L'auteur n'a pas fourni de résumé en français / L'auteur n'a pas fourni de résumé en anglais

Histone lysine methylation is an important epigenetic mark for regulation of gene expression and chromatin organization. Setdb1 (Set domain, bifurcate 1), one of the histone lysine methyltransferases, specifically methylates histone H3 at lysine 9 (H3K9) and participates in transcriptional repression and heterochromatin formation. The major task of my thesis work was to investigate the epigenetic roles of Setdb1 in regulating brain functions. I started my thesis work by examining Setdb1 expression pattern during mouse brain development. The most robust signal of Setdb1 was detected in the fetal brains at embryonic day 12.5, with a ubiquitous distribution in all the proliferative zones, as well as the cortical plate and other regions comprised of postmitotic neurons. The expression of Setdb1 decreased as the brain developed, and this down-regulation profile was correlated to neuronal maturation as examined in a primary culture model of mouse cortical neurons. I then generated CK-Setdb1 transgenic mice, in which a myc-tagged full length mouse Setdb1 was constantly expressed in postmitotic neurons under the control of the CaMK II alpha promoter (CK). The expression of mycSetdb1 was detected in NeuN positive cells throughout most forebrain regions including cerebral cortex, striatum and hippocampus. A sustained increase of Setdb1 in CK-Setdb1 transgenics was verified at both mRNA and protein levels. Furthermore, an increase of H3K9 trimethylation was detected at major satellite DNA repeats in CK-Setdb1forebrains, which indicated that transgene-expressed mycSetdb1 was functionally active in adult brains. The behavioral phenotype of CK-Setdb1 transgenics was examined by using two separate founder lines. Gross neurological functions including body weight, locomotion activity, motor coordination, and breeding behavior were generally normal in CK-Setdb1 mice. CK-Setdb1 mice were further subjected to behavioral paradigms related to mood and cognitive functions. Intriguingly, as compared to the littermate controls, CK-Setdb1 mice represent a lower level of depression as indicated by decreased total immobility in two different behavioral despair tests. Moreover, CK-Setdb1 mice showed an accelerated extinction in the learned helplessness paradigm after a delayed interval (7 days), indicating a faster recovery from an established status of despair. The potential confounding factors, like memory deficits, were ruled out as CK-Setdb1 mice showed normal or even improved performances in different memory-related paradigms. Anxiety scores and stimulant drug response were normal in CK-Setdb1mice. Taken together, these findings suggested that a specific antidepressant-like phenotype was elicited by the over-expression of Setdb1 in adult mice forebrains. To further study the molecular mechanism underlying Setdb1-associated antidepressant-like behavioral changes, I screened for Setdb1-binding sites in a genome-scale by ChIP-on-chip using a tiling microarray from Affymetrix. Unexpectedly, Setdb1 showed a very restricted binding profile with a high specificity towards ionotropic glutamate receptor genes including the NMDA receptor 2B subunit gene Grin2b, which is a new target for the treatment for major depression. An increase of H3K9 dimethylation at Setdb1-binding site on Grin2b locus was detected in CK-Setdb1 hippocampus, which was correlated to a decrease of Grin2b expression as well as an accelerated desensitization of NMDA receptor. Furthermore, Chromosome Conformation Capture (3C) on Grin2b locus revealed a repressive chromatin loop structure, which tethered the distal Setdb1-binding site (~ 32 Kb downstream of transcriptional start site (TSS)) to a proximal intronic region (~12 Kb downstream of TSS) that is enriched for the binding of KAP1, a well-studied Setdb1-interacting transcriptional corepressor. Taken together, our data indicated that Setdb1 repressed Grin2b expression via H3K9 hypermethylation and higher-order chromatin loop formation, which may contribute to the antidepressant-like phenotype we observed in CK-Setdb1mice. The second part of my thesis work was to investigate the role of Setdb1 in the animal model of a neurodevelopmental disorder - Rett syndrome (RTT). Loss-of-function mutations of the gene encoding methyl-CpG binding protein 2 (MECP2) is the primary cause of RTT. There is an overlap between Setdb1- and Mecp2-associated repressive chromatin machineries, which both include histone deacetylase complex, H3K9 methyltransferase, DNA methyltransferase and heterochromatin protein 1 (HP1). Moreover, in contrast to Setdb1, which is downregulated during the cortical neuronal differentiation, Mecp2 is upregulated and the expression level is positively correlated to neuronal maturation. Therefore, we hypothesized that there is a functional redundancy between Setdb1 and Mecp2, and the up-regulation of Setdb1 in mature neurons will compensate for brain deficiency due to the loss of Mecp2. To test this hypothesis, I crossed CK-Setdb1 transgenic mice with nestincre-Mecp2 conditional knockout mice (Mecp2-/y). The behavior changes of CK-Setdb1/Mecp2-/y mice, including body weight, locomotion, motor coordination, and life span, were then compared to Mecp2-/y mice. No significant improvements in behaviors or survival were observed from CK-Setdb1/Mecp2-/y mice. Because the activation of CK promoter is limited to defined population of postmitotic neurons in forebrain, I tested our hypothesis by generating another strain of Setdb1 overexpression mice – tauSetdb1, in which the expression of mycSetdb1 is under the control of an endogenous pan-neuronal active promoter Tau. However, the introduction of tauSetdb1 also failed to rescue Mecp2 deficiency. The life span of tauSetdb1/ Mecp2-/y was even shorter as compared to Mecp2-/y mice (Kaplan-Meier, p=0.07). In conclusion, up-regulation of Setdb1 in adult brain was not sufficient to rescue Mecp2deficiency in the mouse model of RTT. One of the most challenges to study neuronal dysfunctions in brain diseases is the cellular heterogeneity of central nervous system. Current techniques for chromatin studies, including chromatin immunoprecipitation (ChIP) assays, usually lack of single cell resolution and are unable to examine the neurobiological changes in defined cell populations. In the third part of my thesis work, I developed a modified protocol to isolate neuronal nuclei from brain homogenates via Fluorescence-Activated Cell Sorting (FACS). In general, total nuclei was extracted from frozen brains, neuronal nuclei were then immuno-tagged with NeuN and sorted via FACS. Besides the NeuN labeling-FACS protocol, I also generated CK-H2BeGFP transgenic mice, in which a histone H2B-eGFP (enhanced green fluorescent protein) fusion protein was expressed in the nuclei of postmitotic neurons in mouse forebrain. Nuclei extracted from CKH2BeGFP brain were directly applied for FACS sorting. By using this protocol, we routinely got around 6-8 x106neuronal nuclei from one adult mouse forebrain, which was sufficient for ChIP applications followed by single gene PCR and microarray studies. In conclusion, our protocol permits large-scale studies of chromatin modifications or any other nuclei events in defined cell populations from distinct brain regions. Taken together, my dissertation work will lead to a better understanding of the epigenetic roles of histone H3K9 methyltransferase Setdb1 in brain functions, and may provide new targets for the therapeutic treatment of major depression.