Anti-oestrogens and aromatase inhibitors are currently used as endocrine therapies in breast cancer. Despite the significant role that these treatments play in reducing breast cancer mortality, some patients may display intrinsic and acquired therapeutic resistance. Different mechanisms are thought to contribute to endocrine resistance in patients treated with anti-oestrogens such as tamoxifen and aromatase inhibitors such as letrozole. The importance of epithelial-stromal interactions in progression of breast tumour and the potential contribution of these interactions in resistance to tamoxifen have been suggested using data from different studies. Stromal compartment has also been shown to exhibit possible prognostic and therapeutic significance in breast cancer. Therefore it is essential to acquire better understanding of the changes in stromal and epithelial cells that occur during successful and unsuccessful endocrine treatment. In this study an important trial in which patients received tamoxifen for short windows was under investigation. Biopsies were taken from these patients before and after treatment. These biopsies were available as formalin-fixed paraffin-embedded (FFPE) tissue blocks. Responding and non-responding patients were identified according to the Ki67 scores. Laser Capture Microdissection (LCM) was utilised to microdissect epithelial and stromal cells from the biopsies taken after treatment. The samples obtained following microdissection underwent RNA extraction and were subsequently used for gene expression profiling. Several differentially expressed genes in the epithelial and stromal compartments of tamoxifen responding and non-responding cases were identified. To assess the significance of the differentially expressed genes between the non-responding and responding cases, bioinformatics approaches were employed to incorporate the acquired data into functional enrichment analysis, pathway analysis and network construction. To examine the presence of possible predictive markers for therapeutic response and potential targets for therapy in breast cancer within the differentially expressed genes, data was compared to several published gene sets and publicly available datasets. Several significant genes involved in recognised pathways and networks were identified within the list of differentially expressed genes. The presence of previously proposed possible markers within the differentially expressed genes was also confirmed. These findings can be further explored for the future management of breast cancer.

Les cancers du sein luminaux, exprimant le récepteur aux œstrogènes (RE) représentent 65-75% des cancers du sein soit environ 35.000 nouvelles patientes par an en France. Les référentiels thérapeutiques en vigueur recommandent une prescription systématique d’hormonothérapie au stade précoce, et quasiment constante au stade avancé. Néanmoins, il est admis que plus de 20% des patientes au stade précoce, et la quasi-totalité au stade avancé, vont échapper au traitement endocrinien, rendant impératif le développement de modèles précliniques permettant d’étudier les mécanismes d’hormonorésistance. Dans un contexte de modèles cellulaires anciens et très imparfaits (MCF7, T47D), et de quasi absence de modèles murins pertinents, nous avons choisi de développer des modèles murins dérivés de tumeurs fraîches, dits PDX (patient derived xenografts). Nous avons montré que ces modèles, difficiles à obtenir, récapitulaient avec une grande fidélité les caractéristiques morphologiques et biologiques des tumeurs d’origine. Les PDX se distinguent également par une grande stabilité de ces caractéristiques lors des passages successifs, les rendant utilisables au long cours. Nous avons également évalué les modèles obtenus pour leur profil de sensibilité à diverses modalités de traitement hormonal.Dans une seconde étape, nous avons développé des modèles résistants à partir des PDX précédemment obtenues. Quatre modèles ont pu être obtenus, qui nous ont permis d’avoir à disposition des modèles rendant compte de situations cliniques variées. Ces 4 modèles ont fait l’objet d’analyses biologiques extensives visant à identifier les caractéristiques moléculaires potentiellement associées à telle modalité de résistance : nos données suggèrent fortement qu’il y a autant de mécanismes de résistance que de situations, rendant illusoire une définition biologique unifiée de l’hormonorésistance. La reprogrammation fonctionnelle du RE semble être au centre de ces mécanismes.La voie PI3K/mTOR est une des plus fréquemment associée à l’hormonorésistance. De manière originale, nous avons mis en évidence que cette voie était activée aussi bien dans les modèles sensibles que dans les modèles résistants. La troisième étape a consisté à évaluer l’efficacité de l’everolimus, agent ciblant mTORC1. Nous avons pu montrer que l’everolimus était hautement actif dans toutes les situations considérées, sans argument pour une synergie entre everolimus et tamoxifène ou exemestane. En revanche, il existe une nette tendance à la synergie avec le fulvestrant, inhibiteur hautement spécifique du RE entraînant sa dégradation, et faisant suggérer des interactions avec la voie non génomique du RE.Nous testons actuellement des inhibiteurs spécifiques de la PI3KCA grâce à diverses collaborations industrielles qui permettront également de mener des analyses génomiques approfondies. De multiples projets académiques sont en cours. / Luminal breast cancer (ER+, HER2 negative) accounts for 65-75% of all breast carcinomas. Current guidelines strongly recommend endocrine treatment at both the early and advanced stages. However, more than 20% of early stage patients, and all advanced patients will eventually develop endocrine resistance.As most preclinical models (MCF7, T47D) do not recapitulate tumor biology, we have chosen to develop murine models derived from fresh tumors, hence called patient derived xenografts (PDX). We show that these models, although difficult to generate, faithfully exhibit the morphological and biological features of their parental counterpart, with high long term stability. These models have also been evaluated for their sensitivity to various endocrine treatments.In the next step, we developed from these initially endocrine sensitive models new tumors rendered resistant to endocrine therapies. We show that there is no unique biological pattern associated with endocrine resistance, although ER functional reprogramming appears to be critical. We also show that PI3K/mTOR pathway activation, may not be always related to endocrine resistance, and suggest that fulvestrant, an ER down regulator, may be highly synergistic with everolimus in specific cases.Several PI3KCA inhibitors are currently being evaluated in this setting.

Chemotherapy and adjuvant endocrine treatment for breast cancer has been associated with varying degrees of cognitive dysfunction, with 15-50% of women reported to experience subtle cognitive decline. While these treatments may have direct adverse consequences on neurological functioning, cancer diagnosis and treatment is also associated with many health and psychosocial factors that may decrease performance on neuropsychological tests. However, despite a growing body of literature on affected cognitive domains and observable neurological changes after chemotherapy, there has not been a thorough investigation into potentially important psychosocial and physical health mechanisms that may underlie the observed cognitive dysfunction. Therefore, the primary aim of this thesis was to evaluate the relationship between health/ treatment, psychosocial, and endocrine-related factors and cognitive dysfunction after breast cancer treatment. In addition, a smaller secondary aim was to assess the appropriateness of different methods of individual change. Chapter 1 provides a brief overview of the structure and content of the thesis. Chpaters 2 and 3 are review papers that evaluate whether there is evidence that variations in psychosocial adjustment, health and treatment factors result in cognitive changes after chemotherapy. Based on previous research, the mechanisms evaluated are endocrine-related changes (use of adjuvant endocrine treatment and chemotherapy-induced menopause); chemotherapy-induced anaemia; depression; anxiety; fatigue; quality of life; and other treatment factors (e.g. treatment duration, time since treatment, tumour stage, use of concomitant medications and co-morbid medical conditions). The impact of confounding variables such as age of participants, level of baseline functioning and methodological limitations are also considered. These two chapters have been published (refer to Appendix A for a complete list of presentations and publications arising from this thesis). The fourth and fifth chapters are methodological in nature. Chapter 4 describes methods, while Chapter 5 is a brief paper (under review) which examines methodological considerations regarding analysis of individual change in neuropsychological performance over time and across domains for women undergoing treatment for breast cancer. The sixth and seventh chapters involve empirical analyses of the data collected as part of the Cognition in Breast Cancer (CBC) study, a longitudinal study examining the causes of variation in cognitive functioning, health and well-being in women up to 2 years post-chemotherapy. Chapter 6 was an experimental study designed to investigate the acute effects of psychosocial mechanisms on cognitive functioning after chemotherapy in a sample of 157 breast cancer patients. Many of the methodological limitations identified in the review studies were addressed and the neuropsychological performance of two groups was compared, namely recently diagnosed breast cancer patients scheduled for chemotherapy (n = 136) or other forms of treatment (n = 21). Participants were assessed prior to commencing treatment and approximately one month post completion of chemotherapy (or 6 months after the first assessment). Individual cognitive impairment was examined using the Reliable Change Index, while Pearson correlations were utilised in order to investigate the effect of psychosocial and health factors on cognitive change. The results indicated that decline in haemoglobin levels and increased anxiety over the course of chemotherapy significantly predicted impairment in multiple cognitive measures, while change in specific cognitive measures was significantly associated with baseline measures of fatigue, depression and functional well-being. The impact of these findings on rehabilitation strategies for women after chemotherapy was discussed. Chapter 7 investigated whether endocrine-related changes, namely chemotherapy-induced menopause and adjuvant endocrine treatment, resulted in increased cognitive dysfunction. One hundred and thirty-six breast cancer patients were assessed using a comprehensive neuropsychological assessment over three time-points, namely pre-chemotherapy, one month and six months post chemotherapy (or at similar time-points). Linear mixed models evaluated the effects of these two factors, with little evidence found to suggest that endocrine-related factors contribute to cognitive dysfunction in breast cancer patients. Chapter 8 comprises a brief summary and overview of the entire thesis and offers overarching conclusions, strengths and weaknesses, and directions for future research. The findings of the present investigations attempt to elucidate the contributions of potentially important psychosocial and health/ treatment-related mechanisms for cognitive dysfunction after breast cancer treatment. While there was little evidence to suggest endocrine-related changes impacted on cognition, the findings linking chemotherapy-induced anaemia and baseline psychosocial measures may play an important role in identifying and treating at-risk individuals. These findings have potential research implications for the ways data is collected, analysed and presented in empirical research as well as clinical ramifications for how women are affected cognitively as well as psychologically by treatment for breast cancer.