Les modifications post-traductionnelles des histones jouent un rôle majeur dans la régulation de l’expression de gènes impliqués dans la plasticité et la mémoire. Parmi ces modifications, l’acétylation des histones permet le maintien de la chromatine dans un état « permissif », accessible pour la transcription. Nos travaux visent à identifier le rôle joué par l’acétylation de deux histones, H3 et H4, dans la formation de différentes formes de mémoire mettant en jeu les systèmes hippocampique et striatal chez la souris. Nous avons également recherché si des perturbations d’acétylation des histones sont responsables des déficits mnésiques observés au cours du vieillissement. Nous avons utilisé deux types d’apprentissage en piscine de Morris permettant de dissocier la mémoire spatiale, impliquant principalement l’hippocampe et la mémoire procédurale/indicée, impliquant le striatum. Nos résultats mettent en lumière une régulation différentielle de l’acétylation des histones dans l’hippocampe et le striatum selon la nature de la tâche et l’âge des animaux. L’apprentissage spatial induit une augmentation de l’acétylation des histones sélectivement dans l’hippocampe (CA1 et gyrus denté) alors que la tâche indicée augmente l’acétylation des histones spécifiquement dans le striatum. Nous montrons également que des changements opposés de l’acétylation de H3 (augmentation) et de H4 (diminution) dans l’hippocampe pourraient contribuer aux déficits de mémoire spatiale observés chez les souris âgées. Lors d’un test de compétition en piscine de Morris, durant lequel les souris ont le choix entre les stratégies spatiale et indicée pour résoudre la tâche, l’injection intra-hippocampique de Trichostatine A (TSA), un inhibiteur des histones déacétylases, immédiatement après l’apprentissage, perturbe la fonction striatale et favorise l’utilisation préférentielle de la stratégie spatiale hippocampique. Cependant, cet effet de la TSA est absent chez les souris âgées dont la fonction hippocampique est altérée. Dans une dernière série d’expérience, l’analyse des effets d’une injection intra-hippocampique de TSA, après un apprentissage spatial, a permis de préciser les contributions respectives des histones H3/H4 et du facteur de transcription CREB dans les déficits mnésiques associés au vieillissement. Dans leur ensemble, nos travaux apportent des éléments importants concernant l’importance de l’acétylation des histones dans la modulation des interactions entre systèmes de mémoire hippocampique et striatal. / Post-translational modifications of histone proteins play a crucial role in regulating plasticity and memory-related gene expression. Among these modifications, histone acetylation leads to a relaxed or “opened” chromatin state, permissive for transcription. Our work aims to identify the role played by histone H3 and H4 acetylation in the formation of different forms of memory involving hippocampal and striatal systems in mice. We also examined whether alterations of histone acetylation are responsible for age-associated memory deficits. We used two versions of the Morris water maze learning task to dissociate a spatial form of memory that relies on the hippocampus and a procedural/cued memory supported by the striatum. Our results highlight a differential regulation of histone acetylation within the hippocampus and striatum depending on the nature of the task and age of animals. Spatial and cued learning elicited histone acetylation selectively in the hippocampus (CA1 region and dentate gyrus) and the striatum, respectively. Age-related spatial memory deficits were associated with opposite changes in H3 acetylation (increase) and H4 (decrease) selectively in the hippocampus. During a water maze competition task in which mice can choose between spatial and cue-guided strategies, intra-hippocampal injection of Trichostatin A (TSA), an histone deacetylase inhibitor, immediately post-acquisition, impaired striatal function and promoted the use of a hippocampus-based spatial strategy. However, this effect of TSA was absent in old mice in which hippocampal function is impaired. In a final series of experiments, analysis of the effects of intra-hippocampal TSA injection immediately after a spatial training helped to clarify the respective contributions of histone H3/H4 and the transcription factor CREB in spatial memory deficits associated with aging. Taken together, our work provides important information regarding the importance of histone acetylation in modulating interactions between hippocampal and striatal memory systems.

L'obésité maternelle peut prédisposer aux pathologies métaboliques à l'âge adulte. Une perte de poids préconceptionnelle est recommandée aux femmes obèses, mais ses effets sur la croissance fœto-placentaire et la santé de la descendance adulte sont encore peu connus. Les objectifs de cette thèse étaient d’étudier les effets des trajectoires pondérales maternelles sur le phénotype de la descendance à terme et à l’âge adulte, ainsi que sur l’expression génique. Les descendants de mères obèses présentent une restriction de croissance fœtale, associée à des altérations d’expression des gènes de la machinerie épigénétique dans le foie fœtal et le labyrinthe placentaire. Notre étude souligne la sensibilité particulière de la machinerie d’acétylation des histones au métabolisme maternel. À l'âge adulte, les mâles nés des mères obèses développent une obésité aggravée lorsqu'ils sont exposés à un environnement obésogène. La perte de poids maternelle préconceptionnelle améliore la croissance fœtale et normalise le poids à l’âge adulte. Elle est donc bénéfique pour la descendance. Cependant, certains effets de l’obésité, corrigée par l’intervention nutritionnelle, sont conservés car le poids fœtal et l’expression d’une partie de gènes restent altérés. Ce travail apporte des premiers éléments sur les mécanismes du conditionnement développemental par les trajectoires pondérales maternelles. / Maternal obesity (OB) impacts fetal growth and adult offspring phenotype. It is still unknown whether the currently recommended preconceptional weight loss (WL) for obese women is beneficial for feto-placental growth and adult offspring health. The objectives of this thesis were to assess the effects of maternal weight trajectories on offspring phenotype at term and in adulthood, as well as gene expression in placenta and fetal liver. At E18.5, fetuses from obese females presented a fetal growth restriction (FGR); this FGR was almost completely abolished by maternal WL. Placental and hepatic expression of epigenetic machinery genes was affected by maternal OB, especially the histone acetylation pathway. Maternal WL normalized the expression of only a subset of these genes. Males born to OB mothers gained weight faster under high-fat diet than males born to control mothers; maternal WL rescued this phenotype. These results show that expression of epigenetic machinery genes and in particular histone acetylation regulators, is highly sensitive to maternal obesity. Preconceptional WL alleviates the effects of OB on fetal and adult weight but some effects of obesity cured by nutritional intervention were retained in offspring phenotype at term. This study is an important step toward understanding the mechanisms linking maternal nutrition to fetal growth and adult health.

Durant la dernière décennie, la problématique de santé liée aux perturbateurs endocriniens (PE) s’est étendue à la toxicité des mélanges. L’objectif de ce travail était de définir les conséquences d’une exposition continue à des cocktails des PE, à des doses faibles et définies comme «non nocives » par les autorités réglementaires. Des mélanges associant la génistéine, la vinclozoline, et le Bisphénol A, ont fait l’objet d’étude intégrative et multi-générationnelle chez le rat qui prend en compte le comportement maternel, le comportement alimentaire et le développement. Nos résultats montrent que ces mélanges peuvent: a) diminuer le comportement maternel, b) modifier les préférences gustatives (sucré, salé), c) affecter le développement dès la période utérine (malformations) jusqu’à l’âge adulte (surpoids), d) perturber le bilan métabolique (femelles) et l’expression par la glande salivaire de gènes codant des protéines impliquées dans la gustation, d’engendrer des effets épigénétiques sur sur la génération F2 non exposée. L’étude in vitro confirme que la Génistéine et/ou la Vinclozoline, introduites durant l’induction de la différenciation adipocytaire affectent le développement des 3T3-L1et leur activité endocrine (leptine; triglycérides), et révèle que la Vinclozoline potentialise l’effet anti-adipogénique de la Génistéine.En conclusion, ce travail montre qu’une exposition à des mélanges de PE peut altérer le comportement et le développement, et prédisposer l’organisme à développer des maladies métaboliques telles que le diabète et l’obésité, mais que les propriétés hormonales de chaque composant ne sont pas prédictives des effets cocktails / During the last decade, the issue of health-related endocrine disruptors (ED) has been extended to the toxicity of mixtures. The objective of this study was to define the effects of lifelong exposure to ED mixtures, at low doses defined as "non-harmful" by the authorities. In this aim, the effects of mixtures combining genistein, vinclozolin and bisphenol A, have been investigated in the rat by using an integrative and multi-generational experimental approach which takes into account maternal behavior, feeding behavior and development. Our results show that these mixtures could: a) reduce maternal behavior, b) change taste preferences (sweet, salty), c) affect the development from the in utero period (birth defects) up to adulthood (body overweight) d) disrupt the metabolic balance (females) and the salivary gland expression of genes encoding proteins involved in gustation, and e) generate epigenetic effects on the unexposed F2 generation.An in vitro study confirms that Genistein and / or Vinclozolin, introduced at the step of the induction of adipocyte differentiation affect the development and endocrine activity (leptin, triglycerides) of 3T3-L1 cells and reveals that Vinclozolin potentiates the anti-adipogenic effect of Genistein. To conclude, this study shows that exposure to PE mixtures could affect behavior and development, and could predispose the body to develop metabolic diseases such as diabetes and obesity, but the own hormonal properties of component could not be used to predict the cocktails toxicity

Mammals have two orthologs, Mll and Trx2, for the Drososphila protein Trithorax (TRX), which is the founding member of the trithorax group (TrxG) of epigenetic regulators. TrxG proteins are characterized by an evolutionary conserved SET domain. A major function of all SET domain- containing proteins is to modulate gene activity, but the underlying mechanisms are poorly understood. Apparently TRX, Mll and Trx2 are histone H3 lysine 4 specific methyltransferases. So far all evidence points to roles in expression of specific target genes. However, target genes and function of the epigenetic regulator Trx2 were still unknown. Homozygous trx2 mutant embryos arrest in development because of severe and widespread defects {Glaser, 2005 #296}. Thus mouse embryonic stem (ES) cells carrying a null mutation of trx2 were used as an alternative model system to address the implication of Trx2 in differentiation. This study showed that Trx2 is redundant for ES cell self-renewal. Homozygous trx2 knockout ES cells did not exhibit cell cycle defects. However, loss of Trx2 resulted in reduced proliferation and increased apoptosis rates in trx2-/- ES cells. Due to the fact that differentiation requires an appropriate rate of population growth, trx2-/- cells were affected adversely upon in vitro differentiation. Neurogeneic differentiation of trx2 mutant cells generated fewer mature neurons than wild type cells. Moreover a temporal delay in the developmental progression to differentiation became apparent. Cardiac differentiation of trx2-/- cells confirmed the developmental defect and temporal delay. Notably differentiation of trx2-/- cells was merely delayed or impaired but it was not absent, implying that Trx2 is not required for gene expression programs specific for neurons or cardiac myocytes. We propose that differentiation of trx2-/- ES cells is impaired because apoptosis is disturbing differentiation. Apart from analyzing the phenotype of trx2 mutant cells, this work was focused on the identification of Trx2 target genes. Oligonucleotide expression arrays were used to identify genes whose expression levels were affected by the absence of Trx2. In general, loss of Trx2 function resulted in more genes with decreased than increased expression levels. This is consistent with the hypothesis that Trx2 functions as a transcriptional activator. Comparison of gene expression profiles for constitutive and conditional trx2 mutant cells enabled a distinction between direct and indirect target genes for Trx2. As a result Magoh2 was identified as the key candidate target gene for Trx2. Interaction between Trx2 and Magoh2 suggested a potential regulatory role for Trx2 in alternative splicing. Furthermore this work provided evidence that Trx2 could be involved in the maintenance of CpG island promoter gene expression, thus providing a potent regulatory mechanism for ubiquitously expressed genes.

Maternal infection during pregnancy is associated with increased risk of both schizophrenia and autism in offspring. Based on this observation, the maternal immune activation mouse model was developed, in which pregnant rodents are treated with immune-activating agents and the brains and behavior of the adult offspring studied. This model has been found to recapitulate a variety of molecular, cellular, and behavioral abnormalities observed in both schizophrenia and autism. However, despite the abundant evidence provided by these studies that prenatal exposure to inflammation alters brain development and function later in life, the molecular mechanisms by which inflammation mediates these effects remains unclear. It has been suggested that other prenatal risk factors for neuropsychiatric disease may alter brain development, in part, via epigenetic mechanisms such as DNA methylation and histone modification. However, a link between inflammation and epigenetic modification in brain has not been established. Therefore, the focus of my thesis was to examine the effect of inflammation on the histone modification, trimethylated histone H3 lysine 4 (H3K4me3), which has been implicated in both normal brain development and in schizophrenia. In Chapter II, I describe experiments examining the effect of a specific, cytokine, interleukin-6 (IL-6), on H3K4me3 in rat forebrain culture. I show that IL-6 treatment results in altered levels of H3K4me3 at multiple gene promoters, frequently in conjunction with altered mRNA expression levels, and demonstrate that a subset of these alterations appear to be dependent on signaling via the signal transducer and activator of transcription 3 (Stat3) pathway. Furthermore, some of the genes affected by IL-6 also showed altered H3K4me3 levels in autism postmortem brain. Though a direct link still remains to be established, this observation suggests that epigenetic changes observed in neuropsychiatric disease may have been induced by prenatal exposure to inflammation. In Chapter III, I describe in vivo experiments employing the maternal immune activation (MIA) mouse model to examine the effects of prenatal inflammation on H3K4me3 in the brain of the offspring, at both fetal and adult stages. I found that immune activation resulted in increased levels of IL-6 protein in fetal brain, working memory deficits in the adult offspring, and subtle changes in H3K4me3 levels in fetal and adult brain. Taken together, these findings demonstrate that an environmental risk factor for schizophrenia and autism—namely, inflammation—is capable of inducing robust and widespread histone modifications in a model of the central nervous system and smaller changes in vivo. This suggests that prenatal exposure to inflammation in human populations may lead to increased susceptibility for neuropsychiatric disorders, in part, by altering chromatin modifications in developing brain.

DNA double-strand break (DSB) repair is essential for maintenance of genome stability. However, the compaction of the eukaryotic genome into chromatin creates an inherent barrier to any DNA-mediated event, such as during DNA repair. This demands that there be mechanisms to modify the chromatin structure and thus access DNA. Recent work has implicated a host of chromatin regulators in the DNA damage response and several functional roles have been defined. Yet the mechanisms that control their recruitment to DNA lesions, and their relationship with concurrent histone modifications, remain unclear. We find that efficient DSB recruitment of many yeast chromatin regulators is cell-cycle dependent. Furthering this, we find recruitment of the INO80, SWR-C, NuA4, SWI/SNF, and RSC enzymes is inhibited by the non-homologous end joining machinery, and that their recruitment is controlled by early steps of homologous recombination. Strikingly, we find no significant role for H2A.X phosphorylation (γH2AX) in the recruitment of chromatin regulators, but rather that their recruitment coincides with reduced levels of γH2AX. We go on to determine the chromatin remodeling enzyme Fun30 functions in histone dynamics surround a DSB, but does not significantly affect γH2AX dynamics. Additionally, we describe a conserved functional interaction among the chromatin remodeling enzyme, SWI/SNF, the NuA4 and Gcn5 histone acetyltransferases, and phosphorylation of histone H2A.X. Specifically, we find that the NuA4 and Gcn5 enzymes are both required for the robust recruitment of SWI/SNF to a DSB, which in turn promotes the phosphorylation of H2A.X.

Indiana University-Purdue University Indianapolis (IUPUI) / Histone acetylation has been linked to developmental changes in gene expression and is a validated drug target of apicomplexan parasites, but little is known about the roles of individual histone modifying enzymes and how they are recruited to target genes. The protozoan parasite Toxoplasma gondii (phylum Apicomplexa) is unusual among invertebrates in possessing two GCN5-family lysine acetyltransferases (KATs). While GCN5a is required for gene expression in response to alkaline stress, this KAT is dispensable for parasite proliferation in normal culture conditions. In contrast, GCN5b cannot be disrupted, suggesting it is essential for Toxoplasma viability. To further explore the function of GCN5b, we generated clonal parasites expressing an inducible HA-tagged form of GCN5b containing a point mutation that ablates enzymatic activity (E703G). Stabilization of this dominant-negative form of GCN5b was mediated through ligand-binding to a destabilization domain (dd) fused to the protein. Induced accumulation of the ddHAGCN5b(E703G) protein led to a rapid arrest in parasite replication. Growth arrest was accompanied by a decrease in histone H3 acetylation at specific lysine residues as well as reduced expression of GCN5b target genes in GCN5b(E703G) parasites, which were identified using chromatin immunoprecipitation coupled with microarray hybridization (ChIP-chip). We also demonstrate that GCN5b interacts with AP2-domain proteins, which are plant-like transcription factors in Apicomplexa. The interactions between GCN5b, AP2IX-7, and AP2X-8 were confirmed by reciprocal co-immunoprecipitation and revealed a “core complex” that includes the co-activator ADA2-A, TFIID subunits, LEO1 polymerase-associated factor (Paf1) subunit, and RRM proteins. The dominant-negative phenotype of ddHAGCN5b(E703G) parasites, considered with the proteomics and ChIP-chip data, indicate that GCN5b plays a central role in transcriptional and chromatin remodeling complexes. We conclude that GCN5b has a non-redundant and indispensable role in regulating gene expression required during the Toxoplasma lytic cycle.

Indiana University-Purdue University Indianapolis (IUPUI) / Epigenetics is defined as “the study of stable, often heritable, changes that influence gene expression that are not mediated by DNA sequence” (Fingerman et al., 2013). Epigenetic marks such as covalent histone modifications and DNA methylation are important for maintaining chromatin structure and epigenetic inheritance. Several proteins have been found to bind and/ or regulate epigenetic marks. One such protein, CXXC finger protein 1 (Cfp1) is an important chromatin regulator that binds to unmethylated CpG islands. It has been found to be essential for mammalian development. Mice lacking Cfp1 exhibit an embryonic- lethal phenotype. However, the function of Cfp1 can be studied using Cfp1 Null mouse ES cells, which are viable. Thus far, Cfp1 has been shown to be important for cell growth, cytosine methylation, histone modifications, subnuclear localization of Set1A histone H3K4 methyltransferase, and cellular differentiation. When Cfp1 Null ES cells are induced to differentiate by removal of Leukemia Inhibitory Factor (LIF), the cells are not able to turn off pluripotency markers such as Oct4 and alkaline phosphatase and fail to express differentiation markers such as Gata4 and Brachyury. In this study, we used established protocols to further examine the differentiation capacity of Cfp1 Null cells. Specifically, we tested the ability of Cfp1 Null ES cells to retain stem cell properties in the absence of LIF, differentiate into cardiomyocytes in the presence of TGF-β2 and differentiate into neuron precursors in the presence of retinoic acid (RA). While the differentiation effects of RA were inconclusive, Null cells were able to start differentiating in the absence of LIF, either as individual cells or EBs, and the presence of TGF-β2 when seeded on gelatin coated tissue culture dishes. However, no difference was seen between cells treated without LIF and those treated with TGF-β2. In both conditions, only a small portion of cells were able to differentiate, while the majority of the cell population retained stem cell characteristics. Cell growth and the differentiation capacity of Cfp1 Null cells were also compromised in comparison to WT cells. Thus, further supporting the need for the correct epigenetic patterns maintained by Cfp1 during cellular differentiation.

Les mutations somatiques sur le variant d’histone H3.3 et les régulateurs épigénétiques associés à H3.3 ont été identifiés dans 30 % des glioblastomes pédiatriques. Ces mutations sont caractérisées par des substitutions d'acides aminés à des positions spécifiques dans la région N-terminale de l'histone H3.3 telles que la glycine 34 en valine/arginine (G34V/R), l'alanine 29 en proline (A29P), ou une haplo-insuffisance de la protéine Zinc Finger MYND-Type Containing 11 (ZMYND11). ZMYND11 est un co-répresseur de la transcription qui se lie spécifiquement à H3.3K36me3 pour moduler l'activité de l'ARN polymérase II. De plus, il est intéressant de mentionner que l’interaction entre ZMYND11 et H3.3K36me3 est altérée lorsque le résidu G34 est muté en G34V. Récemment, les mutations germinales H3.3G34V, H3.3A29P et ZMYND11 ont été identifiées chez des patients présentant une déficience neurologique. Nous émettons l'hypothèse que les mutants H3.3G34V et H3.3A29P empêchent ZMYND11 de se lier à H3.3K36me3 et pourrait converger mécaniquement avec la perte de fonction de ZMYND11, ce qui perturberait la neurogenèse. À l'aide de la technologie CRISPR Cas9, nous avons généré des modèles mutants isogéniques à partir de cellules souches pluripotentes (iPSC) pour H3F3B-A29P, H3F3B-G34V et ZMYND11-knock-out (KO). Par la suite, nous avons stimulé la différenciation de ces modèles vers des lignées neuronales afin d’identifier si ces mutations affectent la neurogenèse. Enfin, en utilisant des méthodes de séquençage à haut-débit nous avons analysé le profil épigénomique et transcriptomique pour déterminer comment l’interaction entre ZMYND11 et H3K36me3 est perturbée et à quels degrés ces mutations impactent sur les modifications post-traductionnelles des histones. Ce projet permettra de mieux comprendre les fonctions de ZMYND11 sur le remodelage de la chromatine et sa fonction biologique au cours du développement cérébral. / Somatic mutations on the histone 3 variant H3.3 and H3.3-associated chromatin modifiers have been identified in 30% of pediatric high-grade gliomas (pHGG). The mutations are characterized by amino acid substitutions at specific positions within the histone H3.3 tail such as glycine 34 to valine/arginine (G34V/R), alanine 29 to proline (A29P), or haploinsufficiency of the chromatin reader Zinc Finger MYND-Type Containing 11 (ZMYND11). ZMYND11 is a transcriptional co-repressor that specifically reads H3.3K36me3 to modulate RNA polymerase II activity. Notably, binding of ZMYND11 to H3.3K36me3 is altered when G34 residue is mutated to G34V. Recently, germline mutations of H3.3G34V, H3.3A29P, and ZMYND11 have been identified in patients with Intellectual disability. We hypothesize that H3.3 G34V and H3.3A29P mutants impede the binding of ZMYND11 to H3.3K36me3 and may mechanistically converge with ZMYND11 loss-of-function mutation to perturb neurogenesis. Using CRISPR Cas9-mediated gene editing, we will generate isogenic human induced pluripotent stem cell (iPSC) models for H3F3B-A29P, H3F3B-G34V and ZMYND11-KO, and perform in vitro neural differentiation to identify whether specific neural lineages are affected. Next, using epigenomic and transcriptomic profiling we will study whether binding between ZMYND11 and H3K36me3 is disrupted, and the downstream impact on Post-Translational Modifications of histones (PTMs) and transcription. This project will lead to a better understanding of the crucial role of the chromatin reader ZMYND11 on chromatin remodeling and the biological function during neural development.

Au cours du développement précoce de l'embryon, une importante vague de reprogrammation épigénétique efface et rétablit les profils de méthylation d’ADN (metADN) à travers le génome. Cependant, des régions spécifiques telles que les gènes à empreinte doivent échapper à cette vague de reprogrammation et maintenir leurs profils de metADN précis par l’activité constante de l’enzyme DNMT1 (ADN méthyltransférase 1) pour assurer le bon développement embryonnaire. En utilisant un modèle de cellules souches embryonnaires (mES) de souris avec une répression inductible de Dnmt1 (Dnmt1tet/tet), nous avons précédemment montré que la perte temporaire de Dnmt1 déclenche la perte permanente des profils de metADN sur les régions à empreinte et régions similaires, ainsi que sur d'autres régions du génome. Nous ne comprenons toujours pas pourquoi certaines séquences génomiques sont incapables de rétablir leurs profils de metADN normaux après la ré-expression de Dnmt1, et comment d'autres marques épigénétiques (e.g. les modifications des histones) sont altérées. Notre hypothèse est qu’un réarrangement erroné des marques d’histones aux régions promotrices, suivant une perte temporaire du maintien de la méthylation d’ADN par DNMT1, empêchera l’expression normale dans les cellules souches embryonnaires de souris. Pour ce faire, nous avons collecté des cellules mES Dnmt1tet/tet avant l'inactivation de Dnmt1, après l'inactivation de Dnmt1, puis après la réactivation complète de l'expression de Dnmt1. Nous avons ensuite utilisé la technique ChIP-Seq pour les marques d'histones (H3K4me3, H3K27me3, H3K27ac, H3K9me3, H3K4me1), celle de RRBS pour la méthylation de l'ADN et la technique de RNA-Seq pour l'expression des gènes. En définissant une liste de 18 166 promoteurs uniques, nous les avons classés en quatre catégories (Actif, Bivalent, Déplété et Réprimé). Nous montrons que l'inactivation de Dnmt1 mène à une dérégulation drastique des marques d'histones à travers les types de promoteurs. Cependant, lors de la réactivation de Dnmt1, la plupart de ces défauts ont été corrigés. Pourtant, dans l’ensemble des catégories, nous observons des promoteurs avec des dysrégulations persistantes des marques d'histones ainsi qu'un nombre significatif de gènes avec une expression différentielle. Dans l'ensemble, nos résultats montrent qu'une absence temporaire de DNMT1 a un impact plus important sur la conservation des profils des marques d'histones et l'expression des gènes que sur le maintien des profils de metADN sur les régions promotrices, dans les cellules souches embryonnaires de souris. Cela suggère que l'absence temporaire de maintien de la méthylation d’ADN déclenche une série d'événements qui conduisent à des dérégulations permanentes de marques d'histones aux promoteurs, lesquelles ne sont pas directement associés aux altérations sous-jacentes de la méthylation d’ADN dans les régions promotrices. / During early embryo development, a major epigenetic reprogramming wave erases and re-establishes DNA methylation (DNAmet) profiles across the genome. However, specific regions such as imprinting loci must escape this reprogramming wave and maintain their precise DNAmet profiles by constant DNMT1 (DNA methyltransferase 1) activity to ensure the proper development. Using a mouse embryonic stem (mES) cell model with inducible Dnmt1 repression (Dnmt1tet/tet), we previously showed that the temporary loss of Dnmt1 triggers the permanent loss of DNAmet profiles on imprinted and imprinted-like regions, as well as on other regions across the genome. We still do not understand why particular genomic sequences are unable to re-establish their normal DNAmet profiles following Dnmt1 re-expression, and how other epigenetic marks (e.g., histone modifications) are altered. Our hypothesis is that an erroneous rearrangement of histone marks on promoter regions following a temporary lack of DNAmet maintenance by DNMT1 will prevent proper gene expression in mouse embryonic stem cells. To test this, we collected mESDnmt1tet/tet cells prior to Dnmt1 inactivation, after Dnmt1 inactivation, and following complete reactivation of Dnmt1 expression. We then performed ChIP-Seq for histone marks (H3K4me3, H3K27me3, H3K27ac, H3K9me3, H3K4me1), RRBS for DNA methylation and RNA-Seq for gene expression. By defining a list of 18 166 unique promoters we categorized them in four categories (Active, Bivalent, Depleted and Repressed). We show that inactivation of Dnmt1 lead to drastic dysregulation of histone marks across types of promoters. However, upon reactivation of Dnmt1, most of these defects were rescued. Still, across categories, we observe promoters with persistent histone mark dysregulations as well as a significant number of associated genes with differential expression. Overall, our results show that a temporary lack of DNMT1 has a greater impact on the conservation of histone mark profiles and gene expression than it has on the maintenance of DNAmet profiles on promoter regions in mouse embryonic stem cells. This suggests that the temporary lack of methylation maintenance triggers a series of events that leads to the permanent dysregulation of histone marks in promoter regions, which are not directly associated with underlying DNA methylation alterations in the promoter regions.