Epigenetic deregulation contributes significantly to the development of multiple human diseases, including cancer. While great effort has been made to elucidate the underlying mechanism, our knowledge on epigenetic regulation is still fragmentary, an important gap being how the diverse epigenetic events coordinate to control gene transcription. In the first part of our study, we demonstrated an important link between Snail-mediated transcriptional control and epigenetic regulation during cancer development. Specifically, we found that the highly conserved SNAG domain of Snail sequentially and structurally mimics the N-terminal tail of histone H3, thereby functions as a molecular “hook”, or pseudo substrate, for recruiting histone lysine specific demethylase 1 (LSD1) repressor complex to the E-cadherin promoter. Furthermore, we showed that Snail and LSD1 are both required for E-cadherin repression and EMT induction, and their expression is highly correlated with each other in multiple human tumor tissues. Our findings have important clinical ramifications in that compounds mimicking the SNAG domain may disrupt Snail-LSD1 interaction and inhibit EMT and metastasis. In the second part of our study, we designed a batch of compounds based on the structure of the SNAG domain and are currently screening for candidates capable of competing with SNAG peptide for LSD1 binding. In addition, we applied a peptide pulldown/mass spectrometry-coupled analysis to identify SNAG-interacting proteins, among which are many chromatin enzymes and modulators. Functional characterization of these proteins will help to elucidate the Snail-mediated epigenetic regulation process. In the third part of our study, we found that Snail interacts with poly(ADP-ribose) polymerase 1 (PARP1) through a potential pADPr-binding motif and is subject to poly(ADP-ribosyl)ation, which can stabilize the Snail-LSD1 complex for enhanced PTEN suppression under DNA damage condition. Our findings added another layer to the delicate Snail transcriptional machinery, and indicated that PARP inhibitors may be applied in combination with conventional chemotherapies to target cancers with high expression of Snail and LSD1. In summary, we demonstrated that Snail cooperates with multiple epigenetic machineries to induce EMT as well as survival of tumor cells. Our findings contribute to a better appreciation of Snail-mediated epigenetic network as well as diversification of therapeutic strategies against cancer.

京都大学 / 新制・課程博士 / 博士(医学) / 甲第24190号 / 医博第4884号 / 新制||医||1060(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 伊藤 能永, 教授 安達 泰治, 教授 椛島 健治 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

LSD1 et PHF2 sont des déméthylases de lysines capables de déméthyler à la fois les protéines histones qui influencent l’expression génique et les protéines non histones en affectant leurs activités ou stabilités. Des approches fonctionnelles d’inactivation de Lsd1 ou Phf2 chez la souris ont démontré l’implication de ces enzymes dans l'engagement des cellules progénitrices au cours de la différenciation. La myogenèse est l'un des exemples les mieux caractérisés sur la façon dont les cellules progénitrices se multiplient et se différencient pour former un organe fonctionnel. Elle est initiée par une expression temporelle spécifique des gènes régulateurs cibles. Parmi ces facteurs, MYOD est un régulateur clé de l'engagement dans la différenciation des cellules progénitrices musculaires. Bien que l’action de MYOD au cours de la différenciation cellulaire ait été largement étudiée, peu de chose sont connus sur les événements de remodelage de la chromatine associés à l'activation de l'expression de MyoD. Parmi les régions régulatrices de l'expression de MyoD, la région Core Enhancer (CE) qui est transcrite en ARN activateur non codant (CEeRNA) a été démontrée pour contrôler l'initiation de l'expression de MyoD au cours de l'engagement de myoblastes dans la différenciation.Nous avons identifié LSD1 et PHF2 comme des activateurs clés du CE de MyoD. L'invalidation in vitro et in vivo de LSD1 ou l'inhibition de l'activité enzymatique de LSD1 empêche le recrutement de l'ARN PolII sur le CE, empêchant l’expression du CEeRNA. D’après nos résultats, l'expression forcée du CEeRNA restaure efficacement l'expression de MyoD et la fusion myoblastique en l'absence de LSD1. De plus, PHF2 interagit avec LSD1 en régulant sa stabilité protéique.En effet, l'ablation in vitro de PHF2 entraîne une dégradation massive de LSD1 et donc une absence d'expression du CEeRNA. Cependant, toutes les modifications d'histones qui ont lieu dans la région du CE lors de l'activation de la différenciation ne peuvent pas être directement attribuées à l'activité enzymatique de LSD1 ou PHF2. Ces résultats soulèvent la question de l'identité des partenaires de LSD1 et PHF2, qui co-participeraient à l'expression du CEeRNA et donc à l'engagement des myoblastes dans la différenciation cellulaire. / LSD1 and PHF2 are lysine de-methylases that can de-methylate both histone proteins, influencing gene expression and non-histone proteins, affecting their activity or stability. Functional approaches using Lsd1 or Phf2 inactivation in mouse have demonstrated the involvement of these enzymes in the engagement of progenitor cells into differentiation. One of the best-characterized examples of how progenitor cells multiply and differentiate to form functional organ is myogenesis. It is initiated by the specific timing expression of the specific regulatory genes; among these factors, MYOD is a key regulator of the engagement into differentiation of muscle progenitor cells. Although the action of MYOD during muscle differentiation has been extensively studied, still little is known about the chromatin remodeling events associated with the activation of MyoD expression. Among the regulatory regions of MyoD expression, the Core Enhancer region (CE), which transcribes for a non-coding enhancer RNA (CEeRNA), has been demonstrated to control the initiation of MyoD expression during myoblast commitment. We identified LSD1 and PHF2 as key activators of the MyoD CE. In vitro and in vivo ablation of LSD1 or inhibition of LSD1 enzymatic activity impaired the recruitment of RNA PolII on the CE, resulting in a failed expression of the CEeRNA. According to our results, forced expression of the CEeRNA efficiently rescue MyoD expression and myoblast fusion in the absence of LSD1. Moreover PHF2 interacts with LSD1 regulating its protein stability. Indeed in vitro ablation of PHF2 results in a massive LSD1 degradation and thus absence of CEeRNA expression. However, all the histone modifications occurring on the CE region upon activation cannot be directly attributed to LSD1 or PHF2 enzymatic activity. These results raise the question of the identity of LSD1 and PHF2 partners, which co-participate to CEeRNA expression and thus to the engagement of myoblast cells into differentiation.

Le cartilage articulaire est un tissu avasculaire ayant une faible capacité de régénération. Ce tissu est essentiellement constitué d’un type cellulaire, les chondrocytes, inclus dans une matrice extracellulaire abondante et de composition très spécifique. L’arthrose, la maladie touchant le cartilage la plus fréquente, est caractérisée par la perte progressive de cette matrice extracellulaire, ce qui conduit à l’érosion des surfaces articulaires. Les causes sont multiples et encore mal comprises: inflammation, génétique, mécanique etc... Plusieurs études ont récemment mis en évidence l’implication des mécanismes épigénétiques dans la réponse des chondrocytes aux cytokines inflammatoires (contribuant au catabolisme du tissu).Notre but a été d’étudier le rôle encore peu exploré de ces mécanismes dans la synthèse de la matrice extracellulaire du cartilage (contribuant à l'anabolisme). En utilisant des chondrocytes articulaires humains en culture primaire, nous avons identifié des marques de méthylation de l'ADN étroitement associées à l’expression de gènes codant les principaux composants de la matrice cartilagineuse. Ceci apporte un nouvel éclairage sur l’instabilité du phénotype chondrocytaire. De plus, nous décrivons pour la première fois l'implication de la lysine déméthylase LSD1 (une enzyme modifiant l'état de la chromatine dont l’expression est augmentée dans le cartilage arthrosique), dans la régulation génique d'un collagène du cartilage, le collagène de type IX. L’ensemble des résultats met en évidence de nouveaux mécanismes de régulation génique dans les chondrocytes articulaires, qui pourraient être impliqués dans le développement de l’arthrose / The articular cartilage is an avascular tissue displaying a very limited regenerative capacity. This tissue is mainly composed of one cell type, the chondrocytes, which are embedded within an abundant and highly specialized extracellular matrix. Osteoarthritis, which is the most common joint disease, is characterized by the progressive loss of that matrix, leading to the erosion of articular surface. The causes of this pathology are multiple (genetic, biomechanical, inflammatory…) and are still not fully understood. Several studies have recently highlighted the involvement of epigenetic mechanisms in the chondrocyte response to inflammatory cytokines (contributing to cartilage catabolism).The aim of our work was to investigate the unexplored role of the epigenetic mechanisms in the ability of chondrocytes to synthesize the cartilage-specific matrix (contributing to cartilage anabolism). Using primary culture of human articular chondrocytes, we identified a DNA methylation profile closely associated with the expression of the genes encoding the main structural components of the extracellular matrix. These findings bring new insights in the comprehension of chondrocyte phenotype instability. Moreover, we report for the first time the involvement of the lysine demethylase LSD1, a chromatin-modifying enzyme highly expressed in osteoarthritic tissue, in the gene regulation of COL9A1, a cartilage-specific collagen. Altogether, these results highlight new mechanisms of gene regulation in articular chondrocytes, which may be involved in the development of osteoarthritis

<p>The capability to control the progression of the cell cycle, including the means to enter into a stable non-proliferative state, is essential for eukaryotic unicellular and multicellular organisms. A quiescent state similar to G0 of higher eukaryotes can be induced by nitrogen starvation of the fission yeast model organism Schizosaccharomyces pombe. Using high-resolution tiling arrays for genome-wide transcriptional profiling we explore the early transcriptional reprogramming on the route to quiescence. Furthermore, we demonstrate that cells carrying a mutation in the high mobility group (HMG) box of the histone lysine demethylase spLsd1 fail to acquire characteristics of quiescent cells and rapidly lose viability under nitrogen-starved conditions. Since no such defect is seen as a result of catalytic inactivation, the HMG domain of spLsd1 seems to confer a function to the protein that is independent of the histone demethylase activity. We show that the HMG domain of spLsd1 is required for transcriptional activation and repression of a large set of genes, both during vegetative growth and on the route to quiescence. We also confirm that spLsd1 is a repressor of antisense transcription, and that this function is at least partially dependent on the HMG domain of the protein.</p>

The capability to control the progression of the cell cycle, including the means to enter into a stable non-proliferative state, is essential for eukaryotic unicellular and multicellular organisms. A quiescent state similar to G0 of higher eukaryotes can be induced by nitrogen starvation of the fission yeast model organism Schizosaccharomyces pombe. Using high-resolution tiling arrays for genome-wide transcriptional profiling we explore the early transcriptional reprogramming on the route to quiescence. Furthermore, we demonstrate that cells carrying a mutation in the high mobility group (HMG) box of the histone lysine demethylase spLsd1 fail to acquire characteristics of quiescent cells and rapidly lose viability under nitrogen-starved conditions. Since no such defect is seen as a result of catalytic inactivation, the HMG domain of spLsd1 seems to confer a function to the protein that is independent of the histone demethylase activity. We show that the HMG domain of spLsd1 is required for transcriptional activation and repression of a large set of genes, both during vegetative growth and on the route to quiescence. We also confirm that spLsd1 is a repressor of antisense transcription, and that this function is at least partially dependent on the HMG domain of the protein.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第22326号 / 医博第4567号 / 新制||医||1041(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 滝田 順子, 教授 小川 誠司, 教授 遊佐 宏介 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Oral squamous cell carcinoma (OSCC), or oral cancer, accounts for the majority of head and neck cancers. Resistance to therapy is a challenge, and 5-year survival rate remains at ~50 percent. Lysine-specific demethylase 1 (LSD1) plays a crucial role in controlling cell homeostasis in health and disease. LSD1 is elevated in oral cancer and promotes metastasis and correlates with poor prognosis. LSD1 is a nuclear histone demethylase that has been implicated in maintaining the undifferentiated state of cancer-initiating stem cells and promoting OSCC. Large dataset analysis showed that genetic alterations, including upregulation of LSD1, are seen in clinical cancers including OSCC. This study aims to evaluate the unknown mechanism of LSD1 and determine if pharmacologic inhibition of LSD1 has preventative and/or therapeutic applications for OSCC. This study used the 4NQO mouse model to induce OSCC in mice and split the mice into 8 treatment groups. Each group received a different immunotherapy treatment (SP2509, Verteporfin, anti PD-1 and anti PD-L1 alone and in combination). Our results have shown that LSD1 inhibition reduces the development of gross pathologic lesions. LSD1 inhibition has also shown to cause differences in gene expression in preneoplasia and OSCC, attenuating many genes that are part of the pro-oncogenic gene network (LSD1, YAP, EGFR), immune checkpoints (PD-1 and PD-L1), and Hippo signaling effectors (YAP, TAZ). Interestingly, LSD1 has shown a role in regulating the immune microenvironment and promoting antitumor immunity, which led us to investigate LSD1 in combination with immune checkpoint antibodies (anti PD-1 and anti PD-L1). Our results show that LSD1 sensitizes to anti-PD-1 and anti-PD-L1 antibodies to treat mouse tongue OSCC. Thus, we showed for the first time that blocking LSD1 inhibits preneoplasia and OSCC feed-forward loop, which could have implications in OSCC prevention, chemo- and immunotherapeutic combinations.

Les récepteurs nucléaires sont des facteurs de transcription qui exercent leur fonction via le contrôle de la transcription de leurs gènes cibles, une régulation qui est dépendante de cofacteurs associés. Les complexes transcriptionnels ainsi formés dialogueront avec l’environnement chromatinien (méthylation de l’ADN, remodelage des nucléosomes, modifications post-traductionnelles des histones) afin de promouvoir la répression ou l’activation transcriptionnelle des cibles géniques de ces récepteurs. Ce projet a identifié une interaction entre la lysine déméthylase LSD1 et le récepteur nucléaire orphelin ERRα dans des cellules humaines de cancers du sein. LSD1 protège ERRα d’une dégradation protéasomale de manière indépendante de son activité catalytique. Par ailleurs, LSD1 déméthyle H3K9 et H3K4 in vivo, mais est incapable in vitro de déméthyler H3K9. La présence de ERRα révèle cette activité de LSD1 sur H3K9, suggérant que le complexe ERRα -LSD1 agit comme un régulateur positif de la transcription. En ce sens, ERRα et LSD1 régulent un nombre important de gènes communs identifiés par RNAseq. Ainsi, 10 gènes activés ont été sélectionnés et le recrutement de ERRα et LSD1 a été examiné sur ces cibles géniques. En association avec les résultats obtenus in vitro, nous avons observé in vivo qu’en absence de ERRα ou LSD1, les gènes activés par ces deux partenaires présentent une augmentation de la marque répressive H3K9me2 sans affecter H3K4me2 au niveau du site d’initiation de la transcription. En conclusion, LSD1 interagit avec ERRα et inhibe sa dégradation, conduisant à une coopération transcriptionnelle de ces protéines. Pour la première fois, un rôle direct de ERRα sur l’environnement chromatinien a été identifié via l’activité de LSD1 sur des marques répressives d’histones. / Nuclear receptors are transcription factors that cooperate with chromatin associated factors to promote their activities. These transcriptional complexes are able to modulate the chromatin landscape to repress or promote transcription. Interestingly, there is an intricate cross-talk between these complexes and the chromatin environment that can influence each other to coordinate gene expression led by nuclear receptors. Post-translational modifications of histones regulate in part, DNA accessibility and the activities of nuclear receptors. One of these histone modifiers is LSD1, which is known to demethylate lysines 4 (H3K4) and 9 (H3K9) on histone 3. This manuscript focuses on the discovered LSD1-ERRα complex in human cancer cell lines. LSD1 interacts with ERRα, hence, modulates ERRα protein stability via a demethylation independent manner. Moreover, LSD1 is able to demethylate H3K4me2 in vitro but not H3K9me2. Interestingly, we observed that ERRα is able to switch LSD1 activity toward H3K9me2 to promote gene transcription without any additional cofactor in vitro. To confirm this effect in vivo, a transcriptomic analysis on mammary cancer cells was performed and highlights common target genes between ERRα and LSD1. We selected 10 genes activated by both and verified ERRα and LSD1 recruitment on these targets. Moreover, upon knock-down of ERRα or LSD1, the transcriptional start sites of activated genes -bound and regulated by both proteins- are enriched in the repressive mark H3K9me2. Altogether, these results describe a positive regulation of ERRα by LSD1 which in turn, drives the demethylase activity on H3K9me2 to promote transcription. Finally, these data highlight a direct function of ERRα on chromatin landscape.

ERRα est un récepteur nucléaire dont l’activité est controlée par des co-régulateurs transcriptionnels. La forte expression de ERRα dans les cancers est corrélée à un mauvais pronostic. Les mécanismes par lesquels ERRα régule la migration des cellules cancéreuses sont mal compris, tout comme les co-régulateurs impliqués. Nous avons identifié deux enzymes modificatrices d’histone, LSD1 et SET7, agissant comme régulateurs positifs de ERRα.I. ERRα modifie les activités biochimiques de la déméthylase LSD1 vers la déméthylation (activatrice) de H3K9me2. L’activation des cibles de ERRa-LSD1 (identifiées par RNA-Seq) requiert le recrutement de ce complexe aux sites d’initiation de la transcription (TSSs), réalisé par le facteur de transcription NRF1 qui, lui, ne régule pas l’activité enzymatique de LSD1.II. Un autre groupe de cibles de ERRα (identifié par RNA-Seq) est sous le contrôle de l’histone méthyltransférase SET7 qui mono-méthyle H3K4. Le recrutement de SET7 aux TSSs est contrôlé par le facteur de transcription ETS1, qui promeut les interactions entre SET7 et ERRα, conduisant à l’activation de l’expression des gènes en aval.Des analyses par Gene Ontology ont montré que les cibles communes de ERRα/LSD1 et de ERRα/SET7 sont fortement enrichies en termes d’invasion cellulaire. De manière cohérente, la déplétion individuelle de chacun de ces facteurs (et également celle de NRF1 ou ETS1) réduit les capacités d’invasion, observée en tests in vitro (transwell) ou in vivo par xénogreffe sur embryons de poisson-zèbre.En résumé, nos résultats montrent deux réseaux de régulation impliquant des modifications d’histone induites par ERRα, conduisant à l’invasion cellulaire. / ERRα is a nuclear receptor whose activity mainly depends on its interaction with transcription co-regulators. High levels of ERRα are found in various cancer types and correlate with poor prognosis. However, the mechanisms linking ERRα to cancer cell migration as well as the coregulators involved are unclear. In our study, we found two histone-modifying enzymes, LSD1 and SET7, acting as positive regulators of ERRα.I. ERRα impacts the biochemical activities of the LSD1 demethylase. Activation of ERRα -LSD1 targets (identified by RNA-Seq) requires the recruitment of this complex at Transcriptional Start Sites (TSSs), which is achieved by the NRF1 transcription factor. In our study, we have shown several points: NRF1, but not ERRα , is involved in positioning LSD1 to TSS, whereas ERRα , but not NRF1, regulates LSD1 enzymatic activities towards demethylating H3K9me2.II. A distinct group of ERRa target genes (identified by RNA-Seq) is under the control of the histone methyltransferase SET7 which mono-methylates H3K4. Appropriate recruitment of SET7 at TSSs is controlled by the ETS1 transcription factor, promoting the interaction between SET7-ERRa, leading to target gene expression.Gene Ontology analysis revealed that ERRa-LSD1 co-targets, as well as ERRa-SET7 co-targets, are enriched in terms of cell invasion. Consistently, depletion of each of these factors, as well as depletion of NRF1 or ETS1, leads to reduced cell invasion capacities as observed in transwell assays or in vivo, using xenotransplantation in the zebrafish embryo.Altogether, our results show two regulatory networks involving histone modifications induced by nuclear receptors, leading to increased cell invasion.