Transcription lies at the centre of gene expression. In eukaryotes, transcription occurs not only at genes but also across the non-coding portion of the genome, an apparently pervasive process that gives rise to a wide array of different transcripts. In recent years, it has emerged that genes themselves are frequently subject to non-coding transcription of their antisense strand. This antisense transcription is evident in eukaryotes from yeast to mammals; however its general genome-wide role, if indeed it has one, remains elusive. Here, the nature of antisense transcription in the budding yeast Saccharomyces cerevisiae is explored on a genome-wide scale. Antisense transcription is ubiquitous and often abundant, and appears to be driven by a promoter architecture at the 3’ end of genes, one which shows evidence of regulation, and which mirrors that found at the 5’ end. Furthermore, antisense transcription shows evidence of changing gene behaviour. It is associated with a drastically altered chromatin environment at the 5’ promoter and across the gene body; however it is not associated with a change in the level of gene transcription itself. Rather, these chromatin changes appear to enforce a change in the mode of gene transcription, promoting rapid bursts of transcription re-initiation that result in noisier gene expression – a hitherto unknown role of antisense transcription. It is proposed that antisense transcription represents a fundamental layer of gene regulation, and that it should be considered a canonical feature of eukaryotic genes.

Genome transcription is much more widespread than has been traditionally thought because our view of a "gene" or "transcription unit" has changed dramatically over the past 4 to 5 years with the identification of many different non-coding ribonucleic acids. In the yeast, Saccharomyces cerevisiae, meiosis and sporulation are an important part of the life cycle and IME4 gene expression is required for these processes. IME4 sense transcript levels of expression are influenced by the level of its complementary non-coding antisense strand by mechanisms that are currently unknown. The a1-alpha2 heterodimer binding in the downstream 3' region of IME4 is one component required for repression of IME4 antisense transcription. However, this thesis shows that the general regulatory protein Reb1 is also required in this system. Reb1 involvement is most likely to create a nucleosome-free zone in the promoter region of the IME4 antisense strand therefore contributing to transcription.

The human genome contains ~1.5% coding sequence, with the remaining 98.5% being non-coding. The functional potential of the majority of this non-coding sequence remains unknown. Much of this non-coding sequence is derived from transposable element (TE) sequences. These TE sequences contain their own regulatory information, e.g. promoter and transcription factor binding sites. Given the large number of these sequences, over 4 million in the human genome, it would be expected that the regulatory information that they contain would affect the expression of nearby genes. This dissertation describes research that characterizes that alternation of and contribution to the human transcriptome by non-coding elements, including TE sequences.

Bien que très proches dans leur organisation génomique, le rétrovirus HTLV-1 est impliqué dans le développement de la leucémie à cellule T de l’adulte (ATL) alors que l’infection par HTLV-2 n’a jamais été associée à des désordres hématologiques malins. La transformation des cellules infectées par HTLV-1 a longtemps été attribuée uniquement à la protéine virale transactivatrice Tax (Tax-1). Cependant, son expression est très faible dans les cellules ATL. La protéine HBZ a été découverte en 2002. Elle est traduite à partir d'un ARNm transcrit à partir du LTR 3' d'HTLV-1 et est exprimée par les cellules infectées issues de tous les patients HTLV-1 quel que soit leur statut clinique. HBZ participe au maintien du phénotype tumoral en stimulant la prolifération des cellules leucémiques et intervient dans l'échappement du virus au système immunitaire. Des analyses in silico nous avaient permis de détecter un cadre ouvert de lecture sur le brin complémentaire de l’ARN génomique d’HTLV-2. Mon travail de thèse a consisté à amplifier et caractériser d’une part le transcrit APH-2 et d’autre part la protéine qui en est issue. Nous avons démontré dans un premier temps que la transcription d’APH-2 était initié dans le LTR 3’ et que le transcrit APH-2 était épissé et poly-adénylé. Nous avons ensuite mis en évidence l’expression d’APH-2 dans les lignées infectées par HTLV-2, ainsi que dans des cultures de lymphocytes issus de deux porteurs sains africains. La mise au point d’une technique quantitative de RT-PCR nous a permis de détecter APH-2 ex vivo chez 94% des individus d’une série de 51 porteurs sains américains. Nous avons aussi montré que l'expression de cet ARNm était proportionnelle à la charge provirale. APH-2 code une protéine de 183 acides aminés dont nous avons mis en évidence l’expression dans la lignée MO. Mes travaux ont aussi permis de démontrer le rôle inhibiteur d’APH-2 sur la transcription virale malgré l’absence d’un domaine bZip classique, ainsi que son interaction avec le facteur de transcription CREB. Par immunofluorescence, nous avons établi la localisation nucléaire d’APH-2. La protéine semble associée aux corps PML grâce à une région de six arginines comprise entre les résidus 78 et 92. Cependant, contrairement à HBZ, nous n’avons pas observé d’interactions entre APH-2 et les facteurs cJun, JunD ou le cofacteur de transcription CBP/p300. De plus nous avons observé qu’APH-2 était incapable d’induire la prolifération des lymphocytes in vitro alors qu’une lymphocytose est souvent observée chez les porteurs d’HTLV-2. Grâce à une approche comparative, mes travaux ont ainsi permis d’apporter des éléments nouveaux dans la compréhension de la différence de pathogénicité qu’il existe entre HTLV-1 et HTLV-2. / Although they are very similar in their genomic organization, the HTLV-1 retrovirus is involved in the development of adult T cell leukaemia (ATL) while HTLV-2 has not been associated to any malignant haematological disorders. The tumoral transformation of infected cells was widely associated to the viral transactivactor protein Tax (Tax-1), which modulates many cellular functions. However, its expression is slightly in ATL cells. In 2002, the HBZ protein was discovered, encoded from the 3’ LTR by the complementary strand of HTLV-1 and expressed by all HTLV-1 infected people. HBZ participates in the maintenance of the tumoral phenotype by stimulating leukemic cells proliferation and is involved in the immune system escape. We recently detected a coding region by an in silico analysis in the complementary strand of HTLV-2. My work consisted in the characterization of the APH-2 transcript, and in a second part, of the associated protein. At first, we characterized the APH-2 transcription initiation in the 3’LTR and that transcript was spliced and poly-adenylated and demonstrated that the APH-2 expression in all HTLV-2 cell lines and in short cultured lymphocytes from African healthy carriers. We used a quantitative RT-PCR on uncultured cells from 51 American HTLV-2 healthy carriers and we we detected APH-2 expression in 94% of them. We then showed that APH-2 RNA expression is correlated to the HTLV-2 proviral load. The APH-2 transcript encoded a 183 amino acid protein that was shown to be expressed in the HTLV-2 infected Mo cell line. Our work demonstrated the inhibitory functions of APH-2 in the viral transcription and its interaction with the transcriptional cofactor CREB despite the lack of a bZip domain. By an immunofluorescence approach we established the nuclear localisation of APH-2, which is in particular in the PML nuclear bodies. We demonstrated that six arginines in the 78-92 amino acids region is involved in this PML colocalization. Contrary to HBZ, we didn’t observe any interaction with between APH-2 and cJun or JunD factors nor with the transcriptional cofactor CBP/p300. Furthermore we showed that APH-2 is not involved in lymphocyte proliferation in vitro although a lymphocytosis is often observed in HTLV-2 carriers. According to this comparative approach, my work allowed us to better understand the difference of pathogenicity existing between HTLV-1 and HTLV-2.

Le premier membre de la famille des rétrovirus humains HTLV (Virus T-lymphotropique Humain), HTLV-1, a été découvert en 1980 et l’on estime aujourd’hui à plus de 10 millions le nombre d’individus infectés à travers le monde. Après une période de latence d’environ 40 ans, 5% des individus infectés développent des leucémies, des lymphomes adultes de lymphocytes T (ATLL) ou encore une myélopathie associée à HTLV-1/ paraparésie spastique tropicale (HAM/TSP). L’apparition de la maladie serait en grande partie orchestrée par deux protéines virales, soit Tax et HTLV-1 bZIP factor (HBZ). L’expression du génome viral se fait à partir d’un transcrit sens de pleine longueur suite à un épissage alternatif, à l’exception du gène HBZ. HBZ est produite à partir d’un transcrit antisens initié dans la séquence terminale longue répétée (LTR)’3. Elle a été décrite comme étant capable de réguler négativement la transcription virale dépendante de Tax en se dimérisant avec des facteurs de transcription cellulaires tels que CREB-2 et certains membres de la famille Jun. HBZ a aussi un pouvoir prolifératif et bien que nous ne sachions toujours pas le mécanisme moléculaire menant à l’oncogenèse par HBZ, nous savons qu’elle module une multitude de voies de transduction de signaux, dont AP-1. Nous avons récemment mis en évidence un transcrit antisens nommé Antisense Protein of HTLV-2 (APH-2) chez HTLV-2 qui n’est associé qu’à une myélopathie apparentée au HAM/TSP. Ce n’est qu’en 2005 que HTLV-3 et HTLV-4 se sont rajoutés au groupe HTLV. Cependant, aucune corrélation avec le développement d’une quelconque maladie n’a été montrée jusqu’à ce jour. Le premier volet de ce projet de doctorat avait pour objectif de détecter et caractériser les transcrits antisens produits par HTLV-3 et HTLV-4 et d’étudier les protéines traduites à partir de ces transcrits pour ainsi évaluer leurs similitudes et/ou différences avec HBZ et APH-2. Nos études de localisation cellulaire réalisées par microscopie confocale ont montré que APH-3 et APH-4 sont des protéines nucléaires, se retrouvant sous la forme de granules et, dans le cas d’APH-3, partiellement cytoplasmique. Ces granules co-localisent en partie avec HBZ. Les analyses à l’aide d’un gène rapporteur luciférase contenant le LTR 5’ de HTLV-1 ont montré que APH-3 et APH-4 peuvent aussi inhiber la transactivation du LTR 5’ par Tax. Aussi, des études faisant appel au gène rapporteur précédé d’un promoteur de collagénase (site AP-1), ont montré que ces deux protéines, contrairement à HBZ, activent la transcription dépendante de tous les membres des facteurs de transcription de la famille Jun. De plus, les mutants ont montré que le motif fermeture éclair (LZ) atypique de ces protéines est impliqué dans cette régulation. En effet, APH-3 et APH-4 modulent la voie Jun-dépendante en se dimérisant via leur LZ atypique avec la famille Jun et semblent activer la voie par un mécanisme ne faisant pas par d’un domaine activateur autonome. Dans un deuxième volet, nous avions comme objectif d’approfondir nos connaissances sur la localisation nucléolaire de HBZ. Lors de nos analyses, nous avons identifié deux nouveaux partenaires d’interaction, B23 et la nucléoline, qui semblent être associés à sa localisation nucléolaire. En effet, ces interactions sont plus fortes suivant une délétion des domaines AD et bZIP de HBZ qui dans ce cas est localisée strictement au nucléole. De plus, bien que APH-3 et APH-4 puissent se localiser aux nucléoles, HBZ est la seule protéine traduite à partir d’un transcrit antisens pouvant interagir avec B23. Finalement, ces travaux ont clairement mis en évidence que HTLV-3 et HTLV-4 permettent la production de transcrits antisens comme chez d’autres rétrovirus. Les protéines traduites à partir de ces transcrits antisens jouent d’importants rôles dans la réplication rétrovirale mais semblent avoir des fonctions différentes de celles de HBZ au niveau de la régulation de la transcription de la voie Jun. HBZ semble aussi jouer un rôle unique dans le nucléole en ciblant les protéines nucléolaires de la cellule. Ces études démontrent que les protéines produites à partir de transcrits antisens chez les rétrovirus HTLV partagent plusieurs ressemblances, mais démontrent aussi des différences. Ainsi, les APH pourraient, en tant qu’outil comparatif, aider à mieux cibler les mécanismes moléculaires importants utilisés par HBZ pour induire la pathogénèse associée à une infection par HTLV. / The first human T-cell lymphotropic virus (HTLV) family member was discovered in 1980 and it is estimated that approximately 10 million people are infected with HTLV-1 worldwide. After about 40 years, 5% of infected individuals will develop an adult T-cell leukemia/lymphoma (ATLL) while another 4% will develop HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP). It is believed that two viral proteins, Tax and HBZ, together orchestrate the oncogenic process. The viral proteins are expressed from an alternatively spliced sense transcript except for the HBZ gene. HBZ is translated from an antisense transcript initiated in the long terminal repeat (LTR)’3. This viral protein is capable of inhibiting Tax transactivation of the LTR5’ by dimerizing with cellular transcription factors such as CREB-2 and c-Jun. HBZ also has proliferating capacities and while the molecular mechanisms leading to the disease still need to be elucidated, it is well known that HBZ can modulate a multitude of signal transduction pathways like AP-1. We have recently discovered an antisense transcript termed Antisense Protein of HTLV-2 (APH-2) produced in HTLV-2. HTLV-2 is only associated to myelopathies resembling HAM/TSP. HTLV-3 and HTLV-4 were discovered in 2005 and have not been associated with any type of disease thus far. The first goal of this PhD project was hence to detect and characterize the antisense transcripts produced in HTLV-3 and HTLV-4, to study the functions of these translated proteins and to evaluate their similarities and/or differences shared with HBZ and APH-2. Our localization studies using confocal microscopy demonstrated that APH-3 and APH-4 are found in the nucleus as speckles, and for APH-3, also partially cytoplasmic. These two proteins can also partially colocalize with HBZ. Using a luciferase reporter plasmid bearing the HTLV-1 LTR5’, we demonstrated that APH-3 and APH-4 could inhibit Tax transactivation of the LTR5’. We also used a luciferase reporter plasmid bearing the collagenase promoter, which bears an AP-1 site, and demonstrated that both viral proteins could activate transcription in the presence of any of the Jun family of transcription factors. We generated several mutants and the atypical leucine zipper (LZ) found in APH-3 and APH-4 is crucial for this regulation. In fact, APH-3 and APH-4 using their atypical LZ dimerize with Jun family members and activate this pathway using a mechanism other than an autonomous activation domain. Our next goal was to investigate the significance of the HBZ nucleolar localization. During this project, we identified two new interacting partners, B23 and nucleolin, which seem to be associated with its nucleolar localization. In fact, these interactions are stronger when HBZ is deleted of its AD and bZIP domains and hence when HBZ demonstrates a stronger nucleolar distribution. Moreover, while APH-3 and APH-4 are also found in the nucleolus, HBZ is the only antisense protein able to interact with B23. Finally, this work clearly demonstrates that HTLV-3 and HTLV-4 can produce an antisense transcript alike other retroviruses. The encoded proteins play an important role in retroviral replication and seem to regulate Jun-dependant transcription differently than HBZ. HBZ also seems to have a unique role in the nucleoli by targeting specific cellular nucleolar proteins. Similarities but also differences are shared between the antisense proteins. Thus, the APH proteins represent a good comparative tool in order to better understand the molecular mechanisms involved in HTLV induced diseases.

O câncer de próstata é o quinto tipo mais comum de câncer no mundo e o mais comum em homens. Fatores clínicos e anatomopatológicos atualmente usados na clínica não são capazes de distinguir entre a doença indolente e a agressiva. Existe uma grande necessidade de novos marcadores de prognóstico, a fim de melhorar o gerenciamento clínico de pacientes de câncer de próstata. Além das anormalidades em genes codificadores de proteínas, alterações em RNAs não codificadores (ncRNAs) contribuem para a patogênese do câncer e, portanto, representam outra fonte potencial de biomarcadores de câncer de próstata. Entretanto, até o momento, poucos estudos de perfis de expressão de ncRNAs foram publicados. Este projeto teve como principal objetivo identificar perfis de expressão de genes codificadores e não codificadores de proteína correlacionados com recorrência de tumor de próstata, a fim de gerar um perfil prognóstico com potencial uso como biomarcadores e elucidar o possível papel de ncRNAs no desenvolvimento do câncer. Para isso, foram analisados os perfis de expressão de genes codificadores e não codificadores de proteína de um conjunto de 42 amostras de tecido tumoral de câncer de próstata de pacientes de amostras de pacientes submetidos à prostatectomia radical, com longo acompanhamento clínico (cinco anos) e conhecida evolução da doença Nós utilizamos microarranjos por nós desenhados e fabricados pela Agilent sob encomenda, interrogando aproximadamente 18.709 transcritos não codificadores longos (>500 nt), sem evidência de splicing, que mapeiam em regiões intrônicas dentro de 5.660 loci genômicos. Os dados de expressão foram extraídos de cada arranjo, normalizados entre todas as 42 amostras de pacientes. Usando uma estratégia de múltipla amostragem, foi identificado um perfil de expressão de mau prognóstico, contendo 51 transcritos intrônicos não codificadores de proteína. O perfil prognóstico de ncRNAs foi aplicado a um conjunto teste independente de 22 pacientes, classificando corretamente 82% das amostras. Uma análise de Kaplan-Meier dos pacientes do conjunto teste indicou que as curvas de sobrevida dos grupos de alto e baixo risco foram significativamente distintas (Log-rank test p = 0,0009; Hazard ratio = 23,4, 95% CI = 3,62 a 151,2), confirmando assim que este classificador é útil para identificar pacientes com alto risco de recorrência. Além disso, estas descobertas indicam um potencial papel destes RNAs intrônicos não codificadores na progressão do tumor de próstata e apontam para os RNAs intrônicos como potenciais novos marcadores de câncer / Prostate cancer is the fifth most common type of cancer in the world, and the most common in men. Clinical and anatomo-pathological factors currently used in clinic are not able to distinguish between the indolent and the aggressive disease. There is a major need of new prognostic makers in order to improve the clinical management of prostate cancer patients. Apart from abnormalities in protein-coding genes, changes in non-coding RNAs (ncRNAs) contribute to the pathogenesis of cancer and thus represent another potential source of prostate cancer biomarkers. However, few studies of expression profiles of ncRNAs have been published. This project aimed to identify expression profiles of protein-coding and non-coding genes correlated to prostate cancer biochemical recurrence. For this, we analyzed the expression profile of 42 prostate cancer samples from patients undergoing radical prostatectomy, with long follow-up (five years), and know disease outcome. We used a custom microarray designed by us and printed by Agilent, that probes 18,709 long (>500 nt) ncRNAs mapping to intronic regions within 5,660 genomic loci. The expression data were extracted from each array and normalized across all 42 samples. Using a multiple random sampling validation strategy, we identified an expression profile of poor prognosis, comprising 51 ncRNAs. The prognostic profile of ncRNAs was applied to an independent test set of 22 patients, correctly classifying 82% of the samples. A Kaplan-Meier analysis of the test set of patients indicated that the survival curves of high and low risk groups were significantly different (Log-rank test p = 0.0009, Hazard ratio = 23.4, 95% CI = 3.62 to 151.2) thus confirming that this classifier is useful for identifying patients at high risk of recurrence. Furthermore, these findings indicate a potential role of these intronic non-coding RNAs in the progression of prostate tumors and points to the intronic ncRNAs as potential new markers of cancer.

Le génome des rétrovirus existe sous deux formes différentes : sous forme d'ARN simple brin, qui est traduit ou encapsidé, ou sous forme d'ADN double brin intégré dans le génome de la cellule hôte infectée. Cette dernière forme, l'ADN proviral, est indispensable à la production de tous les ARNm viraux nécessaires à la synthèse des protéines virales, qui en retour agissent sur la région promotrice située au niveau du LTR 5'. Cependant, l'ADN proviral possède un second LTR à son extrémité 3', capable de réguler une transcription antisens, orientée dans la direction opposée à celle contrôlée par le LTR 5'. L'ADN proviral a donc deux brins codants, ce qui offre au virus un plus grand potentiel de synthèse protéique. Dans le cas du Virus de l'immunodéficience Humaine de type 1 (VIH-1), la transcription antisens permet la production d'une protéine, appelée ASP (Antisense Protein). Dans ce manuscrit, nous démontrons que cette activité transcriptionnelle antisens s'exprime préférentiellement dans les cellules d'origine monocytaire, en particulier les cellules dendritiques ; une localisation membranaire de la protéine ASP a par ailleurs été mise en évidence dans ce type cellulaire. Nos résultats suggèrent également que la transcription antisens du VIH-1 est indépendante de la protéine Tat, et que par ailleurs les deux types de transcriptions ne sont pas exprimés simultanément au sein d'une même cellule. En outre, nos données soulignent que la séquence codante de la protéine ASP est très fortement conservée parmi les différents isolats viraux. Sur la base de l'ensemble de ces résultats, notre hypothèse est que la protéine ASP du VIH-1 possède des fonctions cruciales dans le cycle réplicatif des rétrovirus, indépendantes de la production virale. / Genome of retroviruses exists in two different forms: as single-stranded RNA that is translated or packaged, or as double-stranded DNA integrated into the genome of the infected host cell. The latter form, the proviral DNA, is essential for the production of all viral mRNAs required for the synthesis of viral proteins, which in turn act on the promoter region located at the 5 '-LTR. However, the proviral DNA has a second LTR at its 3 '-end, capable of regulating antisense transcription oriented in the opposite direction to that controlled by the 5'-LTR. The proviral DNA has then two coding strands, which gives the virus a greater potential for protein synthesis. In the case of the Human Immunodeficiency Virus type 1 (HIV-1), antisense transcription allows the production of a protein called ASP (Antisense Protein). In this manuscript, we demonstrate that this antisense transcriptional activity is preferentially expressed in cells of the monocyte lineage, in particular dendritic cells; a membrane localization of the ASP protein was also observed in this cell type. Our results also suggest that the antisense transcription of HIV-1 is Tat-independent, and what's more that the two types of transcription are not expressed simultaneously within the same cell. In addition, our data highlight that the ASP protein coding sequence is highly conserved among different viral isolates. Based on these results, our hypothesis is that the ASP protein of HIV-1 has critical functions in the replicative cycle of retroviruses, distinct from viral production.

O câncer de próstata é o quinto tipo mais comum de câncer no mundo e o mais comum em homens. Fatores clínicos e anatomopatológicos atualmente usados na clínica não são capazes de distinguir entre a doença indolente e a agressiva. Existe uma grande necessidade de novos marcadores de prognóstico, a fim de melhorar o gerenciamento clínico de pacientes de câncer de próstata. Além das anormalidades em genes codificadores de proteínas, alterações em RNAs não codificadores (ncRNAs) contribuem para a patogênese do câncer e, portanto, representam outra fonte potencial de biomarcadores de câncer de próstata. Entretanto, até o momento, poucos estudos de perfis de expressão de ncRNAs foram publicados. Este projeto teve como principal objetivo identificar perfis de expressão de genes codificadores e não codificadores de proteína correlacionados com recorrência de tumor de próstata, a fim de gerar um perfil prognóstico com potencial uso como biomarcadores e elucidar o possível papel de ncRNAs no desenvolvimento do câncer. Para isso, foram analisados os perfis de expressão de genes codificadores e não codificadores de proteína de um conjunto de 42 amostras de tecido tumoral de câncer de próstata de pacientes de amostras de pacientes submetidos à prostatectomia radical, com longo acompanhamento clínico (cinco anos) e conhecida evolução da doença Nós utilizamos microarranjos por nós desenhados e fabricados pela Agilent sob encomenda, interrogando aproximadamente 18.709 transcritos não codificadores longos (>500 nt), sem evidência de splicing, que mapeiam em regiões intrônicas dentro de 5.660 loci genômicos. Os dados de expressão foram extraídos de cada arranjo, normalizados entre todas as 42 amostras de pacientes. Usando uma estratégia de múltipla amostragem, foi identificado um perfil de expressão de mau prognóstico, contendo 51 transcritos intrônicos não codificadores de proteína. O perfil prognóstico de ncRNAs foi aplicado a um conjunto teste independente de 22 pacientes, classificando corretamente 82% das amostras. Uma análise de Kaplan-Meier dos pacientes do conjunto teste indicou que as curvas de sobrevida dos grupos de alto e baixo risco foram significativamente distintas (Log-rank test p = 0,0009; Hazard ratio = 23,4, 95% CI = 3,62 a 151,2), confirmando assim que este classificador é útil para identificar pacientes com alto risco de recorrência. Além disso, estas descobertas indicam um potencial papel destes RNAs intrônicos não codificadores na progressão do tumor de próstata e apontam para os RNAs intrônicos como potenciais novos marcadores de câncer / Prostate cancer is the fifth most common type of cancer in the world, and the most common in men. Clinical and anatomo-pathological factors currently used in clinic are not able to distinguish between the indolent and the aggressive disease. There is a major need of new prognostic makers in order to improve the clinical management of prostate cancer patients. Apart from abnormalities in protein-coding genes, changes in non-coding RNAs (ncRNAs) contribute to the pathogenesis of cancer and thus represent another potential source of prostate cancer biomarkers. However, few studies of expression profiles of ncRNAs have been published. This project aimed to identify expression profiles of protein-coding and non-coding genes correlated to prostate cancer biochemical recurrence. For this, we analyzed the expression profile of 42 prostate cancer samples from patients undergoing radical prostatectomy, with long follow-up (five years), and know disease outcome. We used a custom microarray designed by us and printed by Agilent, that probes 18,709 long (>500 nt) ncRNAs mapping to intronic regions within 5,660 genomic loci. The expression data were extracted from each array and normalized across all 42 samples. Using a multiple random sampling validation strategy, we identified an expression profile of poor prognosis, comprising 51 ncRNAs. The prognostic profile of ncRNAs was applied to an independent test set of 22 patients, correctly classifying 82% of the samples. A Kaplan-Meier analysis of the test set of patients indicated that the survival curves of high and low risk groups were significantly different (Log-rank test p = 0.0009, Hazard ratio = 23.4, 95% CI = 3.62 to 151.2) thus confirming that this classifier is useful for identifying patients at high risk of recurrence. Furthermore, these findings indicate a potential role of these intronic non-coding RNAs in the progression of prostate tumors and points to the intronic ncRNAs as potential new markers of cancer.

O câncer de mama é o carcinoma que mais acomete mulheres no Brasil. Os tratamentos disponíveis são recomendados a partir da análise de fatores de prognóstico como a classificação pelo sistema TNM, tipo histológico, status de receptores hormonais e marcadores de proliferação tumoral. No entanto, a classificação dos tumores de mama é muito variável e o poder prognóstico dos marcadores tumorais atuais ainda é limitado, levando muitas pacientes à terapia adjuvante desnecessária. Portanto, novos métodos de prognóstico mais sensíveis são necessários para melhorar a tomada de decisão na clínica oncológica de pacientes com câncer de mama. Do ponto de vista de ciência básica, as modificações transcricionais associadas à oncogênese e progressão do câncer de mama ainda são pouco conhecidas. Além da alteração na expressão de genes codificadores para proteínas, evidências recentes sugerem que RNAs não-codificadores (ncRNAs) podem ter um papel importante na transformação maligna. Este projeto teve como principais objetivos: i) investigar a expressão de ncRNAs intrônicos em amostras de adenocarcinoma de mama e ii) identificar assinaturas de expressão gênica associadas a características anatomo-patológicas e clínicas de tumores de mama com potencial aplicação clínica. Para isso, foram comparados os perfis de expressão gênica de 58 amostras de tecido tumoral de mama, com seguimento clínico conhecido, utilizando uma plataforma de microarranjos de cDNA, enriquecida em ncRNAs provenientes de regiões intrônicas de genes humanos conhecidos. 9 Durante o projeto foram testadas diferentes metodologias para análise da expressão gênica utilizando microarranjos de cDNA com uma ou duas cores. O desenho experimental das hibridizações incluiu a co-hibridização de cada microarranjo com alvos fluorescentes representando o transcritoma da amostra de tumor juntamente com um oligonucleotídeo referência complementar a uma região presente em todas as sondas de cDNA (RefOligo). Este desenho experimental permitiu a avaliação de duas abordagens de análise da expressão gênica: a primeira baseada nas intensidades diretas de cada transcrito (One-Color) e a segunda baseada em razões de expressão onde a intensidade de cada transcrito foi normalizada pelo oligonucleotídeo referência (RefOligo). A utilização direta das intensidades se mostrou mais reprodutível e sensível para a detecção de assinaturas de expressão correlacionadas com características das amostras de mama, e essa abordagem foi escolhida para as análises subseqüentes. Os dados provenientes dos experimentos de microarranjos revelaram níveis de expressão ubíqüos dos transcritos intrônicos nas amostras analisadas, extendendo para o câncer de mama a relevância do estudo desta classe de ncRNAs. Além disso, foi identificada uma assinatura contendo 95 transcritos, correlacionada com o status de expressão do receptor de estrogênio (REr), dos quais cerca de 15% correspondem a ncRNAs. Utilizando apenas amostras com seguimento clínico superior a 4 anos, foi identificada uma assinatura com 113 transcritos, dos quais cerca de 30% são ncRNAs intrônicos, capaz de distinguir com 100% de acurácia pacientes que desenvolveram metástase daqueles que permaneceram livres da doença. Além de contribuir com novos candidatos a marcadores de prognóstico no câncer de mama, este estudo aponta para a participação de ncRNAs intrônicos em complexas redes transcricionais, possivelmente modulando a expressão de genes codificadores para proteínas. A caracterização detalhada da função de ncRNAs com expressão correlacionada a características fenotípicas e clínicas dos tumores de mama deverá fornecer novas informações sobre as bases moleculares da tumorigênese e progressão desta neoplasia. / Breast carcinoma is the most frequently occurring cancer amongst women in Brazil. The treatments available for breast cancer are prescribed based on the results of prognostic factors, such as the TNM classification system, histological type, hormonal receptor status and tumoral markers for cell proliferation. Nevertheless, breast cancer classification can be variable and inconsistent, and the prognosis power of tumoral markers is still limited, resulting in many patients unnecessarily undergoing adjuvant therapy. Therefore, there is an urge for new prognosis methods that are more sensitive, as well as accurate, in order to improve treatment decisions for breast cancer patients. From a basic science perspective, transcriptional modifications associated with oncogenesis and breast cancer progression are still poorly understood. Beyond alterations of the expression of protein-coding genes, recent evidences suggest that non-coding RNAs (ncRNAs) might have an important role in malignant transformation. The main goals of this project are: i) to investigate the expression of intronic ncRNAs in breast cancer tissue and ii) to identify gene expression signatures correlated to anatomo-pathological and clinical characteristics of human breast tumors, with a potential clinical aplication. To achieve this, gene expression profiles of 58 breast tumor samples with clinical follow-up were compared using a microarray platform enriched in non-coding RNAs (ncRNAs) derived from intronic regions of known human genes. During this project different gene expression methodologies were tested for the analysis of one- or two-color cDNA microarrays. The experimental design included the co-hybridization of the microarrays with fluorescent targets representing the tumor sample transcriptome with a reference oligonucleotide that is complementary to a 12 common region present in all cDNA probes (RefOligo). This experimental design permited the evaluation of two gene expression analysis approaches: the first based on direct intensities of each transcript (One-Color) and the second based in expression ratios where the intensity of each transcript is normalized by the reference oligonucleotide (RefOligo). One-Color methodology has shown to provide a more reproducible and sensitive gene expression signatures correlated to the breast samples characteristics and, therefore, this approach was chosen for subsequent analysis. The data provided by the microarray experiments revealed that ubiquitous expression of intronic ncRNAs was observed, confirming the relevance of investigating the role of this class of ncRNAs in breast cancer. Furthermore, a gene expression signature comprising 95 transcripts and correlated to the estrogen receptor status of breast tumor samples was identified, from which approximately 15% are ncRNAs. Using only samples from patients with known follow-up, a signature of 113 transcripts was identified, of which 30% are ncRNAs. This gene expression signature was able to distinguish with 100% accuracy patients that developed metastasis from those that remained disease-free up to 4 years after surgery. Besides the contribution of new molecular prognostic markers for breast cancer, the present study indicates that intronic ncRNAs might play a role in complex transcriptional networks, possibly regulating the expression of protein-coding genes. The detailed caracterization of the functional roles of ncRNAs, whose expression levels are correlated to fenotypical and clinical characteristics of breast tumors, is likely to provide new insigths on the molecular basis of tumorigenesis and progression of this neoplasia.

Les rétrotransposons sont des éléments génétiques mobiles qui constituent presque la moitié de notre génome. Seule la sous-famille L1HS appartenant à la classe des Long Interspersed Element-1(LINE-1 ou L1) a gardé une capacité de mobilité autonome chez l’Homme. Leur mobilisation dans la lignée germinale, mais Aussi dans certains tissus somatiques, contribue à la diversité du génome humain ainsi qu’à certaines maladies comme le cancer. Ainsi, de nouvelles copies de L1 peuvent directement s'intégrer dans des séquences codantes ou régulatrices, et altérer leur fonction. De plus, les séquences L1 contiennent elles-mêmes plusieurs éléments cis-régulateurs et leur insertion à proximité ou dans un gène peut produire des altérations génétiques plus subtiles. Afin d'explorer l'ensemble de ces altérations à l'échelle du génome, nous avons développé un logiciel dédié à l’analyse des données de séquençage d'ARN qui permet d'identifier des transcrits chimériques ou antisens impliquant les L1 et d'annoter ces isoformes en fonction des différents événements d’épissage alternatif subits. Au cours de ce travail, il est apparu que la compréhension du lien entre polymorphisme des insertions et phénotype nécessite une vue complète des différentes copies L1HS présentes chez un individu donné. Afin de disposer d'un catalogue aussi complet que possible de ces polymorphismes identifiés dans des échantillons humains sains ou pathologiques et publiés dans des journaux scientifiques, nous avons développé euL1db, la base de données des insertions de rétrotransposon L1HS chez l’Homme. En conclusion, ce travail aidera à comprendre l’impact des L1 sur l’expression des gènes, à l'échelle du génome. / Retrotransposons are mobile genetics elements, which form almost half of our genome. Only the L1HS subfamily of the Long Interspersed Element-1 class (LINE-1 or L1) has retained the ability to jump autonomously in humans. Their mobilization in the germline – but also in some somatic tissues – contributes to human genetic diversity and to diseases, such as cancer. L1 reactivation can be directly mutagenic by disrupting genes or regulatory sequences. In addition, L1 sequences themselves contain many regulatory cis-elements. Thus, L1 insertions near a gene or within intronic sequences can also produce more subtle genic alterations. To explore L1-mediated genic alterations in a genome-wide manner, we have developed a dedicated RNA-seq analysis software able to identify L1 chimeric or antisense transcripts and to annotate these novel isoforms with their associated alternative splicing events. During the course of this work, it appeared that understanding the link between L1HS insertion polymorphisms and phenotype or disease requires a comprehensive view of the different L1HS copies present in a given individual or sample. To provide a comprehensive summary of L1HS insertion polymorphisms identified in healthy or pathological human samples and published in peer-reviewed journals, we developed euL1db, the European database of L1HS retrotransposon insertions in humans. This work will help understanding the overall impact of L1 insertions on gene expression, at a genome-wide scale.