Les troubles des conduites alimentaires (TCA) et les troubles des conduites suicidaires (TCS) sont des pathologies mentales graves. Ces deux troubles représentent un problème de santé publique majeur. En effet, il s’agit de pathologies pour lesquelles les pronostics sont médiocres, ayant un impact lourd sur la vie des patients et de leur entourage et représentant un coût élevé pour la société. Il n’existe actuellement aucun biomarqueur validé dans l’une ou l’autre de ces pathologies. L’objectif de ce travail est de réussir à identifier de potentiels biomarqueurs dans ces deux troubles. L’identification de biomarqueurs potentiels pourrait être utile dans divers domaines : faciliter le diagnostic des patients, améliorer la prise en charge des patients et enfin servir de cibles thérapeutiques dans le but d’élaborer de nouveaux traitements plus efficaces. Nous présenterons dans ce travail l’intérêt et l’utilisation des biomarqueurs en psychiatrie puis deux axes d’exploration des TCA ainsi qu’un axe d’exploration des TCS: l’approche neuropsychologique ainsi que l’approche biologique pour les TCA et l’approche génétique pour les TCS. Concernant les TCA, nos résultats suggèrent un intérêt potentiel des oestro-progestatifs dans l’amélioration des fonctions cognitives des patientes. Au sujet des TCS, il semblerait que deux des polymorphismes génétiques que nous avons étudiés (l’un du gène MOR et l’autre du gène SKA2) soient associés avec le risque d’émergence d’idées suicidaires au cours de l’initiation d’un traitement antidépresseur. Si ces études étaient répliquées, ces divers éléments pourraient s’avérer être des outils diagnostics et thérapeutiques. / Eating Disorders (EDs) and Suicidal Behavior (SB) are severe mental illnesses. Those two disorders represent a major public health problem. Indeed, those pathologies present a poor prognosis, induce a heavy impact on patient and their families and represant an important cost for society. There is currently no validated biomarker in one or the other disorder. The goal of this work is to successfully identify potential biomarkers in both disorders. The identification of potential biomarkers could be useful in various areas: facilitating patient’s diagnosis, improving patient management, and serving as therapeutic targets for the development of new and more effective treatments. We will present in this work the interest and the use of biomarkers in psychiatry then two axes of exploration of EDs and finally an axis of exploration of SB: the neuropsychological approach as well as the biological approach for EDs and genetic approach for SB. Regarding TCA, our results suggest a potential interest of estrogen-progestins in improving the cognitive functions of patients. With regards to SB, it appears that two of the genetic polymorphisms we studied (one from the MOR gene and the other from the SKA2 gene) are associated with the risk of suicidal ideation during the initiation of suicide an antidepressant treatment. If those studies are replicated, those elements could peep out diagnostic and therapeutic tools.

Dans la pratique courante, la posologie des anticancéreux est le plus souvent normalisée parrapport à la surface corporelle du patient (parfois son poids). Cependant, cette adaptation de dose neprend pas en compte la variabilité interindividuelle pharmacologique du médicament. Les facteursmodulant cette dernière sont nombreux, dont notamment les capacités individuelles de métabolisme etd’élimination. A dose égale, l’exposition systémique du patient peut donc varier, et entraîner un risquede surdosage pour certains patients, avec la survenue de toxicités, ou de sous-dosage pour d’autres,associée à une limitation de l’efficacité thérapeutique. Des alternatives à l’utilisation de la surfacecorporelle ont donc été proposées afin de prendre en compte de façon plus importante la variabilitéinterindividuelle des patients, tel que le suivi thérapeutique pharmacologique ou lapharmacogénétique. Les polymorphismes des gènes codants pour des enzymes impliquées dans lemétabolisme ou le transport des cytotoxiques peuvent modifier leur activité et donc influer sur lapharmacocinétique d’un médicament donné. Le développement des connaissances dans ces domainespourrait permettre la mise en place, de façon progressive, de nouvelles méthodes d’ajustement de laposologie.Ce projet de thèse porte sur l’individualisation de la posologie des anticancéreux ens’intéressant aussi bien aux chimiothérapies cytotoxiques, avec pour modèle l’irinotecan, qu’auxthérapies ciblées, en distinguant dans cette même classe les inhibiteurs de tyrosine kinase et lesanticorps monoclonaux (modèle du bévacizumab). Afin d’optimiser l’index thérapeutique de cestraitements, l’adaptation de la posologie a été abordée selon le profil pharmacogénétique maiségalement selon le profil pharmacocinétique.Une étude clinique a été mise en place afin d’évaluer, en terme de toxicité et d’efficacité,l’intérêt d’une adaptation de posologie de l’irinotécan dans le cadre d’une polychimiothérapie dans letraitement du cancer colorectal (FOLFIRI +/- bévacizumab ou cétuximab), en se basant sur unpolymorphisme génétique de l’UGT1A1, enzyme impliquée dans son métabolisme. Les résultats del’étude intermédiaire de cet essai de phase II, mené chez 16 patients, sont en faveur de la poursuite del’étude avec 12,5% de toxicités hématologiques de grade 4 et un taux de réponse objective de 58,3 %.L’objectif secondaire a été l’étude de la pharmacocinétique de l’irinotécan et de ses métabolites lorsd’une des cures de chimiothérapie. Ces dosages ont été réalisés par chromatographie liquide couplée àla spectrométrie de masse. Cette technique, largement utilisée pour la quantification plasmatique desmédicaments, a été développée et validée selon les normes en vigueur. Le dosage du bévacizumab, unanticorps monoclonal associé au FOLFIRI dans le traitement du cancer colorectal, a lui aussi étédéveloppé par cette technique et a fait l’objet d’un travail de mise au point avec l’identification despeptides spécifiques, ainsi que l’optimisation des étapes de prétraitement de l’échantillon et de latechnique analytique.En parallèle, la mise en place du dosage des inhibiteurs de tyrosine kinase par spectrométrie demasse, ainsi que sa validation, ont fait partie de ce travail de thèse, et autorisent son utilisation enroutine hospitalière dans le cadre du suivi thérapeutique pharmacologique.Mots-clés : suivi thérapeutique pharmacologique, pharmacogénétique, irinotécan, chromatographie / Usually, the dosage of anticancer is normalized with used to the patient's body surface area(sometimes its weight). However, this adaptation does not take into account the interindividualpharmacological variability in the pharmacokinetics of the drug. Factors modulating thispharmacokinetics are numerous, including interindividual capacities of metabolism and elimination.At the same dosage, the systemic exposure of the patient can thus be changed, and cause a risk ofoverdosing for some patients, with the development of toxicities, or under dosing for others, withlimiting the therapeutic efficacy. Alternatives to the use of the body surface area have been proposedin order to more take account of inter-individual variability of patients, such as therapeutic drugmonitoring or pharmacogenetics. Genetic polymorphisms of enzymes involved in the metabolism ortransport of a cytotoxic, can change and modify their activity, and impact on drug’s metabolism. Thedevelopment of knowledge in these areas could allow the establishment of new methods to dosage’sadjustment.This PhD project focuses on individualizing dosage of anticancer focusing in both cytotoxicchemotherapies with the model of irinotecan, as targeted therapies, distinguishing in this same classtyrosine kinase inhibitors and monoclonal antibodies (bevacizumab model). To optimize thetherapeutic index of these therapies, the dose adjustment was according to pharmacogenetics profile,but also according to the pharmacokinetics profile.A clinical study was developed to studying, in terms of toxicity and efficacy, the interest ofadapting dosage of irinotecan in combination of chemotherapy to traitement of metastatic colorectalcancer (FOLFIRI +/- bevacizumab or cetuximab), based on a genetic polymorphism of UGT1A1, anenzyme involved in its metabolism. Results of the interim study of this phase II trial, conducted in 16patients, are in favor of study continuation with 12.5% grade 4 hematological toxicities, and objectiveresponse rate of 58.3%. The secondary objective was to study the pharmacokinetics of irinotecan andits metabolites during one cycle. These assays were performed by liquid chromatography-massspectrometry. This technique, widely used for quantification of drug in plasma was developed andvalidated according to standards. The dosage of bevacizumab, a monoclonal antibody used tocolorectal cancer treatment (associated to FOLFIRI) was also developed by this technique, withspecific peptide identification, and optimization of sample preparation and analytical technique.In parallel, the implementation of the determination of tyrosine kinase inhibitors by massspectrometry and its validation were part of this PhD project, and allow its use in hospital routine intherapeutic drug monitoring.

No description available.

No description available.

Dislipidemia é uma desordem multifatorial causada pela interação entre fatores ambientais e genéticos. A identificação dos componentes genéticos responsáveis por essas características tem sido intensamente investigada nos últimos anos. Esses estudos têm enfocado principalmente polimorfismos nos genes que codificam proteínas estruturais e enzimas relacionadas ao metabolismo dos lipídios. Sendo assim, este trabalho teve como objetivos avaliar as freqüências dos polimorfismos EcoRI, TaqI, S447X, -250G>A dos genes APOB, CETP, LPL e LIPC e investigar a interação desses polimorfismos com dados clínicos, bioquímicos e antropométricos em 119 pacientes dislipidêmicos, de uma amostra da população de Porto Alegre. Para analisar os genótipos de cada polimorfismo e sua possível influência sobre a eficácia ao tratamento com estatinas foram analisados 48 pacientes dislipidêmicos. As medidas dos níveis lipídicos foram verificadas ao longo do estudo. Utilizou-se a técnica de PCRRFLP para a realização das análises de biologia molecular. O polimorfismo -250G>A foi associado significativamente com os níveis de HDL-C. Para análises não ajustadas, os portadores do alelo G aumentaram mais o nível de HDL-C (P=0,004) que os indivíduos homozigotos AA, considerando que o genótipo AA apresentou níveis de TG mais elevados (P=0,017) que os indivíduos com genótipo GG ou GA. Para níveis de TG esses resultados mantiveram-se para análises ajustadas, com elevados níveis de TG para indivíduos com genótipo AA em comparação aos indivíduos com genótipo GG ou GA (P=0,073). Diferenças entre os genótipos no percentual de variação nos níveis lipídicos foram observadas para os polimorfismos LIPC e LPL. Depois de ajustadas por covariáveis, os indivíduos com genótipo GA ou AA apresentaram uma redução maior do nível de CT, comparados com os indivíduos com o genótipo GG (-26,4% ± 15,5 vs. -18,2% ± 11,8, P=0,034). Para análises não ajustadas, os indivíduos com o alelo G do polimorfismo LPL S447X mostraram um aumento dos níveis de HDL-C comparados com os indivíduos com o genótipo CC (13,8% VS. 3,3%, P=0,047). Depois de ajustadas por covariáveis, a significância do efeito desse polimorfismo foi observada para o nível de CT. O percentual da média de redução no nível de CT foi maior nos indivíduos homozigotos CC que os indivíduos com o alelo G (-26,6% ± 13,6 vs -20,5% ±13,6, P=0,046). Nossos dados sugerem uma associação do polimorfismo LIPC -250G>A com níveis de HDL-C e TG, para análises não ajustadas, mas para análises não ajustadas por covariáveis a associação manteve-se para níveis de TG. Nós também encontramos associação significante dos polimorfismos LIPC -250G>A e LPL S447X na resposta ao tratamento com estatinas. Esses resultados podem ser explicados através de vários fatores, tais como: gênero, estrogênio, IMC e outras variáveis que possam interferir no efeito dos polimorfismos sobre os níveis lipídicos e resposta ao tratamento. / Dyslipidemia is a multifactorial disorder caused by an interaction between genetic and environmental factors. The identification of the genetic component of this traits have been intensively investigated in the last year. These studies focused mainly on polymorphism in genes coding for structural proteins and enzymes related to lipid metabolism. Therefore, in the present study, we investigated the frequencies of the polymorphisms EcoRI, TaqIB, S447X and (-250G>A) of the APOB, CETP, LPL and LIPC genes with clinical, biochemical, anthropometrics data of the one hundred and nineteen patients with dyslipidemia in a sample of Southern Brazilian population. To determine the genotype association with response to statin treatment, only 48 individuals were enrolled for analysis. Plasma lipids and lipoproteins were measured before and throughout the study. PCR-RFLP method was used for molecular biology analysis. Plasma lipids and lipoproteins were measured before and throughout the study. Baseline lipid and lipoprotein parameters were compared among APOB EcoRI, CETP TaqIB, LPL S447X (G>C) and LIPC -250G>A genotypes after genotyping by PCR and restriction mapping. Data from forty-eight patients with statin treatment were used to pharmacogenetic statistical analyses. The LIPC -250G>A polymorphism was significantly associated with HDL-C. For unadjusted levels, carriers of the G allele had higher HDL-C concentrations (P=0.004) than AA homozygotes, whereas AA genotype had higher TG concentrations (P=0.073) than GG and GA genotypes. For TG levels the same results were observed for adjusted data, with higher TG concentrations in AA homozygotes than GG and GA genotypes (P=0.017). Differences among genotypes in the percentage variation in lipid and lipoprotein concentrations for LIPC and LPL polymorphism were observed. After adjustment for covariates, GA and AA carriers genotypes showed a greater reduction in total cholesterol compared than GG genotype (-26.4% ± 15.5 vs. -18.2% ± 11.8, P=0.034). For unadjusted data, G allele carriers for LPL S447X gene polymorphism showed a greater HDLcholesterol increase compared to CC subjects (13.8% vs. 3.3%, P = 0.047). After adjustment for covariates, a significant effect of this polymorphism was observed for change in TC levels. The mean percent reduction in TC was greater in CC homozygotes than in G carriers (-26.6% ± 13.6 vs. -20.5% ± 13.6, P=0.046). Our data suggest an association of LIPC -250G>A gene polymorphism with HDL-C and TG concentrations for unadjusted data, but not after adjustment for covariates. For TG concentrations the associations was maintained after adjustment. We also found a significant effect dependent of covariates of LIPC and LPL polymorphisms on statin treatment response. These results can be explained on the basis of there being several factors such as gender, estrogens, BMI and other variables that can modulate the effect of gene polymorphisms on the lipid in lipoprotein concentration and treatment response.

Dislipidemia é uma desordem multifatorial causada pela interação entre fatores ambientais e genéticos. A identificação dos componentes genéticos responsáveis por essas características tem sido intensamente investigada nos últimos anos. Esses estudos têm enfocado principalmente polimorfismos nos genes que codificam proteínas estruturais e enzimas relacionadas ao metabolismo dos lipídios. Sendo assim, este trabalho teve como objetivos avaliar as freqüências dos polimorfismos EcoRI, TaqI, S447X, -250G>A dos genes APOB, CETP, LPL e LIPC e investigar a interação desses polimorfismos com dados clínicos, bioquímicos e antropométricos em 119 pacientes dislipidêmicos, de uma amostra da população de Porto Alegre. Para analisar os genótipos de cada polimorfismo e sua possível influência sobre a eficácia ao tratamento com estatinas foram analisados 48 pacientes dislipidêmicos. As medidas dos níveis lipídicos foram verificadas ao longo do estudo. Utilizou-se a técnica de PCRRFLP para a realização das análises de biologia molecular. O polimorfismo -250G>A foi associado significativamente com os níveis de HDL-C. Para análises não ajustadas, os portadores do alelo G aumentaram mais o nível de HDL-C (P=0,004) que os indivíduos homozigotos AA, considerando que o genótipo AA apresentou níveis de TG mais elevados (P=0,017) que os indivíduos com genótipo GG ou GA. Para níveis de TG esses resultados mantiveram-se para análises ajustadas, com elevados níveis de TG para indivíduos com genótipo AA em comparação aos indivíduos com genótipo GG ou GA (P=0,073). Diferenças entre os genótipos no percentual de variação nos níveis lipídicos foram observadas para os polimorfismos LIPC e LPL. Depois de ajustadas por covariáveis, os indivíduos com genótipo GA ou AA apresentaram uma redução maior do nível de CT, comparados com os indivíduos com o genótipo GG (-26,4% ± 15,5 vs. -18,2% ± 11,8, P=0,034). Para análises não ajustadas, os indivíduos com o alelo G do polimorfismo LPL S447X mostraram um aumento dos níveis de HDL-C comparados com os indivíduos com o genótipo CC (13,8% VS. 3,3%, P=0,047). Depois de ajustadas por covariáveis, a significância do efeito desse polimorfismo foi observada para o nível de CT. O percentual da média de redução no nível de CT foi maior nos indivíduos homozigotos CC que os indivíduos com o alelo G (-26,6% ± 13,6 vs -20,5% ±13,6, P=0,046). Nossos dados sugerem uma associação do polimorfismo LIPC -250G>A com níveis de HDL-C e TG, para análises não ajustadas, mas para análises não ajustadas por covariáveis a associação manteve-se para níveis de TG. Nós também encontramos associação significante dos polimorfismos LIPC -250G>A e LPL S447X na resposta ao tratamento com estatinas. Esses resultados podem ser explicados através de vários fatores, tais como: gênero, estrogênio, IMC e outras variáveis que possam interferir no efeito dos polimorfismos sobre os níveis lipídicos e resposta ao tratamento. / Dyslipidemia is a multifactorial disorder caused by an interaction between genetic and environmental factors. The identification of the genetic component of this traits have been intensively investigated in the last year. These studies focused mainly on polymorphism in genes coding for structural proteins and enzymes related to lipid metabolism. Therefore, in the present study, we investigated the frequencies of the polymorphisms EcoRI, TaqIB, S447X and (-250G>A) of the APOB, CETP, LPL and LIPC genes with clinical, biochemical, anthropometrics data of the one hundred and nineteen patients with dyslipidemia in a sample of Southern Brazilian population. To determine the genotype association with response to statin treatment, only 48 individuals were enrolled for analysis. Plasma lipids and lipoproteins were measured before and throughout the study. PCR-RFLP method was used for molecular biology analysis. Plasma lipids and lipoproteins were measured before and throughout the study. Baseline lipid and lipoprotein parameters were compared among APOB EcoRI, CETP TaqIB, LPL S447X (G>C) and LIPC -250G>A genotypes after genotyping by PCR and restriction mapping. Data from forty-eight patients with statin treatment were used to pharmacogenetic statistical analyses. The LIPC -250G>A polymorphism was significantly associated with HDL-C. For unadjusted levels, carriers of the G allele had higher HDL-C concentrations (P=0.004) than AA homozygotes, whereas AA genotype had higher TG concentrations (P=0.073) than GG and GA genotypes. For TG levels the same results were observed for adjusted data, with higher TG concentrations in AA homozygotes than GG and GA genotypes (P=0.017). Differences among genotypes in the percentage variation in lipid and lipoprotein concentrations for LIPC and LPL polymorphism were observed. After adjustment for covariates, GA and AA carriers genotypes showed a greater reduction in total cholesterol compared than GG genotype (-26.4% ± 15.5 vs. -18.2% ± 11.8, P=0.034). For unadjusted data, G allele carriers for LPL S447X gene polymorphism showed a greater HDLcholesterol increase compared to CC subjects (13.8% vs. 3.3%, P = 0.047). After adjustment for covariates, a significant effect of this polymorphism was observed for change in TC levels. The mean percent reduction in TC was greater in CC homozygotes than in G carriers (-26.6% ± 13.6 vs. -20.5% ± 13.6, P=0.046). Our data suggest an association of LIPC -250G>A gene polymorphism with HDL-C and TG concentrations for unadjusted data, but not after adjustment for covariates. For TG concentrations the associations was maintained after adjustment. We also found a significant effect dependent of covariates of LIPC and LPL polymorphisms on statin treatment response. These results can be explained on the basis of there being several factors such as gender, estrogens, BMI and other variables that can modulate the effect of gene polymorphisms on the lipid in lipoprotein concentration and treatment response.

Dislipidemia é uma desordem multifatorial causada pela interação entre fatores ambientais e genéticos. A identificação dos componentes genéticos responsáveis por essas características tem sido intensamente investigada nos últimos anos. Esses estudos têm enfocado principalmente polimorfismos nos genes que codificam proteínas estruturais e enzimas relacionadas ao metabolismo dos lipídios. Sendo assim, este trabalho teve como objetivos avaliar as freqüências dos polimorfismos EcoRI, TaqI, S447X, -250G>A dos genes APOB, CETP, LPL e LIPC e investigar a interação desses polimorfismos com dados clínicos, bioquímicos e antropométricos em 119 pacientes dislipidêmicos, de uma amostra da população de Porto Alegre. Para analisar os genótipos de cada polimorfismo e sua possível influência sobre a eficácia ao tratamento com estatinas foram analisados 48 pacientes dislipidêmicos. As medidas dos níveis lipídicos foram verificadas ao longo do estudo. Utilizou-se a técnica de PCRRFLP para a realização das análises de biologia molecular. O polimorfismo -250G>A foi associado significativamente com os níveis de HDL-C. Para análises não ajustadas, os portadores do alelo G aumentaram mais o nível de HDL-C (P=0,004) que os indivíduos homozigotos AA, considerando que o genótipo AA apresentou níveis de TG mais elevados (P=0,017) que os indivíduos com genótipo GG ou GA. Para níveis de TG esses resultados mantiveram-se para análises ajustadas, com elevados níveis de TG para indivíduos com genótipo AA em comparação aos indivíduos com genótipo GG ou GA (P=0,073). Diferenças entre os genótipos no percentual de variação nos níveis lipídicos foram observadas para os polimorfismos LIPC e LPL. Depois de ajustadas por covariáveis, os indivíduos com genótipo GA ou AA apresentaram uma redução maior do nível de CT, comparados com os indivíduos com o genótipo GG (-26,4% ± 15,5 vs. -18,2% ± 11,8, P=0,034). Para análises não ajustadas, os indivíduos com o alelo G do polimorfismo LPL S447X mostraram um aumento dos níveis de HDL-C comparados com os indivíduos com o genótipo CC (13,8% VS. 3,3%, P=0,047). Depois de ajustadas por covariáveis, a significância do efeito desse polimorfismo foi observada para o nível de CT. O percentual da média de redução no nível de CT foi maior nos indivíduos homozigotos CC que os indivíduos com o alelo G (-26,6% ± 13,6 vs -20,5% ±13,6, P=0,046). Nossos dados sugerem uma associação do polimorfismo LIPC -250G>A com níveis de HDL-C e TG, para análises não ajustadas, mas para análises não ajustadas por covariáveis a associação manteve-se para níveis de TG. Nós também encontramos associação significante dos polimorfismos LIPC -250G>A e LPL S447X na resposta ao tratamento com estatinas. Esses resultados podem ser explicados através de vários fatores, tais como: gênero, estrogênio, IMC e outras variáveis que possam interferir no efeito dos polimorfismos sobre os níveis lipídicos e resposta ao tratamento. / Dyslipidemia is a multifactorial disorder caused by an interaction between genetic and environmental factors. The identification of the genetic component of this traits have been intensively investigated in the last year. These studies focused mainly on polymorphism in genes coding for structural proteins and enzymes related to lipid metabolism. Therefore, in the present study, we investigated the frequencies of the polymorphisms EcoRI, TaqIB, S447X and (-250G>A) of the APOB, CETP, LPL and LIPC genes with clinical, biochemical, anthropometrics data of the one hundred and nineteen patients with dyslipidemia in a sample of Southern Brazilian population. To determine the genotype association with response to statin treatment, only 48 individuals were enrolled for analysis. Plasma lipids and lipoproteins were measured before and throughout the study. PCR-RFLP method was used for molecular biology analysis. Plasma lipids and lipoproteins were measured before and throughout the study. Baseline lipid and lipoprotein parameters were compared among APOB EcoRI, CETP TaqIB, LPL S447X (G>C) and LIPC -250G>A genotypes after genotyping by PCR and restriction mapping. Data from forty-eight patients with statin treatment were used to pharmacogenetic statistical analyses. The LIPC -250G>A polymorphism was significantly associated with HDL-C. For unadjusted levels, carriers of the G allele had higher HDL-C concentrations (P=0.004) than AA homozygotes, whereas AA genotype had higher TG concentrations (P=0.073) than GG and GA genotypes. For TG levels the same results were observed for adjusted data, with higher TG concentrations in AA homozygotes than GG and GA genotypes (P=0.017). Differences among genotypes in the percentage variation in lipid and lipoprotein concentrations for LIPC and LPL polymorphism were observed. After adjustment for covariates, GA and AA carriers genotypes showed a greater reduction in total cholesterol compared than GG genotype (-26.4% ± 15.5 vs. -18.2% ± 11.8, P=0.034). For unadjusted data, G allele carriers for LPL S447X gene polymorphism showed a greater HDLcholesterol increase compared to CC subjects (13.8% vs. 3.3%, P = 0.047). After adjustment for covariates, a significant effect of this polymorphism was observed for change in TC levels. The mean percent reduction in TC was greater in CC homozygotes than in G carriers (-26.6% ± 13.6 vs. -20.5% ± 13.6, P=0.046). Our data suggest an association of LIPC -250G>A gene polymorphism with HDL-C and TG concentrations for unadjusted data, but not after adjustment for covariates. For TG concentrations the associations was maintained after adjustment. We also found a significant effect dependent of covariates of LIPC and LPL polymorphisms on statin treatment response. These results can be explained on the basis of there being several factors such as gender, estrogens, BMI and other variables that can modulate the effect of gene polymorphisms on the lipid in lipoprotein concentration and treatment response.

Ce travail de thèse reposait sur l’hypothèse que la variabilité génétique des protéines « cibles » des médicaments immunosuppresseurs de la famille des inhibiteurs de la calcineurine (ICN ; ciclosporine et tacrolimus) pourrait expliquer une partie de la variabilité observée dans leur efficacité et toxicité. Une revue de la littérature nous a permis de lister un panel de variants génétiques au sein de la voie de la calcineurine, considérés comme étant de bons candidats pour des études en transplantation. Ces variants n’ont pas été associés au risque de rejet aigu ou d’infection grave dans une étude incluant 381 patients transplantés rénaux suivis durant un an après la transplantation. La variabilité pharmacodynamique des ICN a ensuite été explorée au travers des régulations épigénétiques. Une analyse de la méthylation de l’ADN après exposition médicamenteuse a été menée sur deux modèles. Premièrement, la lignée cellulaire JURKAT a été utilisée pour développer la méthode d’immunoprécipitation de l’ADN méthylé (MeDIP). Chez des souris traitées par ciclosporine et tacrolimus durant 3 mois, nous avons ensuite isolé les cellules cibles des médicaments, les lymphocytes T CD4 puis, après immunoprécipitation de l’ADN méthylé et analyse par séquençage pangénomique haut débit (MeDIP-seq, séquençeur Ion Proton), nous avons recherché les régions du génome présentant des différences de méthylation induites par le traitement. L’analyse différentielle bio-informatique a été menée à l’aide des outils SAMtools (Li et col., 2009), BEDtools (Quinlan and Hall, 2010), MACS2 (Zhang et col., 2008) et Diffbind (Stark and Brown, 2011 - Bioconductor). Sur l’ensemble du génome, nous n’avons identifié que 24 régions présentant un niveau de méthylation modifié par l’exposition au tacrolimus. Le promoteur du gène Calm2, codant pour l’isoforme 2 de la calmoduline, semble être davantage méthylé chez les souris traitées. Ces résultats préliminaires semblent prometteurs pour la découverte de biomarqueurs épigénétiques de la réponse thérapeutique aux immunosuppresseurs. / Inter-individual genetic variation might account for diverse efficacy and toxicity of calcineurin inhibitors (cyclosporin and tacrolimus). In particular, some variants located within genes coding for proteins of the calcineurin pathway can explain part of this variability. In this manuscript, a panel of candidate genes was selected based on bibliographic review and tested in a pharmacogenetics study encompassing 381 renal transplants followed for one year after surgery. None of these candidates was associated with the acute rejection or serious infection risks. Furthermore, the pharmacodynamic variability of these drugs was also investigated, exploring the use of epigenomics profiling as proximal readout of the calcineurin inhibition treatment. In particular, we investigated the impact of drug exposure on DNA methylation in two experimental models. Methylated DNA immunoprecipitation followed by high-throughput sequencing (MeDIP-seq, Ion Proton technology) was deployed in JURKAT cell line, used as in vitro model, and in CD4 T lymphocytes isolated from mice treated with either cyclosporin or tacrolimus for three months. After sequencing, the differentiated methylated regions caused by drug exposure were analyzed. Bioinformatics analyses were performed using SAMtools (Li et al., 2009), BEDtools (Quinlan and Hall, 2010), MACS2 (Zhang et al., 2008) and Diffbind (Stark and Brown, 2011 - Bioconductor). Overall, the genome-wide analysis revealed only 24 regions with a differentiated enrichment in DNA methylation after three month-tacrolimus treatment, indicating a targeted effect of these treatments on a subset of key genes. Of note, CALM2 promoter, coding for the calmodulin isoform 2 protein, showed significant hypermethylation in tacrolimus-treated mice. These preliminary results corroborate the interest in using DNA methylation as promising approach to identify candidate biomarkers for therapeutic drug monitoring in calcineurin inhibitor treatments.

INTRODUCTION Des réponses thérapeutiques variables aux glucocorticoïdes (GCs) sont observées parmi les patients atteints de la leucémie lymphoblastique aiguë (LLA). Les protéines Bax et Bim ont déjà montré un rôle important dans l’apoptose des cellules leucémiques. L’expression de Bax était plus basse chez les patients leucémiques résistants au médicament, de même une sensibilité diminuée aux GCs a été associée avec une expression réduite de Bim. La différence dans l’expression pourrait être due à des polymorphismes présents dans ces gènes et donc être associés avec la résistance aux GCs. MÉTHODE Dix-huit polymorphismes en régions régulatrices, 2 polymorphismes exoniques et 7 polymorphismes en région 3’UTR de ces gènes ont été analysés chez les témoins (n=50) et ont permis de déterminer un nombre minimal de polymorphismes suffisants pour définir les haplotypes (tagSNPs). Ces 8 polymorphismes ont ensuite été génotypés chez 286 enfants atteints de la LLA et ont été testés pour l’issue de la maladie par l’analyse de survie. RÉSULTATS Une survie sans évènement et une survie sans rechute diminuées ont été observées pour l’haplotype 3 (p=0,03 et p=0,02). Une survie globale diminuée a été associée avec l’homozygotie pour l’allèle exonique T298C>T (p=0,03), de même que pour les haplotypes 1 et 4 (p=0,04 et p=0,02) du gène Bim. CONCLUSION Les polymorphismes ont été associés avec une survie diminuée chez des enfants atteints de LLA. Il reste à tester d’autres polymorphismes présents dans ces deux gènes ainsi qu’à définir leurs fonctions afin de comprendre leurs rôles dans la réponse aux GCs. / INTRODUCTION Variable therapeutic responses to glucocorticoids (GCs) are observed for acute lymphoblastic leukemia (ALL) patients. Proteins Bax and Bim have already shown to play a major role in mediating GC-induced apoptosis in leukemia cells. Bax expression was lower in drug-resistant leukemia samples; likewise lower sensitivity to GC was associated with reduced Bim expression. The difference in the expression can be due to polymorphisms in these genes and therefore associated to GC resistance. METHOD Eighteen polymorphisms in the regulatory region, two exonic polymorphisms and seven polymorphisms in 3’UTR of these genes were analysed in controls (n=50) and have permitted to determine a minimal number of polymorphisms sufficient to define haplotypes (tagSNPs). These 8 single nucleotide polymorphisms (SNPs) were then genotyped in 286 LLA children and were tested for disease outcome by survival analysis. RESULTS A diminished event free survival and a diminished relapse free survival were observed for haplotype 3 (p=0,03 and p=0,02). A diminished overall survival was associated with the exonic T298C>T allele (p=0,03) and with haplotypes 1 and 4 (p=004 and p=0,02) of Bim gene. CONCLUSION Bax and Bim were associated with a diminished survival in LLA children. We still have to test other polymorphisms located in these genes and to define their functions in order to understand their roles in GC response.

A varfarina é o anticoagulante oral mais prescrito no mundo todo. Algoritmos farmacogenéticos têm sido desenvolvidos para estimar a dose de varfarina. Os principais objetivos deste estudo foram desenvolver um algoritmo farmacogenético estimador de dose de varfarina e comparar o algoritmo desenvolvido neste trabalho com algoritmos disponíveis na literatura. Para atingir os objetivos foram incluídos dois grupos de pacientes tratados com varfarina (primeira coorte, n = 832; e segunda coorte, n = 133). Foram realizadas as genotipagens dos polimorfismos CYP2C9*2, CYP2C9*3 e VKORC1 (c.G1639A). A derivação do algoritmo foi realizada utilizando os dados dos pacientes da primeira coorte com dose estável (n=368) e foi replicado utilizando os dados dos pacientes provenientes da segunda coorte (n=133). Como resultado o algoritmo desenvolvido neste trabalho alcançou um coeficiente de determinação de 40%, incluindo as variáveis: idade, sexo, peso, altura, raça autodeclarada, uso de amiodarona, uso de indutores enzimáticos, os genótipos na VKORC1 (c.G1639A) e os fenótipos de acordo com polimorfismos CYP2C9. Os dados sugerem que o nosso algoritmo desenvolvido é mais acurado do que o algoritmo IWPC (The International Warfarin Pharmacogenetics Consortium) quando a aplicação é focada em pacientes brasileiros. Os algoritmos farmacogenéticos estimadores de dose de varfarina desenvolvidos para uma população específica podem ser mais efetivos para a terapia com varfarina em comparação com o uso de algoritmos estimadores de dose atualmente disponíveis / Warfarin is the most prescribed oral anticoagulant in the world. Pharmacogenetic algorithms have been developed to estimate the dose of warfarin. The main aims of the present study were to develop a pharmacogenetic-based warfarin dosing algorithm and compare algorithm developed in this study with others algorithms available in the literature. We included two patient cohorts treated with warfarin (first cohort, n = 832; and second cohort, n = 133). Polymorphisms were genotyped in the CYP2C9 * 2, CYP2C9 * 3 and VKORC1 (c.G1639A). The derivation of the algorithm was performed using the data from the patients in the first cohort with a stable dose (n = 368) and was replicated using data from patients from the second cohort (n = 133). Our algorithm achieved a determination of coefficient of 40%, including variables: age, sex, weight, height, self-declared race, use of amiodarone, use of enzymatic inducers, genotypes in VKORC1 (c.G1639A) and phenotypes according to CYP2C9 polymorphisms. Our data suggest that the developed algorithm is more accurate than the IWPC algorithm when the application is focused on patients from the Brazilian population. Pharmacogenetics- based warfarin dosing algorithms developed for a specific population may lead to improved performance compared use dosing algorithms currently available