O emprego da metodologia de microarranjos no estudo do câncer tem permitido a identificação de genes com alterações em seus perfis de expressão, os quais estão direta ou indiretamente envolvidos na etiologia dessa doença. O crescente número de publicações de experimentos de expressão gênica em repositórios de dados de microarranjos demonstra que, em geral, a limitação não está na quantidade ou qualidade desses experimentos, mas no processamento desses dados. Dessa forma, há a necessidade de desenvolver metodologias de bioinformática que sejam capazes de analisar tais perfis de forma integrada, incluindo no contexto da análise, dados provenientes de outros experimentos. O desenvolvimento do câncer tem sido associado principalmente a distúrbios nos mecanismos de controle do ciclo celular, cujas vias são dependentes de uma maquinaria transcricional e de seus elementos regulatórios; entre estes últimos, os fatores de transcrição (FTs) têm sido estudados como potenciais alvos para a terapia molecular. No presente trabalho, um conjunto de dados de microarranjos realizados a partir de amostras de glioblastoma foi obtido em repositórios públicos (GEO e ArrayExpress). O teste estatístico SAM foi aplicado aos dados e os genes diferencialmente expressos (FDR 0,05), induzidos em glioblastoma (1.830 genes), foram submetidos a uma análise de associação a FTs (programa FatiGO+), a qual associou os FTs HNF1A, IRF7 e E2F (p 0,05) aos genes induzidos. Adicionalmente, a lista de genes foi submetida a uma análise de associação a um banco de dados de assinaturas transcricionais do software TBrowser, extraídas de diversos experimentos de microarranjos do GEO. Como resultado, 7.910 assinaturas foram associadas (p 0,01), sendo que as cem primeiras também foram relacionadas a FTs por meio de ferramentas disponibilizadas no próprio TBrowser. Os FTs E2F1 e E2F4 foram associados a mais de 80% das assinaturas analisadas. Dessa forma, ambas as análises apontaram o FT E2F, mais especificamente E2F1 e E2F4, como associados à lista inicial de genes fornecida pelo SAM. A expressão de E2F1 e E2F4 foi avaliada por meio da técnica de RT-PCR quantitativa em tempo real em amostras de RNA de sete linhagens de glioblastoma (T98, U251, U138, U87, U343, MO59J e MO59K). Interessantemente, ambos os genes foram apontados como superexpressos em todas as linhagens, a maioria com significância estatística. A família de FTs E2F apresenta papéis importantes no controle da proliferação celular e apoptose. Alguns membros atuam como oncogenes, e outros, como genes supressores de tumor, dependendo do contexto molecular nos quais se encontram. Muitos trabalhos da literatura têm apontado a importância desses FTs, especialmente E2F1, no processo de itumorigênese. Com base nos resultados obtidos no presente trabalho, o status de expressão (indução) de E2F1 e E2F4 é provavelmente associado ao glioblastoma. Esses FTs podem influenciar a expressão de uma série de genes cruciais, frequentemente alterados nessa doença, o que ainda requer um estudo mais detalhado, visando à validação desses FTs como biomarcadores, ou até mesmo como alvos moleculares que possam ser empregados no estabelecimento de novas modalidades de terapias. / The application of DNA microarrays in cancer research has provided the identification of abnormal expression profiles of a series of genes, which may be directly involved in the etiology of such disease. The increasing number of publications related to gene expression profile in microarray repositories has demonstrated that the limitation of such experiments is not associated with the quantity neither the quality of such experiments, but with data processing. Therefore, there is a great interest in the development of methodologies in bioinformatics that allow the integrated analysis of such profiles, including data set from other experiments. Cancer development has been associated mostly with changes in cell cycle control, whose intracellular pathways are dependent on the transcriptional machinery. Regulatory elements involved in transcription, which include transcriptional factors (TFs), can be potential targets for molecular therapy. In the present study, a set of microarray data achieved from glioblastoma samples were obtained from public repositories (GEO and ArrayExpress). Data were submitted to statistical analysis using SAM, and 1,830 up-regulated genes (FDR 0.05) were submitted to TFs association analysis (FatiGO+ program). Those genes were associated to the TFs HNF1A, IRF7 and E2F (p 0.05). Moreover, the set of genes was submitted to a transcription signature bank data association analysis using the software TBrowser, extracted from GEO, which provided a wide range of data sets from microarray experiments. The analysis led to 7,910 associated signatures (p 0.01); out of them, the first 100 were related to TFs according to the tools available in TBrowser. The TFs E2F1 and E2F4 were associated to more than 80% from the analyzed signatures. Therefore, both analysis suggested that the TF E2F, specifically E2F1 and E2F4, were associated to the initial gene set obtained by the SAM analysis. The expression of E2F1 and E2F4 was evaluated by quantitative real-time PCR using RNA samples from seven glioblastoma cell lines (T98, U251, U138, U87, U343, MO59J and MO59K). Interestingly, both genes were found up-regulated in all cell lines. The E2F TFs family members present important roles in cell proliferation control and apoptosis. Depending on the molecular context involved, some members act as oncogenes and others as tumor suppressor genes. A great number of studies has suggested the importance of such TFs , specially E2F1, in the tumorigenesis process. According to the present results, the expression status (induced) of E2F1 and E2F4 may be associated with glioblastoma. These TFs may influence the abnormal expression of important genes in cancer, even though detailed studies should be necessary in order to validate these TFs as biomarkers or molecular targets for therapeutical intervention.

O emprego da metodologia de microarranjos no estudo do câncer tem permitido a identificação de genes com alterações em seus perfis de expressão, os quais estão direta ou indiretamente envolvidos na etiologia dessa doença. O crescente número de publicações de experimentos de expressão gênica em repositórios de dados de microarranjos demonstra que, em geral, a limitação não está na quantidade ou qualidade desses experimentos, mas no processamento desses dados. Dessa forma, há a necessidade de desenvolver metodologias de bioinformática que sejam capazes de analisar tais perfis de forma integrada, incluindo no contexto da análise, dados provenientes de outros experimentos. O desenvolvimento do câncer tem sido associado principalmente a distúrbios nos mecanismos de controle do ciclo celular, cujas vias são dependentes de uma maquinaria transcricional e de seus elementos regulatórios; entre estes últimos, os fatores de transcrição (FTs) têm sido estudados como potenciais alvos para a terapia molecular. No presente trabalho, um conjunto de dados de microarranjos realizados a partir de amostras de glioblastoma foi obtido em repositórios públicos (GEO e ArrayExpress). O teste estatístico SAM foi aplicado aos dados e os genes diferencialmente expressos (FDR 0,05), induzidos em glioblastoma (1.830 genes), foram submetidos a uma análise de associação a FTs (programa FatiGO+), a qual associou os FTs HNF1A, IRF7 e E2F (p 0,05) aos genes induzidos. Adicionalmente, a lista de genes foi submetida a uma análise de associação a um banco de dados de assinaturas transcricionais do software TBrowser, extraídas de diversos experimentos de microarranjos do GEO. Como resultado, 7.910 assinaturas foram associadas (p 0,01), sendo que as cem primeiras também foram relacionadas a FTs por meio de ferramentas disponibilizadas no próprio TBrowser. Os FTs E2F1 e E2F4 foram associados a mais de 80% das assinaturas analisadas. Dessa forma, ambas as análises apontaram o FT E2F, mais especificamente E2F1 e E2F4, como associados à lista inicial de genes fornecida pelo SAM. A expressão de E2F1 e E2F4 foi avaliada por meio da técnica de RT-PCR quantitativa em tempo real em amostras de RNA de sete linhagens de glioblastoma (T98, U251, U138, U87, U343, MO59J e MO59K). Interessantemente, ambos os genes foram apontados como superexpressos em todas as linhagens, a maioria com significância estatística. A família de FTs E2F apresenta papéis importantes no controle da proliferação celular e apoptose. Alguns membros atuam como oncogenes, e outros, como genes supressores de tumor, dependendo do contexto molecular nos quais se encontram. Muitos trabalhos da literatura têm apontado a importância desses FTs, especialmente E2F1, no processo de itumorigênese. Com base nos resultados obtidos no presente trabalho, o status de expressão (indução) de E2F1 e E2F4 é provavelmente associado ao glioblastoma. Esses FTs podem influenciar a expressão de uma série de genes cruciais, frequentemente alterados nessa doença, o que ainda requer um estudo mais detalhado, visando à validação desses FTs como biomarcadores, ou até mesmo como alvos moleculares que possam ser empregados no estabelecimento de novas modalidades de terapias. / The application of DNA microarrays in cancer research has provided the identification of abnormal expression profiles of a series of genes, which may be directly involved in the etiology of such disease. The increasing number of publications related to gene expression profile in microarray repositories has demonstrated that the limitation of such experiments is not associated with the quantity neither the quality of such experiments, but with data processing. Therefore, there is a great interest in the development of methodologies in bioinformatics that allow the integrated analysis of such profiles, including data set from other experiments. Cancer development has been associated mostly with changes in cell cycle control, whose intracellular pathways are dependent on the transcriptional machinery. Regulatory elements involved in transcription, which include transcriptional factors (TFs), can be potential targets for molecular therapy. In the present study, a set of microarray data achieved from glioblastoma samples were obtained from public repositories (GEO and ArrayExpress). Data were submitted to statistical analysis using SAM, and 1,830 up-regulated genes (FDR 0.05) were submitted to TFs association analysis (FatiGO+ program). Those genes were associated to the TFs HNF1A, IRF7 and E2F (p 0.05). Moreover, the set of genes was submitted to a transcription signature bank data association analysis using the software TBrowser, extracted from GEO, which provided a wide range of data sets from microarray experiments. The analysis led to 7,910 associated signatures (p 0.01); out of them, the first 100 were related to TFs according to the tools available in TBrowser. The TFs E2F1 and E2F4 were associated to more than 80% from the analyzed signatures. Therefore, both analysis suggested that the TF E2F, specifically E2F1 and E2F4, were associated to the initial gene set obtained by the SAM analysis. The expression of E2F1 and E2F4 was evaluated by quantitative real-time PCR using RNA samples from seven glioblastoma cell lines (T98, U251, U138, U87, U343, MO59J and MO59K). Interestingly, both genes were found up-regulated in all cell lines. The E2F TFs family members present important roles in cell proliferation control and apoptosis. Depending on the molecular context involved, some members act as oncogenes and others as tumor suppressor genes. A great number of studies has suggested the importance of such TFs , specially E2F1, in the tumorigenesis process. According to the present results, the expression status (induced) of E2F1 and E2F4 may be associated with glioblastoma. These TFs may influence the abnormal expression of important genes in cancer, even though detailed studies should be necessary in order to validate these TFs as biomarkers or molecular targets for therapeutical intervention.

The regulation of gene expression at the right time, place, and degree is crucial for many cellular processes such as proliferation and development. In addition, in order to maintain cellular life, cells must rapidly and appropriately respond to various environmental stimuli. Sequence-specific transcription factors (TFs) can recognize functional regulatory DNA elements in a sequence-specific manner so that they can regulate only a specific group of genes, a process which enables cells to cope with diverse internal and external stimuli. Human has approximately 1,400 sequence-specific TFs whose aberrant expression causes a wide range of detrimental consequences including developmental disorders, diseases, and cancers; therefore, it is pivotal to identify the binding sites of each sequence-specific TF in order to unravel its roles in and mechanisms of gene regulation. Even though some TFs have been intensively studied, the majority of TFs still remain to be studied, particularly the tasks of identifying their genome-wide target genes and deciphering their biological roles in specific cellular contexts. Many questions remain unanswered: how many sites on the human genome a sequence-specific TF can bind; whether all TF-bound sites are functional; how a TF achieves binding specificity onto its targets; how and to what extent a TF is involved in gene regulation. Comprehensive identification of the binding sites of sequence-specific TFs and follow-up molecular studies including gene expression microarrays will provide close answers to these questions. Chromatin Immunoprecipitation coupled with recently developed high-throughput sequencing (ChIP-seq) allows us to perform genome-scale unbiased identification of the binding sites of sequence-specific TFs. Here, to gain insight into gene regulatory functions of TFs as well as their influences on gene expression, we conducted, in diverse cell lines, genome-wide identification of the binding sites of several sequence-specific TFs (CTCF, E2F4, MYC, Pol II) that are involved in a wide range of biological functions, including cell proliferation, development, apoptosis, genome stability, and DNA repair. Analysis of ChIP-seq data provided not only comprehensive binding profiles of those TF across the genome in diverse cell lines, but also revealed tissue-specific binding of CTCF, MYC, and Pol II as well as combinatorial usage among these three factors. Analyses also showed that some CTCF binding sites were inherited from parents to children and regulated in an individual-specific as well as allele-specific manner. Finally, genome-wide target identification of several TFs will broaden our understanding of the gene regulatory roles of these sequence-specific TFs. / text

Ischemic stroke occurs as a result of a lack or severe reduction of blood supply to the brain. Presently therapeutic interventions are limited and there is a need to develop new and efficacious stroke treatments. To this end, a great deal of research effort has been devoted to studying the potential molecular mechanisms involved in ischemic neuronal death. Correlative evidence demonstrated a paradoxical activation of the cell cycle machinery in ischemic neurons. The levels and activity of key cell cycle regulators including cyclin D1, Cdk2 and Cdk4 are upregulated following ischemic insults. However, the functional relevance of these various signals following ischemic injury was unclear. Accordingly, the research described in this thesis address the functional relevance of the activation of the cell cycle machinery in ischemic neuronal death. The data indicate that the inhibition of Cdk4 protects neurons from ischemia-induced delayed death, whereas abrogation of Cdk5 activity prevents excitotoxicity-induced damage in vitro and in vivo. Examination of upstream activators of mitotic-Cdks showed that Cdc25A is a critical mediator of delayed ischemic neuronal death. Investigation of the potential molecular mechanism by which cell cycle regulators induced neuronal death revealed perturbations in the levels and activity of key downstream targets of Cdk4. The retinoblastoma protein family members, pRb and p130 are increasingly phosphorylated following ischemic stresses. Importantly, p130 and E2F4 proteins are drastically reduced following ischemic insults. Additionally, E2F1 association with promoters of pro-apoptotic genes are induced while that of E2F4 is reduced. These changes appear to be important determinants in ischemic neuronal death. Cumulatively, the data supports the activation of the cell cycle machinery as a pathogenic signal contributing to ischemic neuronal death. The development of neuroprotectant strategies for stroke has been hampered in part by its complex pathophysiology. Previous research indicated that flavopiridol, a general CDK-inhibitor, is unable to provide sustained neuroprotection beyond one week following cerebral ischemia. The potential benefit of combining flavopiridol with another neuroprotectant, minocycline, was explored. The data indicate that while this approach provided histological protection 10 weeks after insult, the protected neurons are not functional due to progressive dendritic degeneration. This evidence indicates that targeting cell cycle pathways in stroke while important must be combined with other therapeutic modalities to fully treat stroke-induced damage.

Ischemic stroke occurs as a result of a lack or severe reduction of blood supply to the brain. Presently therapeutic interventions are limited and there is a need to develop new and efficacious stroke treatments. To this end, a great deal of research effort has been devoted to studying the potential molecular mechanisms involved in ischemic neuronal death. Correlative evidence demonstrated a paradoxical activation of the cell cycle machinery in ischemic neurons. The levels and activity of key cell cycle regulators including cyclin D1, Cdk2 and Cdk4 are upregulated following ischemic insults. However, the functional relevance of these various signals following ischemic injury was unclear. Accordingly, the research described in this thesis address the functional relevance of the activation of the cell cycle machinery in ischemic neuronal death. The data indicate that the inhibition of Cdk4 protects neurons from ischemia-induced delayed death, whereas abrogation of Cdk5 activity prevents excitotoxicity-induced damage in vitro and in vivo. Examination of upstream activators of mitotic-Cdks showed that Cdc25A is a critical mediator of delayed ischemic neuronal death. Investigation of the potential molecular mechanism by which cell cycle regulators induced neuronal death revealed perturbations in the levels and activity of key downstream targets of Cdk4. The retinoblastoma protein family members, pRb and p130 are increasingly phosphorylated following ischemic stresses. Importantly, p130 and E2F4 proteins are drastically reduced following ischemic insults. Additionally, E2F1 association with promoters of pro-apoptotic genes are induced while that of E2F4 is reduced. These changes appear to be important determinants in ischemic neuronal death. Cumulatively, the data supports the activation of the cell cycle machinery as a pathogenic signal contributing to ischemic neuronal death. The development of neuroprotectant strategies for stroke has been hampered in part by its complex pathophysiology. Previous research indicated that flavopiridol, a general CDK-inhibitor, is unable to provide sustained neuroprotection beyond one week following cerebral ischemia. The potential benefit of combining flavopiridol with another neuroprotectant, minocycline, was explored. The data indicate that while this approach provided histological protection 10 weeks after insult, the protected neurons are not functional due to progressive dendritic degeneration. This evidence indicates that targeting cell cycle pathways in stroke while important must be combined with other therapeutic modalities to fully treat stroke-induced damage.