Chez les mammifères, une centaine de gènes sont soumis à une régulation épigénétique où seule la copie maternelle, ou paternelle, est exprimée. Ce phénomène appelé empreinte parentale alimente encore différentes théories liées à la reproduction, notamment celles du conflit parental et de la coadaptation entre mère et enfant. Pw1/Peg3 est un gène d'empreinte paternellement exprimé. Cependant, à l'aide de deux modèles de souris bien distincts, une souris rapporteur (Pw1IRESnLacZ) et une nouvelle souris knockout pour Pw1/Peg3, nous avons détecté des transcrits Pw1/Peg3 maternels dans le cerveau périnatal. Plus précisément, nous avons mis en évidence une expression bi-allélique du gène rapporteur Pw1IRESnLacZ restreinte aux deux futures niches de cellules souches neurales adultes. In vitro, nous avons conclu, via des cultures primaires de cellules souches neurales, que l'expression bi-allélique endogène de Pw1/Peg3 est un évènement ponctuel rare. D'ailleurs lors de la caractérisation de notre modèle de souris Pw1/Peg3 knockout, nous avons observé un retard de croissance uniquement lors de la délétion de l'allèle Pw1/Peg3 paternel. Ce phénotype n'est pas lié à un problème de prise alimentaire chez les nouveaux-nés et contrairement à ce qui a été précédemment décrit, nous n'avons détecté aucun défaut de comportement maternel chez les femelles mutantes pour Pw1/Peg3. La lactation n'est pas non plus impactée par la délétion de Pw1/Peg3. Ces résultats démontrent que Pw1/Peg3 favorise intrinsèquement la croissance postnatale et que, désormais, ce gène d'empreinte ne peut plus être utilisé afin d'illustrer la théorie de coadaptation entre mère et enfant. / In mammals, a hundred of genes are preferentially expressed from one specific parental allele; a phenomenon referred as genomic imprinting. Establishing theories to explain the emergence of such a gene dosage strategy is challenging. Pw1/Peg3 is a paternally expressed gene. Using both a reporter mouse model and a novel constitutive knockout mouse model, we detected Pw1/Peg3 transcription from the maternal allele, which is normally silent, in the perinatal brain. Specifically, we observed that a putative Pw1/Peg3 bi-allelic expression is mainly restricted to the two future adult neural stem cells niches. In vitro experiments on primary neural stem cells allowed us to conclude that imprinting relaxation of the Pw1/Peg3 maternal allele is a rare event. Whether it affects the mouse phenotype is currently under investigation. In parallel, consistent with previously established mutant mouse models we confirmed that paternal Pw1/Peg3 deletion leads to growth retardation. However we did not find any impairment in maternal behaviors upon heterozygous or homozygous loss of Pw1/Peg3. Lactation was also not disrupted and mutant pups exhibited a normal suckling ability. Taken together, PW1/PEG3 promotes growth intrinsically and can no longer be used to illustrate the popular coadaptation theory between mother and infant.

No description available.

L'empreinte génomique parentale est un mécanisme de régulation épigénétique conduisant à la répression d'un allèle d'un gène en fonction de son origine parentale. Ce mécanisme affecte un nombre restreint de gènes chez les mammifères métathériens et euthériens. Ces gènes, dits gènes soumis à empreinte (GSE), ont des fonctions moléculaires variées et sans lien apparent. Cependant, deux thèmes reviennent de manière récurrente dans leurs fonctions: le contrôle de la croissance embryonnaire et la tumorigenèse. Ma thèse a consisté à démontrer l'existence d'un lien fonctionnel entre les GSE. Nous montrons que les GSE s'inscrivent dans un même réseau de co-expression transcriptionnelle et qu'ils sont co-régulés dans différentes situations biologiques lors des transitions entre les différents états cellulaires. En effet, une induction coordonnée de la plupart des GSE a lieu lors des sorties du cycle cellulaire, réversibles (quiescence) ou non (différenciation). Les perturbations individuelles de l'expression de plusieurs GSE dans le modèle des pré-adipocytes 3T3-L1 confirment un rôle du réseau des GSE dans le contrôle des transitions entre prolifération, quiescence et différenciation. De plus, l'analyse des gènes bi-alléliques inclus dans le même réseau de co-régulation que les GSE montre un enrichissement en gènes de la matrice extracellulaire. La fonction associée à ce réseau serait donc le contrôle des transitions entre les différents états cellulaires, via le remodelage de la matrice extracellulaire. Pour conclure, outre l'identification d'une fonction commune aux GSE, nos résultats suggèrent un scénario pour le ciblage de ces gènes par l'empreinte génomique parentale au cours de l'évolution des mammifères. / Genomic imprinting is an epigenetic mechanism leading to the repression of one allele of a gene, depending on its parental origin. This mechanism affects a small number of genes in metatherian and eutherian mammals. These genes, named imprinted genes (IGs), display various molecular functions and thus seem unrelated. However, their alterations are frequently associated with the control of embryonic growth and tumorigenesis. My PhD project has consisted in demonstrating a functional link between IGs. We show that IGs are frequently co-expressed and belong to a common gene network. They are co-regulated in biological situations corresponding to the transitions between different cellular states. Coordinated induction of most IGs takes place at the outputs of the cell cycle. Loss and gain of function experiments of several IGs in the 3T3-L1 pre-adipocyte model demonstrate a role of the IG network in controlling transitions between cellular states (proliferation, quiescence and differentiation). In addition to IGs, this network also includes bi-allelic genes, with many extracellular matrix genes. Therefore, the function associated with the IG network could be the fine control of transitions between cellular states through a remodeling of the extracellular matrix.To conclude, in addition to the identification of a common cellular function for IGs, our results suggest a possible scenario for the targeting of these genes by parental genomic imprinting during mammalian evolution.

O imprinting genômico é um processo regulado epigeneticamente que faz com que os alelos sejam expressos de acordo com a sua origem parental. No cromossomo 11 (11p15.5), existem duas regiões controladoras de imprinting (ICR1 e ICR2), que controlam a expressão de genes marcados (imprinted). Os padrões de metilação dessas regiões podem ser alterados pela dissomia uniparental (DUP), que ocorre quando parte de ou um cromossomo inteiro do mesmo par de homólogos é herdado de somente um genitor. Erros mitóticos podem gerar mosaicismo com uma linhagem de células com DUP e a outra biparental. As síndromes de Silver-Russell (SSR) e Beckwith-Wiedemann (SBW) são doenças de alterações do imprinting genômico, envolvendo os cromossomos 7 (SSR) e 11 (SSR e SBW). A Hemihiperplasia Isolada (HHI) parece corresponder a uma forma mais leve da SBW.. No presente trabalho, foi realizada uma varredura in silico para busca de novos microssatélites nos cromossomos 7 e 11, e selecionados seis do tipo tetra ou pentanucleotídeos, no cromossomo 7, e 12, no cromossomo 11. O perfil de metilação nas ICRs foi verificado por três técnicas distintas: MS-MLPA, DESM-RT e por uma nova estratégia desenvolvida neste trabalho denominada DESM-QFPCR. Foram avaliados 32 pacientes com SBW, 16 HHI, 20 com SSR e seus pais, quando disponíveis, além de um paciente com fenótipo aparentemente normal com cariótipo 46,XX/46,XY e cuja placenta apresentou displasia mesenquimal placentária (DMP) a qual está associada à SBW. Os novos marcadores apresentaram alta taxa de heterozigose (média de 70%), e ausência das características indesejáveis dos dinucleotídeos predominantemente utilizados para detecção de DUP. Seis marcadores estão entre genes controlados pelas ICRs 1 e 2. A DUP paterna do cromossomo 11 (DUPpat Cr11), sempre restrita a 11p15.5, foi responsável por 13% dos casos de HHI e 19% dos de SBW. As alterações estruturais foram confirmadas por minissequenciamento quantitativo de SNPs e por MS-MLPA. Um paciente apresentou duplicação paterna abrangendo ambas as ICRs. Uma deleção não descrita anteriormente no gene CDKN1C foi observada em uma paciente e sua mãe. Para os pacientes com DUPpat Cr11, foram investigados microssatélites em 13 autossomos e nos cromossomos sexuais para detecção de mosaicismo global. Apenas o paciente com DMP apresentou mosaicismo [células androgenéticas (25-30%) e biparentais], sugerindo evento de dupla fertilização. Nos pacientes com SSR, foi observada hipometilação na ICR1 em 25% dos casos. Para a SBW, foi observada hipermetilação na ICR1 e hipometilação na ICR2 em 6% e 42% dos casos, respectivamente. Os casos com DUPpat Cr11 apresentaram alteração de metilação em ambas as ICRs. As frequências de alterações (epi) genéticas encontradas foram semelhantes às previamente descritas na literatura para as SBW, SSR e HHI. Neste trabalho, foi desenvolvida uma nova técnica para estudo de metilação do DNA de ICRs e testados marcadores microssatélites inéditos na região 11p15, que quando comparados com metodologias mais tradicionais de avaliação, como DESM-RT e MS-MLPA, mostraram elevada correlação dos resultados. Os achados mostram a complexidade da etiologia das doenças estudadas no presente trabalho e os dados moleculares serão imprescindíveis para o aconselhamento genético adequado para cada caso em particular e suas famílias. / Genomic imprinting is a epigenetically regulated process where the alleles are expressed in terms of their parental origin. On chromosome 11 (11p15.5) there are two regions controlling imprinting (ICR1 and ICR2), which control imprinted gene expression. The methylation patterns in these regions may be altered by uniparental disomy (UPD), which occurs when part or whole chromose is inherited from only one parent. Mitotic errors can lead to mosaicism with a cell line with DUP and other, biparental. The Silver-Russell syndrome (SRS) and Beckwith-Wiedemann syndrome (BWS) are diseases of abnormal genomic imprinting, involving chromosomes 7 (SSR) and 11 (SRS and BWS). The Isolated Hemihiperplasia (IHH) seems to correspond to a milder form of the SBW. In the present study, we performed an in silico scan to search for new microsatellites on chromosomes 7 and 11, and selected six tetra- and/or pentanucleotides on chromosome 7, and 12 on chromosome 11. The pattern of methylation in ICRs was verified by three different techniques: MS-MLPA, DESM-RT and a new strategy developed in this work called DESM-QFPCR. We evaluated 32 patients with BWS, HHI 16, with 20 SSR and their parents, when available, and one patient with apparently normal phenotype with karyotype 46, XX/46, XY and whose placenta showed placental mesenchymal dysplasia (PMD) which is associated with SBW. The new markers showed a high heterozygosity rate (average 70%), and absence of undesirable characteristics of dinucleotides, predominantly used for detection of DUP. Six markers spans genes controlled by the ICRs 1 and 2. The paternal UPD for chromosome 11 (UPDpat Cr11), all restricted to 11p15.5, was responsible for 13% of cases of HHI and 19% of the SBW. Structural changes were confirmed by quantitative SNaPshot sequencing of SNPs and MS-MLPA. One patient had paternal duplication encompassing both ICRs. A not previously described deletion in the gene CDKN1C was observed in one patient and her mother. For patients with DUPpat Cr11, microsatellites were investigated in 13 autosomes and sex chromosomes to detect wide mosaicism. Only patients with DMP showed mosaicism [androgenetic cells (25-30%) and biparental], suggesting double fertilization. In patients with SRS, ICR1 hypomethylation was observed in 25% of cases. For BWS, ICR1 hypermethylation and in ICR2 hypomethylation were observed 6% and 42% of cases, respectively. All cases with UPDpat Cr11 presented abnormal methylation in both ICRs. The (epi) genetic change frequencies were similar to those previously described in the literature for BWS, SRR andIHH. In the present work, we developed a new technique to study DNA methylation of ICRs and tested novel microsatellite markers in the 11p15 region, which showed high correlation of results, when compared with more traditional methods such as RT-DESM and MS-MLPA. The results show the complex etiology of these diseases and the molecular data are essential for appropriate patient and families genetic counseling.

A maioria dos casos de puberdade precoce central (PPC) em meninas permanece idiopática. A hipótese de uma causa genética vem se fortalecendo após a descoberta de alguns genes associados a este fenótipo, sobretudo aqueles implicados com o sistema kisspeptina (KISS1 e KISS1R). Entretanto, apenas casos isolados de PPC foram relacionados à mutação na kisspeptina ou em seu receptor. Até recentemente, a maioria dos estudos genéticos em PPC buscava genes candidatos selecionados com base em modelos animais, análise genética de pacientes com hipogonadismo hipogonadotrófico, ou ainda, nos estudos de associação ampla do genoma. Neste trabalho, foi utilizado o sequenciamento exômico global, uma metodologia mais moderna de sequenciamento, para identificar variantes associadas ao fenótipo de PPC. Trinta e seis indivíduos com a forma de PPC familial (19 famílias) e 213 casos aparentemente esporádicos foram inicialmente selecionados. A forma familial foi definida pela presença de mais de um membro afetado na família. DNA genômico foi extraído dos leucócitos do sangue periférico de todos os pacientes. O estudo de sequenciamento exômico global realizado pela técnica ILLUMINA, em 40 membros de 15 famílias com PPC, identificou mutações inativadoras em um único gene, MKRN3, em cinco dessas famílias. Pesquisa de mutação no MKRN3 realizada por sequenciamento direto em duas famílias adicionais (quatro pacientes) identificou duas novas variantes nesse gene. O MKRN3 é um gene de um único éxon, localizado no cromossomo 15 em uma região crítica para a síndrome de Prader Willi. O gene MKRN3 sofre imprinting materno, sendo expresso apenas pelo alelo paterno. A descoberta de mutações em pacientes com PPC familial despertou o interesse para a pesquisa de mutações nesse gene em 213 pacientes com PPC aparentemente esporádica por meio de reação em cadeia de polimerase seguida de purificação enzimática e sequenciamento automático direto (Sanger). Três novas mutações e duas já anteriormente identificadas, incluindo quatro frameshifts e uma variante missense, foram encontradas, em heterozigose, em seis meninas não relacionadas. Todas as novas variantes identificadas estavam ausentes nos bancos de dados (1000 Genomes e Exome Variant Server). O estudo de segregação familial em três dessas meninas com PPC aparentemente esporádica e mutação no MKRN3 confirmou o padrão de herança autossômica dominante com penetrância completa e transmissão exclusiva pelo alelo paterno, demonstrando que esses casos eram, na verdade, também familiares. A maioria das mutações encontradas no MKRN3 era do tipo frameshift ou nonsense, levando a stop códons prematuros e proteínas truncadas e, portanto, confirmando a associação com o fenótipo. As duas mutações missenses (p.Arg365Ser e p.Phe417Ile) identificadas estavam localizadas em regiões de dedo ou anel de zinco, importantes para a função da proteína. Além disso, os estudos in silico dessas duas variantes demonstraram patogenicidade. Todos os pacientes com mutação no MKRN3 apresentavam características clínicas e hormonais típicas de ativação prematura do eixo reprodutivo. A mediana de idade de início da puberdade foi de 6 anos nas meninas (variando de 3 a 6,5) e 8 anos nos meninos (variando de 5,9 a 8,5). Tendo em vista o fenômeno de imprinting, análise de metilação foi também realizada em um subgrupo de 52 pacientes com PPC pela técnica de MS-MLPA, mas não foram encontradas alterações no padrão de metilação. Em conclusão, este trabalho identificou um novo gene associado ao fenótipo de PPC. Atualmente, mutações inativadoras no MKRN3 representam a causa genética mais comum de PPC familial (33%). O MKRN3 é o primeiro gene imprintado associado a distúrbios puberais em humanos. O mecanismo preciso de ação desse gene na regulação da secreção de GnRH necessita de estudos adicionais / Most cases of central precocious puberty (CPP) in girls remain idiopathic. The hypothesis of a genetic cause has been strengthened after the discovery of some genes associated with this phenotype, particularly those involved with the kisspeptin system (KISS1 and KISS1R). However, genetic defects in KISS1 and its receptor are rare and have been identified in only a few patients with CPP.over the past years. To date, most genetic studies in CPP was based mainly on a candidate gene approach, including genes selected in animal studies, human models of patients with hypogonadotropic hypogonadism or in genome wide association studies. In the present study, we used whole exome sequencing, a more advanced method of sequencing, to identify variants associated with CPP. Thirty-six patients with the familial form of CPP (19 families) and 213 apparently sporadic cases were initially selected. The familial form was defined by the presence of more than one member affected in the family. Genomic DNA was extracted from peripheral blood leukocytes in all patients. Whole exome sequencing performed by ILLUMINA technique in 40 members of 15 families with CPP, identified inactivating mutations in a single gene, MKRN3, in five out of these families. Analysis of MKRN3 mutations performed by automatic sequencing in two additional families (four patients) identified two novel mutations. MKRN3 is an introless gene located on chromosome 15, in the Prader Willi syndrome critical region, and it is expressed only by the paternal allele due to the maternal imprinting. Following the initial findings, we searched for MKRN3 mutations in 213 patients with apparently sporadic CPP using polymerase chain reaction followed by direct enzymatic purification and automated sequencing (Sanger). Three new mutations and two previously reported, including four frameshifts and one missense variant was identified in six unrelated girls with CPP. All variants were not described in the two databases (1000 Genomes and Exome Variant Server). The familial segregation analysis performed in three out of these girls with apparently sporadic CPP and MKRN3 mutations confirmed the autosomal dominant inheritance with complete penetrance and exclusive transmission through the paternal allele, revealing familial inheritance in apparently sporadic cases. Most of these MKRN3 mutations were frameshifts or nonsense, leading to premature stop codons and truncated proteins, thus demonstrating positive genotype- phenotype correlation. The two missense mutations (p.Arg365Ser and p.Phe417Ile) identified were located within zinc finger motifs, regions predicted to be essential for the protein function. Besides that, all missense mutations were predicted to be pathogenic by in silico analysis. All patients carrying MKRN3 mutations exhibited typical clinical and hormonal features of premature activation of the reproductive axis. The median age of puberty onset was 6.0 years in girls (ranging from 3.0 to 6.5) and 8.0 years in boys (ranging from 5.9 to 8.5). In view of the imprinting phenomenon, methylation analysis was also performed in a subgroup of 52 patients with CPP by MSMLPA technique, but no methylation abnormalities were detected. In conclusion, our work has identified a new gene associated with CPP. Currently, inactivating mutations in MKRN3 represent the most common genetic cause of familial CPP (33%). MKRN3 is the first imprinted gene associated with pubertal disorders in humans. However, its precise mechanism of action in the regulation of GnRH secretion needs further studies

A idade normal para início da puberdade em meninas varia bastante, de 8 a 13 anos, e os genes envolvidos nesse controle são parcialmente conhecidos. Fatores ambientais, como alimentação e exposição a disruptores endócrinos, contribuem para essa variabilidade, de modo que genes modulados epigeneticamente podem justificar parte da complexidade desse processo. O termo epigenética se refere às modificações na expressão gênica que não são causadas por alterações na sequência do DNA. A metilação do DNA é o mecanismo epigenético mais bem estudado. Na última década surgiram evidências demonstrando a relação entre metilação do DNA e desenvolvimento puberal. Em fêmeas de roedores, a hipermetilação do DNA levou à puberdade precoce. Em humanos, a puberdade precoce central (PPC) familial causada por mutações nos genes MKRN3 e DLK1 é considerada um defeito do imprinting, fenômeno epigenético no qual apenas um dos alelos parentais é expresso, estando o outro metilado e inativo. Além disso, um conceito atual propõe que o início da puberdade requer a repressão epigenética de fatores inibidores do eixo gonadotrófico. Recentemente, genes zinc finger (ZNF) foram relacionados ao processo puberal, e muitos deles codificam repressores transcricionais. Neste trabalho, estudamos a metilação do DNA do sangue periférico de 10 pacientes do sexo feminino com PPC familial (casos índices) e 33 meninas com desenvolvimento puberal normal (15 pré-púberes e 18 púberes), usando a plataforma Human Methylation 450 BeadChip. Duas pacientes tinham PPC de causa genética (uma com mutação no MKRN3 e outra com deleção no DLK1) e oito tinham PPC idiopática, sem mutações identificadas pelo sequenciamento exômico global. Cento e vinte regiões diferencialmente metiladas foram identificadas entre as meninas saudáveis pré-púberes e púberes, estando 74% delas no cromossomo X. Apenas uma região mostrou-se hipometilada no grupo púbere e, de maneira importante, contém a região promotora do ZFP57, fator necessário para manutenção do imprinting. Uma vez que a hipermetilação nas regiões promotoras dos genes é relacionada à inibição transcricional, o achado de hipermetilação global do DNA na puberdade sugere que haja inibição de fatores inibidores do eixo gonadotrófico, o que resultaria no início do processo puberal. O receptor estrogênico destacou-se como um fator transcricional que se liga a sete genes diferencialmente metilados entre os controles pré-púberes e púberes. As pacientes com PPC apresentaram mais sítios CpG hipermetilados tanto na comparação com as meninas pré-púberes (81%) quanto púberes (89%). Há doze genes ZNF contendo sítios CpG hipermetilados na PPC. Não foram encontradas anormalidades de metilação nos genes MKRN3 e DLK1 nem em suas regiões regulatórias. Em conclusão, este estudo evidenciou hipermetilação global do DNA em meninas com puberdade normal e precoce, sugerindo que esse padrão é uma marca epigenética da puberdade. Pela primeira vez, mudanças no metiloma de pacientes com PPC foram descritas. Modificações na metilação de vários genes ZNF parecem compor a complexa rede de mecanismos que leva ao início da puberdade humana / Normal puberty initiation varies greatly among girls, from 8 to 13 years, and the genetic basis for its control is partially known. Environmental factors, such as nutrition and exposure to endocrine disruptors, contribute to this variance, and epigenetically modulated genes may justify some of the complexity observed in this process. Epigenetics refers to alterations in gene expression that are not caused by changes in DNA sequence itself. DNA methylation is the best studied epigenetic mechanism. In the last decade, evidence has emerged showing the relationship between DNA methylation and pubertal development. In female mice, DNA hypermethylation led to precocious puberty. In humans, familial central precocious puberty (CPP) caused by mutations in the MKRN3 and DLK1 genes is considered a disorder of imprinting, an epigenetic phenomenon in which only one parental allele is expressed, and the other allele is methylated and inactive. In addition, animal studies indicated that pubertal timing requires epigenetic repression of inhibitory factors of the gonadotrophic axis. Recently, zinc finger genes (ZNF) were related to pubertal development, many of which encode transcriptional repressors. In the present study, we analyzed the DNA methylation of peripheral blood samples from 10 female patients with familial CPP (index cases) and 33 girls with normal pubertal development (15 pre-pubertal and 18 pubertal), using the Human Methylation 450 BeadChip assay. Genetic CPP was diagnosed in two patients (one with a MKRN3 mutation and the other with a DLK1 deletion). The remaining eight cases with idiopathic CPP were previously evaluated by whole exome sequencing and no causative mutations were identified so far. We evidenced 120 differentially methylated regions between pre-pubertal and pubertal healthy girls, and 74% of them were located at the X chromosome. Only one genomic region was hypomethylated in the pubertal group. Of note, it contains the promoter region of ZFP57, an important factor for imprinting maintenance. As DNA hypermethylation in gene promoters is related to gene silencing, the finding of global DNA hypermethylation in puberty suggests inhibition of inhibitory factors of the hypothalamic-pituitary-gonadal axis that results in puberty onset. Importantly, the estrogen receptor was identified as a transcriptional factor that binds to seven differentially methylated genes associated with pubertal process. Patients with CPP exhibited more hypermethylated CpG sites compared to both pre-pubertal (81%) and pubertal (89%) controls. Twelve ZNF genes were recognized as having hypermethylated CpG sites in CPP. The methylation analyses of MKRN3 and DLK1 genes showed no abnormalities. In conclusion, this study revealed a widespread DNA hypermethylation in girls with normal and precocious puberty, suggesting that this pattern can be an epigenetic signature of puberty. For the first time, changes in the methylome of patients with CPP were described. We highlight that alterations in methylation levels of several ZNF genes may impact the onset of human puberty

O imprinting genômico é um processo epigenético essencial para o desenvolvimento normal de mamíferos com placenta e refere-se à expressão gênica alelo-específica, de acordo com a origem parental. A expressão dos genes marcados por imprinting é controlada por regiões diferencialmente metiladas (DMRs), situadas em regiões controladoras de imprinting (ICRs). O cromossomo 29 de Bos taurus possui dois domínios cromossômicos semelhantes à região 11p15.5 de humanos, que são denominados KvDMR1 (na ICR2) e H19DMR (na ICR1). Essas ICRs controlam um cluster de genes importantes para o crescimento e desenvolvimento, sendo a KvDMR1 metilada no alelo materno e a e H19DMR metilada no alelo paterno. No presente trabalho, foi verificado o padrão de metilação da KvDMR1 e da H19DMR em oócitos não maturados (Vg) e maturados in vitro (MII) e nos blastocistos inicial (Bi) e expandido (Bx) bovinos e em placentas bovinas e humanas de primeiro trimestre. Foram coletados oócitos e embriões pré-implantação no estágio de blastocisto produzidos pela técnica de Fertilização in vitro. Também foram coletados o tecido placentário e de um feto bovino de 49 dias e de uma placenta humana, com idade gestacional de 12 semanas. O DNA genômico foi extraído e modificado com bissulfito de sódio. O padrão de metilação das regiões KvDMR1 e H19DMR foi verificado por meio de clonagem e seqüenciamento do DNA modificado com bissulfito de sódio. Para as análises de expressão gênica nos oócitos e blastocistos, foi realizada a extração do RNA e em seguida o cDNA foi produzido para a quantificação relativa da expressão gênica por meio da técnica de PCR em tempo real. Os resultados de metilação para a amostra controle de espermatozóide apresentaram um perfil hipometilado para a KvDMR1 e hipermetilado para a H19DMR. Os oócitos Vg e MII mostraram um perfil hipermetilado para a KvDMR1 e nos Bi e Bx foi observado um perfil hipermetilado e hipometilado, respectivamente. Para a H19DMR, foi observado um perfil hipermetilado para as amostras Vg e MII, sendo que para os Bi foi observado um perfil hipometilado e, para os Bx, um perfil monoalélico de expressão. A expressão dos genes LIT1 e IGF2 foi relativamente baixa nas amostras analisadas, sendo que o gene LIT1 foi expresso nos mesmos níveis para todas as amostras e o IGF2 não foi expresso nos Bi e Bx. Os oócitos MII apresentaram altos níveis de expressão do IGF2 quando comparados com os oócitos Vg. Nas placentas precoces de bovinos, a porcentagem de metilação para a KvDMR1 variou entre os cotilédones de 39,6%. e 88,9%. A porcentagem de metilação para a H19DMR nos cotilédones variou entre 35,0% e 57,0% , sendo que apenas uma amostra apresentava-se completamente demetilada para esta ICR. Nas análises das vilosidades humanas, foi observado um perfil hipermetilado em todas as amostras analisadas, em que as porcentagens de metilação para a KvDMR1 e H19DMR variaram entre 84,4% e 97,9%. Os resultados mostram um perfil alterado de metilação nos oóctios MII para as duas regiões analisadas, e uma alteração nas amostras de oócitos Vg para a H19DMR. Para os blastocistos, o esperado seria um perfil monoalélico para as duas regiões, no entanto, esse resultado só foi encontrado para os Bx na H19DMR. Em bovinos, as DMRs apresentaram um funcionamento antagônico, enquanto a KvDMR1 tende a uma hipermetilação a H19DMR tende a uma hipometilação. O resultado das análises comparativas das placentas bovina e humana não sugerem uma relação do padrão de metilação dessas regiões entre essas duas espécies, no entanto, servem de base para o conhecimento do imprinting na placenta. Os estudos nos oócitos e blastocistos realizados representam um passo inicial na investigação da influencia das tecnologias de reprodução assistida no desenvolvimento embrionário, sendo o primeiro relato do funcionamento dessas DMRs nas amostras de oócitos e embriões pré-implantação bovinos. / Genomic imprinting is an epigenetic process that plays an essential role in the development of placental mammals with a parent-of-origin-specific manner of gene expression, in which only one allele is expressed. The imprinted gene expression is controlled by differentially methylated regions (DMRs), located in imprinting control regions (ICRs). In Bos Taurus, chromosome 29 presents two imprinted domains similar to human 11p15.5 region which are named KDMR1 (in the ICR2) and H19DMR (in the ICR1). Several genes that play an essential role in growth and development are under the control of the ICRs, in which the KvDMR1 is methylated on the maternal allele and the H19DMR is methylated on the paternal allele. In this study, the DNA methylation status of the KvDMR1 and H19DMR was verified in bovine non-matured germinative vesicle (GV) in vitro matured (MII) oocytes, as well in early (EA) and expanded (EX) blastocysts, and in bovine and human early placenta. The oocytes and blastocysts were collected after in vitro fertilization (IVF) techniques. Tissues from bovine placenta and fetus with 49 days of gestational age and human placenta with 12 weeks of gestational age were also collected. The DNA was extracted and modified by sodium bisulfite. The methylation pattern of KvDMR1 e H19DMR was verified by cloning and bisulfite sequencing. RNA extraction and cDNA synthesis for the relative quantification of gene expression by real time PCR were performed for oocytes and blastocysts. The methylation profile for the control sample of sperm was hypomethylated for KvDMR1 and hypermethylated for H19DMR. The GV and MII oocytes showed a hypermethylated pattern for KvDMR1 and in the EA and EX was hypermethylated and hypomethylated, respectively. The H19DMR displayed a hypermethylated pattern for GV and MII oocytes. For EA was observed a hypomethylated profile and EX presented a monoallelic expression. The LIT1 and IGF2 gene expression were low for all samples, however the LIT1 had the same level of expression in all samples while the IGF2 was not expressed in EA and EX. The MII oocytes showed high levels of IGF2 gene expression when compared with GV oocytes. The methylation levels for KvDMR1 in bovine early placenta varied between the cotyledons (39,6% and 88,9%). The percentage of methylation for H19DMR in cotyledons varied between 57.0% to 35.0%. Only one sample was not methylated for this ICR1. In human villous, a hypermethylated profile was observed for all samples, and the percentage of methylation for KvDMR1 and H19DMR varied between 84.4% e 97.9%. The results show an altered methylation profile in MII oocytes for two analysed regions and an alteration of H19DMR in GV oocytes. The expected for blastocysts was a monoallelic profile for KvDMR1 and H19DMR, however these results were observed only in EX for H19DMR. In bovine, the methylation levels of KvDMR1 and H19DMR seems to work antagonistically. While the KvDMR1 tended to hypermethylation, the ICR1 tended to a hypomethylation. The comparative analysis of bovine and human early placentas does not suggest a relationship between the methylation patterns of these regions in these two species. However, these studies provide the basis for the understanding of imprinting in the placenta. The study in oocytes and blastocysts represent an initial investigation in understanding how the assisted reproductive technologies affect the embryo growth and development, and this is the first report on the methylation patterns of KvDMR1 and H19DMR in bovine oocytes and blastocysts.

Síndrome de Angelman (AS - MIM 105830) é causada pela ausência de função do gene UBE3A que codifica uma proteína ubiquitina - ligase (E6-AP). Esse gene é expresso bialelicamente em vários tecidos exceto no cérebro, onde a expressão é preferencialmente materna. O RNA anti-senso de UBE3A é considerado o regulador dessa expressão diferencial entre os alelos, e faz parte de um transcrito grande que só o alelo paterno é expresso devido ao imprinting genômico; no cérebro, esse transcrito se entende até a região anti-senso de UBE3A, mas nos demais tecidos o transcrito é menor e não engloba a região anti-senso. Este trabalho visa obter um modelo para estudo da AS. Células-tronco da polpa do dente (SHEDs) de pacientes com deleção do segmento 15q11-q13 ou mutação no gene UBE3A foram caracterizadas e submetidas à diferenciação neuronal. A diferenciação foi analisada através do estudo de RNA e proteínas para marcadores neuronais e, também, por testes funcionais. As SHEDs são células-tronco mesenquimais e constituem uma população heterogênea. Essas células ou algumas dessas células já expressam algumas proteínas neuronais ou de células excitáveis como nestina, β-tubulina III, MAP2 e proteína de canais dependentes de voltagem de sódio e potássio. Um ponto interessante é que as SHEDs apresentam baixa expressão do UBE3A anti-senso e a expressão do UBE3A nas células de pacientes é menor que 50% da expressão encontrada nas células de controles, que pode indicar a ocorrência de expressão preferencial materna desse gene em outros tipos celulares além de neurônios maduros. Quando induzidas à diferenciação neurogênica, a maioria das linhagens controles apresentou aumento da expressão de MAP2 e, principalmente, β-tubulina III; e a maioria das linhagens de pacientes com AS não apresentou aumento notável na expressão dessas proteínas, exceto uma linhagem de paciente que aumentou a expressão de β-tubulina III. As células induzidas à diferenciação apresentaram aumento estatisticamente significativo da condutância de sódio através de canais de sódio dependentes de voltagem. Com a análise de expressão de UBE3A e do UBE3A anti-senso é possível afirmar que a expressão deles não alterou com a diferenciação neuronal. Assim, é possível concluir que as células-tronco da polpa do dente, com o protocolo de diferenciação neurogênica, progrediram na via de diferenciação, mas a maioria das células não atingiu o estágio de maturação necessário para que ocorresse o imprinting do UBE3A ou a via de diferenciação não ia em direção a neurônios que apresentam imprinting do UBE3A. / Angelman syndrome (AS - MIN 105830) is caused by the loss of function of the maternal UBE3A gene, which encodes an ubiquitin protein ligase (E6-AP). UBE3A displays biallelic expression in most of tissues, but maternal predominant expression is observed in the brain. A RNA antisense that is paternally expressed in some regions in the brain is considered to be responsible for this tissue-specific imprinting; UBE3A antisense is part of a large transcript that starts at SNURF-SNRPN gene and is paternally expressed, and in the brain this transcript includes UBE3A antisense region however in other tissues this region is not included. The aim of the present study is to develop a new model for studying AS. Dental pulp stem cells (SHEDs) were characterized and differentiated by an already described protocol. SHEDs intrinsically express some neuronal proteins as nestin, β-tubulin III, MAP2 and voltage-gated sodium channels and potassium channels. Interestingly, SHEDs also present a low expression of UBE3A antisense, and UBE3A expression in cells from patients with AS is lower than 50% of the cells from normal control, so it is possible that preferential maternal expression of this gene might occur in some cells beyond mature neurons. After the neuronal differentiation, most control lineages and one lineage of AS patients had an increase of MAP2 and β-tubulin III expression. Two control lineages and most lineages from AS patients did not have a notable increase of expression of these proteins. Neuronal differentiated cells displayed an increase in conductance through voltage-gated sodium channels. Analysis of UBE3A and UBE3A antisense expression in SHEDs and cells induced to differentiate into neurons indicated no changes in their expression. Thus, after neuronal differentiation induction, dental pulp stem cells progressed through neuronal differentiation pathway. However, most cells did not reach the stage which UBE3A imprinting occurs or the neuronal differentiation is resulting in a cell that do not present UBE3A imprinting.

Upon fertilization the paternal and maternal genomes unite, giving rise to the embryo, with its unique genetic code. All cells in the human body are derived from the fertilized ovum: hence they all contain (with a few exceptions) the same genetic composition. However, by selective processes, genes are turned on and off in an adaptable, and cell type-specific, manner. The aim of this thesis is to investigate how signals coming from outside the cell and epigenetic factors residing in the cell nucleus, cooperate to control gene expression. The transforming growth factor-β (TGF-β) superfamily consists of around 30 cytokines, which are essential for accurate gene regulation during embryonic development and adult life. Among these are the ligands TGF-β1 and bone morphogenetic (BMP) -7, which interact with diverse plasma membrane receptors, but signal via partly the same Smad proteins. Smad4 is essential to achieve TGF-β-dependent responses. We observed that by regulating transcription factors such as Id2 and Id3 in a specific manner, TGF-β1 and BMP-7 achieve distinct physiological responses. Moreover, we demonstrate that CTCF, an insulator protein regulating higher order chromatin conformation, is able to direct transcription by recruiting RNA polymerase II to its target sites. This is the first mechanistic explanation of how an insulator protein can direct transcription, and reveals a link between epigenetic modifications and classical regulators of transcription. We also detected that DNA loci occupied by CTCF replicate late. The timing of replication is a crucial determinant of gene activity. Genes replicating early tend to be active, whereas genes replicating late often are silenced. Thus, CTCF can regulate transcription at several levels. Finally, we detected a substantial cross-talk between CTCF and TGF-β signaling. This is the first time that a direct interplay between a signal transduction pathway and the chromatin insulator CTCF is demonstrated.

The imprinting status of the SLC22A18AS gene, located in the human chromosome segment 11p15.5, was studied using PCR-SSCP analysis and fetal tissues from a battery of 17 abortuses. This gene showed monoallelic expression (genomic imprinting) in different tissues from two abortuses which were heterozygous for an SNP (c.473G>A) in its coding region. This gene was found to be paternally imprinted (maternally expressed) in five tissues namely lung, liver, brain, kidney and placenta from an abortus. The parental origin of the expressed allele could not be determined in the second abortus as both the mother and the abortus were heterozygous for the SNP. Since paternal blood samples from none of the 17 abortuses could be collected for DNA isolation, the mother's genotype was used to find the origin of the expressed allele. In order to understand the mechanism underlying imprinting of this gene, it was important to understand the nature of the epigenetic marks (imprints) on the two alleles of this gene. Since these epigenetic marks are generally observed in promoters or CpG islands associated with the imprinted genes, the promoters of the SLC22A18AS gene was characterized using transient transfection of putative SLC22A18AS promoter fragments cloned in the pGL3-Basic vector in human cells followed by luciferase reporter assay. Since the promoter of a gene lies upstream to the transcription start site (TSS), TSS of this gene was mapped. In silico approach revealed an EST (CB129046) which had an additional 39 bases upstream to the known mRNA sequence. TSS was then identified by the 5’ primer extension analysis. TSS was found to be 166 bases upstream to the 5’ end of this EST. In order to select cell lines for transient transfection of putative promoter constructs for promoter charaterization, RT-PCR analysis was used to see the expression of this gene in the following available cell lines in the lab: HuH7, HepG2, A549, HeLa, LNCaP and PC3. This gene was found to be maximally expressed in HepG2 cells. Expression of this gene was also observed in A549, HeLa, LNCaP and PC3 cells. HuH7, on the other hand, did not show any detectable expression of this gene. Based on the above data, HepG2 and A549 cells were selected for promoter characterization. Seven putative promoter constructs were transiently transfected in these cells and the promoter activity of different constructs was measured by luciferase assay. The assay identified two promoters for the gene: P1 promoter in a region from -855 to -254 bp and P2 promoter in a region from -1441 to -855. In order to see the presence of putative transcription factor binding sites in the upstream region of the gene, the MatInspector Professional program was used. The gene was found to be devoid of TATA and CCAAT boxes. Most of the putative transcription factor binding sites were present in a region from -855 to -254 bp which spans the P1 promoter, including a binding site for the Sp1 transcription factor. In order to see if Sp1 binds to the promoter of this gene, ChIP assay was performed. Sp1 was shown to bind the region harboring the P1 promoter. In order to see if Sp1 has a role in the regulation of this gene, Sp1 constructs were co-transfected with the SLC22A18AS P1 promoter construct in HepG2 and Sp1-null Drosophila SL2 cells. The results showed that the Sp1 has a positive regulation on the SLC22A18AS promoter activity. As stated earlier, epigenetic marks such as differential methylation of CpG dinucleotides in two alleles are associated with promoters of the genes. Since the promoters for SLC22A18AS were characterized, the presence of allele-specific differentially methylated regions (DMRs) associated with the promoters was investigated. In order to differentiate the two alleles in the promoter regions by SNPs, DNA sequence analysis of the promoter regions was performed in a battery of 17 abortuses to search for SNPs. Abortus no. 3 showed heterozygosity for a C to A change at nucleotide position -445 in the P1 promoter region, while abortus no. 2 showed heterozygosities for G to A and A to G changes at nucleotide positions -919 and -1321 respectively in the P2 promoter region. The alleles in the abortus no. 3 were designated as allele C and allele A. The alleles in abortus no. 2 were designated as allele GG and allele AA. Once the two alleles were differentiated by these SNPs, identification of DMRs was performed using sodium bisulfite genomic DNA sequencing. Genomic DNA from the abortus no. 3 was taken for the identification of DMR in the P1 promoter region, while genomic DNA from abortus no. 2 was taken for the identification of DMR in the P2 promoter region. Sodium bisulfite genomic DNA sequencing of the P1 promoter region showed heavy methylation of both the alleles. No DMR was observed in this region. Sodium bisulfite genomic DNA sequencing of the P2 promoter region using DNA from abortus no. 2 did not show any differential methylation of the two alleles. However, like the P1 promoter region, the P2 promoter region was also heavily methylated. In order to see the methylation status of both the promoter regions in human sperms, sperm DNA from an unrelated healthy volunteer was also subjected to sodium bisulfite genomic sequencing. A dense methylation was observed in both the promoter regions of the gene. Heavy methylation of CpG dinucleotides in these regions corroborates the imprinting result for this gene. Since the methylation epigenetic mark is also known to be associated with CpG islands, CpG Plot/CpG Report analysis was used to identify CpG islands in this gene. The analysis showed the presence of two CpG islands, CpG I and CpG II, in the second intron of the gene. As the CpG I island is known to lack methylated CpGs (Ali et al., unpublished result from our lab), a DMR was sought for the CpG II island region. Heterozygosity was ascertained in this region by sequencing DNA from 17 abortuses. However, none of the abortuses showed heterozygosity. It was reasoned that if there is a differential methylation of the two alleles in this region, half of the clones (alleles) should be unmethylated, and the other half should show methylation. Therefore, DNA from abortus no. 3 was randomly chosen for sodium bisulfite genomic sequence anaylsis to identify DMR. The CpG II island showed heavy methylation. However, a DMR was not identified. In order to see the methylation status of the CpG II island in human sperms, sperm DNA from an unrelated healthy volunteer was also subjected to sodium bisulfite genomic sequencing. Almost all the CpG sites showed methylation. The observation of a dense methylation of both the promoters and CpG II island suggested that methylation has a role in the expression of this gene. In order to confirm this observation, A549 and HuH7 cells were treated with a methyltransferase inhibitor, 5-aza-2’-deoxycytidine. 5-Aza-2’-deoxycytidine treatment in HuH7 cells restored the expression of this gene. Further, the expression of this gene was increased in A549 cells following the drug treatment. These results suggested that DNA methylation has a definite role in the modulation of expression of the SLC22A18AS gene. Histone acetylation is another key epigenetic player which is known to have a role in the expression of genes. In order to study the role of the histone acetylation, HuH7 and A549 cells were treated with TSA, a histone deacetylase inhibitor. Treatment of HuH7 and A549 cells with TSA didn’t have any effect on the expression of this gene. On the other hand, the expression of TPA, a gene shown to be regulated by TSA earlier, increased following the TSA treatment in both cell lines. These results suggested that histone acetylation doesn’t have any effect on the expression of this gene. Based on this observation, it was reasoned that histone acetylation is not associated with the imprinting of this gene. Therefore, we did not look for the allele-specific acetylation of histones in this gene. The SLC22A18AS gene has a weak ORF of 253 amino acids as the translation intiation site does not contain a consensus Kozak sequence for efficient translation. In order to determine if it codes for a protein, Western blot analysis was performed using lysates from A549 cells and human fetal liver tissue, and a polyclonal antibody raised in a rabbit against a bacterially expressed SLC22A18AS protein fragment from amino acids 138 to 245. The Western blot result was negative. It was reasoned that this gene might be expressed at a low level and therefore its expression could not be detected by Western blot analysis. Immunoprecipitation was then performed to enrich the SLC22A18AS protein in the lysates followed by Western blot analysis. SLC22A18AS was shown to be expressed as a 30 kDa band in the immunoprecipiates from A549 cell and human fetal liver tissue lysates. The subcellular localization of this gene was studied by immunofluorescence. The fluorescence immunolocalization was performed on A549 cells with anti-SLC22A18AS antibody. The SLC22A18AS protein was found to be localized in the cytoplasm of A549 cells.