Ce travail de thèse a permis, dans un premier temps, de mettre en évidence de nouveaux grandsréarrangements dans trois pathologies étudiées au laboratoire : la mucoviscidose, la pancréatitechronique et l’hémochromatose. En particulier, ces travaux ont permis de trouver de nouveaux CNVs(Copy Number Variations) pathologiques dans le gène CFTR (Cystic Fibrosis Transmembraneconductance Regulator), de mieux comprendre les mécanismes d’un remaniement complexe dansPRSS1 (Protease Serine 1) et d’aider à caractériser finement un réarrangement dans HFE(Hemochromatosis). Ces études ont donc servi de preuve de concept pour l’utilisation de puces à ADN àl’échelle d’un gène et dans des zones difficiles car riches en séquences répétées.Dans un second temps, la recherche de facteurs de susceptibilité génétiques aux infarctus cérébraux(AICs) du sujet jeune a été réalisée chez 168 cas et 200 témoins âgés de moins de 40 ans. Dans notrepopulation, l’hypertension, les migraines, le tabac, et la prise de stupéfiants sont des facteurs de risqueimportants, multipliant respectivement par 35, 3,8, 4 et 2,8 le risque d’AIC. Notre étude pangénomiquepar CGH-array (Comparative Genomic Hybridization-array) a mis en évidence 98 régionspolymorphiques dans le génome humain. Parmi elles, la délétion d’une partie du gène NOTCH2, pourraitjouer un rôle protecteur dans la survenue des AICs (OR=0,11 [0,01-0,87] ; p=0,013) mais qui ne dépassepas le seuil fixé par la correction de Bonferroni). Ce travail a également mis en évidence environ 400CNVs rares, dont deux récurrents chez les cas, l'un portant les gènes CRELD2 (cysteine-rich with EGFlikedomains 2) et AGL12 (asparagine-linked glycosylation 12, alpha-1, 6-mannosyltransferase) (p=0,02)et le deuxième situé en 5’ du gène VBP1 (von Hippel-Lindau binding protein 1) (p=0,04). Enfin, uneapproche gènes candidats a été effectuée sur les gènes NOTCH2 et ALOX5AP (5-lipoxygenaseactivating protein) sans donner de résultats significatifs. Ceci a également été réalisé sur les mutationsprincipales de trois gènes de la coagulation (Facteur II, Facteur V Leiden et MTHFR). Une associationsignificative a été mise en évidence entre la C677T du gène MTHFR (5,10-methyltetrahydrofolate) et lesinfarctus cérébraux du sujet jeune (OR=2,39, p=0,02 pour le génotype TT). Ce travail de thèse a permisde confirmer l’existence de facteurs de risque environnementaux et génétiques déjà connus mais surtoutd’émettre de nouvelles hypothèses génétiques dans la survenue des AICs du sujet jeune. / The use of locus-specific array-CGH (Comparative Genomic Hybridization) has allowed us to detect largerearrangements in three pathologies of our laboratory: cystic fibrosis, chronic pancreatitis andhemochromatosis. We successfully observed new pathological CNV (Copy Number Variations) in theCFTR (Cystic Fibrosis Transmembrane conductance Regulator) gene and characterized complex eventsin PRSS1 (Protease Serine 1) and HFE (Hemochromatosis) genes, showing that the use of thistechnique is possible even in regions with high sequence homologies.We also confirmed that hypertension, migraine, tobacco and drugs are high significant risk factors forischemic strokes (IS) in young population (under 40 years) (OR=35, 3.8, 4 and 2.8, respectively). Then,we tried to identify new genetic susceptibility loci using a pangenomic approach. Among the 98 copynumber polymorphisms (CNP) observed, an interstitial NOTCH2 deletion is candidate for a protective rolein IS (OR=0.11 [0.01-0.87] ; p=0.013 before Bonferonni correction). We also observed approximately 400uncommon CNV, two of them being particularly reccurent in patients: a 22q13.31 duplication containingCRELD2 (cysteine-rich with EGF-like domains 2) and AGL12 (asparagine-linked glycosylation 12, alpha-1, 6-mannosyltransferase) genes (p=0.02) and a Xq28 deletion localised in the 5’ region of the VBP1 (vonHippel-Lindau binding protein 1) gene (p=0.04). We also applied a candidate-gene approach onNOTCH2, ALOX5AP (5-lipoxygenase activating protein) and coagulation genes (Factor II, Factor VLeiden and MTHFR). A significant association was found for the C677T in the MTHFR gene (5,10-methyltetrahydrofolate) and young ischemic strokes (OR=2.39, p=0.02 for TT genotype). In conclusion,this study confirmed the implication of environmental and genetic factors in ischemic strokes before 40years and suggests new genetic risk factors for IS.

O genoma humano varia entre os indivíduos não somente na forma de sequência, mas também estruturalmente. Originalmente, organismos diploides possuem duas cópias de cada região autossômica, uma por cromossomo. Entretanto, com o avanço das técnicas moleculares de identificação do DNA, foram descritas sequências que se repetem em diferentes regiões do genoma em número maior ou menor do que as duas cópias esperadas. Essas variantes são denominadas copy number variants (CNVs) e definidas como segmentos genômicos, geralmente maiores do que 1 kilobase (kb), que variam em número de cópias em comparação com o genoma de referência. As CNVs podem contribuir para a variabilidade do risco entre os indivíduos na etiologia de doenças complexas. Nesse contexto, o Lúpus Eritematoso Sistêmico (LES) é uma doença autoimune com forte componente genético, caracterizada por inflamação crônica e produção de autoanticorpos. Estudos de associação genômica em larga escala (GWAS) identificaram vários loci associados ao LES que contribuem para a susceptibilidade à patogênese. Entretanto, as pesquisas atuais com CNVs e LES focalizam apenas a análise individual de algumas variantes. O objetivo do presente trabalho foi conduzir o primeiro estudo de CNVs em larga escala em pacientes com LES. A detecção de CNVs foi feita por ensaio de Hibridação Genômica em arrays, utilizando a plataforma Affymetrix GeneChip® CytoScan HD em amostra de 23 pacientes com LES. Foram identificadas 406 CNVs distribuídas em todos os cromossomos, exceto no Y. A média foi de 18 CNVs por paciente. As deleções foram mais frequentes do que as duplicações, 311 e 95, respectivamente. O perfil de CNVs revelou 269 CNVs envolvendo genes, 152 CNVs únicas e 59 regiões de CNVs (CNVRs). Nove CNVs identificadas não haviam sido descritas em bancos de dados de variantes estruturais. Adicionalmente, encontramos CNVs em cinco genes previamente associados com LES: CFHR4, CFHR5, STAT4, MECP2 e HLA-DPB2. CNVs nestes genes foram reportadas em pacientes com LES pela primeira vez. O conhecimento das CNVs associadas com LES e autoimunidade podem contribuir para o entendimento da etiologia da doença. Em conclusão, o presente estudo foi o primeiro delineamento em larga escala de CNVs do genoma completo em pacientes com LES. / The human genome varies between individuals not only at the sequence level but also structurally. Originally, diploid organisms have two copies of each autosomal region, one per chromosome. Advances in molecular-based techniques for DNA identification enabled the description of many repeated sequences with higher or lower copy number than that two copies expected. Those sequences are termed copy number variants (CNVs) and are defined as genomic segments, usually greater than 1 kilobase (kb) in size, ranging in copy number when compared to reference genome. CNVs can contribute to risk variability among individuals in complex diseases etiology. In this context, Systemic Lupus Erythematosus (SLE) is an autoimmune disease with strong genetic component and is characterized by chronic inflammation and autoantibodies production. To date, genome-wide association studies (GWAS) have identified several loci associated with SLE that contribute to pathogenesis susceptibility. However, current CNVs studies associated with SLE focus only in few variants analysis. The aim of the present study was to conduct the first genome-wide CNVs study in SLE patients. CNVs detection was performed by high-resolution array Genomic Hybridization Assay, using the Affymetrix GeneChip® CytoScan HD platform, in 23 SLE patients samples. We identified 406 CNVs distributed in all chromosomes, except Y. The average was 18 CNVs per patient. Deletions were more frequent than duplications, 311 and 95, respectively. CNV profile showed 269 CNVs overlapped by genes, 152 unique CNVs and 59 CNV regions (CNVRs). Nine CNVs were never described in structural variants databases. We found CNVs in five genes previously associated with SLE: CFHR4, CFHR5, STAT4, MECP2 and HLA-DPB2. CNVs in these genes were reported in SLE patients for the first time. Knowledge of CNVs associated with SLE risk and autoimmunity could also improve our understanding of disease etiology. In conclusion, the present study was the first effort to search for CNVs in whole genome of SLE patients.

O genoma humano varia entre os indivíduos não somente na forma de sequência, mas também estruturalmente. Originalmente, organismos diploides possuem duas cópias de cada região autossômica, uma por cromossomo. Entretanto, com o avanço das técnicas moleculares de identificação do DNA, foram descritas sequências que se repetem em diferentes regiões do genoma em número maior ou menor do que as duas cópias esperadas. Essas variantes são denominadas copy number variants (CNVs) e definidas como segmentos genômicos, geralmente maiores do que 1 kilobase (kb), que variam em número de cópias em comparação com o genoma de referência. As CNVs podem contribuir para a variabilidade do risco entre os indivíduos na etiologia de doenças complexas. Nesse contexto, o Lúpus Eritematoso Sistêmico (LES) é uma doença autoimune com forte componente genético, caracterizada por inflamação crônica e produção de autoanticorpos. Estudos de associação genômica em larga escala (GWAS) identificaram vários loci associados ao LES que contribuem para a susceptibilidade à patogênese. Entretanto, as pesquisas atuais com CNVs e LES focalizam apenas a análise individual de algumas variantes. O objetivo do presente trabalho foi conduzir o primeiro estudo de CNVs em larga escala em pacientes com LES. A detecção de CNVs foi feita por ensaio de Hibridação Genômica em arrays, utilizando a plataforma Affymetrix GeneChip® CytoScan HD em amostra de 23 pacientes com LES. Foram identificadas 406 CNVs distribuídas em todos os cromossomos, exceto no Y. A média foi de 18 CNVs por paciente. As deleções foram mais frequentes do que as duplicações, 311 e 95, respectivamente. O perfil de CNVs revelou 269 CNVs envolvendo genes, 152 CNVs únicas e 59 regiões de CNVs (CNVRs). Nove CNVs identificadas não haviam sido descritas em bancos de dados de variantes estruturais. Adicionalmente, encontramos CNVs em cinco genes previamente associados com LES: CFHR4, CFHR5, STAT4, MECP2 e HLA-DPB2. CNVs nestes genes foram reportadas em pacientes com LES pela primeira vez. O conhecimento das CNVs associadas com LES e autoimunidade podem contribuir para o entendimento da etiologia da doença. Em conclusão, o presente estudo foi o primeiro delineamento em larga escala de CNVs do genoma completo em pacientes com LES. / The human genome varies between individuals not only at the sequence level but also structurally. Originally, diploid organisms have two copies of each autosomal region, one per chromosome. Advances in molecular-based techniques for DNA identification enabled the description of many repeated sequences with higher or lower copy number than that two copies expected. Those sequences are termed copy number variants (CNVs) and are defined as genomic segments, usually greater than 1 kilobase (kb) in size, ranging in copy number when compared to reference genome. CNVs can contribute to risk variability among individuals in complex diseases etiology. In this context, Systemic Lupus Erythematosus (SLE) is an autoimmune disease with strong genetic component and is characterized by chronic inflammation and autoantibodies production. To date, genome-wide association studies (GWAS) have identified several loci associated with SLE that contribute to pathogenesis susceptibility. However, current CNVs studies associated with SLE focus only in few variants analysis. The aim of the present study was to conduct the first genome-wide CNVs study in SLE patients. CNVs detection was performed by high-resolution array Genomic Hybridization Assay, using the Affymetrix GeneChip® CytoScan HD platform, in 23 SLE patients samples. We identified 406 CNVs distributed in all chromosomes, except Y. The average was 18 CNVs per patient. Deletions were more frequent than duplications, 311 and 95, respectively. CNV profile showed 269 CNVs overlapped by genes, 152 unique CNVs and 59 CNV regions (CNVRs). Nine CNVs were never described in structural variants databases. We found CNVs in five genes previously associated with SLE: CFHR4, CFHR5, STAT4, MECP2 and HLA-DPB2. CNVs in these genes were reported in SLE patients for the first time. Knowledge of CNVs associated with SLE risk and autoimmunity could also improve our understanding of disease etiology. In conclusion, the present study was the first effort to search for CNVs in whole genome of SLE patients.

In this thesis we have studied genetic elements potentially contributing to the pathophysiology of psychiatric disorders, focusing on different sources of human genome variability, including SNPs and CNVs, which can affect not only coding genes but also RNA regulatory elements, such as miRNAs. First, we have interrogated different candidate genes for psychiatric disorders overlapping with known CNVs, finding 14 different genes variable in copy number in psychiatric disorders but not in control individuals. Then, narrowing the analysis on mood disorders, we explored GSK3β gene considering both SNPs and a partially overlapping CNV. The GSK3β promoter and intron 1 region was found significantly associated with an earlier onset of the major depressive disorder. Finally, we have found evidence possibly pointing to a precise post-transcriptional regulation of circadian rhythms by miRNAs in mood disorder patients. Concretely, a variant in the precursor form of miR-182 could play an important role in fine-tuning its target sites involved in the control of sleep/wake cycles. Overall, we have provided evidence of different types of genome variation on neuronal genes or miRNA regulatory regions that can potentially contribute to the development of psychiatric disorders. / En aquesta tesi hem estudiat elements genètics que podrien contribuir potencialment en la fisiopatologia dels trastorns psiquiàtrics, centrant-nos en diferents fonts de variabilitat genòmica humana, incloent els SNPs i els CNVs, els quals poden afectar no només a gens codificants sinó també a elements reguladors, com els miRNAs. Primer, vam interrogar diferents gens candidats per trastorns psiquiàtrics solapats amb CNVs coneguts, trobant que 14 gens eren variables en el número de còpia en pacients però no en individus controls. Després, restringint l'anàlisi a trastorns afectius, vam explorar el gen GSK3β considerant SNPs així com també un CNV que se solapa parcialment amb el gen. Vam trobar la regió del promotor i de l'intró 1 del gen GSK3β associada de manera significativa amb una inferior edat d'inici del trastorn de depressió major. Finalment, hem trobat evidències que possiblement indiquen una precisa regulació post-transcriptional dels ritmes circadians per miRNAs en pacients amb trastorns afectius. Concretament, una variant en la forma precursora del miR-182 podria jugar un paper important en la fina regulació dels seus gens diana implicats en el control dels cicles de son i vigília. En general, hem aportat evidències de què diferents tipus de variació genòmica en gens neuronals o regions reguladores com els miRNAs podrien contribuir potencialment en el desenvolupament de trastorns psiquiàtrics.