Abstract ED02-04: Identifying targets for breast and ovarian cancer prevention

Abstract

Abstract Treatment trials have provided a platform to test the influence of anti-estrogens and confirm the value of estrogen receptor (ER) as a target for effectiveness. Relative resistance to treatment of tumors that are also HER2/neu-positive, identified a second important breast cancer target. Therapeutic interventions related to ER AND HER2/neu have prepared the way for exploiting these same targets in breast cancer prevention. Phase III prevention studies with SERMs have demonstrated risk reduction of ER-positive breast cancer. It is expected that AIs will also join the list of targeted drugs indicated for breast cancer risk reduction. Treatment studies with kinase inhibitors, trastuzumab, and anti-HER2 vaccines likewise anticipate a role for breast cancer risk reduction by targeting the EGFR family. This handful of targets is only a small fraction of the list of targeted interventions that have been proposed for breast cancer prevention. Although a broad spectrum of agents may modulate signaling pathways, it is difficult to predict, a priori, when modulation can translate into breast cancer incidence reduction. Even with agents like tamoxifen or raloxifene, a greater understanding of molecular susceptibility to cancer formation and response to pharmaceutical intervention is needed to allow highly accurate screening of subjects for intervention and prediction of risk reduction at the individual level. Of all the breast cancer prevention targets, ER and its associated pathways are characterized in greatest detail. Details of estrogen-related interventions, allows the identification of cancer prevention concerns that still need to be addressed, suggesting overarching issues that also apply to other prevention targets under consideration. The Breast Cancer Prevention Trial with tamoxifen (BCPT) will serve as an example. In comparison with treatment studies, clinical Phase III prevention trials are resource intensive, requiring large numbers of participants for an incidence endpoint. A further disadvantage of these trials is that the individuals benefiting from intervention cannot be readily identified. Both of these points speak to the need for robust surrogate markers of breast cancer risk (tissue-based targets) that correspond to followable disease in treatment trials. In the initial report of results from the BCPT, women taking tamoxifen who avoided a breast cancer diagnosis remain unknown, but out of 6681 randomized to tamoxifen of the 89 who developed invasive breast cancer, 41 had ER-positive tumors.1 Despite limited numbers, information is needed about patterns related to factors such as tamoxifen metabolism or tumor molecular profile. Working hypotheses to account for lack of benefit from tamoxifen include failure to metabolize the prodrug2 as well as the lack of effect on a molecular subtype of ER-positive breast cancer. Although current risk models allow identification of women who have increased absolute risk of developing invasive breast cancer at a certain level (e.g. >1.67% in 5 years), a further difficulty is prediction in advance of diagnosis of which women will have tumors that are not ER-positive. Based on the availability of drugs that effectively target HER2/neu-positive breast cancer, it is expected that the development of anti-HER2 therapies for breast cancer prevention will be possible. Although a “Gail Model” population as has been used in previous large prevention trials, testing of anti-HER2/neu therapies like trastuzumab is linked to overexpression HER2 in breast tissue. It follows that HER2-positive DCIS is an appropriate setting to study the effect of trastuzumab for a high risk, tissue abnormality short of invasive cancer, allowing the correlation of biomarkers with response to intervention. This question is being addressed in NSABP B-43. Another trial is examining the activity of lapatinib, a dual kinase inhibitor, for its effect on DCIS positive for HER2 and/or EGFR. Ultimately, a vaccination prevention strategy is of interest.3 The gaping hole in identifying agents for breast cancer prevention is created by tumors that cannot be positively classified in the conventional clinical context; e.g. those tissue abnormalities that are referred to as ER-negative. In this case, the empiric approach has been to work with agents that do not specifically relate to ER pathways, but instead reduce proliferation or other processes that are thought to drive carcinogenesis. Using this approach in pilot prevention trials, mechanistic information is obtained that verifies the presence and modulation of a biomarker in tissue, but the predictive link to breast cancer incidence, particularly ER-negative breast cancer is missing. Another question related to targeting is whether drugs that are less specifically targeted might also be effective. Curcumin provides an extreme example of a multi-targeted drug.4 Given the need to manage investigational complexity, the identification of highly specific surrogate endpoint biomarkers needs to remain a prime focus. Returning to observational studies, the epidemiology of ER-negative breast cancer is most specifically associated with BRCA1 mutation carriers, who develop basal-like ER-negative breast cancer about 80% of the time. Preliminary data suggests that histologically normal, ALDH+ lobules in breast biopsies from BRCA1 mutation carriers may predict a very high risk of developing ER-negative breast cancer.5 Using this lead, it may be possible to demonstrate the effect of PARP inhibitors on this tissue with the goal of eliminating a molecularly defined abnormality and potentially a new surrogate endpoint biomarker. Ultimately, an understanding of the molecular biology of ER-negative lesions from BRCA1 mutation carriers could yield critical insights about the etiology and prevention of sporadic ER-negative breast cancer.6 The demonstration of ovarian cancer risk reduction by pharmacologic intervention is undeveloped with the best observational model provided by reduction in ovarian cancer incidence in association with oral contraceptive use.7 Progesterone is identified with this association, but a definitive phase III trial using a progestin to test this hypothesis is unlikely. Taking a leaf from the story of ER-negative breast cancer, it may be possible to conduct ovarian cancer prevention studies in BRCA mutation carriers with familial risk for ovarian cancer. A potential precursor from the fallopian tube of BRCA mutation carriers has been identified.8 Modulation of this lesion might be predictive of ovarian cancer risk reduction so it would appear that targeted drugs of interest to breast cancer prevention in BRCA mutation carriers might also be of interest to ovarian cancer prevention. The potential role of PARP inhibitors in eliminating molecularly defined risk of breast and ovarian cancer particularly as it relates to cancer progenitor cells and their precursors raises the question of whether the expanding wealth of information from molecular biology might harbor other examples where precursor cells are vulnerable. The role of anti-Mullerian hormone in breast and ovarian cancer will be reviewed in support of AMHR2 as a hypothetical target for ovarian cancer prevention. As molecular details expand testable hypotheses, preclinical drug screening technologies are needed to prioritize targeted agents for translational evaluation. Few studies compare human tissue and tumors in animal models to identify conserved gene expression features across species. One example is the comparative analysis of gene expression in tumors from thirteen murine models vs. subtypes of human breast cancer.9 This analysis suggests that using the TgC3(1)-Tag mouse model to test prevention agents may inform the selection of agents to go forward in development for the prevention of basal-like breast cancer. A leading animal model of ovarian cancer, the white leghorn hen, could benefit from a cross-species comparison. Citation Information: Mol Cancer Ther 2009;8(12 Suppl):ED02-04.