Breast cancer is the most prevalent form of cancer in women and accounts for 519,000 annual deaths (WHO Statistics). It has long been established that oestrogen (E2) stimulates tumour growth of oestrogen receptor (ER) positive breast cancer and is involved in the pathogenesis of the disease. Consequently, therapeutic approaches targeting the ER were developed. The use of endocrine therapy is an integral component in treating breast cancer however resistance to such drugs is a major limitation. Unfortunately, even initially responding tumours eventually develop resistance - acquired resistance. The aim of this study was to determine which intracellular pathways may be important in conferring acquired endocrine resistance. In order to do so, a three-stage MCF-7 cell model emulating the clinical development of acquired endocrine was used. MCF-7/LCC1 (LCC1) and MCF-7/LCC9 (LCC9) cells lines were derived from the oestrogen dependent and antioestrogen sensitive MCF-7 cell line. LCC1 cells remain responsive to endocrine therapies but their growth is not dependent on oestrogenic stimulus. LCC9 cells, on the other hand are fully resistant to endocrine therapies and completely oestrogen independent. A number of different cell membrane receptors and intracellular pathways have been implicated in endocrine resistance including HER receptor family, PI3K/Akt & MEK/ERK pathways. These pathways are of particular interest since they are able to activate ER in the absence of oestrogenic stimulus. It is likely that several pathways may be important in conferring resistance to endocrine therapies therefore the experiments in this study focussed on the transcriptional regulation of HER receptors, the activation of the Akt pathway and its implication to basic cellular processes. Following E2 treatment (48h), HER2/3/4 mRNA and protein levels were reduced in MCF- 7 and LCC1 but not in the endocrine-resistant LCC9 cell line as measured by QRT-PCR and Western blotting. The anti-estrogen fulvestrant (ICI 182,780) reversed the E2 modulation. A previous study has shown that ER and the HER2 promoter compete for limiting amounts of SRC-1 in oestrogen-responsive ZR-75-1 cells, causing HER2 repression after E2 stimulation (Newman et al.,Oncogene, 19, 490-7, 2000). ER RNAi abolished E2 repression of HER2 in MCF-7 and LCC1 cells. Furthermore, LCC9 cells have reduced SRC-1 recruitment to ER (assessed by ChIP) allowing SRC-1 to bind to the HER2 promoter. SRC-1 RNAi reduced HER2 transcription in MCF7 cells in a manner similar to E2 whilst it did not restore E2 repression in LCC9 suggesting that the latter cells have alternative mechanisms regulating HER2 transcription. RNAis against the other two p160 co-activators TIF2 and AIB1 did not restore E2 mediated HER2 repression in LCC9 cells. The importance of redundancy between p160 co-activators was also determined by performing double knockouts. SRC-1/TIF2 and TIF2/AIB1 double siRNAs had little effect on HER2 mRNA levels however SRC-1/AIB1 siRNA restored oestrogen mediated downregulation of HER2 transcription in LCC9 cells. This data indicates that SRC-1 and AIB1 co-activators play a role in the transcriptional regulation of HER receptor particularly in MCF-7 and LCC1 cells. The regulation of this transcriptional mechanism is altered in resistant LCC9 cells but, as evidenced by the double knockouts, p160 coactivators are still able to affect HER expression in these cells. This mechanism was further studied in primary breast cancer tumour material. The importance of the Akt pathway in this cell line model was also investigated as phospho-Akt levels are elevated in LCC1 and LCC9 cells. This in turn was shown to activate mTOR and ER (Ser167 residue phosphorylation) thereby contributing to increased growth and ligand independent activation of the oestrogen receptor respectively. Activation of PI3K and PTEN is unchanged in LCC1 and LCC9 cells suggesting that these proteins are not responsible for elevated Akt phosphorylation. In contrast, these cells do express higher levels of phospho-IGFR due to the high availability of receptor ligands (IGFI & IGFII). This is likely to be, at least partially, responsible for the elevated Akt activation. Moreover, the role of Akt isoforms was also determined as they are known to have different functions. The levels of Akt 2 phosphorylation are higher in endocrine resistant cell lines in comparison to parental MCF-7 cells. Interestingly, the Akt 3 phosphorylation is present in all cell lines whilst Akt 1 phosphorylation is minimal. Nevertheless, Akt RNAi studies reveal that Akt 1 and 2 siRNA dramatically reduce growth in MCF-7, LCC1 and LCC9 cells. These results suggest that Akt 2 phosphorylation may play a part in conferring endocrine resistance but the other isoforms are also important for normal cellular growth. The cell cycle profiles of LCC1 and LCC9 are very similar to MCF-7. Similarly, migration levels are unchanged in endocrine resistant cell lines. However, in the presence of antioestrogenic drugs, apoptosis in LCC1 and LCC9 cells in reduced in comparison to the parental MCF-7 cell line. Furthermore, LCC1 and LCC9 cells have higher invasion rates. The deregulation of HER receptor expression and elevated Akt activation may together confer survival advantage in LCC1 and LCC9 cells whilst also increasing their invading potential.

The transcription factor c-Myc is an early downstream target of oestrogen action in breast cancer cells in culture and it has been speculated that aberrant c-Myc expression may mediate antioestrogen resistance. However, studies of c-Myc protein expression as either a prognostic or predictive marker in human breast cancer have been limited and contradictory, as have been studies of c-Myc expression during breast cancer evolution. In order to assess the relationship between c-Myc protein expression and outcome from breast cancer, a representative cohort of 292 women with invasive ductal carcinoma (IDC) and linked clinicopathological data was assembled and tissue microarrays (TMA) generated from the archived breast cancer specimens. Detailed assessments of the expression of cyclin D1, cyclin E, p21WAF1/Cip1 and p27Kip1 were also conducted and analysed in relation to c-Myc expression using immunohistochemistry. Changes in c-Myc protein expression in a TMA model of breast cancer evolution were also conducted. Finally the cell-cycle effects of low-level constitutive c-Myc expression and high-level inducible c-Myc expression were evaluated in MCF-7 cells in vitro. Key novel results obtained were that c-Myc protein expression changed from preferentially nuclear to preferentially cytoplasmic during the evolution of breast cancer. In women with early invasive breast carcinoma, a "high-risk" cytoplasmic predominant c-Myc expression pattern was defined (~13% of cases) that independently predicted for poor outcome generally, among ER positive cases and in ER postive cases treated with endocrine therapy. In vitro studies confirmed that c-Myc overexpression was associated with resistance to the anti-proliferative effects of anti-oestrogens with persistence of both cyclin D1-cdk4 and cyclin E-cdk2 activities in the face of anti-oestrogen treatment. Further novel findings were that high cyclin D1 expression (upper 10% of expressors) was an independent predictor of poor outcome among ER positive breast cancer cases. Amongst ER + PR positive cases, both "high-risk" c-Myc expression and high level cyclin D1 expression were independent predictors of poor outcome. In summary, these data indicate that aberrant expression of the cell cycle proteins c-Myc and cyclin D1 may result in poor breast cancer outcomes in hormone receptor positive breast cancer and reinforces the importance of the cell cycle as a potential site of therapeutic manipulation in endocrine-resistant breast cancer.

The transcription factor c-Myc is an early downstream target of oestrogen action in breast cancer cells in culture and it has been speculated that aberrant c-Myc expression may mediate antioestrogen resistance. However, studies of c-Myc protein expression as either a prognostic or predictive marker in human breast cancer have been limited and contradictory, as have been studies of c-Myc expression during breast cancer evolution. In order to assess the relationship between c-Myc protein expression and outcome from breast cancer, a representative cohort of 292 women with invasive ductal carcinoma (IDC) and linked clinicopathological data was assembled and tissue microarrays (TMA) generated from the archived breast cancer specimens. Detailed assessments of the expression of cyclin D1, cyclin E, p21WAF1/Cip1 and p27Kip1 were also conducted and analysed in relation to c-Myc expression using immunohistochemistry. Changes in c-Myc protein expression in a TMA model of breast cancer evolution were also conducted. Finally the cell-cycle effects of low-level constitutive c-Myc expression and high-level inducible c-Myc expression were evaluated in MCF-7 cells in vitro. Key novel results obtained were that c-Myc protein expression changed from preferentially nuclear to preferentially cytoplasmic during the evolution of breast cancer. In women with early invasive breast carcinoma, a "high-risk" cytoplasmic predominant c-Myc expression pattern was defined (~13% of cases) that independently predicted for poor outcome generally, among ER positive cases and in ER postive cases treated with endocrine therapy. In vitro studies confirmed that c-Myc overexpression was associated with resistance to the anti-proliferative effects of anti-oestrogens with persistence of both cyclin D1-cdk4 and cyclin E-cdk2 activities in the face of anti-oestrogen treatment. Further novel findings were that high cyclin D1 expression (upper 10% of expressors) was an independent predictor of poor outcome among ER positive breast cancer cases. Amongst ER + PR positive cases, both "high-risk" c-Myc expression and high level cyclin D1 expression were independent predictors of poor outcome. In summary, these data indicate that aberrant expression of the cell cycle proteins c-Myc and cyclin D1 may result in poor breast cancer outcomes in hormone receptor positive breast cancer and reinforces the importance of the cell cycle as a potential site of therapeutic manipulation in endocrine-resistant breast cancer.

Introduction Recent advances in microarray technology have allowed more understanding of the complex molecular biology of breast cancer. The traditional prognostic information afforded by hormone receptor status and pathology variables is being supplemented and superseded by gene signatures predictive of risk of recurrence and response to treatments. Approximately 75% of breast cancers are oestrogen receptor positive (ER+) and can be treated by drugs that block oestrogen production such as letrozole. However not all ER+ tumours respond and even those that initially respond can develop resistance. Treating patients with neoadjuvant letrozole affords a unique opportunity to sample the same tumour in vivo at different time points reducing any potential inter-patient and inter-tumour variability. The molecular effects of drugs can be assessed long before clinical outcome is apparent. Underlying genetic differences or characteristics of the patient, tumour or sample may affect the molecular response to treatment. This project set out to use sequential patient-matched samples to evaluate molecular changes in breast tumours in the presence or absence of endocrine treatment in different subtypes, defined by histology or mutation status and to assess molecular variation between primary tumour and nodal metastasis. Methods RNA was extracted and processed to generate whole transcriptome Illumina Beadarray gene expression data from four unique cohorts of patients. Clinical data on treatments, recurrence and survival was collected from medical records. The first cohort compared 25 breast cancer patients with matched samples at diagnosis and at surgery, 14-35 (median 23) days later, with no intervening treatment; with 36 patients treated with neoadjuvant letrozole. A PCR assay to detect 8 known PIK3CA mutations and assessment of PTEN status was performed at both the primary and secondary event in a second cohort of 120 patients with endocrine treated disease who relapsed with either recurrence, lymph node metastases, a new second primary or progression of disease on primary endocrine therapy. The third cohort compared the molecular response to neoadjuvant letrozole in 14 patients with invasive lobular cancer (ILC) and 14 patients with invasive ductal cancer (IDC). A fourth cohort of women with node positive disease at diagnosis were assessed for variations in gene expression profiles between primary tumour and synchronous metastatic axillary lymph nodes (68 samples from 31 patients). Results The genomic profile of the no intervening treatment cohort did not differ significantly. Some changes in inflammatory genes were evident. This reassures us that changes seen during treatment are truly due to drug effect. This validates the use of a second biopsy to explore prediction of response. PIK3CA mutation status is maintained in the majority of patients with endocrine resistant disease and changed in only 15.7%. Where there was a change in PIK3CA this was significantly more likely to be a second primary breast cancer rather than a recurrence or progression of the primary cancer. PTEN status was also maintained in most patients. This does not support the theory that acquisition of a PIK3CA mutation is responsible for developing endocrine resistance. Novel PI3K inhibitor drugs may still be suitable in endocrine-resistant disease if activation of the pathway develops by other mechanisms. Consistent with previous studies, significant molecular differences were observed between ILC and IDC pre-treatment. Over half of these molecular differences were maintained after 3 months of letrozole. However, changes over time in individual tumours in response to letrozole were highly consistent in both ILC and IDC. When comparing primary with synchronous metastatic nodes only 39% of tumours clustered together with their matched primary or node. The molecular subtype of the node was often a poorer prognosis than the primary. There were also differences in subtype between nodes in a small cohort of patients with 2 involved nodes. Conclusions We have demonstrated that neoadjuvant window studies are a valid model for assessment of drug effects and evaluated differences in histology and mutation status. Endocrine resistance in breast cancer is rarely related to acquisition of PIK3CA mutations. Synchronous lymph node metastases can differ greatly from their matched primary. These findings are highly relevant when considering prescribing (neo)/adjuvant therapy and have significantly improved our understanding of breast cancer as we strive towards personalised medicine.

As a robust system, cells can wisely choose and switch between different signaling programs according to their differentiation stages and external environments. Cancer cells can hijack this plasticity to develop drug resistance. For example, breast cancers that are initially responsive to endocrine therapy often develop resistance robustly. This process is dynamically controlled by interactions of genes, proteins, RNAs and environmental factors at multiple scales. The complexity of this network cannot be understood by studying individual components in the cell. Systems biology focuses on the interactions of basic components, so as to uncover the molecular mechanism of cell physiology with a systemic and dynamical view. Mathematical modeling as a tool in systems biology provides a unique opportunity to understand the underlying mechanisms of endocrine response and resistance in breast cancer. In Chapter 2, I focused on the experimental observations that breast cancer cells can switch between estrogen receptor α (ERα) regulated and growth factor receptor (GFR) regulated signaling pathways for survival and proliferation. A mathematical model based on the signaling crosstalk between ERα and GFR was constructed. The model successfully explains several intriguing experimental findings related to bimodal distributions of GFR proteins in breast cancer cells, which had been lacking reasonable justifications for almost two decades. The model also explains how transient overexpression of ERα promotes resistance of breast cancer cells to estrogen withdrawal. Understanding the non-genetic heterogeneity associated with this survival-signaling switch can shed light on the design of more efficient breast cancer therapies. In Chapter 3, I utilized a novel strategy to model the transitions between the endocrine response and resistance states in breast cancer cells. Using the experimentally observed estrogen sensitivity phenotypes in breast cancer (sensitive, hypersensitive, and supersensitive) as example, I proposed a useful framework of modeling cell state transitions on the energy landscape of breast cancer as a dynamical system. Grounded on the most possible routes of transitions on the breast cancer landscape, a state transition model was developed. By analyzing this model, I investigated the optimum settings of two intuitive strategies, sequential and intermittent treatments, to overcome endocrine resistance in breast cancer. The method used in this study can be generalized to study treatment strategies and improve treatment efficiencies in breast cancer as well as other types of cancer. / Ph. D.

Les œstrogènes, et en particulier l’œstradiol E2, régulent un nombre considérable de fonctions physiologiques au sein de l’organisme et permettent notamment l’établissement et le maintien des fonctions reproductives chez tous les vertébrés. L’E2 agit localement dans de multiples organes cibles via l’intermédiaire de ses récepteurs : ERα et ERβ. Par son action proliférative contribuant au renouvellement de l’épithélium mammaire, l’E2 ainsi que son récepteur ERα ont été associés au développement pathologique de tumeurs mammaires. Celles-ci sont qualifiées d’hormono-dépendantes car elles répondent pour la majorité d’entre elles à l’utilisation de l’hormonothérapie visant à bloquer leur croissance. Malheureusement, on estime que 30 à 40% des tumeurs mammaires finissent par présenter une résistance aux traitements anti-oestrogéniques, par des mécanismes extrêmement complexes. Les travaux présentés dans ce manuscrit ont pour objectifs de mieux comprendre les mécanismes moléculaires et cellulaires impliqués dans le phénomène d’échappement des cellules tumorales mammaires au contrôle hormonal. Dans le cadre de cette thèse, nous nous sommes intéressés à deux facteurs capables de moduler l’activité d’ERα : l’hypoxie, qui désigne l’appauvrissement en oxygène du microenvironnement cellulaire, et la voie RhoA/MKL1 fréquemment mise en place au cours de la transition épithélio-mésenchymateuse. L’hypoxie est une caractéristique majeure des tumeurs solides, et des études lui suggèrent un rôle dans l’apparition de résistance endocrine. Nous montrons que le stress hypoxique inhibe fortement l’expression d’ERα, principalement au niveau protéique, et qu’il abolit la prolifération et la survie cellulaire induites par l’E2. L’analyse transcriptomique démontre qu’un certain nombre de gènes cibles d’ERα sont également régulés par l’hypoxie, qui peut soit réprimer (CXCL12…) ou bien augmenter leur expression (AREG…). Par ailleurs, l’analyse du cistrome d’ERα démontre une perte massive du nombre d’ERBSs (Estrogen Receptor Binding Site) par l’hypoxie, mais également une apparition d’ERBSs hypoxie-spécifiques. Nos résultats suggèrent que le fort recouvrement de régulation entre ERα et l’hypoxie puisse moduler l’efficacité des thérapies antihormonales. Enfin, l’équipe a démontré que l’activation de la voie RhoA/MKL1 provoque une forte inhibition de la fonction AF1 d’ERα. Afin de mieux appréhender les effets de cette voie de signalisation sur l’activité d’ERα, une lignée cellulaire MCF7 exprimant stablement un mutant constitutivement actif du facteur MKL1 a été générée. Nous montrons que son expression modifie profondément le contexte cellulaire en provoquant le basculement d’un phénotype luminal vers un phénotype basal-like. L’analyse transcriptomique de la réponse à l’E2 montre que le changement d’orientation cellulaire induit par MKL1 abolit toute régulation transcriptionnelle des gènes cibles d’ERα. Ce changement d’orientation cellulaire s’accompagne d’une reprogrammation massive du cistrome d’ERα avec une perte importante de ses sites de fixation à la chromatine, mais également de façon inattendue, un enrichissement en nouveaux ERBSs. Enfin, nous montrons une forte augmentation des interactions « non-génomiques » d’ERα avec des partenaires cytoplasmiques tels que PI3K, MSK1 et Src. Ces données suggèrent que dans des cellules agressives de type mésenchymal exprimant ERα, l’activité du récepteur repose majoritairement sur son action « non-génomique ». De façon intéressante, l’utilisation de l’anti-œstrogène pur ICI 182 780 n’a aucun effet inhibiteur sur ces interactions, pour lesquelles un rôle fonctionnel reste à établir. / Estrogens, and in particular estradiol E2, regulate a considerable number of physiological functions in the body and allow the establishment and maintenance of reproductive functions in all vertebrates. E2 acts locally in multiple target organs via its receptors: ERα and ERβ. By its proliferative action contributing to the renewal of the mammary epithelium, E2 as well as its ERα receptor have been associated with the pathological development of mammary tumors. These are qualified as hormone-dependent because they, for the majority of them, respond to the use of hormone therapy to block their growth. Unfortunately, it is estimated that 30-40% of mammary tumors end up with resistance to anti-estrogen treatments, through extremely complex mechanisms. The work presented in this manuscript aims to better understand the molecular and cellular mechanisms involved in the escape of mammary tumor cells to hormonal control. In this thesis, we looked at two factors that can modulate the ERα activity: hypoxia, which refers to oxygen depletion in the cellular microenvironment, and the RhoA/MKL1 pathway that is frequently activated during the epithelial-mesenchymal transition. Hypoxia is a major feature of solid tumors, and studies suggest a role in the development of endocrine resistance in breast cancer. We show that hypoxic stress strongly inhibits the expression of ERα, mainly at the protein level, and that it abolishes E2-induced cell proliferation and survival. Transcriptomic analysis shows that a certain number of ERα target genes are also regulated by hypoxia, which can either repress (CXCL12) or increase their expression (AREG ...). Moreover, the analysis of the ERα cistrome demonstrates a massive loss of the number of ERBSs (Estrogen Receptor Binding Site) by hypoxia, but also an appearance of hypoxia-specific ERBSs. Our results suggest that the strong regulatory overlap between ERα and hypoxia may modulate the efficacy of anti-hormonal therapies. Finally, the team demonstrated that the activation of the RhoA/MKL1 pathway causes a strong inhibition of the ERα AF1 function. In order to better understand the effects of this signaling pathway on ERα activity, an MCF7 cell line stably expressing a constitutively active mutant of the MKL1 factor was generated. We show that its expression profoundly modifies the cellular context by causing the switch from a luminal phenotype to a basal-like phenotype. The transcriptomic analysis of the E2 response shows that the MKL1 induced change in cell fate abolishes any transcriptional regulation of ERα target genes. This change in cellular orientation is accompanied by massive reprogramming of the ERα cistrome with a significant loss of its chromatin binding sites, but also unexpectedly, an enrichment of new ERBSs. Finally, we show a strong increase of "non-genomic" ERα interactions with cytoplasmic partners such as PI3K, MSK1 and Src. These data suggest that in aggressive mesenchymal cells expressing ERα, the receptor activity is mainly based on its "non-genomic" action. Interestingly, the use of pure anti-estrogen ICI 182 780 has no inhibitory effect on these interactions, for which a functional role remains to be established.