La variabilité spatiale et temporelle des biomarqueurs entre tumeur primitive et métastase est une question clé à une époque où une meilleure individualisation des thérapeutiques est vivement attendue par les cliniciens, les chercheurs et les patients. Le projet de la thèse est de préciser les profils moléculaires différentiels entre tumeur primitive et métastases. Des échantillons conservés congélation issus de deux cohortes de patients ont été considérés : cancer bronchique non à petites cellules (15 patients) et cancer colorectal (13 patients, dont 11 patients avec contrôle de tissus sains). Ces prélèvements ont été analysés par séquençage haut débit et en expression génique afin d'étudier l'hétérogénéité tumorale lors de la progression métastatique et ainsi que les voies potentiellement impliquées dans le processus métastatique.Une forte conservation des profils génomiques est observée à la première progression métastatique pour les gènes récurrents des cancers bronchiques non à petites cellules et des cancers colorectaux. Si les études d'expression génique ne permettent pas de mettre en évidence de profil spécifique de la métastase, elles apportent des éléments utiles sur la conservation de certaines voies importantes dans les processus oncogéniques, identifient des gènes d'intérêt particulier dans l'étude de la progression métastatique et soulignent l'impact de l'effet tissu pour les analyses d'expression. / The question of spatial and temporal variability of biomarkers in solid tumors is a key issue in an era where personalized therapy is strongly advocated by clinicians, researchers and patients. The purpose of the work is to compare molecular profiles of primary tumor and matched metastasis in order to precise tumor heterogeneity during metastatic progression and to investigate pathways potentially involved in the metastatic process.Frozen samples from two cohorts of patients were considered: non-small cell lung cancer (15 patients) and colorectal cancer (13 patients, 11 patients with healthy control tissues). These samples were analyzed by high-throughput sequencing and gene expression. Highly conserved genomic profiles were observed in the first metastatic progression for known recurrent genes in non-small cell lung cancer and colorectal cancer. If gene expression studies do not highlight specific profile of metastasis, they provide useful data on the conservation of some important oncogenic pathways; identify genes of interest in the study of metastatic progression and highlight the putative impact of the effect of healthy tissue for expression analysis.

Magister Scientiae (Medical Bioscience) - MSc(MBS) / Breast cancer (BC) is one of the most common malignancies in women. Death results from treatment failure and metastatic disease. Thousands of lives might be saved if it was possible to detect and eliminate occult metastatic cells before they become clinically evident. Therefore, there is a critical need to identify new markers to improve treatment options for these patients. Genomic instability is the earliest indication of breast cancer and the use of genomic methodologies is a progress towards early detection and treatment, through the identification of biomarkers that can be translated into novel therapy targets. The interferon regulatory factor-1(IRF-1) gene, localized on chromosome 5q31.1, is believed to act as a tumor suppressor gene in breast cancer. The IRF-1 was found to be inactivated by single nucleotide polymorphism (SNP) in breast cancer suggesting that the loss of its function might be critical to the development of the disease. The phosphatidylinositol 3-kinase (PIK3) signaling pathway mediates key cellular functions and alterations of genes in this pathway, including PIK3CA, serine-threonine protein kinases (AKT1and AKT2), phosphatase and tensin homolog (PTEN), fibroblast growth factor receptor 2 (FGFR2) and ERBB2, whose expression have been demonstrated to be altered in breast cancer patients. In addition, these genes are linked to treatment resistance. vi In this study, we have investigated allelic loss of IRF-1 gene in primary tumors obtained from patients undergoing mastectomy at Groote Schuur hospital (Cape Town, South Africa). These samples were then further analyzed for the DNA copy number changes of specific genes involved in the PIK3/AKT signaling pathway. Statistical analysis has been performed in order to correlate genomic findings with clinical-histopathological and follow up information from the patients and to establish whether these genes can predict prognosis. Our data analysis has indicated that 46 cases (45.5%) out of 101 cases were informative for the IRF-1 dinucleotide marker used for LOH analysis (Figure 3.1). LOH was detected in 23 of these informative cases (23/46; 50%). No statistical significance was found between LOH at the IRF-1 locus and age (≤50 years or >50 years) (P value = 1.0000) and earlier stage (Stages I and II) (P value= 0.4982) based on Fisher’s exact test. Patients presented a high level of DNA copy number changes in genes involved in the PIK3/AKT pathway. The most frequent changes were observed in the PIK3CA and PTEN genes. PIK3CA presented high copy number in 36.8% of the cases. PTEN was observed with low copy number in 47.5% of the cases. This dissertation shows the effectiveness of genomic methodologies as means for the detection of early breast cancer progression in South African women. The PIK 3/AKT genes can validate the usefulness of breast cancer therapies.