BRCA1/2 MUTATIONS AND TRIPLE NEGATIVE BREAST CANCERS

Abstract

INTRODUCTION

Approximately 5–10% of all newly diagnosed breast cancers in Western nations are hereditary, attributable primarily to inherited mutations in the BRCA1 and BRCA2 (BRCA1/2) genes. According to a recent meta-analysis, BRCA1/2 gene mutations are associated with a 40–57% lifetime risk of female breast cancer and an 18–40% lifetime risk of ovarian cancer [10]. It is also well established that BRCA1/2 carriers with breast cancer have markedly elevated risks of contralateral breast cancer of approximately 50% at 25 years post-diagnosis [21,22].

In general, management for mutation carriers with breast cancer includes consideration of risk reducing bilateral mastectomies or alternatively, breast cancer screening with annual mammography and magnetic resonance imaging (MRI) starting at age 25 years [37]. Bilateral salpingo-oophorectomy is recommended upon completion of childbearing not only to reduce ovarian cancer risk, but when performed premenopausally, it also reduces breast cancer risk, and has been shown to reduce mortality from both of these cancers [12,37].

Newly diagnosed breast cancer patients may take advantage of rapid genetic counseling and testing to inform surgical decisions if they test positive for a BRCA1/2 mutation in that they may opt for bilateral mastectomies instead of breast conservation or unilateral mastectomy [41]. In addition, as reviewed by Rodler et al. in this volume, emerging data suggest that BRCA1/2 carriers with TN disease may respond favorably to cisplatin-targeted therapy [6,8,42]. In addition, Poly(ADP-ribose) polymerase-1 (PARP-1) inhibitors appear to be particularly effective in BRCA1/2 mutation carriers [44], and may also play a role in the treatment of sporadic TN disease [31]. Several ongoing trials of cisplatin and PARP inhibitors in BRCA1/2 mutation carriers are currently in progress. In addition to the significant risks of developing cancer faced by probands (i.e., the first individual in a family to undergo genetic testing) with a positive test result, first-degree relatives of individuals with positive results have a 50% chance of inheriting the identified BRCA1 or BRCA2 mutation. Thus, the clinical and public health implications of identifying BRCA1/2 mutation carriers are quite significant.

Potential candidates for BRCA1/2 testing are typically identified based on a qualitative and/or quantitative analysis of personal and family history of cancer. However, given the limitations that may be encountered when interpreting a family history, increasing attention has been paid to whether breast cancer markers may provide an independent indication of women who are at increased risk of carrying a mutation. Specifically, the triple negative (TN) phenotype [i.e., defined by the absence of three markers: estrogen and progesterone receptors, and human epidermal growth factor (HER)-2 expression] is the most common histological subtype observed in BRCA1 carriers. As described below, the prevalence of TN disease in BRCA1 carriers is considerably higher than the general population prevalence of 11–20% [19].

It is important to note that many features of TN cancers overlap with those observed in basal-like cancers, but they are not identical [11]. While TN breast cancers are determined based on the assessment of the three key biomarkers mentioned above, basal-like cancers are defined based on gene expression assays and immunophenotypic profiles of cytokeratins. For the purposes of this review, we will focus primarily on TN breast cancers only, given that the pathologic markers used to determine this phenotype are routinely used in practice relative to cytokeratin studies, and therefore provide more accessible data for risk assessment.

TRIPLE NEGATIVE BREAST CANCER IN BRCA1/2 MUTATION CARRIERS

IDENTIFICATION OF BRCA1/2 MUTATION CARRIERS

One of the challenges faced by clinicians is the identification of individuals for whom genetic testing is appropriate. Several key factors are used to aid in this process, including the age at diagnosis for the proband (e.g., breast cancer diagnosed < 50 years); ethnicity (e.g., presence of Ashkenazi Jewish ancestry); and family history, with the presence of epithelial ovarian cancer being a strong predictor of identifying a BRCA1/2 mutation [17]. Increasingly, integration of breast cancer pathology, particularly TN status, is being used to refine estimates of BRCA1 carrier probability.

Most genetic testing guidelines consist of criteria that include ethnicity as well as an individual’s personal and family history of breast and genetically associated cancers, such as ovarian, prostate, and pancreatic [1,37,45]. However, as age of breast cancer diagnosis alone may be an indicator of hereditary risk, it is appropriate to offer BRCA1/2 testing to young affected breast cancer patients even if they have no or minimal family history [37]. In many instances, though, these women will not test positive for BRCA1/2 mutations [5,20,53]. In evaluating candidates for genetic testing, clinicians need to consider whether family history is limited or adequate enough to properly assess hereditary cancer risk [50].

Quantitative models are often useful adjuncts to qualitative pedigree interpretations to provide mathematical estimates of BRCA1/2 carrier probability and associated a priori risks of breast and ovarian cancer, although such models have varying degrees of discriminatory accuracy [2,33,38]. Commonly used models in current clinical practice to assess probands with breast cancer include Myriad data [36], Penn II [47], BRCAPRO [4], BOADICEA [46], and Manchester [13,15]. These models vary in how they consider personal and family history of breast, ovarian, and other cancers; age at diagnosis of affected individuals; presence of and age of unaffected relatives; oophorectomy status; bilaterality of breast cancer; race/ethnicity; gene mutation penetrance; and contribution of genes other than BRCA1/2 to hereditary breast cancer.

As mentioned, it is increasingly recognized that incorporating breast cancer pathology may improve BRCA1/2 carrier risk prediction, particularly for breast cancer patients with less striking family histories. Specifically, the histopathology and immunohistochemistry of BRCA1-associated breast cancers may deviate in notable ways from patterns observed in sporadic (i.e., non-hereditary) cases [24]. Comparable to women with sporadic breast cancer, the most frequent histopathology in BRCA1 mutation carriers is invasive ductal adenocarcinoma; however, medullary carcinoma and atypical medullary carcinoma are over-represented in BRCA1 carriers. As previously discussed, the tumors in BRCA1 carriers are more frequently high grade and are negative for estrogen and progesterone receptors, and HER-2. Additionally, a high proportion of BRCA1 tumors express basal markers such as the high molecular weight cytokeratins (CK), including CK5/6 and CK14 [24], although these markers are not currently routinely assessed in clinical practice. Conversely, studies have not demonstrated consistent findings regarding whether BRCA2-associated breast cancers exhibit a different histological profile relative to sporadic cases; however, they are more frequently moderately to poorly differentiated owing to less tubule formation, and have a higher proportion of continuous pushing margins [24]. Molecular profiles with regard to hormone receptor status, HER-2 protein expression, and basal markers in BRCA2-associated breast cancers appear to be similar to profiles observed in sporadic cases [24].

Thus, given the unique molecular phenotype noted in BRCA1 carriers with breast cancer, in particular, the high prevalence of TN tumors, such information could be useful in improving a priori carrier prediction. Several studies have demonstrated that the use of tumor morphology, namely ER status and histological grade, can increase the sensitivity of BRCA1 predictions even when family history is not considered [9,16,29,32]. Not surprisingly, more recent studies have shown that the inclusion of TN and basal markers is likely to be highly predictive of BRCA1 status, particularly when basal markers are assessed in TN tumors [49].

Further, quantitative models have begun to integrate breast tumor markers in assessing BRCA carrier probabilities. Mavaddat et al. [34] used several tumor characteristics (e.g., receptor and triple negative status, basal markers) to improve the accuracy of BOADICEA. This modification is expected to become available on the BOADICEA web interface [34]. The Manchester scoring system was recently modified to include several pathological features, including grade 3 triple negativity, which significantly improved the model’s ability to predict BRCA1 status within a validation study [14]. The other model, which is the most widely used in the United States, BRCAPRO, considers positive or negative status for the following markers: ER, PR, and cytokeratins (CK5/6 and CK14) [43]. In a recent study, BRCAPRO outperformed five other models in selecting individuals for BRCA1/2 mutation testing in clinical practice [25]. However, using tumor grade and hormone receptor status for an affected proband significantly improved the ability to discriminate BRCA1 positive families, but not BRCA2 positive families [25].

Finally, a recent Markov Monte Carlo simulation demonstrated that performing BRCA1/2 testing in women with TN breast cancers diagnosed younger than age 50 is cost-effective, and if systematically applied in the U.S., could reduce the number of new breast cancers by 23% and reduce the number ovarian cancers by 41% [27]. The authors estimate that only 7–10% of unselected women with breast cancer will test positive for a BRCA1 or BRCA2 mutation, whereas up to 25% of women with TN breast cancer unselected for family history will test positive [27].

As an example of how consideration of TN status affects carrier probability, Figure 1 shows a hypothetical pedigree of a newly diagnosed breast cancer patient who presents for genetic counseling. She was diagnosed with breast cancer at age 40 and her mother was diagnosed with breast cancer at age 45. The pedigree was entered into CancerGene version 5.1 [48]. This computing platform includes the BRCAPRO model, which considers tumor pathology in addition to family history to calculate BRCA1/2 carrier probabilities. As mentioned, the breast cancer tumor parameters that can be entered are ER, PR, CK5/6 and CK14 status (not tested/unknown, positive, or negative). Of note, the BRCAPRO model does not consider HER-2 status because it was not shown to be predictive of tumor marker status after accounting for ER [43]. When all markers are unknown, the proband’s probability of testing positive for a BRCA1 or BRCA2 mutation is only 3.6%. However, when her negative ER and PR status is entered, the carrier probability rises to 10.3% (i.e., 8.5% chance of a BRCA1 mutation specifically), and when positive CK status for both markers is also included, her probability is a remarkable 50.6% (with most of that risk, 49.6%, accounting for a BRCA1 mutation).

Open in a separate window

FIG. 1

Pedigree of newly diagnosed breast cancer patient

In summary, the integration of extensive tumor pathology including TN and basal markers appears to improve the ability to predict BRCA1 carrier probability compared to standard risk factor analysis. In some cases, the resulting change in risk assessment with molecular markers could influence recommendations for genetic testing or the patient’s decision to pursue genetic testing. It is also possible that consideration of molecular markers will become formally integrated into genetic testing criteria, including those used by insurance companies. However, it is important to note that at the present time, while markers for TN status are routinely assessed, CK markers are not.

SUMMARY

The ability to identify women at increased risk for carrying a BRCA1 or BRCA2 mutation remains an important priority for clinicians and researchers. In addition to traditional parameters used in risk assessment, such as age at cancer diagnosis, family history, and ethnicity, the addition of pathological markers often improves prediction of carrier probability. Specifically, the recognition that BRCA1 carriers are more likely to develop TN breast cancers and express basal markers are important considerations in risk assessment, particularly for patients with less striking family histories. Quantitative models of BRCA1/2 probability are increasingly integrating breast tumor pathology and are expected to undergo continued validation in the future. Once BRCA1/2 mutation carriers are identified, management focuses on reducing the risk of breast and ovarian cancer, and extending genetic counseling and testing to interested at-risk relatives. Although knowledge of their BRCA1/2 status may already influence surgical decision making in newly diagnosed breast cancer patients, such information may eventually play a more central role in systemic treatment decisions as well.

Acknowledgments

Footnotes

Contributor Information

Beth N. Peshkin, Fisher Center for Familial Cancer Research, Georgetown University, Lombardi Comprehensive Cancer Center, 3300 Whitehaven Street, NW, Suite 4100, Washington, DC 20007-2401, Phone: 202.687.2716, Fax: 202.687.0305.

Michelle L. Alabek, Norton Cancer Institute, 3991 Dutchmans Lane, Suburban Plaza II, Suite 405, Louisville, KY 40207, Phone: 502.899.6818, Fax: 502.899.6763.

Claudine Isaacs, Fisher Center for Familial Cancer Research, Georgetown University, Lombardi Comprehensive Cancer Center, 3800 Reservoir Road, NW, Washington, DC 20007, Phone: 202.444.3677, Fax: 202. 444.9429.

References