Ovarian cancer is one of the five most common causes of cancer death in women in the USA and UK. It is usually diagnosed when it is well established beyond the ovary in the peritoneum. Intravenous injection of cisplatin is a common palliative therapy for ovarian cancer patients. Intraperitoneal therapy has been shown to improve survival for patients. Poly(lactide-co-glycolide) (PLGA) is a biodegradable polyester which has been proven safe for medical implantation. PLGA microspheres or fibres have been considered in this work as depots for delivering intraperitoneal cisplatin directly to the tumour site. The aims of this work were (1) to develop microsphere depot formulations with improved drug release profiles compared to previous work; (2) Novel cisplatin containing solid and hollow fibres were to be developed and investigated as alternative structures for depot devices; (3) The drug release profiles were to be examined using mathematical models to allow rational comparison of the devices. It was found that cisplatin containing PLGA 65:35 solid and hollow fibres represent a novel, reproducible formulation for encapsulating higher amounts of cisplatin for an equivalent mass of excipient than other polymer formulations. The fibres developed in this study were able to maintain elevated concentrations of unbound cisplatin in the presence of a biological matrix for approximately 100 hours in vitro.

New drugs are needed for better cancer management. Clinical trials are currently underway to assess the use of flavonoids (natural polyphenols) as anticancer agents. Among them, myricetin has been shown to induce cell cycle arrest and apoptosis in pre-clinical cancer models. We hypothesised that myricetin-derived novel flavonoids designed to enhance this natural potential and improve on the drug-likeness limitations of myricetin might have increased potential for their application in the management of breast and ovarian cancer. The effect of a library of novel flavonoids was screened on 3 panels of breast and ovarian cancer cell lines, representing different molecular subtypes and phenotypes, to assess their potency. The second-generation bi-methoxylated analogue AO-1530-OMe (Oncamex) was identified as the most effective candidate in the library, with sub-micromolar concentrations exerting a strong antiproliferative effect across almost all models studied. Results suggested that changes in the hydroxylation profile, the addition of methoxylations and a decyl alkyl chain were some of the structure-activity relationships contributing to this improved efficacy. Plate assays showed 8 h treatment with Oncamex reduced cell viability and induced cytotoxicity and apoptosis, concomitant with caspase activation and PARP cleavage. Pre-incubation with an antioxidant partially blocked these effects, suggesting the possible involvement of ROS modulation in the mechanism of action of Oncamex. Fluorescence microscopy reported the quick and stable delivery of Oncamex to the mitochondria. Fluorescent probes showed that Oncamex can induce mitochondrial superoxide production at concentrations associated with its antiproliferative effects. Study of the electrochemical properties of Oncamex by cyclic voltammetry supported this. Differential gene expression analysis following a microarray experiment showed Oncamex induces changes in the expression of genes controlling cell cycle and apoptosis. Together with previous results, the findings from this analysis led to the postulation of a model for the mechanism of action of Oncamex: due to its enhanced reactivity and mitochondrial targeting, Oncamex can generate mitochondrial superoxide, leading to mitochondrial dysfunction, membrane permeabilisation and the activation of the JNK pathway and the transcription factor FOXO3, which together contribute to the induction of intrinsic apoptosis and the inhibition of proliferation. Further proliferation assays on cell culture models also reported enhanced effect of Oncamex when administered in combination with paclitaxel and TRAIL. These improved responses were observed in breast and ovarian cancer models, including cells lines characterised by their treatment-resistant phenotype. Cotreatment with Oncamex also improved the effect of tamoxifen on anti-oestrogen resistant LCC9 breast cancer cells. Results from preliminary in vivo studies in mice implanted with the MDA-MB-231 breast cancer xenograft were consistent with an antiproliferative effect of Oncamex (25mg/kg/day) in vivo, as treatment inhibited tumour growth and reduced the expression of the marker of proliferation Ki-67 without signs of systemic toxicity. Tissues from this experiment also allowed for preliminary in vivo validation of the proposed mechanism of action of Oncamex by immunohistochemistry. The in vivo cytostatic effect of Oncamex was confirmed in a second in vivo experiment, which also investigated the effect of Oncamex at higher doses or in combination with paclitaxel. In conclusion, the novel flavonoid Oncamex has shown a promising antiproliferative effect in pre-clinical models of breast and ovarian cancer, including models of treatment-resistant cancers. Preliminary in vivo studies have demonstrated a partial recapitulation of the effect of Oncamex. A mechanistic model has been proposed by which Oncamex induces intrinsic apoptosis through its redox reactivity and mitochondrial targeting. These results support the potential of this prototypic candidate, although possible work in the structure and formulation of this candidate and further study and validation of its mechanism of action is needed for its continued development as an anticancer agent.

Le cancer épithélial de l’ovaire (EOC) a la particularité d’être diagnostiqué à un stade avancé chez 75% des patientes et de récidiver dans un grand nombre de cas malgré une bonne réponse initiale à la chimiothérapie, expliquant ainsi son pronostic sombre. Le rôle du microenvironnement tumoral semble être de premier plan dans le développement et la survie des cellules cancéreuses mais il existe encore peu de données concernant les cellules mésenchymateuses souches (MSC). Dans ce travail nous avons donc cherché à déterminer les mécanismes moléculaires entre les MSC et les cellules tumorales ovariennes. Dans la première partie de ce travail, nous avons mis en évidence l’émergence d’un profile pro-métastatique des cellules tumorales ovariennes après contact avec les MSC. Nous avons ensuite développé un modèle d’infiltration tumorale 3D révélant que les MSC augmentaient la dissémination tumorale ovarienne par la sécrétion d’IL6. Enfin nous avons démontré que les MSC étaient capables d’induire chez les cellules tumorales ovariennes un phénotype thermotolérant lié à la sécrétion CXCL12. Ces données vont donc toutes dans le même sens en démontrant les propriétés pro-tumorales des MSC et ouvrent de nouvelles perspectives de thérapies ciblant les interactions entre le stroma et la tumeur. / Patients with epithelial ovarian cancer (EOC) are diagnosed with advanced stage in 75% of cases and most of them will relapse despite a good primary response to chemotherapy, thus explaining the bad prognosis of EOC. While tumor microenvironment seems to play an important role for the development and survival of cancer cells, there is only few data regarding the mesenchymal stem cells (MSC) in EOC. In this work we therefore aimed at identifying the molecular determinant between MSC and ovarian cancer cells. In the first part of this work, we demonstrated that ovarian cancer cells acquired pro-metastatic profile upon contact with MSC. We then showed that MSC could enhance ovarian cancer cells infiltration through IL6 secretion in an amniochorionic membrane based 3D model. Finally we showed that MSC could protect ovarian cancer cells from hyperthermia through CXCL12 secretion. Taken together, our data are concordant to reveal the pro-tumoral properties of MSC. Cytokine inhibitors interrupting the cross-talk between OCC and MSC should now be tested as new therapies for EOC.

Le microenvironnement, composé de différents éléments cellulaires et de la matrice extracellulaire, joue un rôle primordial dans le développement tumoral et la dissémination métastatique. Ainsi, l’étude de ces interactions cellulaires est importante pour que des thérapies ciblées luttent contre la chimiorésistance des cellules tumorales. Ce travail de thèse a pour but d’étudier le rôle du facteur de croissance IGF-I dans la chimiorésistance des cellules du cancer de l’ovaire et des leucémies myéloïdes présente au sein du microenvironnement.Dans un premier temps, nous avons mis en évidence que la chimiorésistance des cellules du cancer de l’ovaire, acquise grâce aux cellules hôtes (hospicells), est liée à la sécrétion de IGF-I par ces cellules. Nous avons également montré que IGF-I est impliqué dans la régulation de certains gènes ABC (MDR-1, MRP1, MRP2, et BCRP) via les voies de STAT3, Jak2, PI3K et ERK.Dans les leucémies myéloïdes, nous avons montré que IGF-I a un effet sur la prolifération des cellules tumorales. Il induit l’expression de la protéine P-gp ainsi que la chimiorésistance des cellules sensibles à la chimiothérapie. Nous avons également déterminé le rôle de IGF-I dans la résistance des cellules leucémiques en présence des hospicells. Ces dernières ont une activité hyperangiogénique in vivo, lié à l’HIF-1 et au VEGF, et inhibent les réponses immunes des lymphocytes T par production de NO.Nous avons déterminé le rôle crucial de MMP-9 dans la migration des cellules résistantes du cancer du sein exprimant la protéine P-gp et dans la formation d’un réseau tubulaire, suggérant un lien existant entre l’expression de P-gp et de MMP-9. / The microenvironment, composed of several cellular elements and extracellular matrix, plays an important role in tumor development and metastasis. Thus, the study of these interactions is important for cell targeted therapies fighting against chemoresistant tumor cells. This thesis aims to investigate the role of growth factor IGF-I in the chemoresistance of ovarian cancer cells and myeloid leukemia, present in the microenvironment.As a first step, we demonstrated that drug resistance of ovarian cancer cells gained by host cells (hospicells) is related to the secretion of IGF-I by these cells. We have also demonstrated that IGF-I is involved in the regulation of genes ABC (MDR-1, MRP1, MRP2, and BCRP) via STAT3, Jak2, PI3K et ERK signaling pathways.In myeloid leukemia, we have shown that IGF-I has an effect on cell proliferation. It induces the expression of P-gp protein and chemoresistance of cells sensitive to chemotherapy. We also determined the role of IGF-I in the resistance of leukemic cells in the presence of hospicells. These cells have an in vivo hyperangiogenic activity, related to HIF-1 and VEGF, and inhibit immune responses of T cells by NO production.We determined the crucial role of MMP-9 in resistant cells migration of breast cancer expressing P-gp protein and in the formation of a tubular network, suggesting a link between the expression of P-gp and MMP-9.

Epithelial ovarian cancer frequently metastasizes to the omentum, a process that requires pro-angiogenic activation of local microvascular endothelial cells (ECs) by tumour-secreted factors. We have previously shown that ovarian cancer cells secrete factors, other than vascular endothelial growth factor (VEGF), with possible roles in metastatic angiogenesis including the lysosomal proteases cathepsin L (CathL) and cathepsin D (CathD), and insulin-like growth factor binding protein 7 (IGFBP7). However, the mechanisms by which these factors may contribute to omental endothelial angiogenic changes are unknown. Therefore the aims of this thesis were a) to examine disease relevant human omental microvascular endothelial cell (HOMEC) proliferation, migration and angiogenesis tube-formation induced by CathL, CathD and IGFBP7; b) to investigate whether CathL and CathD act via a proteolytic or non-proteolytic mechanism; c) to identify activated downstream intracellular signalling cascades in HOMECs and their activation in proliferation and migration; and finally d) to identify activated cell surface receptors by these factors. CathL, CathD and IGFBP7 significantly induced proliferation and migration in HOMECs, with CathL and CathD acting in a non-proteolytic manner. Proteome-profiler and ELISA data identified increased phosphorylation of the ERK1/2 and AKT (protein kinase B) pathways in HOMECs in response to these factors. CathL induced HOMEC proliferation and migration via the ERK1/2 pathway, whereas, although CathD-induced proliferation was mediated by activation of ERK1/2, its migratory effect was dependent on both ERK1/2 and AKT pathways. Interestingly, CathL induced secretion of galectin-1 (Gal1) from HOMECs which in turn significantly induced HOMEC proliferation via ERK1/2. However, none of the ERK1/2 or AKT pathways was observed to be active in Gal1-induced HOMEC migration. Interestingly, Gal1-induced proliferation and migration were significantly inhibited by L-glucose, suggesting a role for a receptor with extracellular sugar moieties. IGFBP7-induced migration was shown to be mediated via activation of the ERK1/2 pathway only. CathL, Gal1 and IGFBP7 significantly induced angiogenesis tube-formation in HOMECs which was not observed in CathD-treated cells. Receptor tyrosine kinase array revealed activation of Tie-1 and VEGF receptor type 2 (VEGFR2) in CathL and IGFBP7-treated HOMECs respectively. In conclusion, all CathL, CathD, Gal1 and IGFBP7 have the potential to act as proangiogenic factors in the metastasis of ovarian cancer to the omentum. These in vitro data suggest all four factors activate intracellular pathways which are involved in well-known angiogenesis models.

Cancer, which is characterised by aggressiveness and increased capacity for metastatic spread still requires basic researchers and clinicians to direct enormous efforts toward the development of novel therapeutic targets. Potential novel targets can be identified and exploited in combination with currently existing therapeutic approaches to improve their efficacy and overcome treatment resistance of tumour cells, protecting the patient from recurrence. To achieve this, different strategies and techniques can be proposed to identify the most promising candidate molecules for further exploitation as therapeutic targets. Human epidermal growth factor receptors (HERs) and NF-E2-related factor 2 (NRF2) are regulators of cellular proliferation and determinants of cancer initiation and progression. NRF2 and HERs confer cancers with resistance to several therapeutic agents. Nevertheless, there is limited understanding of the regulation of HER expression and activation, and the link between NRF2 and HER signalling pathways. This research has demonstrated that pharmacological activation of NRF2 by tert-butyl hydroquinone (tBHQ) upregulates the expression of HER family receptors, HER1 and HER4, elevates phospho protein kinase B (pAKT) levels, and enhances the proliferation of ovarian cancer cells. Pharmacological inhibition using retinoic acid (RA) and bexarotene and genetic inhibition using small interfering RNA (siRNA), did the opposite. Further, tBHQ caused transcriptional induction of HER1 and HER4 with different levels of expression, while siRNA-mediated knockdown of NRF2 prevented this and further caused transcriptional repression. A panel of potent NRF2 inhibitors were screened with the hope of finding the most potent for further investigation. Bexarotene was found to be the most potent and was used either alone, or in combination with lapatinib or erlotinib. The use of these drugs in combination with bexarotene resulted in the repression of HER1, HER2, HER3 and HER4 expression, inhibition of NRF2, elevation of ROS, depletion of glutathione and enhanced cytotoxicity in PEO1, OVCAR3, SKOV3 and MCF7-AREc32 cell lines. This explained the crosstalk mechanism between HER receptor family and NRF2 and the role of NRF2 in drug resistance and as a relevant anti-cancer target which opens up novel avenues of targeting HER receptor kinase family and NRF2 pathways for improving cancer therapy.

Survival rates in cancers have improved vastly over the years. However, some survival rates remain extremely low, as is the case for ovarian and lung cancer. The lack of robust and reliable biomarkers is strongly reflected in the absence of pre-screening programs, and as such, most patients in these cancer types are diagnosed only in advanced stages, leaving few treatment options. Moreover, relapse and resistance to therapies adds to the complexities of treating these diseases, even in the era of targeted drug development. Research has shown the presence of cancer material, in the form of circulating cancer cells (CTCs) and genomic material in the blood of patients, opening the possibility of 'liquid biopsies'. Liquid biopsies allow sampling of the disease to provide phenotypic and genomic data on the cancer in real-time and on a routine basis. Moreover, they overcome obstacles currently faced by conventional tissue biopsies. In this work we evaluate the use of a novel CTC imaging flow-cytometry platform, and report the ability to characterise and quantify these cells in blood samples. Moreover, we report significantly higher levels of CTCs in cancer patients compared to controls, and found them to be associated with a poorer prognosis. In particular, in lung cancer we observe these findings even in the early stages, suggesting a potential diagnostic use for this assay. We detect a similar trend in when analysing the ctDNA and suggest the possibility of using this technique with a prognostic value in the advanced setting. We also report on the analysis of existing microarray data by use of unique gene regulatory networks to identify biomarkers of interest. RAD51AP1 was identified by this process. Clinical validation revealed an over-expression of this gene in both tissue and blood of ovarian and lung cancers. Moreover, the gene over-expression was associated with a poor overall survival. Functional analysis in vitro revealed silencing RAD51AP1 suppressed tumour growth, in addition, various tumorigenic proteins were down-regulated, whilst apoptotic and immune genes were up-regulated. These results suggest a role for RAD51AP1 as a potential therapeutic target. In this study, we also demonstrate the ability to further exploit tumour genomic material in the blood by means of RNAseq, cancer panels, and CNI scoring to identify novel markers, that play an important role in disease genesis and evolution. RNAseq analysis identified XIST as a gene up-regulated in the blood and tissue of lung cancers. The ovarian cancer panels revealed 2 unique gene signatures in the ovarian cancer patients. With the CNI analyses also highlighting chromosomal aberrations from plasma analysis of cancer patients. Collectively, the use of all these techniques and exploitation of available blood based biomarkers could see significant improvements to survival rates in these, currently devastating diseases.

Le cancer du sein est le cancer le plus fréquent et la première cause de mortalité par cancer chez la femme dans le monde [1]. De nombreux facteurs participent au développement de cette maladie et les gènes BRCA1 et BRCA2 sont particulièrement impliqués. En effet, des mutations dans ces deux oncosuppresseurs sont responsables de 5 à 10% des cancers du sein héréditaires [2]. De plus, une baisse de leur expression est retrouvée dans un grand nombre de cancers du sein sporadiques [3]. Les mutations héréditaires des gènes BRCA1 et BRCA2 sont également à l’origine de cancers de l’ovaire [4]. Ce cancer est beaucoup moins fréquent que le cancer du sein, mais il est associé à un mauvais pronostic. En plus de ces facteurs génétiques, des facteurs hormonaux semblent également intervenir dans les processus de carcinogenèse mammaire et ovarienne, mais aussi des facteurs environnementaux et plus particulièrement l’alimentation. En effet, la consommation de soja, fréquente dans certaines régions de l’Asie serait responsable d’une diminution du risque de développer un cancer du sein dans les pays Asiatiques par rapport aux pays Occidentaux. Ce sont les phyto-oestrogènes contenus dans le soja qui agiraient, grâce à leur similarité de structure avec le 17-β-oestradiol de la femme [5]. Les phyto-oestrogènes du soja pourraient également agir sur le développement du cancer de l’ovaire puisque celui-ci est un cancer oestrogéno-dépendant, comme le cancer du sein. L’équipe Nutrition et Cancer du Département d’Oncogénétique du Centre Jean Perrin étudie les effets potentiellement préventifs des phyto-oestrogènes du soja dans le processus de cancérogenèse. Une première étude, menée au sein de l’équipe, a montré que l’expression des gènes BRCA1 et BRCA2 dans la glande mammaire pouvait être modulée par la consommation de soja chez des rates ovariectomisées [6]. Aussi, des études transcriptomiques, ont montré que les conséquences de l’inactivation des oncosuppresseurs BRCA1 et BRCA2 par l’utilisation d’un petit ARN interférent dans les cellules mammaires pouvaient être contrées par un traitement avec les phyto-oestrogènes du soja [7, 8]. Suite à l’émergence de travaux montrant des effets des phyto-oestrogènes du soja sur la méthylation de l’ADN, et la présence de méthylation dans le promoteur des gènes BRCA1 et BRCA2 dans les cancers sporadiques du sein, nous avons voulu voir si les phyto-oestrogènes du soja pourraient agir directement sur la méthylation de ces deux oncosuppresseurs, que nous avons au préalable mis en évidence dans les cancers du sein et de l’ovaire. / Breast cancer is the most common cancer and the leading cause of cancer death among women worldwide [1]. Many factors contribute to the development of this disease and the BRCA1 and BRCA2 genes are particularly involved. Indeed, mutations in these two oncosuppressors are responsible for 5 to 10% of hereditary breast cancers [2]. In addition, a decrease in their expression is found in a large number of sporadic breast cancers [3]. Hereditary mutations of the BRCA1 and BRCA2 genes are also at the origin of ovarian cancers [4]. This cancer is much less common than breast cancer, but it is associated with a poor prognosis. In addition to these genetic factors, hormonal factors also seem to be involved in the processes of breast and ovarian carcinogenesis, but also environmental factors and more particularly food. Soybean consumption, which is common in some parts of Asia, is thought to reduce the risk of developing breast cancer in Asian countries compared to Western countries. It is the phytoestrogens contained in soy that act, thanks to their similarity of structure with the 17-β-estradiol of the woman [5]. Soy phytoestrogens may also affect the development of ovarian cancer since it is an estrogen-dependent cancer, such as breast cancer. The Nutrition and Cancer team of the Department of Oncogenetics at the Jean Perrin Center is studying the potentially preventative effects of soy phytoestrogens in the carcinogenesis process. A first study, conducted within the team, showed that the expression of BRCA1 and BRCA2 genes in the mammary gland could be modulated by the consumption of soy in ovariectomized rats [6]. Also, transcriptomic studies have shown that the consequences of the inactivation of BRCA1 and BRCA2 oncosuppressors by the use of a small interfering RNA in mammary cells could be countered by treatment with soy phytoestrogens [7, 8]. Following the emergence of studies showing the effects of soy phytoestrogens on DNA methylation, and the presence of methylation in the BRCA1 and BRCA2 gene promoter in sporadic breast cancers, we wanted to see if the soy phytoestrogens could directly affect the methylation of these two oncosuppressors, which we have previously identified in breast and ovarian cancers.

Ras genes are frequently mutated in human cancers and present compelling targets for therapeutic intervention. While previous attempts to directly inhibit oncogenic Ras function have largely been unsuccessful use of targeted agents to inhibit the three primary oncogenic pathways activated by mutated Ras: RalGEF-Ral, PI3K-Akt and Raf- MEK-Erk, is an area of intense investigation. Here, we describe the ability of a novel pharmacological inhibitor of geranylgeranyltransferase I, GGTI-2417, to inhibit Ral prenylation and localization. We further used a Ral rescue system to selectively preserve RalA and RalB function and localization during GGTI-2417 treatment and determine the precise roles for inhibition of Ral prenylation in the GGTI anti-cancer response. Specifically, we determined inhibition of RalA is required for GGTI-attenuation of anchorage independent growth whereas inhibition of RalB is required for inhibition of proliferation, induction of apoptosis, suppression of survivin and induction of p27Kip1. We next determined the role of RalGEF-Ral signaling as well as PI3K-Akt and Raf-MEKErk signal transduction pathways in an in vitro model of human ovarian surface epithelial (T80 HOSE) cell Ras-dependent transformation. Using both small interfering RNA (siRNA) and pharmacological inhibitors of Ral, PI3K and MEK we determined that Ras signaling via Ral and PI3K but not MEK is required for ovarian oncogenesis. Furthermore, stable expression of Ras mutants unable to activate Raf-MEK-Erk signaling were able to robustly transform T80 cells. Since we had confirmed the importance of Ral proteins to human epithelial malignancies we next sought to explore the molecular interactions governing Ral transformation using a proteomics approach to rapidly identify proposed Ral interacting partners. Using immunoprecipition of transiently overexpressed FLAG-tagged RalA and RalB followed by 1D-gel separation and tandem MS/MS analysis we determined a database of proposed Ral interacting proteins. One of these, RACK1, is a validated RalA and RalB interacting protein which is at least partially required for Ras and Ral transformation. These results provide both a strong impetus and a solid basis for future studies into the mechanisms of RalA- and RalB- dependent transformation.

Ovarian cancer is a deadly disease that kills an estimated 15,000 women annually in the United States. It is estimated that approximately 10% of ovarian cancers are due to familial inheritance. The most commonly mutated genes in familial ovarian cancer are BRCA1 and BRCA2. It has been reported that cells carrying the BRCA1 185delAG mutation undergo an enhanced caspase-3 mediated apoptotic response. Here, we report on the transfection of cDNA coding for the putative truncated protein product of the BRCA1 185delAG mutant gene into BRCA1 wild-type human immortalized ovarian surface epithelial (IOSE) cells and ovarian cancer cells. Cells transfected with the BRCA1 185delAG truncation protein (BRAt) showed increased levels of active caspase 3, increased cleavage of caspase 3 substrates, PARP and DFF45, and decreased XIAP and cIAP1 following staurosporine (STS) treatment. BRAt also reduced Akt phosphorylation and over expression of activated Akt in BRAt cells restored caspase-3 activity to that seen in wild type cells. Further, BRAt expression increased chemosensitivity in platinum resistant ovarian cancer cells. Similarly, maspin protein has been shown to sensitize breast carcinoma cells to STS-induced apoptosis. We provide the first evidence that BRAt is sufficient to induce maspin protein in IOSE cells. IOSE cell lines carrying the BRCA1 185delAG mutation showed higher maspin levels than wild-type BRCA1 IOSE cell lines. BRCA1 wild-type IOSE cells were transfected with BRAt protein and showed increased maspin mRNA levels and increased nuclear maspin protein levels as compared to control cells. Additionally, both heterozygous carriers of the BRCA1 185delAG mutation and cells transfected with BRAt protein show an increased ability to activate the maspin promoter as compared to control cells. The transcription factor AP1 is at least partially required for full activation of the maspin promoter in BRAt cells, as siRNA directed towards c-jun decreased activation of the full-length maspin promoter. Taken together, our data demonstrate that truncated proteins arising from BRCA1 185delAG mutation increase Akt-mediated apoptosis by increasing nuclear maspin expression, suggesting a possible mechanism by which ovarian cancer patients with germline BRCA1 mutations may respond better to chemotherapy.