HDAC3-Y298F mutant could largely complement for the loss of endogenous HDAC3 ( Fig. 5i ). Collectively, our results show that HDAC3 is a potent regulator of naïve pluripotency genes and is important for efficient exit from naïve pluripotency. Our complementation experiments found that both catalytically dependent and independent functions regulate naïve pluripotency genes and that HDAC3-Y298F mutant allows for functional exit from naïve pluripotency. Genetic ablation of HDAC3 results in embryonic lethality at the time of implantation Lastly, we asked to what extent our results recapitulate HDAC3’s role in development. To address this question, we generated Hdac3 maternal and zygotic knockout (KO) embryos. We crossed heterozygous Hdac3 +/- male mice with Hdac3 loxP/loxP females also carrying oocyte-specific ZP3-cre ( Fig. 6a ) 74 , 75 . This strategy allows for the generation of Hdac3 KO and heterozygous (Het) embryos devoid of maternal protein contribution, which could potentially mask early embryonic phenotypes. Previous studies have reported that genetic ablation of HDAC3 results in embryonic lethality by E9.5 19 , 23 . To characterize the molecular events that precede lethality, we studied the effects of HDAC3 loss at the pre-implantation expanded blastocyst stage of the embryonic (E) day 4.5 as well as at the onset of gastrulation at E6.5. Download figure Open in new tab Figure 6. HDAC3 is essential for early mouse embryonic development. a , Schematic of breeding strategy and genotypes used to generate Hdac3 heterozygous (Het) and knockout (KO) embryos, with Zp3-cre for conditional deletion in oocytes. Specifically, male Hdac3 +/- mice were mated with female Hdac3 loxP/loxP R26:Zp3Cre+ve mice, and embryos were collected at E4.5 and at E6.5 stages. b , Quantification of epiblast volumes at E6.5 for Hdac3- Het and Hdac3- KO embryos. Line indicates the median value and statistical analysis was performed using unpaired t-test. c , Representative images of dissected Hdac3- Het and Hdac3- KO embryos (male and female) by brightfield microscopy. d , Genotype distribution of embryos at E4.5 and E6.5, showing expected Mendelian ratios at E4.5 and reduced KO embryos at E6.5 for both sexes. P- values were calculated using Chi-squared test. e , PCA plot of transcriptomic profiles from male and female Hdac3- Het and Hdac3- KO embryos at E4.5 and E6.5. f , Volcano plot analysis of gene expression levels in E6.5 male Hdac3- KO vs. Hdac3- Het epiblasts with shrunk log 2 (fold change) and -log10(adjusted p- value) calculated using DESeq2. Differentially expressed genes with adjusted p-value<0.05 are highlighted in color. g , Violin plot showing gene expression levels in the Hdac3- Het male E6.5 epiblasts. Plotted are upregulated, downregulated and unchanged genes in Hdac3- KO vs. Hdac3- Het E6.5 male epiblasts. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P- values were corrected using the Benjamini–Hochberg method. h , GSEA of differentially expressed genes in E6.5 Hdac3- KO vs. Hdac3- Het male epiblasts for genes associated with GO terms of biological processes. i , Euler plots showing overlap of up- and downregulated genes in Hdac3- KO male E6.5 epiblasts and in epiblast-like cells (EpiLCs), with calculated odds ratios. When focusing on E4.5 stage, we observed normal morphology and expected Mendelian ratios ( Fig. 6d ). On the other hand, by E6.5, we found a partially penetrant loss of Hdac3-KO embryos with multiple empty deciduae detected ( Fig. 6d ). This wave of embryo loss affects both sexes, but females are more severely affected, as assessed by the deviation from the Mendelian ratios ( Fig. 6d ). Indeed, we estimate an implantation failure of approximately 81% in female and 52% in male Hdac3-KO embryos, respectively. Of the rare embryos that escape this implantation defect, Hdac3-K O E6.5 embryos are significantly smaller, including in terms of epiblast size ( Fig. 6b-c , Extended Data Fig. 8a-b ). Interestingly, despite having smaller epiblasts, surviving Hdac3-KO embryos can still sequester NANOG expression at the posterior part of the epiblast, which implies that they are capable of anterior-posterior axis patterning ( Extended Data Fig. 8a ) 76 – 78 . Overall, we find that while the maternal pool of HDAC3 is dispensable for early development, zygotically encoded HDAC3 is vital for efficient implantation and post-implantation embryo growth, with female embryos being more severely affected than males. To better understand the role of HDAC3 during implantation, we performed full-length transcriptome analysis using FLASH-seq 79 , on whole E4.5 blastocysts and dissected epiblasts at E6.5. Our results found that no major transcriptional changes occur by E4.5, as assessed by PCA and DGE analysis ( Fig. 6e , Extended Data Fig. 8h-i ). On the other hand, at the E6.5 stage, female Hdac3-KO epiblasts exhibit dramatic differences in transcript levels compared to Hdac3-Het ( Fig. 6e , Extended Data Fig. 8c-d ). This striking phenotype in female embryos is likely a composite effect of defects in XCI as well as an altered autosomal gene expression program. To focus on autosomal gene regulation, we next analyzed male E6.5 epiblasts. DGE analysis identified 471 upregulated genes and 408 downregulated genes ( Fig. 6f ). Among the downregulated genes was Hdac3 , supporting the robustness of our approach. Our in vitro work revealed that in pluripotent stem cells, HDAC3 limits the expression of the most actively transcribed genes. Strikingly, we observed that genes upregulated in male E6.5 Hdac3-KO epiblasts compared to Hdac3-Het had higher baseline transcript levels than HDAC3-independent genes ( Fig. 6g ). This finding indicates that, also in vivo , HDAC3 prevents gene overactivation. Additionally, GSEA for GO terms of biological processes showed that upregulated genes in male E6.5 epiblasts are enriched in genes related to housekeeping functions, similar to what we observed in vitro ( Fig. 6h ). Furthermore, despite stark experimental differences, there is significant overlap between not only E6.5 DE genes and the RNA-sequencing results from Day 2 EpiLCs, but also with our nascent transcriptomics datasets ( Fig. 6i , Extended Data Fig. 8f-g, Supplementary Table 3 ). All in all, HDAC3 loss results in embryonic lethality at a stage earlier than expected—around the time of implantation. In addition, while we observe a more severe effect in female embryos, HDAC3 loss also affects male embryos in terms of survival, morphology, and gene expression. Finally, the phenotype observed in surviving E6.5 embryos particularly affects highly abundant transcripts related to housekeeping biological processes, in accordance with results from our in vitro models. Discussion Gene regulation in all cellular contexts extends far beyond simple binary on/off mechanisms. Instead, precise control of transcriptional dosage underpins the establishment of gene regulatory networks. This fine-tuned regulation enables the formation of attractor cellular states that define cellular identity. At the same time, these states remain inherently poised for change—partially destabilized to allow cells to respond dynamically to external cues, particularly during developmental transitions. Transcriptional dosage is crucial and therefore is established at multiple, parallel levels, including through TFs’ cooperativity and chromatin modifiers. In this study, we find that transcriptional dosage is fine-tuned by epigenetic repressors. Indeed, we reveal that co-binding of transcriptional co-activators such as P300 and co-repressors like HDAC3 to the most active subset of enhancer elements is a key mechanism for achieving appropriate transcriptional output. Our in vivo and in vitro loss-of-function experiments support a model where HDAC3 acts as a molecular “speed limit”, constraining the expression of highly active pluripotency and housekeeping genes. This inherent destabilization of transcriptional output at the most active genes renders them responsive to signalling cues, enabling timely and coordinated developmental progression. In line with this model, we found that in naïve ESCs, HDAC3 is bound to approximately 18% of putative active regulatory elements, primarily at non-promoter regions. Indeed, we previously reported HDAC3 binding to regulatory elements on the X chromosome and proposed that HDAC3 is pre-bound in an inactive form, awaiting repressor cues such as Xist RNA binding 44 . However, these HDAC3-bound loci are also present on autosomes and likely represent a subset of highly active enhancer elements, as indicated by strong enrichment for H3K27ac and H3K4me1, along with a concomitant depletion of the promoter-specific H3K4me3 mark compared to non-HDAC3-bound elements. Apart from the few HDAC3-bound promoters, the remaining H3K4me3 marking is consistent with prior observations that highly active enhancer elements can also be partially decorated by this modification 51 . The majority of HDAC3-bound putative regulatory elements are located relatively close to the TSS, i.e., 50% of them are within a 19 kb window, and are enriched for binding motifs of key pluripotency TFs. Moreover, genes located near HDAC3-bound regulatory elements are associated with pluripotency, housekeeping functions, and pre-gastrulation processes such as embryonic patterning. Supporting our findings, recent proximity labelling experiments identified NCOR1/2 and HDAC3 as being highly enriched near the co-activator P300 in ESCs 52 . Furthermore, during rapid adipocyte differentiation, following super-enhancer activation, there is recruitment not only of transcriptional co-activators but also of co-repressors 80 . Indeed, in brown adipose tissue, HDAC3 is recruited to putative regulatory elements. However, these sites do not correspond to the most active enhancers as indicated by H3K27ac 81 . In such cellular context, HDAC3 recruitment to regulatory elements can lead to their upregulation, reflecting a non-canonical function. By contrast, in the liver, HDAC3 is in part recruited through ligand-free nuclear receptors to maintain gene silencing 35 , 38 . Thus, specifically in ESCs, HDAC3 is bound to the most active putative enhancers to regulate their activity. HDAC3 binding to the most active putative enhancers raises the question of whether it plays a role under steady state conditions or remains inactive, awaiting signaling cues to initiate gene repression. By leveraging HDAC3 degron ESC lines, nascent transcriptomics, and the observation that a large portion of regulatory elements are transcribed—and that their transcriptional activity reflects their regulatory function 54 – 56 —we were able to test these hypotheses. Our integrated data establish HDAC3 as a negative regulator of transcriptional activity in steady-state pluripotent stem cells, where HDAC3 prevents overactivation of reRNAs and the genes located nearby. Indeed, under these steady-state conditions, HDAC3 is functional and helps establish the transcriptional dosage of many highly active genes. These results contrast with previous findings in the liver, where the HDAC3–NCOR1/2 complex functions as a canonical repressor maintaining gene silencing 35 , 38 . Indeed, this canonical mode of action appears to dominate in the trophoblast lineage, as HDAC3 loss leads to the upregulation of initially silent genes. This aligns with a recent report showing that NCOR1/2 is recruited by FGF-dependent factor ERF to decommission regulatory elements in TSCs 41 . Additionally, the role of HDAC3 in ESCs is distinct from that reported for other class I HDACs, namely HDAC1 and HDAC2. Specifically, rapid loss of HDAC1 in the context of Hdac2-KO ESCs leads, among other phenotypes, to the downregulation of highly expressed naïve pluripotency genes 11 . In line with this observation, short-term treatment of ESCs with trichostatin A (TSA), a potent HDAC inhibitor, results in loss of pluripotency and significant 3D chromatin rearrangement 82 . Moreover, TSA treatment in breast cancer cell lines also led to drastic downregulation of reRNAs 83 . In light of these findings, we propose that HDAC3 plays a role that opposes the function of HDAC1/2 by inherently destabilizing naïve pluripotency. The cellular phenotypes observed upon HDAC3 loss could be linked to its chromatin binding, functioning in cis at proximal enhancers. Alternatively, trans- acting functions of HDAC3 may also contribute to the observed effects. Two lines of evidence suggest that HDAC3 primarily acts in cis to reduce the dosage of naïve pluripotency and housekeeping genes. First, we find that loci encoding reRNAs upregulated upon HDAC3 loss are also enriched for HDAC3 binding, supporting a cis- regulatory role. Second, our dCas9 experiments demonstrate that tethering HDAC3 in cis to a regulatory element is sufficient to inhibit transcription. Surprisingly, our nascent transcriptomic studies also identified a number of genes that are downregulated upon HDAC3 loss. However, these genes are not located near HDAC3-regulated reRNA loci or HDAC3 peaks and are expressed at significantly lower levels than other genes in unperturbed cells. This is contrary to other systems where HDAC3 is a cis- acting transcriptional activator 81 . Our data could therefore indicate a potential trans-acting role for HDAC3, such as regulating the acetylation status of TFs. Indeed, SIRT1, another deacetylase, has been shown to modulate the activity of the key pluripotency TF SOX2 84 Alternatively, the downregulation of lowly transcribed genes upon HDAC3 loss might result from hyperactivation of already highly active genes. Given the short timepoints used and the absence of a global increase in transcription (as assessed by 4sU-staining), we speculate that hyperactivation of active genes leads to a redistribution of limited transcriptional resources, thereby reducing pervasive transcription of lowly expressed genes. Another open question is the extent to which HDAC3 regulates gene expression through its enzymatic activity. Our complementation assays, combined with dCas9 tethering experiments, suggest that both catalytic and non-catalytic functions of HDAC3 are important. Indeed, the well-characterized HDAC3 Y298F mutant could partially complement the loss of endogenous HDAC3 under steady-state conditions, and to a greater extent during the exit from naïve pluripotency. In line with this finding, tethering the HDAC3 Y298F mutant to a reporter gene in ESCs led to modest transcriptional attenuation, unlike wild-type HDAC3. Consistent with these observations, in Drosophila melanogaster , enzymatically inactive HDAC3 can partially rescue the developmental phenotype caused by HDAC3 loss 85 . Similarly, in the mouse liver, loss of endogenous HDAC3 can be partially complemented by the HDAC3 Y298F mutant 27 . These non-catalytic functions of HDAC3 are likely mediated through protein–protein interactions. Overall, our nascent transcriptomic datasets uncovered HDAC3 as a potent thresholding mechanism for the most active putative enhancers in pluripotent stem cells. Differentiation and in vivo experiments further confirmed that this function is crucial for cell state transitions. Indeed, genetic ablation of HDAC3 during mouse embryogenesis leads to acute implantation defects, affecting both sexes but with a more severe impact on females. These data imply an important role of HDAC3 in extraembryonic lineages 41 , 42 . Remarkably, this phenotype closely mirrors our in vitro findings in terms of timing, as the implanting epiblast corresponds to ESCs exiting naïve pluripotency. In surviving male embryos, we observed upregulation of highly transcribed genes and a significant overlap in differentially expressed genes between the in vivo and in vitro systems. These knockout embryos also exhibit growth defects, including reduced size and epiblast volume, but retain the ability to form the anterior–posterior axis. These findings are broadly consistent with previously reported Hdac3 knockout lethality prior to E9.5 19 , 23 . The observed sex differences are likely due to the inability of female embryos to efficiently silence the X chromosome, a process that occurs at implantation 43 , 72 . Interestingly, the female embryo phenotype is more severe than that observed in Xist knockouts, which successfully implant but show developmental delay by E6.5 86 . This suggests that the female phenotype results from a combination of autosomal and XCI deregulation. Our study provides new insights into how chromatin-modifying enzymes influence the execution of transcriptional programs in a context-specific manner. We underscore the importance of enhancer utilization in these processes and demonstrate how HDAC3 can serve as a rheostat for specific gene sets. Future work should elucidate how chromatin-modifying enzymes are recruited to different genomic regions and how differences between cellular contexts contribute to the diversity of transcriptional outcomes. Methods Cell culture and cell line generation E14Tg2a (129/Ola, XY) wild-type mouse ESCs, kindly offered by the lab of Joshua Brickman, were used for all experiments, apart from CRISPR-dCas9 GCN 4 -based tethering experiments, in which we used the reporter mCherry2-H2B ES cell line (129/B6, XY) developed by Policarpi et al. 61 . E14Tg2a cells were grown on gelatin, cultured in Serum/LIF conditions, specifically GMEM (ThermoFisher) supplemented with 10% FBS (Merck), 1,000 U mL −1 LIF (in-house production), 0.1 mM 2-mercaptoethanol, 0.1 mM non-essential amino acids, 2mM L-alanyl-L-glutamine equivalent (GlutaMax, ThermoFisher), 1 mM sodium pyruvate. For transition to naïve conditions, ESCs were cultured for 12 days on fibronectin-coated plates (Sigma Aldrich), in N2B27 basal culture medium: DMEM/F12 medium (ThermoFisher) and Neurobasal medium (ThermoFisher) 1:1, with the addition of N2 supplement 1/200 (ThermoFisher), B-27 Serum-Free Supplement 1/100 (ThermoFisher), 2mM L-alanyl-L-glutamine (GlutaMax, ThermoFisher), 0.1 mM 2-mercaptoethanol (Merck), supplemented with 1 µM PD0325901, 3 μM CHIR99021, 1,000 U mL −1 LIF (in-house production), as described by Mulas et al. 87 . After transition, cells were maintained and regularly passaged in these conditions. For mCherry2-H2B reporter line, cells were cultured on gelatin-coated plates in N2B27 basal culture medium (NDiff by Takara, y40002), supplemented with 1 µM PD0325901, 3 μM CHIR99021, 1,000 U mL −1 LIF (in-house production), 1% FBS (Millipore) and 1% penicillin-streptomycin (ThermoFisher). For generation of transgenic cell lines, ESCs growing in Serum/LIF were electroporated using the P3 Primary Cell 4D-Nucleofector™ X Kit (Amaxa - Lonza) according to manufacturer’s instructions. For experiments using piggyBAC system cells were transfected using Lipofectamine 2000 according to manufacturer’s instructions for ES cells. ESC Differentiation EpiLC induction EpiLC inductions was performed according to Hayashi et al. 68 . Briefly, naïve ESCs were washed 3 times in PBS and then passaged into N2B27 medium supplemented with bFGF (12 ng/mL) and Activin A (20 ng/mL) and 1% KSR for 2 days. Medium was changed daily. TSC transdifferentiation Naïve ESCs growing on fibronectin-coated plates were lipofected with esiRNAs for Pou5f1 (400ngr per 2LxL10L cells), using Lipofectamine 2000 in Optimem. Transfected cells were incubated for 48Lh without media change. Cells were then passaged into TX medium (DMEM/F12 without HEPES and L-glutamine, supplemented with TGF-ß1 (2 ng/mL), FGF4 (25 ng/mL), 1 mg/mL Heparin) on laminin-coated plates and maintained until confluency with regular media changes. Emerging iTSC colonies were transferred to FAXY medium (Fgf2 - 12 ng/mL, Activin A - 20ng/mL, XAV939 - 10µM, Y27632 - 5µM) on fibronectin and maintained in the same conditions. Alkaline phosphatase Alkaline phosphatase staining was performed as described in Mulas et al. 87 . Briefly cells were converted into EpiLCs, detached with Accutase (Stem Cell Technologies) for 1 min at 37°C, counted and seeded at low density into laminin-coated plates in 2iLIF conditions. Medium was changed every two days and cells were stained after 5 days using the Alkaline Phosphatase Detection Kit for leukocytes (Merck) according to manufacturer’s instructions. Plates were stored O/N at 4oC, in the dark and colonies were counted the day after. Immunofluorescence In embryos Embryos were fixed in 4% PFA with 0.001% PVA for 20Lmin at 37L°C, then washed 3 times in PBS-T (0.1% Tween-20, 0.001% PVA) for 10Lmin at room temperature (RT). Permeabilization was performed for 10–15Lmin in 0.1LM glycine, 0.3% Triton X-100, and 0.001% PVA, followed by three additional PBS-T washes. Embryos were blocked for 4–5Lh at RT in blocking buffer (1% BSA, 0.1% Tween-20, 10% normal serum, 0.001% PVA in PBS), then incubated with primary antibodies (in blocking buffer) for 1–2 days at 4L°C, rocking. After washing (3 times, 10Lmin each, PBS-T - 0.001% PVA), embryos were incubated with secondary antibodies overnight at 4L°C in the dark. Embryos were washed (3 times, 10Lmin each, PBS-T - 0.001% PVA - DAPI was added during the second wash step at a final concentration of 1µg/ml), and mounted on poly-lysine–coated glass-bottom plates in Vectashield® Antifade Mounting Medium for imaging (Vector Laboratories). Imaging was performed at a Leica Stellaris Microscope. In cells Cells were seeded in fibronectin-coated Ibidi mini-dishes and cultured in them. The following day, cells were fixed in 4% PFA for 10Lmin at RT, permeabilized with 0.25–0.5% Triton X-100 for 10Lmin, and blocked for ≥3Lh at RT (1% BSA, 0.1% Tween-20, 10% normal serum in PBS). Primary antibodies were applied in blocking buffer and incubated overnight at 4L°C. After three PBS-T washes, secondary antibodies were applied for 3Lh at RT. Cells were washed again (DAPI added in the second wash) and mounted in Vectashield®. Imaging was performed at a Leica Stellaris Microscope. Western Blotting Protein extraction Cells were detached in 1 mL TryPLE, neutralized in equal volume medium and washed twice in PBS and snap-frozen as dry pellets in dry ice. Cell pellets were lysed in RIPA buffer (1x PBS, 150 mM NaCl, 1% NP40, 0.1% SDS, 0.5% Sodium Deoxycholate, 50 mM Tris-HCl, pH 8.0) supplemented with proteinase inhibitors (cOmplete TM , Roche), incubated for 15 minutes on ice, and sonicated using a Bioruptor Plus (Diagenode), set to high power for 10 cycles (30Ls ON, 30Ls OFF). Protein quantification was performed using Pierce BCA kit (ThermoFisher). SDS-PAGE and immunoblotting Protein concentration was equalized across samples and 20µg were loaded on a NuPAGE 4-12% Bis-Tris (ThermoFisher) pre-cast gel, for ∼180 minutes at 120-140V, in NuPAGE MED SDS running buffer. Proteins were transferred into a nitrocellulose mebrane 0.45 µm (Amersham TM Protran ® ) for 70 min at 350 mA, blocked for 1 hour with 5% milk in PBS + 0.3% NP40. Primary antibodies were incubated O/N in 1% milk in PBS + 0.3% NP40 at 4°C at indicated concentrations ( Supplementary Table 2 ). Membranes were washed 3 times in PBS + 0.3% NP40, incubated for an hour with HRP-conjugated secondary antibody (1:10.000 in 1% milk in PBS + 0.3% NP40). Real Time RT-qPCR RNA isolation and RT–qPCR Total RNA was extracted using RNAeasy Plus Mini kit with DNA digestion and equal amounts of RNA were retrotranscribed into cDNA using Superscript III Reverse Transcriptase. Real Time RT–qPCR was performed using the Power Up Master Mix (Applied Biosystems) in a Light Cycler thermocycler (Roche). Primer efficiency was assessed by standard curves of serially diluted, purified amplicon fragment. All samples were run in quadruplicates; data were analyzed using the 2 -ΔΔCt method and normalized to the geometric mean of the signal of housekeeping genes i.e. Tbp and Gapdh . ChIP sequencing ChIP sequencing Analysis of chromatin occupancy for HDAC3 was performed as described in Zylicz et al. 44 . Briefly, naïve ESCs were grown to ∼80% confluency, washed 3 times in PBS and switched to N2B27 medium supplemented with Fgf2 (12 ng/mL), Activin A (20ng/mL) and KSR 1% for 3 hours. Cells were incubated for 30Lmin at room temperature (RT) in 1.5LmM ethylene glycol bis(succinimidyl succinate) (EGS; Thermo Fisher) freshly diluted in PBS and then crosslinked with freshly opened formaldehyde for 10Lmin at a final concentration of 1.88%. Crosslinking was quenched with 0.11LM Glycine for 5Lmin at RT. Cells were washed twice in ice-cold PBS, scraped, pelleted by centrifugation (500 xg, 5Lmin, 4L°C), and snap-frozen in liquid nitrogen. Frozen pellets were thawed on ice and lysed in Nuclear Lysis Buffer (5LmM MgCl₂, 10LmM Tris-HCl pH 8.0, 1LmM EDTA, 0.1LM sucrose, 0.5% Triton X-100, supplemented with protease inhibitors – cOmplete TM , Roche) for 10Lmin on ice. Cells were homogenized using a dounce homogenizer with 20 strokes each of pestle A and B to release nuclei, which were then pelleted (500g, 15Lmin, 4L°C) and resuspended in Chromatin Lysis Buffer (150LmM NaCl, 10LmM Tris-HCl pH 8.0, 1LmM EDTA, 0.5LmM EGTA, 0.5% N-lauroylsarcosine, with protease inhibitors). Chromatin was sheared using a Bioruptor Plus (Diagenode), set to high power for 12 cycles (30Ls ON, 30Ls OFF) to obtain fragments of ∼300–400Lbp. Lysates were centrifuged (500g, 15Lmin), and the supernatant was diluted in Dilution Buffer (150LmM NaCl, 10LmM Tris-HCl pH 8.0, 6LmM EDTA, 0.125% sodium deoxycholate, 1.25% Triton X-100, with protease inhibitors). A small aliquot was reserved as input control, and the remainder was incubated overnight at 4L°C with 10Lμg of anti-Flag M2 antibody and protein G Dynabeads (Invitrogen) pre-washed in buffer. Immunocomplexes were washed five times (10Lmin, 4L°C, rotation) in each of two wash buffers: LiCl Wash Buffer (0.5LM lithium chloride, 10LmM Tris-HCl pH 8.1, 1LmM EDTA, 1.1% sodiumdeoxycholate, 1% NP-40), High Salt Wash Buffer (1LM NaCl, 20LmM Tris-HCl pH 8.1, 2LmM EDTA, 0.1% sodium deoxycholate,0.1% SDS, 1% Triton X-100). Final washes were performed in TE-low EDTA buffer (10LmM Tris-HCl pH 8.0, 0.1LmM EDTA), and DNA was recovered in Proteinase K buffer (20LmM HEPES pH 7.9, 1LmM EDTA, 0.5% SDS, 0.8Lmg/mL Proteinase K) for 30Lmin at 56L°C. Samples were further digested by addition of Proteinase K and NaCl (final concentration 0.3LM NaCl) for 2Lh at 56L°C, followed by 4Lh at 68L°C. DNA purification DNA was purified using paramagnetic beads. Equal volumes of Ampure XP beads and PEG/NaCl solution (final: 13% PEG8000, 0.75 M NaCl) were added to the ChIP reaction, incubated for 15 min at RT, and beads were separated using a magnetic rack. Beads were washed twice with 80% ethanol, air-dried, and DNA was eluted in TE-low EDTA buffer (10 mM Tris-HCl pH 8.0, 0.1 mM EDTA). DNA concentration was measured using Qubit dsDNA HS Assay (Invitrogen) and fragment size assessed using a TapeStation D5000 assay (Agilent). Libraries were performed using the Ovation® Ultralow System V2 1-16 library preparation kit by NuGEN and single-end sequencing was performed P2-100bp NextSeq2000 kit. Data processing Raw sequencing data were demultiplexed and converted into FASTQ files using Illumina bcl2fastq Conversion Software (v2.20.0.422). Resulting FASTQ files were processed using nf-core/chipseq v2.1.0 88 of the nf-core collection of workflows 89 with parameters ’--macs_gsize 2652783500 --narrow_peak --macs_fdr 0.01’. In brief, adapters and low-quality reads were removed using Trim Galore (v0.6.7), trimmed reads were mapped to the mouse reference genome (GRCm38, GENCODE vM15, primary assembly) using BWA 90 . Reads mapping to blacklisted regions were removed, as well as those marked as duplicates (Picard v3.2.0), not uniquely mapped (SAMtools ’-F 0×004 -F 0×400 -q 1’), or containing > 4 mismatches. Narrow peaks were called by MACS3 (v 3.0.1) 91 for each sample separately (each with own input as control) and subsequently merged to form the final peakset. These peaks were annotated relative to gene features using HOMER’s annotatePeaks.pl (v 4.11) 92 , and reads within them counted with featureCounts (v4.0.1) 93 . For the previously published ChIP-seq dataset from 44 , FASTQ files were downloaded from GEO GSE116990 using nf-core/fetchngs, and reprocessed with nf-core/chipseq pipeline v2.1.0 with parameters (--macs_gsize 2652783500 --narrow_peak --macs_fdr 0.0001 -- min_reps_consensus 2). For H3K27ac and H3K4me1 this used the consensus peaks of 2 replicates against input (WT_0h), and for HDAC3 peaks this just used ChIPseq_FLAG_TX1072_Hdac3-AID-FLAG_0h_Dox_Rep1 against ChIPseq_FLAG_TX1072_WT_0h_Dox_Rep1. Consensus peaks for H3K27ac and H3K4me1 were further overlapped using BEDTools intersect 94 , identifying 47,507 putative active regulatory elements. These were further classified based on presence of HDAC3 peaks within 100 bp distance (using bedtools window -w 100 -u) and motif analysis was performed using HOMER’s findMotifsGenome.pl (mm10 -size given -mask) on those with overlapping or nearby HDAC3 peaks versus those without. Transcriptomics and nascent transcriptomics TT chem -seq Analysis of nascent transcriptome was performed as described by Gregersen et al. 53 . Briefly, cells were labelled for 15 minutes with 1LmM of the nucleotide analogue 4-thiouridine (4sU), followed by RNA extraction using TRIzol and isopropanol precipitation. Extracted RNA (100 µg) was mixed to 1 µg of Saccharomyces cerevisiae RNA, labelled for 5 minutes with 5 mM 4-thiouracil (4tU). RNA mixture was fragmented with 0.2LM NaOH on ice for 20Lmin, neutralized, and purified using phenol:chloroform:isoamylalkohol extraction and isopropanol precipitation. Fragmented RNA was biotinylated using MTSEA-biotin-XX (Insight Biotechnologies, cat.no.: 90066, dissolved in DMF) and purified by phenol:chloroform extraction and isopropanol precipitation. Biotin-labeled RNA was enriched using streptavidin beads (Miltenyi μMACS), washed, and eluted with DTT. RNA was purified with RNeasy MinElute Cleanup kit (QIAGEN)columns and quality checked with TapeStation (Agilent). Libraries were prepared using the NEBNext Ultra II Directional RNA Kit and dual-index UMI adapters (NEB), following the rRNA-depleted FFPE RNA protocol. Single-end sequencing was performed in a P3-50bp kit on a NestSeq2000 sequencer. Data processing Raw sequencing data were demultiplexed and converted into FASTQ files using Illumina bcl2fastq Conversion Software (v2.20.0.422). UMI-tools extract (v 1.1.4) 95 was used to extract the UMI barcode from read2 and add it to the read name of read1, adapter sequences and low-quality bases were trimmed using Trim Galore. Reads were first mapped to a single copy of the ribosomal DNA repeat unit (GenBank: BK000964.3 ) to remove related transcribed sequences. Reads that did not map to rDNA were then mapped to yeast genome (sacCer3/R64 assembly), and the unmapped reads subsequently mapped to the mouse genome (GRCm38, GENCODE vM15, primary assembly) using STAR (v2.7.2d) 96 . Resulting bam files were deduplicated using UMI-tools dedup, and lastly count matrices for genes (as annotated in GENCODE vM15) were generated using featureCounts (-s 2 -t gene). To identify non-genic transcripts from nascent RNA sequencing reads, HOMER (v4.11) 92 was used to first combine deduplicated bam files from 2iLIF conditions (+/-IAA) into one tag directory and then run findPeaks on it with parameters ’-style groseq -rev -minBodySize 100’. Subsequently, all identified transcripts overlapping the sense strand of annotated genes (GENCODE vM15) or 5kb downstream were removed using bedtools window (v2.31.0) 94 , and remaining transcripts within 100bp were merged (bedtools merge -s -d 100), resulting in 30,195 transcripts. Overlapping reads were counted with featureCounts (-s 2). RNA-sequencing RNA was extracted using RNAeasy Plus Mini kit with DNA digestion (QIAGEN). RNA integrity was assessed using Fragment Analyzer (AATI). For library preparation we used the Universal Plus™ Total RNA-Seq with NuQuant ® (TECAN) following manufacturer’s instructions. Single-end sequencing was performed using Illumina P3-50bp kit on a NextSeq2000 sequencer. Mouse breeding, embryo collection and dissection Timed natural matings were used for all experiments. Noon of the day when the vaginal plugs of mated females were identified was scored as E0.5. A published conditional allele of Hdac3 loxP was used 97 . To obtain Hdac3 Het or Hdac3 KO embryos, Hdac3 loxP/loxP Zp3-Cre +ve females were crossed with Hdac3 +/- males 74 . All work was carried out in accordance with European legislation, authorized by the Danish National Animal Experiments Inspectorate (Dyreforsøgstilsynet, license no. 2020-15-0201-00609 and 2020-15-0201-00608) and performed according to national guidelines. Flash-seq RNA extraction was performed using the Arcturus PicoPure ® (Applied Biosystems) from E4.5 blastocysts or E6.5 dissected epiblasts according to Chenoweth and Tesar 98 . RNA was eluted in 10 µL, 4 µL of which were used for library construction. Libraries were prepared as described here https://www.protocols.io/view/flash-seq-protocol-kxygxzkrwv8j/v3?version_warning=no 79 after volume adjustment. Paired-end sequencing was performed using P1-100bp & P2-100bp Illumina sequencing kit on a NextSeq2000 sequencer. Embryo genotyping Whole embryos were lysed in 10 µL of Genotyping Buffer (200 mM NaCl, 100 mM Tris-HCl, 0.5% Tween 20) + 0.8Lmg/mL Proteinase K by incubation for 4 hours at 56°C followed by 20 minutes at 95°C. Equal volume of PCR-grade water was added to each sample and the lysates were genotyped by PCR, with primers amplifying the Hdac3 locus, the Zp3-cre locus, and sex chromosome-linked loci. Alternatively, cDNA generated following the Flash-seq protocol was used as a template for Real Time RT-qPCR, with primers detecting the Hdac3 wild-type or mutated allele. Sex distinction was performed by primers detecting the Xist RNA or the transcript of the Y-linked Eif2s3y gene. Data processing Raw sequencing data were demultiplexed and converted into FASTQ files using Illumina bcl2fastq Conversion Software (v2.20.0.422). FASTQ files were processed according to the FLASH-seq UMI protocol V.3 [ https://www.protocols.io/view/flash-seq-umi-protocol-bp2l619rdvqe/v3?step=12 ], focusing on the internal reads only, with read counting done on the level of ENSEMBL gene IDs as opposed to gene symbols. In brief, UMI-containing reads were identified using umi_tools extract and filtered out using BBTools’ filterbyname.sh, trimmed using BBDuk, mapped using STAR (v2.7.2d) against GRCm38, with parameters as given in the protocol. Lastly a count matrix was generated using subread’s featureCount (v2.0.6) with annotations from GENCODE release M15 primary assembly (--fracOverlap 0.25). Data Analysis - Bioinformatics We performed Differential Gene Expression Analysis using DESeq2 (1.46.0) 99 . The variance-stabilized transformed counts that were used for the Principal Component Analysis (PCA) plot were corrected using limma R package (3.62.2) 100 for batch effects introduced for either the sequencing run (FLASH-seq) or the collection day (replicate - TT chem -seq and RNA-seq runs). Raw uncorrected counts were used as input for DESeq2, and the batch parameter was taken into account in the model. Moreover, we used independent filtering from the in-built DESeq2 function with alpha set at 0.05. Gene expression changes were expressed as shrunk log2(Fold changes) using lfcShrink (type=”normal”). Transcript levels were calculated as sizeFactor-normalized fragments per kilobase per million reads (FPKM) using the argument “robust=TRUE”. For Gene Set Enrichment analysis, we used the collapsePathways function (fgsea R package 1.32.0) 101 and the lists were ranked based on the stat value generated from DESeq2. Differential Expression (DE) analysis for HOMER-identified transcripts was performed with DESeq2 using size factors calculated from genic reads. DE analysis for ChIP-seq and for Proteomics was performed using the edgeR package in r (4.4.2) 102 . Heatmaps were visualized using Easeq – Interactive Software for analysis and visualization of ChIP-seq data ( https://easeq.net ) 103 . Extended Data Figure Legends Extended Data Fig 1. HDAC3 resides at highly active enhancers. a–b , Heatmaps of all putative regulatory elements (±5,000 bp) bound ( a ) or not bound ( b ) by HDAC3. Shown is enrichment of HDAC3-FLAG and no FLAG control as well as H3K4me1, H3K4me3, and H3K27ac. Data was downloaded from Ż ylicz et al. 44 and originates from female ESCs. Regions are sorted in descending order of HDAC3 signal. Color intensity indicates reads per million per 1 kb, binned at 100 bp intervals. c , Bar plot showing genomic location of putative regulatory elements bound or not bound by HDAC3. Bar lengths represent percentages. Statistical significance was assessed using the chi-square test. d , Violin plot showing H3K27ac ChIP signal normalized to input at putative regulatory elements bound or not bound by HDAC3. Ratios were calculated from RPKMs. Statistical analysis was performed using the Wilcoxon–Mann–Whitney test. e , Proteins identified in P300 proximity ligation experiment in undifferentiated ESC. Shown are reanalyzed normalized LFQ intensities extracted from Stelloo et al. 52 . Statistical analysis was performed using DESeq2. Extended Data Fig 2. HDAC3 regulates highly active enhancers and their nearby genes . a , Immunoblot analysis of whole-cell extracts from TIR1 only or HDAC3-mAID-FLAG degron ESCs treated with indole acetic acid (IAA) for varying durations. Blots were probed with anti-HDAC3, anti-FLAG, anti-LAMIN B1 and anti-VINCULIN antibodies. b , Euler plot illustrating overlap of H3K27ac and H3K4me1 enriched putative regulatory elements with HOMER-transcript loci and/or HDAC3 peaks. c , GSEA for HDAC3-bound reRNAs against a ranked list of all identified HOMER transcripts. Ranking was performed using the stat value from DESeq2. d , GSEA of genes differentially expressed in ESCs in 2iLIF with or without IAA, for genes associated with GO terms of biological processes. e , Violin plot showing the distance of differentially expressed genes (as in d) from the nearest HDAC3 peak, categorized as upregulated, downregulated, or unchanged. Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. f , Quantification of transcriptional activity in ESCs treated with or without IAA for 3 hours, measured by EU incorporation and normalized to DAPI intensity. Statistical analysis was performed using the Mann–Whitney test on two experimental replicates and two independent clones (N=4). Extended Data Fig 3. HDAC3 catalytic activity is involved in its function in cis . a , Box plots showing transcription levels of genes upregulated upon HDAC3 loss in ESCs and quantified in ESCs treated with or without IAA, and complemented by HDAC3-WT, HDAC3-Y298F, or empty vector. b , MA-plot showing the gene shrunk log2(fold change) between HDAC3-mAID-FLAG degron 2iLIF ESCs complemented with HDAC3 Y298F vs empty vector. X axis shows nascent transcription levels in unperturbed 2iLIF degron ESCs. c , Violin plot showing nascent transcription levels in the untreated 2iLIF ESCs. Plotted are upregulated, downregulated and unchanged genes upon HDAC3 Y298F expression versus empty vector control. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. d-e , GO-term analysis of genes upregulated (d) or downregulated (e) as described in ( b ). f, MA-plot as described in ( b ) illustrating differences between HDAC3-mAID-FLAG degron 2iLIF ESCs complemented with HDAC3 WT vs empty vector. g , Violin plot showing nascent transcription levels in the untreated 2iLIF ESCs. Plotted are upregulated, downregulated and unchanged genes upon HDAC3 WT expression versus empty vector control. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. h , Euler plot showing the overlaps between genes upregulated or downregulated in the following conditions: HDAC3-mAID-FLAG degron 2iLIF ESCs ±IAA, HDAC3-mAID-FLAG degron 2iLIF ESCs complemented with HDAC3 WT vs an empty vector or complemented with HDAC3 Y298F vs an empty vector. Extended Data Fig 4. HDAC3 role in enhancer regulation is specific to pluripotent cells. a , Immunofluorescence images of TSCs, stained for TSC markers CDX2 and EOMES, extraembryonic mesoderm marker T (BRACHYURY), and pluripotency marker POU5F1. b , MA-plot of shrunk log2(fold change) for differentially expressed genes in IAA-treated versus untreated TSCs against nascent transcription in unperturbed degron TSCs. c , GSEA of genes differentially expressed in TSCs in FAXY with or without IAA, for genes associated with GO terms of biological processes. d , Venn diagram showing overlap of DE genes in TSCs±IAA and DE genes in ESCs treated with or without IAA for 3 hours. e , Venn diagram showing overlap between the top quartile of most expressed genes in ESCs and TSCs. f , Histogram of common highly expressed genes from ( e ) showing the distribution of log 2 (fold changes) after IAA treatment in ESCs and TSCs. Extended Data Fig 5. HDAC3 regulates highly active enhancers and their nearby genes. a , Volcano plot showing distribution of expression changes for DE genes in ESCs switched to FA medium for 3 hours. b , Heatmap illustrating expression changes of selected genes as in ( a ). c , Violin plot showing basal transcription levels of DE genes in ESCs induced with FA medium for 3 hours, with or without IAA. Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. d , GSEA of genes differentially expressed in FA-induced ESCs with or without IAA, for genes associated with GO terms of biological processes. Extended Data Fig 6. HDAC3 prevents exit from naïve pluripotency and capacitation. a , Heatmap illustrating expression changes of selected naïve, core and primed pluripotency markers in ESCs and EpiLCs. b , Venn diagram showing the overlap of DE genes between identified by RNA-seq in the 48-hour EpiLC±IAA treatment and the DE genes identified by TT chem -seq in the 3-hour FA±IAA treatment. c-d , Immunofluorescence images ( c ) and quantification ( d ) of ESCs induced into EpiLCs with or without IAA, stained for a naïve pluripotency marker TBX3 and core pluripotency marker POU5F1. Statistics were performed using the the Wilcoxon–Mann–Whitney test. e , Schematic of the workflow for short reversion experiments after introduction of vectors encoding HDAC3-WT, HDAC3-Y298F, or the empty vector. Cells were induced in FA for 24 hours and reverted to 2iLIF for 24 hours. f , Violin plot of ratios between bootstrapped medians of TBX3 immunofluorescence signal for ESCs that underwent short reversion. Bootstrapping was performed 5,000 times. Statistics used Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. g , Violin plots of per plate colony count in long term reversion experiments (as in Fig. 5h ). Each dot represents an independent experimental replicate. Representative images of plates used for quantifications. Statistical analysis was performed using Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. Extended Data Fig 7. HDAC3 prevents exit from naive pluripotency. a , Schematic of workflow for short reversion experiment: cells induced in FA for 24 hours, then reverted to 2iLIF for 24 hours. b–c , Immunofluorescence images (c) of cells as in (a), and quantification of TBX3 signal normalized to DAPI intensity (b). Two experimental replicates are shown (top and bottom). Statistical analysis was performed using Mann-Whitney test. Extended Data Fig 8. HDAC3 is required for early mouse development. a , Representative wholemount immunofluorescence images of Hdac3 -Het or Hdac3 -KO mouse E6.5 embryos. DAPI signal is cyan; NANOG is magenta. Epiblast volume is highlighted in white. b , Quantification of the ratio of NANOG signal to epiblast volume, using Imaris Software. Statistical analysis was performed using unpaired t-test. c , Volcano plot analysis of gene expression levels in E6.5 female Hdac3- KO vs. Hdac3- Het epiblasts with shrunk log 2 (fold change) and -log10(adjusted p-value) calculated using DESeq2. Differentially expressed genes with adjusted p-value<0.05 are highlighted in color. d , Violin plot showing gene expression levels in the Hdac3- Het female E6.5 epiblasts. Plotted are upregulated, downregulated and unchanged genes in Hdac3- KO vs. Hdac3- Het E6.5 female epiblasts. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. e , GSEA of differentially expressed genes in E6.5 Hdac3- KO vs. Hdac3- Het female epiblasts for genes associated with GO terms of biological processes. f–g , Venn diagrams showing overlap of DE genes changing in E6.5 Hdac3- KO vs. Hdac3- Het male E6.5 epiblasts and DE genes changing in 2iLIF±IAA and FA±IAA conditions, respectively. h-i , Volcano plot analysis of gene expression levels in E4.5 male (h) and female (i) Hdac3- KO vs. Hdac3- Het embryos with shrunk log 2 (fold change) and -log10(adjusted p-value) calculated using DESeq2. Differentially expressed genes with adjusted p-value<0.05 are highlighted in color. Shown are also violin plots quantifying gene expression levels in the Hdac3- Het male(h) and female(i) E4.5 embryos. Plotted are upregulated, downregulated and unchanged genes in Hdac3- KO vs. Hdac3- Het embryos. Transcript levels are size-factor normalized fragments per kilobase per million reads (FPKMs). Statistical analysis was performed using the Kruskal–Wallis rank sum test followed by Dunn’s test. P -values were corrected using the Benjamini–Hochberg method. Authors Contributions Statement Conceptualization: N.S., J.J.Z.; Data curation: N.S., A.W., J.J.Z.; Formal analysis: N.S., A.W., J.J.Z.; Funding acquisition: L.H.G., J.A.H., J.J.Z.; Investigation (omics experiments): N.S., S.K., G.N. K.K.U., A.W. J.J.Z..; Investigation ( in vivo experiments): N.S., S.B.A., J.J.Z.; Investigation (stem cell experiments): N.S.; Project administration: J.J.Z; Resources: L.H.G., J.J.Z.; Software: N.S., K.K.U., A.W.; Supervision: L.H.G., J.A.H., J.J.Z; Validation: N.S.; Visualization: N.S.; Writing, reviewing and editing of the manuscript: N.S., A.W, K.K.U., L.H.G., J.A.H., J.J.Z. Competing interests Statement All authors declare no competing interests. Acknowledgments This work was supported by grants from: Novo Nordisk Fonden [NNF21CC0073729(JJZ), NNF20OC0059959 (LHG)], Danmarks Frie Forskningsfond [0169-00031B(JJZ), 0134-00190B (JJZ), 0165-00092B(LHG)], Danish National Research Foundation [DNRF154(LHG)] and European Research Council [ChroMeta: 101077271(JJZ), TranscriptStress: 101076758(LHG)]. We are thankful to Dr François Dossin for reading and reviewing the manuscript, and also for sharing the HDAC3-mAID-FLAG vector. We are grateful to Prof. Mitchell Lazar for sharing the Hdac3 mouse line. We are also specially grateful for the critical feedback of Joshua Brickman. We thank reNEW platforms for technical expertise, support, and use of equipment in particular: H. Wollmann, M. Michaut, J. Bulkescher, and A. Kalvisa. We also thank EMBL Rome core facilities, particularly Gerald Pfister (FCF). Language in some paragraphs was proof-read with Copilot and Perplexity; however no passage was written by and no references were added using large language models. Funder Information Declared Novo Nordisk Foundation, https://ror.org/04txyc737 , NNF21CC0073729 , NNF20OC0059959 Danmarks Frie Forskningsfond , 0169-00031B , 0134-00190B , 0165-00092B Danish National Research Foundation, https://ror.org/00znyv691 , DNRF154 European Research Council, https://ror.org/0472cxd90 , 101077271 , 101076758 References 1. ↵ Zabidi , M. A. & Stark , A . Regulatory Enhancer–Core-Promoter Communication via Transcription Factors and Cofactors . Trends Genet . 32 , 801 – 814 ( 2016 ). OpenUrl CrossRef PubMed 2. ↵ Cramer , P . Organization and regulation of gene transcription . Nature 573 , 45 – 54 ( 2019 ). OpenUrl CrossRef PubMed 3. ↵ Malik , S. & Roeder , R. G . Regulation of the RNA polymerase II pre-initiation complex by its associated coactivators . Nat. Rev. Genet . 24 , 767 – 782 ( 2023 ). OpenUrl PubMed 4. ↵ Larke , M. S. C. et al. Enhancers predominantly regulate gene expression during differentiation via transcription initiation . Mol. Cell 81 , 983 – 997 .e7 ( 2021 ). OpenUrl CrossRef PubMed 5. ↵ Bannister , A. J. & Kouzarides , T . Regulation of chromatin by histone modifications . Cell Res . 21 , 381 – 395 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 6. ↵ Lee , K. K. & Workman , J. L . Histone acetyltransferase complexes: one size doesn’t fit all . Nat. Rev. Mol. Cell Biol . 8 , 284 – 295 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 7. Marmorstein , R. & Zhou , M.-M . Writers and Readers of Histone Acetylation: Structure, Mechanism, and Inhibition . Cold Spring Harb. Perspect. Biol . 6 , a018762 – a018762 ( 2014 ). OpenUrl Abstract / FREE Full Text 8. ↵ Seto , E. & Yoshida , M . Erasers of histone acetylation: the histone deacetylase enzymes . Cold Spring Harb. Perspect. Biol . 6 , a018713 ( 2014 ). 9. ↵ Dovey , O. M. , Foster , C. T. & Cowley , S. M . Histone deacetylase 1 (HDAC1), but not HDAC2, controls embryonic stem cell differentiation . Proc. Natl. Acad. Sci . 107 , 8242 – 8247 ( 2010 ). OpenUrl Abstract / FREE Full Text 10. Dovey , O. M. et al. Histone deacetylase 1 and 2 are essential for normal T-cell development and genomic stability in mice . Blood 121 , 1335 – 1344 ( 2013 ). OpenUrl Abstract / FREE Full Text 11. ↵ English , D. M. et al. Rapid degradation of histone deacetylase 1 (HDAC1) reveals essential roles in both gene repression and active transcription . Nucleic Acids Res . 53 , ( 2025 ). 12. Haberland , M. , Mokalled , M. H. , Montgomery , R. L. & Olson , E. N . Epigenetic control of skull morphogenesis by histone deacetylase 8 . Genes Dev . 23 , 1625 – 1630 ( 2009 ). OpenUrl Abstract / FREE Full Text 13. Hagelkruys , A. et al. A single allele of Hdac2 but not Hdac1 is sufficient for normal mouse brain development in the absence of its paralog . Development 141 , 604 – 616 ( 2014 ). OpenUrl Abstract / FREE Full Text 14. Kelly , R. D. W. et al. Histone deacetylases maintain expression of the pluripotent gene network via recruitment of RNA polymerase II to coding and noncoding loci . Genome Res . 34 , 34 – 46 ( 2024 ). OpenUrl Abstract / FREE Full Text 15. ↵ Lagger , G . Essential function of histone deacetylase 1 in proliferation control and CDK inhibitor repression . EMBO J . 21 , 2672 – 2681 ( 2002 ). OpenUrl Abstract / FREE Full Text 16. LeBoeuf , M. et al. Hdac1 and Hdac2 Act Redundantly to Control p63 and p53 Functions in Epidermal Progenitor Cells . Dev. Cell 19 , 807 – 818 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 17. Ma , P. & Schultz , R. M . HDAC1 and HDAC2 in mouse oocytes and preimplantation embryos: Specificity versus compensation . Cell Death Differ . 23 , 1119 – 1127 ( 2016 ). OpenUrl CrossRef PubMed 18. Montgomery , R. L. et al. Histone deacetylases 1 and 2 redundantly regulate cardiac morphogenesis, growth, and contractility . Genes Dev . 21 , 1790 – 1802 ( 2007 ). OpenUrl Abstract / FREE Full Text 19. ↵ Montgomery , R. L. et al. Maintenance of cardiac energy metabolism by histone deacetylase 3 in mice . J. Clin. Invest . 118 , 3588 – 3597 ( 2008 ). OpenUrl CrossRef PubMed Web of Science 20. Montgomery , R. L. , Hsieh , J. , Barbosa , A. C. , Richardson , J. A. & Olson , E. N . Histone deacetylases 1 and 2 control the progression of neural precursors to neurons during brain development . Proc. Natl. Acad. Sci . 106 , 7876 – 7881 ( 2009 ). OpenUrl Abstract / FREE Full Text 21. Norwood , J. , Franklin , J. M. , Sharma , D. & D’Mello , S. R . Histone Deacetylase 3 Is Necessary for Proper Brain Development . J. Biol. Chem . 289 , 34569 – 34582 ( 2014 ). OpenUrl Abstract / FREE Full Text 22. ↵ Trivedi , C. M. et al. Hdac2 regulates the cardiac hypertrophic response by modulating Gsk3β activity . Nat. Med . 13 , 324 – 331 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 23. ↵ Bhaskara , S. et al. Deletion of histone deacetylase 3 reveals critical roles in S phase progression and DNA damage control . Mol. Cell 30 , 61 – 72 ( 2008 ). OpenUrl CrossRef PubMed Web of Science 24. ↵ Wen , Y.-D. et al. The histone deacetylase-3 complex contains nuclear receptor corepressors . Proc. Natl. Acad. Sci . 97 , 7202 – 7207 ( 2000 ). OpenUrl Abstract / FREE Full Text 25. ↵ Mottis , A. , Mouchiroud , L. & Auwerx , J . Emerging roles of the corepressors NCoR1 and SMRT in homeostasis . Genes Dev . 27 , 819 – 835 ( 2013 ). OpenUrl Abstract / FREE Full Text 26. ↵ Guenther , M. G. , Barak , O. & Lazar , M. A . The SMRT and N-CoR corepressors are activating cofactors for histone deacetylase 3 . Mol. Cell. Biol . 21 , 6091 – 6101 ( 2001 ). OpenUrl Abstract / FREE Full Text 27. ↵ Sun , Z. et al. Deacetylase-Independent Function of HDAC3 in Transcription and Metabolism Requires Nuclear Receptor Corepressor . Mol. Cell 52 , 769 – 782 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 28. ↵ You , S.-H. et al. Nuclear receptor co-repressors are required for the histone-deacetylase activity of HDAC3 in vivo . Nat. Struct. Mol. Biol . 20 , 182 – 187 ( 2013 ). OpenUrl CrossRef PubMed 29. ↵ Guenther , M. G. et al. A core SMRT corepressor complex containing HDAC3 and TBL1, a WD40-repeat protein linked to deafness . Genes Dev . 14 , 1048 – 1057 ( 2000 ). OpenUrl Abstract / FREE Full Text 30. Oberoi , J. et al. Structural basis for the assembly of the SMRT/NCoR core transcriptional repression machinery . Nat. Struct. Mol. Biol . 18 , 177 – 184 ( 2011 ). OpenUrl CrossRef PubMed 31. ↵ Zhang , J. , Kalkum , M. , Chait , B. T. & Roeder , R. G . The N-CoR-HDAC3 Nuclear Receptor Corepressor Complex Inhibits the JNK Pathway through the Integral Subunit GPS2 . Mol. Cell 9 , 611 – 623 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 32. ↵ Rosenfeld , M. G. , Lunyak , V. V. & Glass , C. K . Sensors and signals: a coactivator/corepressor/epigenetic code for integrating signal-dependent programs of transcriptional response . Genes Dev . 20 , 1405 – 1428 ( 2006 ). OpenUrl Abstract / FREE Full Text 33. ↵ Evans , R. M. & Mangelsdorf , D. J. Nuclear Receptors, RXR, and the Big Bang . Cell 157 , 255 – 266 ( 2014 ). OpenUrl CrossRef PubMed Web of Science 34. ↵ Emmett , M. J. & Lazar , M. A . Integrative regulation of physiology by histone deacetylase 3 . Nat. Rev. Mol. Cell Biol . 20 , 102 – 115 ( 2019 ). OpenUrl CrossRef PubMed 35. ↵ Feng , D. et al. A Circadian Rhythm Orchestrated by Histone Deacetylase 3 Controls Hepatic Lipid Metabolism . Science 331 , 1315 – 1319 ( 2011 ). OpenUrl Abstract / FREE Full Text 36. Knutson , S. K. et al. Liver-specific deletion of histone deacetylase 3 disrupts metabolic transcriptional networks . EMBO J . 27 , 1017 – 1028 ( 2008 ). OpenUrl Abstract / FREE Full Text 37. Sun , Z. et al. Hepatic Hdac3 promotes gluconeogenesis by repressing lipid synthesis and sequestration . Nat. Med . 18 , 934 – 942 ( 2012 ). OpenUrl CrossRef PubMed 38. ↵ Yin , L. & Lazar , M. A . The Orphan Nuclear Receptor Rev-erbα Recruits the N-CoR/Histone Deacetylase 3 Corepressor to Regulate the Circadian Bmal1 Gene . Mol. Endocrinol . 19 , 1452 – 1459 ( 2005 ). OpenUrl CrossRef PubMed Web of Science 39. ↵ Ferrari , A. et al. HDAC3 is a molecular brake of the metabolic switch supporting white adipose tissue browning . Nat. Commun . 8 , ( 2017 ). 40. ↵ Richter , H. J. et al. Balanced control of thermogenesis by nuclear receptor corepressors in brown adipose tissue . Proc. Natl. Acad. Sci . 119 , ( 2022 ). 41. ↵ Lackner , A. et al. The Fgf/Erf/NCoR1/2 repressive axis controls trophoblast cell fate . Nat. Commun . 14 , 2559 ( 2023 ). OpenUrl PubMed 42. ↵ Olivieri , D. et al. Cooperation between HDAC3 and DAX1 mediates lineage restriction of embryonic stem cells . EMBO J . 40 , e106818 ( 2021 ). OpenUrl PubMed 43. ↵ McHugh , C. A. et al. The Xist lncRNA interacts directly with SHARP to silence transcription through HDAC3 . Nature 521 , 232 – 236 ( 2015 ). OpenUrl CrossRef PubMed 44. ↵ Żylicz , J. J. , et al. The Implication of Early Chromatin Changes in X Chromosome Inactivation . Cell 176 , 182 – 197 .e23 ( 2019 ). OpenUrl CrossRef PubMed 45. ↵ Kafkia , E ., et al. TCA cycle rewiring underpins implantation and histone acetylation programming . BioRxiv Prepr . ( 2025 ) doi: 10.1101/2025.04.25.650409 . OpenUrl Abstract / FREE Full Text 46. ↵ Creyghton , M. P. et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state . Proc. Natl. Acad. Sci . 107 , 21931 – 21936 ( 2010 ). OpenUrl Abstract / FREE Full Text 47. ↵ Heintzman , N. D. et al. Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome . Nat. Genet . 39 , 311 – 318 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 48. ↵ Rada-Iglesias , A. et al. A unique chromatin signature uncovers early developmental enhancers in humans . Nature 470 , 279 – 283 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 49. ↵ Barski , A. et al. High-Resolution Profiling of Histone Methylations in the Human Genome . Cell 129 , 823 – 837 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 50. De Santa , F. et al. A Large Fraction of Extragenic RNA Pol II Transcription Sites Overlap Enhancers . PLoS Biol . 8 , e1000384 ( 2010 ). OpenUrl CrossRef PubMed 51. ↵ Pekowska , A. et al. H3K4 tri-methylation provides an epigenetic signature of active enhancers: Epigenetic signature of active enhancers . EMBO J . 30 , 4198 – 4210 ( 2011 ). OpenUrl Abstract / FREE Full Text 52. ↵ Stelloo , S. et al. Deciphering lineage specification during early embryogenesis in mouse gastruloids using multilayered proteomics . Cell Stem Cell 31 , 1072 – 1090 .e8 ( 2024 ). OpenUrl CrossRef PubMed 53. ↵ Gregersen , L. H. , Mitter , R. & Svejstrup , J. Q . Using TTchem-seq for profiling nascent transcription and measuring transcript elongation . Nat. Protoc . 15 , 604 – 627 ( 2020 ). OpenUrl CrossRef PubMed 54. ↵ Arner , E. et al. Transcribed enhancers lead waves of coordinated transcription in transitioning mammalian cells . Science 347 , 1010 – 1014 ( 2015 ). OpenUrl Abstract / FREE Full Text 55. Carullo , N. V. N. et al. Enhancer RNAs predict enhancer–gene regulatory links and are critical for enhancer function in neuronal systems . Nucleic Acids Res . 48 , 9550 – 9570 ( 2020 ). OpenUrl CrossRef PubMed 56. ↵ Mikhaylichenko , O. et al. The degree of enhancer or promoter activity is reflected by the levels and directionality of eRNA transcription . Genes Dev . 32 , 42 – 57 ( 2018 ). OpenUrl Abstract / FREE Full Text 57. ↵ Martinez-Ara , M. , Comoglio , F. , Van Arensbergen , J. & Van Steensel , B . Systematic analysis of intrinsic enhancer-promoter compatibility in the mouse genome . Mol. Cell 82 , 2519 – 2531 .e6 ( 2022 ). OpenUrl CrossRef PubMed 58. ↵ Peng , T. et al. STARR-seq identifies active, chromatin-masked, and dormant enhancers in pluripotent mouse embryonic stem cells . Genome Biol . 21 , 243 ( 2020 ). 59. Vos , E. S. M. et al. Interplay between CTCF boundaries and a super enhancer controls cohesin extrusion trajectories and gene expression . Mol. Cell 81 , 3082 – 3095 .e6 ( 2021 ). OpenUrl CrossRef PubMed 60. ↵ Xie , L. et al. A dynamic interplay of enhancer elements regulates Klf4 expression in naïve pluripotency . Genes Dev . 31 , 1795 – 1808 ( 2017 ). OpenUrl Abstract / FREE Full Text 61. ↵ Policarpi , C. , Munafò , M. , Tsagkris , S. , Carlini , V. & Hackett , J. A . Systematic epigenome editing captures the context-dependent instructive function of chromatin modifications . Nat. Genet . 56 , 1168 – 1180 ( 2024 ). OpenUrl CrossRef PubMed 62. ↵ Ohinata , Y. & Tsukiyama , T . Establishment of Trophoblast Stem Cells under Defined Culture Conditions in Mice . PLoS ONE 9 , e107308 ( 2014 ). OpenUrl CrossRef PubMed 63. ↵ Tanaka , S. , Kunath , T. , Hadjantonakis , A.-K. , Nagy , A. & Rossant , J . Promotion of Trophoblast Stem Cell Proliferation by FGF4 . Science 282 , 2072 – 2075 ( 1998 ). OpenUrl Abstract / FREE Full Text 64. Kubaczka , C. et al. Derivation and Maintenance of Murine Trophoblast Stem Cells under Defined Conditions . Stem Cell Rep . 2 , 232 – 242 ( 2014 ). OpenUrl 65. Niwa , H. , Miyazaki , J. & Smith , A. G . Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells . Nat. Genet . 24 , 372 – 376 ( 2000 ). OpenUrl CrossRef PubMed Web of Science 66. ↵ Niwa , H. et al. Interaction between Oct3/4 and Cdx2 Determines Trophectoderm Differentiation . Cell 123 , 917 – 929 ( 2005 ). OpenUrl CrossRef PubMed Web of Science 67. ↵ Boroviak , T. , Loos , R. , Bertone , P. , Smith , A. & Nichols , J . The ability of inner-cell-mass cells to self-renew as embryonic stem cells is acquired following epiblast specification . Nat. Cell Biol . 16 , 516 – 528 ( 2014 ). OpenUrl CrossRef PubMed 68. ↵ Hayashi , K. , Ohta , H. , Kurimoto , K. , Aramaki , S. & Saitou , M . Reconstitution of the Mouse Germ Cell Specification Pathway in Culture by Pluripotent Stem Cells . Cell 146 , 519 – 532 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 69. Huang , Y. , Osorno , R. , Tsakiridis , A. & Wilson , V . In Vivo Differentiation Potential of Epiblast Stem Cells Revealed by Chimeric Embryo Formation . Cell Rep . 2 , 1571 – 1578 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 70. ↵ Kalkan , T. et al. Tracking the embryonic stem cell transition from ground state pluripotency . Development 144 , 1221 – 1234 ( 2017 ). OpenUrl Abstract / FREE Full Text 71. ↵ Buecker , C. et al. Reorganization of Enhancer Patterns in Transition from Naive to Primed Pluripotency . Cell Stem Cell 14 , 838 – 853 ( 2014 ). OpenUrl CrossRef PubMed Web of Science 72. ↵ Zylicz , J. J. et al. Chromatin dynamics and the role of G9a in gene regulation and enhancer silencing during early mouse development . eLife 4 , e09571 ( 2015 ). OpenUrl CrossRef PubMed 73. ↵ Hamilton , W. B. et al. Dynamic lineage priming is driven via direct enhancer regulation by ERK . Nature 575 , 355 – 360 ( 2019 ). OpenUrl CrossRef PubMed 74. ↵ de Vries , W. N. et al. Expression of Cre recombinase in mouse oocytes: a means to study maternal effect genes . Genes. N. Y. N 2000 26 , 110 – 112 ( 2000 ). OpenUrl 75. ↵ Mullican , S. E. et al. Histone deacetylase 3 is an epigenomic brake in macrophage alternative activation . Genes Dev . 25 , 2480 – 2488 ( 2011 ). OpenUrl Abstract / FREE Full Text 76. ↵ Barral , A. , et al. Nanog regulates Pou3f1 expression at the exit from pluripotency during gastrulation . Biol. Open 8 , bio046367 ( 2019 ). OpenUrl Abstract / FREE Full Text 77. Hatano , S. et al. Pluripotential competence of cells associated with Nanog activity . Mech. Dev . 122 , 67 – 79 ( 2005 ). OpenUrl CrossRef PubMed Web of Science 78. ↵ Sun , L. T. et al. Nanog co-regulated by Nodal/Smad2 and Oct4 is required for pluripotency in developing mouse epiblast . Dev. Biol . 392 , 182 – 192 ( 2014 ). OpenUrl CrossRef PubMed 79. ↵ Hahaut , V. et al. Fast and highly sensitive full-length single-cell RNA sequencing using FLASH-seq . Nat. Biotechnol . 40 , 1447 – 1451 ( 2022 ). OpenUrl CrossRef PubMed 80. ↵ Siersbæk , R. et al. Dynamic Rewiring of Promoter-Anchored Chromatin Loops during Adipocyte Differentiation . Mol. Cell 66 , 420 – 435 .e5 ( 2017 ). OpenUrl CrossRef PubMed 81. ↵ Emmett , M. J. et al. Histone deacetylase 3 prepares brown adipose tissue for acute thermogenic challenge . Nature 546 , 544 – 548 ( 2017 ). OpenUrl CrossRef PubMed 82. ↵ Paldi , F. et al. Transient histone deacetylase inhibition induces cellular memory of gene expression and three-dimensional genome folding . Preprint at doi: 10.1101/2024.11.21.624660 ( 2024 ). OpenUrl Abstract / FREE Full Text 83. ↵ Greer , C. B. et al. Histone Deacetylases Positively Regulate Transcription through the Elongation Machinery . Cell Rep . 13 , 1444 – 1455 ( 2015 ). OpenUrl CrossRef PubMed 84. ↵ Bone , R. A. et al. Altering metabolism programs cell identity via NAD+-dependent deacetylation . EMBO J . 44 , 3056 – 3084 ( 2025 ). OpenUrl PubMed 85. ↵ Tang , M. et al. Separation of transcriptional repressor and activator functions in Drosophila HDAC3 . Development 150 , dev201548 ( 2023 ). OpenUrl PubMed 86. ↵ Marahrens , Y. , Panning , B. , Dausman , J. , Strauss , W. & Jaenisch , R . Xist-deficient mice are defective in dosage compensation but not spermatogenesis . Genes Dev . 11 , 156 – 166 ( 1997 ). OpenUrl Abstract / FREE Full Text 87. ↵ Mulas , C. et al. Defined conditions for propagation and manipulation of mouse embryonic stem cells . Development 146 , dev173146 ( 2019 ). OpenUrl Abstract / FREE Full Text 88. ↵ Harshil Patel et al. nf-core/chipseq: nf-core/chipseq v2.1.0 - Platinum Willow Sparrow . Zenodo doi: 10.5281/ZENODO.3240506 ( 2024 ). OpenUrl CrossRef 89. ↵ Ewels , P. A. et al. The nf-core framework for community-curated bioinformatics pipelines . Nat. Biotechnol . 38 , 276 – 278 ( 2020 ). OpenUrl CrossRef PubMed 90. ↵ Li , H. & Durbin , R . Fast and accurate short read alignment with Burrows–Wheeler transform . Bioinformatics 25 , 1754 – 1760 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 91. ↵ Zhang , Y. et al. Model-based Analysis of ChIP-Seq (MACS) . Genome Biol . 9 , R137 ( 2008 ). 92. ↵ Heinz , S. et al. Simple Combinations of Lineage-Determining Transcription Factors Prime cis-Regulatory Elements Required for Macrophage and B Cell Identities . Mol. Cell 38 , 576 – 589 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 93. ↵ Liao , Y. , Smyth , G. K. & Shi , W . featureCounts: an efficient general purpose program for assigning sequence reads to genomic features . Bioinformatics 30 , 923 – 930 ( 2014 ). OpenUrl CrossRef PubMed Web of Science 94. ↵ Quinlan , A. R. & Hall , I. M . BEDTools: a flexible suite of utilities for comparing genomic features . Bioinformatics 26 , 841 – 842 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 95. ↵ Smith , T. , Heger , A. & Sudbery , I . UMI-tools: modeling sequencing errors in Unique Molecular Identifiers to improve quantification accuracy . Genome Res . 27 , 491 – 499 ( 2017 ). OpenUrl Abstract / FREE Full Text 96. ↵ Dobin , A. et al. STAR: ultrafast universal RNA-seq aligner . Bioinformatics 29 , 15 – 21 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 97. ↵ McQuown , S. C. et al. HDAC3 Is a Critical Negative Regulator of Long-Term Memory Formation . J. Neurosci . 31 , 764 – 774 ( 2011 ). OpenUrl Abstract / FREE Full Text 98. ↵ Isolation and Maintenance of Mouse Epiblast Stem Cells. in Methods in Molecular Biology 25 – 44 (Humana Press, Totowa, NJ, 2010 ). doi: 10.1007/978-1-60761-691-7_2 . OpenUrl CrossRef PubMed 99. ↵ Love , M. I. , Huber , W. & Anders , S . Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 . Genome Biol . 15 , ( 2014 ). 100. ↵ Ritchie , M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies . Nucleic Acids Res . 43 , e47 – e47 ( 2015 ). OpenUrl CrossRef PubMed 101. ↵ Korotkevich , G. et al. Fast gene set enrichment analysis . Preprint at doi: 10.1101/060012 ( 2016 ). OpenUrl Abstract / FREE Full Text 102. ↵ Chen , Y. , Chen , L. , Lun , A. T. L. , Baldoni , P. L. & Smyth , G. K . edgeR v4: powerful differential analysis of sequencing data with expanded functionality and improved support for small counts and larger datasets . Nucleic Acids Res . 53 , gkaf018 ( 2025 ). OpenUrl CrossRef PubMed 103. ↵ Lerdrup , M. , Johansen , J. V. , Agrawal-Singh , S. & Hansen , K . An interactive environment for agile analysis and visualization of ChIP-sequencing data . Nat. Struct. Mol. Biol . 23 , 349 – 357 ( 2016 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted August 12, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following HDAC3 prevents enhancer hyperactivation to enable developmental transitions Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share HDAC3 prevents enhancer hyperactivation to enable developmental transitions Nikolaos Stamidis , Anne Wenzel , Konrad Kacper Uściło , Smaragda Kompocholi , Sandra Bages-Arnal , Gemma Noviello , Lea Haarup Gregersen , James Alexander Hackett , Jan Jakub Żylicz bioRxiv 2025.08.09.669452; doi: https://doi.org/10.1101/2025.08.09.669452 Share This Article: Copy Citation Tools HDAC3 prevents enhancer hyperactivation to enable developmental transitions Nikolaos Stamidis , Anne Wenzel , Konrad Kacper Uściło , Smaragda Kompocholi , Sandra Bages-Arnal , Gemma Noviello , Lea Haarup Gregersen , James Alexander Hackett , Jan Jakub Żylicz bioRxiv 2025.08.09.669452; doi: https://doi.org/10.1101/2025.08.09.669452 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Developmental Biology Subject Areas All Articles Animal Behavior and Cognition (7633) Biochemistry (17680) Bioengineering (13889) Bioinformatics (41928) Biophysics (21445) Cancer Biology (18585) Cell Biology (25491) Clinical Trials (138) Developmental Biology (13374) Ecology (19897) Epidemiology (2067) Evolutionary Biology (24308) Genetics (15606) Genomics (22496) Immunology (17736) Microbiology (40385) Molecular Biology (17175) Neuroscience (88583) Paleontology (666) Pathology (2830) Pharmacology and Toxicology (4822) Physiology (7641) Plant Biology (15149) Scientific Communication and Education (2045) Synthetic Biology (4293) Systems Biology (9822) Zoology (2271)