Breast Cancer Program
Recent accomplishments of the Breast Cancer Program include:
- Dietary and physical intervention successfully modified pregnancy hormone levels and gestational diabetes - factors associated with increased breast cancer risk (Hilakivi-Clarke) (Kinnunen et al., Eur J Clin Nutr, 2007; Lof et al., Acta Obstet Gynecol Scand, 2008)
- First implication of ERRγ in endocrine resistance in invasive lobular breast cancers (Riggins and Clarke) (Riggins et al., Cancer Res, 2008)
- Identification of the role of EGFRvIII in breast cancer (Tang) (Yu et al., Cancer Biol Ther, 2008)
Additionally, there are many ongoing studies:
- Role of AIB1 in Mammary Carcinogenesis (Anna Riegel, PhD)
- Identification of a novel signaling network associated with endocrine resistance (Robert Clarke, PhD, DSc)
- EGFRvIII Signaling and Breast Cancer (Careen Tang, PhD)
- Role of heavy metals and breast cancer risk (Mary Beth Martin, PhD)
- Timing of Dietary Exposures and Breast Cancer Risk: An Inter-Programmatic Translational Research Award (Leena Hilakivi-Clarke, PhD)
- Molecular predictors of taxane responsiveness (Minetta Liu, MD)
- Breast cancer screening in genetically predisposed women (Claudine Isaacs, MD)
- Phase III trial of paroxetine sulfate for the treatment of hot flashes (Claudine Isaacs, MD)
- Role of circulating tumor cells as predictors of disease response in metastatic breast cancer (Minetta Liu, MD)
Dr Riegel's laboratory is focused on the role of hormone and growth factor signaling cross talk in the progression of breast cancer as well as its role in resistance to estrogen and HER2 targeted therapies. Studies have been conducted in collaboration with Dr. Wellstein. One focus of the Riegel lab is on the nuclear receptor coactivator AIB1 (amplified in breast cancer 1) that is overexpressed in human breast cancer. Dr Riegel's laboratory has determined that overexpression of AIB1 may sensitize cells to the estrogenic effects of tamoxifen (and thus resistance to such therapy) through regulation of TGF beta (Lauritsen et al., Oncogene, 2002). Dr Riegel's lab was one of the first to demonstrate that AIB1 had a major role in IGF-1 signaling in breast cancer mediated through regulation of the PI3 kinase and Jun kinase pathways (Oh et al., Cancer Res, 2004). In collaboration with Dr. Furth, the lab was able to show that low levels of expression of an AIB1 isoform in transgenic mice drives preneoplasia in mouse mammary neoplasia (Tilli et al., Mol Endo, 2005), higher levels of induction of AIB1 are oncogenic in the mammary gland. In collaboration with Dr. Schlegel, it was demonstrated that amplification of AIB1 is not the only mechanism of maintaining high levels of AIB1 since control of the proteasomal degradation of AIB1 through E6AP can also maintain AIB1 in cancer cells (Mani et al., Cancer Res, 2006). More recently the Riegel lab has been the first lab to report that AIB1 is critical to the regulation of the HER family of growth factor receptors. Loss of AIB1 leads to dephosphorylation of HER1 (EGFR), HER2 and HER3 (Lahusen et al., Cancer Res, 2007) and may explain why AIB1 has such a pervasive effect in many epithelial tumors. In mice with loss of only one allele of AIB1 there is a highly significant reduction on HER2 induced mammary tumor formation and loss of phosphorylation of HER2 in mammary tumors (Fereshteh et al., Cancer Res, 2008) suggesting that AIB1 once again affects the activating pathways of HER family members. Finally the Riegel lab has reported that HER/IGF-1 and estrogen activation of Abl kinase can lead to tyrosine phosphorylation of AIB1, an event critical to tumorigenesis (Oh et al., Mol Cell Biol, 2008). The current aims of the lab are to pursue the role of Abl kinase and tyrosine phosphorylation in early stage high grade DCIS where HER2 signaling and tyrosine phosphorylation of AIB1 are very high (unpublished). The possibilities of using imatinib and HER 2 inhibitors in early stage disease are being investigated. These studies are being conducted in collaboration with Dr. Liu, an interaction facilitated by the Breast Cancer Program discussions and fully consistent with the goal of translating laboratory-based research into the clinical setting.
Dr. Clarke's laboratory is involved in studying the molecular basis of acquired endocrine resistance. Most recently, his team has begun to identify a novel signaling network that appears to contribute to the ability of selective estrogen receptor modulators (SERMs), selective estrogen receptor down-regulators (SERDs), and aromatase inhibitors (AIs) to signal from ERÎ± to regulate cell fate. The initial network as currently derived is based on serial analysis of gene expression and gene expression microarray studies (Gu et al., Cancer Res 2002; Clarke et al., Oncogene, 2003). The nodes linked by solid lines reflect functional studies in both estrogen independence and antiestrogen resistance that show key mechanistic roles for IRF-1, NFkB, and XBP1. The identification and subsequent functional validation of these nodes is done in close collaboration with Drs. Liu and Riggins.
As studies have begun to proceed from node-to-node, Dr. Clarke's team now knows that elevated NFkB activation is associated with increased expression of its immediate regulator NEMO/IKKγ (Riggins et al, Mol Cancer Ther, 2005). The research team has now clearly implicated NFkB in acquired estrogen independence (analogous to aromatase inhibitor resistance) (Pratt et al., Mol Cell Biol, 2003) and in acquired resistance to SERMs. For example, the small molecular inhibitor of NFkB (i.e., parthenolide) can reverse acquired resistance to Faslodex in vitro (Riggins et al, Mol Cancer Ther, 2005). Studies of the IRF-1 node have now shown that a dominant negative IRF-1 construct can confer antiestrogen resistance (Bouker et al., Cancer Res 2004). Finally, studies of the XBP1 node have shown that the unconventional splicing of XBP1 (occurs in the cytosol) by IRE1Î± can confer both estrogen-independence and resistance to both Tamoxifen and Faslodex (Gomez et al., FASEB J, 2007). In translational studies, coexpression of these nodes, as measured by immunohistochemistry in human breast tumors, has also been established (Zhu et al., Int J Oncol, 2006a). Work in progress also shows that the expression of several of these genes is associated with TAM responsiveness in breast cancer patients.
Studies with computer scientists and engineers at Virginia Polytechnic Institute and State University have begun to develop new computational methods to extract additional network nodes (Wang et al., Bioinformatics, 2007; Clarke et al., Nat Rev Cancer, 2008; Wang et al., BMC Bioinformatics, 2008; Zhang et al., Bioinformatics, 2009). Dr. Weiner has recently initiated collaboration with Drs. Clarke and Johnson to design, build and apply a new siRNA library to experimentally discover functional estrogen receptor signaling nodes and to interrogate further the molecular determinants of resistance to antiestrogens and aromatase inhibitors.
Studies on the roles of growth factors and hormones have been ongoing in the Breast Cancer Program since its inception. Most recently, Dr. Careen Tang has been studying the functional role of the constitutively active EGFR variant III (EGFR vIII). EGFRvIII was first identified in human glioblastomas and is now known to be expressed in several other human cancers, including breast cancer. When over-expressed, Dr. Tang's group has shown that EGFRvIII enhances tumorigenicity and metastatic potential in breast cancer cells and can induce Tamoxifen and Herceptin resistance. These effects may be mediated through up-regulation of CXCR4, and her current research has now extended into investigating the cross-talk between other ErbB family receptors and CXCR4 in cancer (Yu et al., Cancer Biol Ther, 2008; Luo et al., Int J Cancer, 2004). More recently, Dr. Tang's team has shown that hypophosphorylation of tyrosine residue 1045 is likely to be the cause for EGFRvIII escape from c-Cbl-induced ubiquitination and degradation. Inefficient degradation was only observed in cells expressing a gefitinib resistant EGFR kinase mutant; whereas, a gefitinib sensitive EGFR kinase mutant exhibited similar ubiquitination and degradation patterns as the wild-type EGFR. Collectively, different EGFR mutations exert various negative mechanisms that have the potential to modify receptor internalization and degradation, and may play a key role in resistance to tyrosine kinase inhibitory treatments. More recently, this work has begun to incorporate studies of the putative tumor suppressor gene, Tob-1. Tob-1 directly associates with ErbB-2, another member of the EGFR super-family. Very recently, Dr. Tang has shown that Tob-1 expression is inversely correlated with breast and lung cancer progression in clinical specimens. Currently, this team is investigating the role and function of Tob-1 in human cancer metastasis, as well as drug resistance in vitro and in vivo.
Dr. Martin, in collaboration with Drs. Johnson, Hilakivi-Clarke and Clarke has made seminal observations on the roles of heavy metals and other environmental exposures in regulating estrogen receptor alpha (ERα) action and affecting mammary gland development and breast cancer risk (Johnson et al, Nat Med, 2003). Similar to estradiol, these metalloestrogens induce the growth of breast cancer cells, induce the estrogen-regulated genes progesterone receptor and pS2, and activate ERÎ± in transient transfection experiments. The ability of metalloestrogens to mimic the functions of estradiol is due to their ability to form a high affinity complex within the hormone binding domain of ERα. More importantly, Dr. Martin has collaborated with Drs. Johnson, Clarke, Dakshanamurthy, Hilakivi-Clarke and Wellstein, to show that cadmium at a dose similar to the World Health Organization (WHO) recommended Provisional Tolerable Weekly Intake (PTWI) mimics the effects of estrogen in vivo in target organs of ovariectomized animals and following in utero exposure. Similar to estrogens, exposure of ovariectomized animals to a low dose of cadmium resulted in a proliferative response in the endometrium and mammary gland that was blocked by the antiestrogen. In utero exposure to cadmium also mimicked the effects of in utero exposure to estradiol. There was an earlier onset of puberty and altered mammary gland development in the female offspring. This work has now expanded to include studies of the effects of nitrites (Veselik et al., Cancer Res, 2008) and the metal contaminants in tobacco smoke condensates on ER action (Martin et al., Endocrinology, 2007).
Following on from the studies in experiental models, Dr. Martin is now working with Drs. Byrne and Haddad to design population studies to determine whether environmental exposure to metallo-estrogens is a risk factor for breast cancer and contributes to the failure of endocrine therapy. To determine whether metals are associated with an increased risk of breast cancer, 5,000 women undergoing screening mammography at Georgetown University's Ourisman Breast Health Center and the Capital Breast Care Center will be recruited and their breast density and body burden of metals will be measured. Dr. Byrne and others have shown that mammographic breast density is an intermediate biomarker and a strong independent risk factor of breast cancer. Using mammographic density as an intermediate biomarker, new risk factors can be identified and their role in the etiology of the disease firmly established.
Dr. Martin's laboratory has also shown that cadmium activates the androgen receptor. Preliminary data suggest that polymorphisms in the androgen receptor influence its response to metals. Interactions among environmental exposure to metals and polymorphisms in the steroid receptor superfamily of genes will also be measured to determine their effects on the incidence of hormone dependent cancers.
This award represents a funded P01-style grant focused specifically on breast cancer. Consistent with the translational, interdisciplinary, and international approach to research in the BC Program, the U54 includes significant collaborations within the BC Program, with other Lombardi programs including Drs. Isaacs, Byers, Martin, Albanese, and Byrne, with local institutions outside Georgetown University (e.g., Virginia Tech; University of Illinois), and with colleagues in Europe. The program's central hypothesis is that the timing of dietary exposures is critical for determining later breast cancer risk. Research teams in this program propose that the breast is most sensitive to dietary factors at times when it is undergoing extensive growth, i.e., fetal period, puberty, pregnancy, in response to changes in various hormones and growth factors. Dietary factors that modify hormone-driven signaling are anticipated to alter later breast cancer risk by inducing epigenetic changes in the genes that regulate mammary cell proliferation, differentiation, and survival.
This strongly translational U54 research program developed from a series of still ongoing animal studies from Dr. Hilakivi-Clarke's laboratory (e.g., De Assis et al., Int J Cancer, 2006; De Assis et al., J Nutr, 2006; Yu et al., Int J Cancer 2006; Olivo et al., Cancer Prev Res, 2008; Wang et al., Breast Cancer Res Treat, 2008), and from epidemiological studies she performed with a group of international collaborators including Drs. Riitta Luoto (University of Tampere, Finland), Tarja Kinnunen (UKK Institute, Finland), Marie Lov (Karolinska Institutet, Sweden), Karin Michels (Harvard University), Elisabete Weiderpass (Norwegian Cancer Registry), Johan Eriksson (Public Health Institute, Helsinki, Finland) and David Barker (MRC Environmental Epidemiology Unit, University of Southampton, UK). Studies in women have been funded by multiple grants including an R01 and research grants from the Susan G. Komen Breast Cancer Foundation and the Breast Cancer Research Foundation. These studies, conducted in Finland at Maternity Center registries, school registries, and the National Finnish Cancer Registry, found that high birth weight, high body weight during childhood, and excessive weight gain during pregnancy, each affect later breast cancer risk (Hilakivi-Clarke et al., 2001; Kinnunen et al., 2004; Hilakivi-Clarke et al., 2005; Silva et al., PloS Medicine, 2008). Up to 17,000 Finnish pre- and post-menopausal women were studied. This research team is currently addressing the role of maternal diet during pregnancy in affecting the risk of breast and endometrial cancer in mothers and their daughters (Xue et al., Cancer Epi Biomarker Prev, 2008a, Xue et al., Cancer Epidemiol Biomarker Prev, 2008b; Kinnunen et al., Eur J Clin Nutr, 2007; Kinnunen et al., BMC Women's Health, 2004; Hilakivi-Clarke, et al., J Reprod Med, 2005).
The U54 program also studies mechanisms mediating the effects of timing of exposures to nutritional components that interact with nuclear hormone receptors on later breast cancer risk. In three projects by Drs. Hilakivi-Clarke (Project 1), Byers (Project 2), and Makela (University of Turku, Finland), these investigators are exploring whether dietary exposures to phytoestrogens, n-3 polyunsaturated fatty acids, and vitamins A and D during fetal life, puberty, or pregnancy alter mammary tumorigenesis. Animal models (i.e., carcinogen-induced mammary tumors in wildtype and Brca1+/- mice or spontaneously arising mammary tumors in Pten+/- mice) that mimic both sporadic and familial breast cancers are used; technologies applied include gene expression microarrays, for which new methods for data analysis also are being developed (BBSR). Mechanisms under investigation include changes in ERα and ERβ, BRCA1, and vitamin D receptor (Pishvaian & Byers, Cancer Epidemiol Biomarker Prev, 2007).
Two epidemiologic projects examine whether excess weight at different ages and the level of physical activity affects the risk of developing familial breast cancer - Project 4 is led by Drs. Isaacs and Byrne. The FCR has been utilized to accrue over 600 women for this project. Over 500 were already enrolled in the FCR and data on weight and physical activity levels at different ages were obtained by a modification of the FCR's annual follow up mechanism. The FCR also recruited approximately 100 new subjects for this study. Pilot Project 1, led by Dr. Luoto, is examining whether dietary intervention to prevent excessive weight gain during pregnancy and post-partum weight retention reduces breast cancer risk (University of Tampere, Finland). This project has led to several publications, e.g., Kinnunen et al., Eur J Clin Nutr 2007; and Hilakivi-Clarke et al., submitted; Luoto et al., submitted). Pilot Project 2 is led by Dr. Martin to study the roles of ERÎ±, ERÎ², and BRCA1 in mediating the effects of prepubertal exposure to selenium on the breast.
An R01-funded project, closely related to research within the U54 grant, was to determine whether maternal diet during pregnancy correlates with pregnancy hormonal environment and biomarkers of increased breast cancer risk measured in nipple aspirate fluids (NAFs). Investigators monitored dietary intakes of 300 pregnant Swedish women and obtained multiple blood and NAF samples throughout pregnancy and for the first 12 months postpartum. Initial studies have led to several key publications (Lof et al., Acta Obstet Gynecol Scand, 2008; Lof et al. Br J Cancer, 2007a; Lof et al. Br J Cancer, 2007b).
Dr. Liu recently led a research team that included Drs. Clarke, Isaacs, and Warren, in which they completed a pilot clinical trial wherein 40 chemotherapy-naϊve breast cancer patients with high-risk disease received up to four cycles of single agent Paclitaxel prior to surgery and/or the administration of anthracycline based chemotherapy. Serial core needle biopsies of the primary breast tumor were acquired for gene expression analysis at three required and one optional time point. All subjects were assessed clinically and radiographically for tumor response. Completion of accrual to this study demonstrates the ability of Dr. Liu and her collaborators to:
- Recruit patients to studies that require additional breast biopsies for correlative science;
- Obtain serial biopsies during the course of chemotherapy with minimal inconvenience and morbidity to patients;
- Process the tissue without any detriment to the timeliness of a patient's diagnosis or treatment;
- Obtain serial tissue samples for high quality and tumor RNA for gene expression studies; and
- Establish productive collaborations with clinicians in other disciplines.
Approximately 70% of the patients achieved a clinical response with Paclitaxel, with two pathologic complete responders after Paclitaxel followed by Adriamycin/Cytoxan. An average of three core biopsy specimens were obtained from each subject at each required time point, and approximately 50% of these biopsies contain >70% tumor DNA, with at least one adequate tissue sample for approximately 69% of the time points at which biopsies were obtained. Notably, the number of adequate samples is highest for those collected at baseline and fewest after the completion of therapy; this is expected given that the amount of viable tumor is expected to decrease with each successive cycle of chemotherapy. All of the tissue is currently stored as OCT-embedded samples at -70°C in RNALaterTM using a protocol previously developed by BC Program members (Ellis et al., Clin Cancer Res 2002). For quality control purposes, RNA extraction and gene expression analysis were delayed until all patients completed active study participation. The last protocol directed breast biopsy has been collected and gene expression analysis is ongoing, in collaboration with Dr. Clarke. Gene expression profiles will be generated for each tumor at each time point, and these will be used as an independent data set to test neural network classifiers of taxane non-responsiveness. The neural networks are being built from cell line and xenograft data acquired through collaborative efforts between Dr. Liu and Dr. Clarke. To improve classification accuracy, the team has recently developed a powerful new method that outperforms other machine-learning classification methods (Wang et al. Bioinformatics, 2006). In addition, clinical and gene expression data are currently being imported into the Georgetown Database of Cancer (G-DOC) for analysis such as Kaplan-Meier and ROC curves.
The team also has explored the clinical relevance of some of the cell lines used in the xenograft studies and in the endocrine responsive studies described above. Using gene expression microarray analyses, the researchers showed that the transcriptomes of ERα+ breast cancers and ERα+ breast cancer cell lines show notable similarities (Zhu et al, Int J Oncol, 2006b).
Women with a genetic predisposition to breast cancer face a markedly elevated risk of developing early onset disease. Based on single institution data suggesting that magnetic resonance imaging (MRI) may be helpful in detecting clinical and mammographic occult breast cancer in BRCA1/2 carriers and other genetically high risk women, Claudine Isaacs, MD (Clinical PI), in collaboration with Constance Lehman, MD (Radiology PI, University of Washington), developed and led a multi-institutional trial of breast MRI, ultrasound, and mammography in such high risk women. 195 women, from six institutions that were members of the Cancer Genetics Network (CGN), were enrolled in the study over a 6-month period. Twenty nine women were recruited from CGN and Familial Cancer Registry participants at Lombardi. Such rapid accrual of these women was only possible because of their prior enrollment in the registries. CRMO provided research coordinator support for this study. Six cancers were detected (3.5% cancer yield). MRI detected all six cancers, while mammography and ultrasound both detected two cancers. MRI resulted in greater numbers of biopsies. Characteristics of the various imaging modalities were also examined. After controlling for menopausal status, the only significant difference noted was that BRCA1 and BRCA2 carriers were less likely to have extremely dense breast tissue on mammography (Lehman et al, Radiology, 2007).
Hot flashes are a common problem in breast cancer survivors and negatively affect their quality of life (QOL). A pilot study at Lombardi indicated that paroxetine sulfate significantly reduces the incidence and severity of hot flashes. Based on these promising findings, Drs. Isaacs and Gehan initiated a multi-institution stratified, randomized, double-blind, placebo controlled, cross over trial to evaluate the efficacy of paroxetine in ameliorating hot flashes in breast cancer patients and in other women who could not or did not wish to take estrogen replacement therapy and who were experiencing at least 14 bothersome hot flashes weekly. 151 women, of which 86 were from Lombardi, were randomized to receive 4 weeks of paroxetine at 10 mg or 20 mg followed by 4 weeks of placebo, or to receive placebo for 4 weeks followed by paroxetine 10 mg or 20 mg for 4 weeks. Participants completed baseline daily hot flash diaries for one week prior to the start of the study and throughout the duration of the study, and QOL questionnaires at baseline, week 5 and week 9. Paroxetine 10 mg and 20 mg were both significantly more effective than placebo in terms of a reduction in hot flash frequency and composite score. Paroxetine 10 mg was better tolerated than the 20 mg dose. This study demonstrated that paroxetine is an effective and well-tolerated treatment of hot flashes in breast cancer survivors (Stearns et al, J Clin Oncol 2005).
Circulating tumor cells (CTC) were first identified in the mid 19th century from blood samples taken after the death of a patient with an undiagnosed malignancy. The CellSearch (Veridex, LLC, Warren, NJ) technology now allows for the reliable and reproducible identification, capture, and enumeration of these cells. CTC are undetectable in healthy individuals but detectable in at least two-thirds of patients with metastatic breast cancer. Initial work demonstrated that the enumeration of >5 CTC per 7.5 mL of blood predicts for poorer progression free survival (PFS) in patients with metastatic breast cancer.
Dr. Liu, in collaboration with Drs. Shields, Isaacs, Eng-Wong, and Warren, is conducting a prospective studyÂ that validates these findings and, more importantly, demonstrates a strong correlation between CTC results and radiographic disease progression at the time of, and in advance of, imaging. This correlation is independent of type of therapy (i.e., chemotherapy or endocrine therapy). Peripheral blood was collected for CTC enumeration at baseline and at 3- to 4-week intervals in patients starting a new treatment regimen for progressive, radiographically measurable metastatic breast cancer. Clinical outcomes were based on radiographic studies performed in 9- to 12-week intervals: 68 patients were evaluable for the CTC-imaging correlations, and 74 patients were evaluable for the PFS analysis (59 of which were accrued at Lombardi). Median follow-up was 13.3 months. The CRMO provided support for the conduct of this trial and BBSR conducted the statistical analyses. A statistically significant correlation was seen between CTC levels and radiographic disease progression in patients receiving chemotherapy or endocrine therapy. This correlation applies to CTC results obtained at the time of imaging (odds ratio [OR] 6.3), 3-5 weeks prior to imaging (OR 3.1), and 7-9 weeks prior to imaging (OR 4.9). Analyses stratified by type of therapy remain statistically significant. Shorter PFS was observed for patients with >5 CTCs at 3-5 weeks and 7-9 weeks after the start of treatment. These findings support the role of CTC enumeration as an adjunct to standard methods of monitoring disease status in metastatic breast cancer. A revised manuscript has been resubmitted to the Journal of Clinical Oncology.