De nombreux travaux ont démontré la modulation, par les tumeurs, de certaines fonctions des cellules du système immunitaire. Dans le cadre de notre travail, nous avons étudié les lymphocytes T CD4+, cellules clefs de la réponse immune spécifique, chez des patientes atteintes d’un cancer du sein.<p>Sur base de l’établissement des profils d’expression génique des lymphocytes T infiltrant les tumeurs, nous avons dérivé la « tumor-infiltrating CD4+ signature » (TICD4S) composée de 61 gènes immuns et qui reflète l’état d’activation immunitaire. Cette signature présente une valeur prédictive chez les patientes porteuses de tumeurs ERBB2-positives et ER-négative/PR-négative/ERBB2-négative: une plus forte expression de ces gènes est associée à une meilleure survie.<p>Nous avons également étudié conjointement les profils géniques établis au départ des lymphocytes T CD4+ de la tumeur, du ganglion axillaire et du sang de dix patientes. Nous avons constaté que ces profils d’expression génique des TIL CD4+ diffèrent selon le statut ER de la tumeur qu’ils infiltrent. Les lymphocytes T ganglionnaires CD4+ subissent également les effets de la masse tumorale et, tout comme les TIL, sont moins activés chez les patientes porteuses de tumeurs ER-négatives. Par contre, les lymphocytes T sanguins semblent subir dans une moindre mesure les effets de la tumeur et peu de différences ont été notées par rapport à leurs homologues isolés chez des donneuses saines.\ / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

New technologies to probe the global output of the normal and cancer genomes have recently reached widespread use. The resulting genomewide gene expression profiles, e.g, a gene expression measurement per gene and per tissue sample, remain challenging to analyze and interpret, but have already provided new insights into the pathophysiology of cancer and towards personalized care.<p><p><p>In vitro cell culture-based experimental models are used to elucidate cancer onset and progression because experimentation in humans is difficult practically and ethically unacceptable, and because they provide simplified, reproducible and controlled systems to test hypotheses. The thyroid tumors and their in vitro experimental models are particularly suited to compare the molecular phenotypes of experimental models and tumors. From one type of cell, the thyrocyte, at least five distinct benign and malignant tumors can arise. In addition, many immortalized tumor-derived cell lines and primary cultures models of these cells exist.<p><p><p>This thesis has focused on the bioinformatic comparison of these in vitro models to the in vivo tumors, from the point of view of their gene expression profiles, to gain insight into the pathogenesis of thyroid tumors, and of tumors in general.<p><p><p>In a first study, we showed that primary cultures of freshly isolated normal thyroid cells where proliferation and differentiation through the TSHR/cAMP pathway was chronically activated experimentally resemble specifically the autonomous thyroid adenomas, a type of benign thyroid tumor, and provide insight into a general mechanism of tumor progression: the suppression of negative feedbacks that normally restrain excessive cell division.<p><p><p>Subsequently, we found that immortalized thyroid tumor-derived cell lines have converged to a common phenotype regardless of their tumor subtype of origin. A TSHR/cAMP thyroid cell differentiation signature, derived from data obtained for the first study, was used to show that the cell lines were dedifferentiated. Accordingly, we showed that the cell lines resemble most the phenotype of the more dedifferentiated, clinically aggressive anaplastic thyroid cancers.<p><p><p>Finally, using large databases of gene expression profiles publicly available, we extended the comparison of cell lines and tumors to cancers of five other organs: breast, colon, kidney, ovary and lung. We discuss the correct use of these models and advance an hypothesis regarding the nature of the state to which these cells have converged: they could represent a surviving subpopulation of tumors cells, cancer stem cells, capable of initiating and maintaining tumor growth.<p><p><p>As other technologies designed to perturb the genome in experimental models are emerging, careful characterization and validation of the experimental models are needed to extrapolate the results in vivo.<p> / Doctorat en sciences biomédicales / info:eu-repo/semantics/nonPublished

Les tumeurs thyroïdiennes constituent les tumeurs endocrines les plus fréquentes. Parmi celles-ci, on distingue les adénomes, tumeurs bénignes et encapsulées, et les carcinomes, tumeurs malignes. Ceux-ci sont eux-mêmes subdivisés, principalement sur base histologique, en carcinomes papillaires ou folliculaires, qui conservent certaines caractéristiques de différenciation des cellules thyroïdiennes initiales dont ils dérivent, et qui peuvent évoluer en carcinomes anaplasiques, totalement dédifférenciés. Les carcinomes différenciés de la thyroïde sont généralement de bon pronostic, contrairement aux cancers anaplasiques qui sont nettement plus agressifs, avec un taux de survie à 5 ans inférieur à 5%. <p>La technologie des microarrays permet d’analyser simultanément l’expression de milliers de gènes dans différentes cellules et différentes conditions physiologiques, pathologiques ou toxicologiques. Au cours de cette thèse de doctorat, nous avons déterminé le profil d’expression génique des carcinomes papillaires de la thyroïde à l’aide de la technique des microarrays en utilisant une plateforme contenant plus de 8000 gènes. Douze des 26 cancers papillaires étudiés étaient issus de patients habitant la région de Tchernobyl lors de l’explosion de la centrale nucléaire de 1986 et sont considérés comme des cancers radio-induits. Les 14 tumeurs restantes proviennent de patients habitant la France. Leur étiologie n’étant pas connue, ils sont considérés comme des cancers sporadiques. <p>La réalisation de ces expériences nous a permis d’identifier des signatures moléculaires entre des sous-types de cancers papillaires. Premièrement, nous avons montré que malgré un profil d’expression génique global similaire entre les cancers papillaires sporadiques et radio-induits, une signature multigénique permet de les séparer, indiquant que des subtiles différences existent entre les deux types de tumeurs. Deux autres signatures indépendantes, l’une liée aux agents étiologiques présumés de ces tumeurs (radiation vs. H2O2), l’autre liée aux mécanismes de recombinaison homologue de l’ADN, permettent également de séparer les cancers post-Tchernobyl des cancers sporadiques. Nous avons interprété ces résultats comme une différence de susceptibilité à l’irradiation entre ces deux types de tumeurs. D’autre part, nous avons pu identifier une liste de gènes permettant de séparer les cancers papillaires à variante classique des autres sous-types de cancers papillaires. L’analyse de cette liste de gènes a permis de mettre en relation cette signature avec l’important remodelage de cette variante histologique par rapport aux autres. <p>Ces expériences ont aussi abouti à l’obtention d’une liste de gènes différentiellement exprimés entre les cancers papillaires et leur tissu normal adjacent. Une analyse minutieuse de cette liste à l’aide d’outils statistiques a permis de mieux comprendre la physiopathologie de ces tumeurs et d’aboutir à différentes conclusions :(1) un changement de population cellulaire est observé, avec une surexpression de gènes liés à la réponse immune, reflétant l’infiltration lymphocytaire de ces tumeurs par rapport au tissu normal adjacent (2) la voie de signalisation JNK est activée par surexpression de ses composants (3) la voie de signalisation de l’EGF, également par une surexpression de ses composants, complémente les altérations génétiques des cancers papillaires pour l’activation constitutive de la voie ERK1/2 (4) une sousexpression des gènes de réponse précoce est observée (5) une surexpression de nombreuses protéases, d’inhibiteurs de protéases et de protéines de la matrice extracellulaire permet d’expliquer l’important remodelage des cancers papillaires (6) le profil d’expression génique des cancers papillaires peut être corrélé avec un mode de migration collectif de ces tumeurs. <p>Finalement, dans la dernière partie de la thèse, nous avons déterminé le profil d’expression génique des cancers anaplasiques de la thyroïde et l’avons comparé à celui des cancers papillaires. Nous avons montré que les deux types de tumeurs présentent des profils moléculaires globaux distincts, reflétant leur comportement tumoral très différent. <p> / Doctorat en Sciences biomédicales et pharmaceutiques / info:eu-repo/semantics/nonPublished

Colorectal cancer (CRC) is the third leading cause of cancer related deaths in the United States. The nuclear factor κB (NF-κB) is an important family of transcription factors whose aberrant activation has been found in many types of cancer, including CRC. Therefore, understanding the regulation of NF-κB is of ultimate importance for cancer therapy. Using a novel validation-based insertional mutagenesis (VBIM) strategy, our lab has identified the novel adenomatous polyposis coli (APC) like protein (ALP) gene as a negative regulator of NF-κB. Preliminary studies from our lab demonstrated that overexpression of ALP led to decreased NF-κB activity by κB reporter assay and electrophoresis mobility gel shift assay (EMSA). The current project aims to further evaluate the role of ALP in the regulation of NF-κB signaling in CRC cells. We found that overexpression of ALP in human CRC HT29 cells greatly reduced both the number and the size of colonies that were formed in a soft agar assay. ALP overexpression also decreased the cell growth rate and cell migration ability, while shRNA mediated knockdown of ALP showed opposite effects, confirming that ALP is a tumor suppressor in CRC HT29 cells. Overexpression of ALP led to decreased NF-κB activity by κB reporter assay and condition media assay in CRC HT29 cells. Furthermore, immunohistochemical analysis with human colon vii tissues revealed that there is a gradual loss of ALP protein with tumor progression. We also found that ALP predominantly localizes in the cytoplasm, and binds to the p65 subunit of NF-κB, and might be functioning downstream of IκB kinase (IKK). In summary, in this study, we provide evidence regarding the tumor suppressor role of ALP in CRC by functioning as novel negative regulator of NF-κB. This discovery could lead to the establishment of ALP as a potential biomarker and therapeutic target in CRC.

Indiana University-Purdue University Indianapolis (IUPUI) / Cancer stem cells (CSCs) are proposed to play a major role in tumor progression, metastasis, and recurrence. The Human Epidermal growth factor Receptor 2 (HER2) gene is amplified and/or its protein product overexpressed in approximately 20% of breast cancers. HER2 overexpression is associated with increased CSCs, which may explain the aggressive phenotype and increased likelihood of recurrence for HER2+ breast cancers. Telomerase is reactivated in tumor cells, including CSCs, but has limited activity in normal tissues, providing support for the use of telomerase inhibition in anti-cancer therapy. Telomerase inhibition via an antagonistic oligonucleotide, imetelstat (GRN163L), has been shown to be effective in limiting cell growth in vitro and limiting tumor growth. Moreover, we have previously shown imetelstat can decrease metastases to the lungs, leading us to question if this is due to imetelstat targeting the CSC population. In this thesis, we investigated the effects of imetelstat on CSC and non-CSC populations of HER2+ breast cancer cell lines, as well as a triple negative breast cancer cell line, which lacks HER2 overexpression. Imetelstat inhibited telomerase activity in both CSC and non-CSC subpopulations. Moreover, imetelstat treatment alone and in combination with trastuzumab significantly reduced the CSC fraction and inhibited CSC functional ability, as shown by a significant decrease in mammosphere counts and invasive potential. Tumor growth rate was slower in combination treated mice compared to either drug alone. Additionally, there was a trend toward decreased CSC marker expression in imetelstat treated xenograft cells compared to vehicle control. The decrease in CSC marker expression we observed occurred prior to and after telomere shortening, suggesting imetelstat acts on the CSC subpopulation in telomere length dependent and independent mechanisms. Our study suggests addition of imetelstat to trastuzumab may enhance the effects of HER2 inhibition therapy.

Indiana University-Purdue University Indianapolis (IUPUI) / Although rare, classification of over 200 hereditary cancer susceptibility syndromes accounting for ~5-10% of cancer incidence has enabled the discovery and understanding of cancer predisposition genes that are also frequently mutated in sporadic cancers. The need to prevent or delay invasive cancer can partly be addressed by characterization of cells derived from healthy individuals predisposed to cancer due to inherited "single-hits" in genes in order to develop patient-derived samples as preclinical models for mechanistic in vitro studies. Here, we present microarray-based transcriptome profiling of Li-Fraumeni syndrome (LFS) patient-derived unaffected breast epithelial cells and their phenotypic characterization as in vitro three-dimensional (3D) models to test pharmacological agents. In this study, the epithelial cells derived from the unaffected breast tissue of a LFS patient were cultured and progressed from non-neoplastic to a malignant stage by successive immortalization and transformation steps followed by growth in athymic mice. These cell lines exhibited distinct transcriptomic profiles and were readily distinguishable based upon their gene expression patterns, growth characteristics in monolayer and in vitro 3D cultures. Transcriptional changes in the epithelial-to-mesenchymal transition gene signature contributed to the unique phenotypes observed in 3D culture for each cell line of the progression series; the fully transformed LFS cells exhibited invasive processes in 3D culture with disorganized morphologies due to cell-cell miscommunication, as seen in breast cancer. Bioinformatics analysis of the deregulated genes and pathways showed inherent differences between these cell lines and targets for pharmacological agents. After treatment with small molecule APR-246 that restores normal function to mutant p53, we observed that the neoplastic LFS cells had reduced malignant invasive structure formation from 73% to 9%, as well as an observance of an increase in formation of well-organized structures in 3D culture (from 27% to 91%) by stereomicroscopy and confocal microscopy. Therefore, the use of well-characterized and physiologically relevant preclinical models in conjunction with transcriptomic profiling of high-risk patient derived samples as a renewable laboratory resource can potentially guide the development of safer and more effective chemopreventive approaches.

Indiana University-Purdue University Indianapolis (IUPUI) / Reliable detection of low-frequency single nucleotide variants (SNVs) carries great significance in many applications. In cancer genetics, the frequencies of somatic variants from tumor biopsies tend to be low due to contamination with normal tissue and tumor heterogeneity. Circulating tumor DNA monitoring also faces the challenge of detecting low-frequency variants due to the small percentage of tumor DNA in blood. Moreover, in population genetics, although pooled sequencing is cost-effective compared with individual sequencing, pooling dilutes the signals of variants from any individual. Detection of low frequency variants is difficult and can be cofounded by multiple sources of errors, especially next-generation sequencing artifacts. Existing methods are limited in sensitivity and mainly focus on frequencies around 5%; most fail to consider differential, context-specific sequencing artifacts. To face this challenge, we developed a computational and experimental framework, RareVar, to reliably identify low-frequency SNVs from high-throughput sequencing data. For optimized performance, RareVar utilized a supervised learning framework to model artifacts originated from different components of a specific sequencing pipeline. This is enabled by a customized, comprehensive benchmark data enriched with known low-frequency SNVs from the sequencing pipeline of interest. Genomic-context-specific sequencing error model was trained on the benchmark data to characterize the systematic sequencing artifacts, to derive the position-specific detection limit for sensitive low-frequency SNV detection. Further, a machine-learning algorithm utilized sequencing quality features to refine SNV candidates for higher specificity. RareVar outperformed existing approaches, especially at 0.5% to 5% frequency. We further explored the influence of statistical modeling on position specific error modeling and showed zero-inflated negative binomial as the best-performed statistical distribution. When replicating analyses on an Illumina MiSeq benchmark dataset, our method seamlessly adapted to technologies with different biochemistries. RareVar enables sensitive detection of low-frequency SNVs across different sequencing platforms and will facilitate research and clinical applications such as pooled sequencing, cancer early detection, prognostic assessment, metastatic monitoring, and relapses or acquired resistance identification.

This thesis addresses the use of machine learning techniques to develop clinical diagnostic tools for breast cancer using molecular data. These tools are designed to assist physicians in their evaluation of the clinical outcome of breast cancer (referred to as prognosis).<p>The traditional approach to evaluating breast cancer prognosis is based on the assessment of clinico-pathologic factors known to be associated with breast cancer survival. These factors are used to make recommendations about whether further treatment is required after the removal of a tumor by surgery. Treatment such as chemotherapy depends on the estimation of patients' risk of relapse. Although current approaches do provide good prognostic assessment of breast cancer survival, clinicians are aware that there is still room for improvement in the accuracy of their prognostic estimations.<p>In the late nineties, new high throughput technologies such as the gene expression profiling through microarray technology emerged. Microarrays allowed scientists to analyze for the first time the expression of the whole human genome ("transcriptome"). It was hoped that the analysis of genome-wide molecular data would bring new insights into the critical, underlying biological mechanisms involved in breast cancer progression, as well as significantly improve prognostic prediction. However, the analysis of microarray data is a difficult task due to their intrinsic characteristics: (i) thousands of gene expressions are measured for only few samples; (ii) the measurements are usually "noisy"; and (iii) they are highly correlated due to gene co-expressions. Since traditional statistical methods were not adapted to these settings, machine learning methods were picked up as good candidates to overcome these difficulties. However, applying machine learning methods for microarray analysis involves numerous steps, and the results are prone to overfitting. Several authors have highlighted the major pitfalls of this process in the early publications, shedding new light on the promising but overoptimistic results. <p>Since 2002, large comparative studies have been conducted in order to identify the key characteristics of successful methods for class discovery and classification. Yet methods able to identify robust molecular signatures that can predict breast cancer prognosis have been lacking. To fill this important gap, this thesis presents an original methodology dealing specifically with the analysis of microarray and survival data in order to build prognostic models and provide an honest estimation of their performance. The approach used for signature extraction consists of a set of original methods for feature transformation, feature selection and prediction model building. A novel statistical framework is presented for performance assessment and comparison of risk prediction models.<p>In terms of applications, we show that these methods, used in combination with a priori biological knowledge of breast cancer and numerous public microarray datasets, have resulted in some important discoveries. In particular, the research presented here develops (i) a robust model for the identification of breast molecular subtypes and (ii) a new prognostic model that takes into account the molecular heterogeneity of breast cancers observed previously, in order to improve traditional clinical guidelines and state-of-the-art gene signatures./Cette thèse concerne le développement de techniques d'apprentissage (machine learning) afin de mettre au point de nouveaux outils cliniques basés sur des données moleculaires. Nous avons focalisé notre recherche sur le cancer du sein, un des cancers les plus fréquemment diagnostiqués. Ces outils sont développés dans le but d'aider les médecins dans leur évaluation du devenir clinique des patients cancéreux (cf. le pronostique).<p>Les approches traditionnelles d'évaluation du pronostique d'un patient cancéreux se base sur des critères clinico-pathologiques connus pour être prédictifs de la survie. Cette évaluation permet aux médecins de décider si un traitement est nécessaire après l'extraction de la tumeur. Bien que les outils d'évaluation traditionnels sont d'une aide importante, les cliniciens sont conscients de la nécessité d'améliorer de tels outils.<p>Dans les années 90, de nouvelles technologies à haut-débit, telles que le profilage de l'expression génique par biopuces à ADN (microarrays), ont été mises au point afin de permettre aux scientifiques d'analyser l'expression de l'entièreté du génôme de cellules cancéreuses. Ce nouveau type de données moléculaires porte l'espoir d'améliorer les outils pronostiques traditionnels et d'approfondir nos connaissances concernant la génèse du cancer du sein. Cependant ces données sont extrêmement difficiles à analyser à cause (i) de leur haute dimensionalité (plusieurs dizaines de milliers de gènes pour seulement quelques centaines d'expériences); (ii) du bruit important dans les mesures; (iii) de la collinéarité entre les mesures dûe à la co-expression des gènes.<p>Depuis 2002, des études comparatives à grande échelle ont permis d'identifier les méthodes performantes pour l'analyse de groupements et la classification de données microarray, négligeant l'analyse de survie pertinente pour le pronostique dans le cancer du sein. Pour pallier ce manque, cette thèse présente une méthodologie originale adaptée à l'analyse de données microarray et de survie afin de construire des modèles pronostiques performants et robustes. <p>En termes d'applications, nous montrons que cette méthodologie, utilisée en combinaison avec des connaissances biologiques a priori et de nombreux ensembles de données publiques, a permis d'importantes découvertes. En particulier, il résulte de la recherche presentée dans cette thèse, le développement d'un modèle robuste d'identification des sous-types moléculaires du cancer du sein et de plusieurs signatures géniques améliorant significativement l'état de l'art au niveau pronostique. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

Indiana University-Purdue University Indianapolis (IUPUI) / Mammary ductal epithelial cell growth is controlled by microenvironmental signals in serum under both normal physiological settings and during breast cancer progression. Importantly, the effects of several of these microenvironmental signals are mediated by the activities of the tumor suppressor protein kinases of the Hippo pathway. Canonically, Hippo protein kinases inhibit cellular growth through the phosphorylation and inactivation of the oncogenic transcriptional co-activator Yes-Associated Protein (YAP). This study defines an alternative mechanism whereby Hippo protein kinases induce growth arrest via the phosphorylation of the long isoform of Angiomotin (Amot130). Specifically, serum starvation is found to activate the Hippo protein kinase, Large Tumor Suppressor (LATS), which phosphorylates the adapter protein Amot130 at serine-175. Importantly, wild-type Amot130 potently inhibits mammary epithelial cell growth, unlike the Amot130 serine-175 to alanine mutant, which cannot be phosphorylated at this residue. The growth-arrested phenotype of Amot130 is likely a result of its mechanistic response to LATS signaling. Specifically, LATS activity promotes the association of Amot130 with the ubiquitin ligase Atrophin-1 Interacting Protein 4 (AIP4). As a consequence, the Amot130-AIP4 complex amplifies LATS tumor suppressive signaling by stabilizing LATS protein steady state levels via preventing AIP4-targeted degradation of LATS. Additionally, AIP4 binding to Amot130 leads to the ubiquitination and stabilization of Amot130. In turn, the Amot130-AIP4 complex signals the ubiquitination and degradation of YAP. This inhibition of YAP activity by Amot130 requires both AIP4 and the ability of Amot130 to be phosphorylated by LATS. Together, these findings significantly modify the current view that the phosphorylation of YAP by Hippo protein kinases is sufficient for YAP inhibition and cellular growth arrest. Based upon these results, the inhibition of cellular growth in the absence of serum more accurately involves the stabilization of Amot130 and LATS, which together inhibit YAP activity and mammary epithelial cell growth.

Indiana University-Purdue University Indianapolis (IUPUI) / Break-induced replication (BIR) is a mechanism to repair double-strand breaks (DSBs) that possess only a single end that can find homology in the genome. This situation can result from the collapse of replication forks or telomere erosion. BIR frequently produces various genetic instabilities including mutations, loss of heterozygosity, deletions, duplications, and template switching that can result in copy-number variations (CNVs). An important type of genomic rearrangement specifically linked to BIR is half crossovers (HCs), which result from fusions between parts of recombining chromosomes. Because HC formation produces a fused molecule as well as a broken chromosome fragment, these events could be highly destabilizing. Here I demonstrate that HC formation results from the interruption of BIR caused by a defective replisome or premature onset of mitosis. Additionally, I document the existence of half crossover instability cascades (HCC) that resemble cycles of non-reciprocal translocations (NRTs) previously described in human tumors. I postulate that HCs represent a potent source of genetic destabilization with significant consequences that mimic those observed in human diseases, including cancer.