Breast cancers arising in women with germline BRCA1 mutations often have distinct histopathologic features. The majority are high-grade invasive ductal carcinomas with pushing borders and lymphocytic infiltration and, when compared with sporadic breast cancers, have higher mitotic counts, a greater degree of nuclear pleomorphism and less tubule formation.3,8,13–14 Eighty to 90% of these tumors lack expression of estrogen receptor (ER) and progesterone receptor (PR) and do not show human epidermal growth factor receptor 2 (HER2) protein overexpression (so-called “triple-negative” breast cancers).13 In addition, most breast cancers in BRCA1 mutation carriers cluster among the basal-like subgroup defined by gene expression profiling5,9,11 and express a variety of markers associated with a “basal” phenotype on immunohistochemical analysis including basal cytokeratins 5/6, 14, and 17, epidermal growth factor receptor (EGFR), c-kit, p53, and P-cadherin.2,4,6,13,16
Several prior studies have suggested that the immunohistochemical expression of biomarkers characteristic of basal-like breast cancers can help to define a subset of women who are likely to harbor BRCA1 mutations.9,12–13 However, whether BRCA1 mutation carriers can be reliably identified among a cohort of women with triple-negative cancers using this approach remains an unresolved issue.7
The annotated Dana-Farber/Harvard Cancer Center Breast Cancer SPORE specimen bank was queried for women with first primary invasive ER, PR, and HER2-negative breast cancer diagnosed between 2000 and 2006 as part of a protocol approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board. Patients were included in the study if demographic and family history information were available, a blood specimen for BRCA1 mutation analysis and a pretreatment formalin-fixed, paraffin-embedded breast cancer block were available, and if the patient had provided written informed consent for research. Out of the 274 women diagnosed with triple-negative breast cancer identified from the tumor bank database, 109 patients were excluded either because they had no blood specimen for DNA extraction (22), no pathology specimen available (62), or logistical constraints (such as no consent obtained) (25), leaving a total of 165 cases eligible for inclusion.
Demographics (age at breast cancer diagnosis, race, Ashkenazi Jewish ancestry), breast cancer pathology information (tumor size, axillary node status, presence of synchronous bilateral or inflammatory disease), personal history of ovarian cancer, and self-reported family cancer history information (number and age at diagnosis of first, second, and third degree relatives with breast and/or ovarian cancer) were extracted from the tumor bank database. Data were originally extracted from self-administered questionnaires completed by the subjects. Information on healthy family members was not available.
Hematoxylin and eosin-stained sections and paraffin blocks (or unstained slides if blocks were not available) from the breast cancer specimens collected before any treatment were assembled. Slides were reviewed by 2 pathologists (L.C.C., A.J.M.) blinded to the results of BRCA1 mutation analysis. Histologic tumor type and Nottingham combined histologic grade were recorded. A representative paraffin block was selected from each case when available (72% of cases), and these blocks were used to construct a tissue microarray (TMA) by obtaining 3 0.6-mm tissue cores from each donor block and embedding them in a recipient block. Slides cut from the TMA were immunostained for ER, PR, and HER2 to confirm triple-negative status. In the 28% of cases in which paraffin blocks were not available for TMA construction, the immunostains were performed on unstained slides. Additionally, immunostains for basal cytokeratins (CK 5/6, CK14, and CK 17), EGFR, P-cadherin, and p53 were performed on TMA sections or unstained slides after heat-induced epitope retrieval in citrate buffer (pH 6.0) for 30 minutes (Table 1). Immunostained TMA and whole section slides were evaluated for ER and PR expression, HER2 protein overexpression, and expression of cytokeratins (CK5/6, 14, and 17) and EGFR in each core/section. Tumor cells that showed complete absence of nuclear staining for ER or PR were considered ER negative or PR negative. Tumor cells were considered negative for HER2 protein overexpression when there was no staining or fewer than 10% of the cells showed moderate or strong membrane staining. Cases were considered basal CK-positive or EGFR-positive if any cytoplasmic and/or membranous staining was detected in the tumor cells, even if focal. These latter criteria are similar to those previously used for scoring these markers in invasive basal-like cancers.7,13,14 The Fisher exact test was used to compare the ratios.
Lymphocyte DNA was extracted from whole blood samples by standard procedures. Germline BRCA1 sequence analysis was performed by high-throughput heteroduplex detection. This method is based on conformation-specific gel electrophoresis and has been adapted for use on the ABI377 sequencing instrument.10 The entire coding region was amplified with sets of primer pairs using polymerase chain reaction (PCR). After amplification, PCR products were combined into sets of 18 and electrophoresed on the ABI377 instrument. PCR fragments with aberrant mobility were sequenced. This method has been validated at our institution and has shown a sensitivity of 97.4% in detecting BRCA1 sequence changes (Miron et al, manuscript in preparation). Mutations were classified according to the Breast Cancer Information Core database. Testing for genomic rearrangements was not performed.
The final study population consisted of 144 cases in which triple-negative status was confirmed on repeat ER, PR, and HER2 testing, there was sufficient material for basal cytokeratin and EGFR immunostains, and in which the BRCA1 mutation status was determined. After ER, PR, HER2, basal cytokeratin, and EGFR immunostains were performed, there was sufficient material remaining for P-cadherin and p53 immunostains in 123 cases.
Most of tumors were high grade (89%), invasive ductal (96%) carcinomas. The remainder were invasive lobular carcinoma (2), metaplastic carcinoma (2), and adenoid cystic carcinoma (1) (Table 2). Fifty-seven percent of the cancers had a central fibrotic focus or areas of geographic necrosis (Fig. 1A), 46% had a pushing tumor margin, and 23% of cases had a prominent associated lymphoid infiltrate. Among these 144 cases, the tumor cells in 97 (67%) expressed 1 or more of the basal cytokeratins (Fig. 1B) and 102 (71%) showed EGFR expression (Fig. 1C).
Germline BRCA1 mutations were identified in 20 of the 144 patients (14%). The frequency of expression of basal cytokeratins, EGFR, P-cadherin, and p-53 according to BRCA1 mutation status is shown in Table 3. There were no statistically significant differences between the proportion of BRCA1 mutation carriers expressing basal cytokeratins, EGFR, P-cadherin, or p53 when compared with BRCA1-negative cases expressing these markers. To minimize the possibility that some of the carcinomas among the BRCA1 mutation carriers were sporadic, a feature which could possibly confound the results, we also performed the analysis after excluding women over the age of 50 years (Table 4). Although the frequency of both basal cytokeratin and EFGR expression was higher among BRCA1 mutation carriers less than 50 years of age than among those without mutations, this difference was not statistically significant.
The use of pathologic features or biomarkers to help identify women with breast cancer who could potentially carry deleterious BRCA1 germline mutations is an important clinical goal.7,8,12,13,17 However, whether or not expression of basal cytokeratins and other markers associated with basal differentiation can be used to identify breast cancer patients likely to harbor BRCA1 mutations is an unresolved issue and seems to depend at least in part on the nature of population studied.
This study endeavored to assess the value of basal cytokeratins, EGFR, and other markers associated with a basal phenotype for identifying patients likely to harbor BRCA1 germline mutations in a series of women with triple-negative breast cancers. This is the population of breast cancer patients in which BRCA1 mutations are most likely to be identified as more than 80% of BRCA1-associated breast cancers are triple negative. Although our results indicate that basal cytokeratin and EGFR expression are both highly prevalent among triple-negative breast cancers, the frequency of expression of these markers was similar among women with triple-negative breast cancers with and without BRCA1 mutations. This observation is consistent with the notion that sporadic triple-negative breast cancers are phenocopies of BRCA1-associated cancers.15 Thus, although basal cytokeratin and/or EGFR expression seems to be of value in defining a subset of triple-negative breast cancers that have a basal-like phenotype, they are not sufficient to identify women with triple-negative breast cancers who are likely to harbor BRCA1 germline mutations.
Our ability to distinguish triple-negative breast cancers associated with a BRCA1 mutation was not improved by the addition of other biomarkers associated with the basal phenotype including P-cadherin, a transmembrane glycoprotein involved in cell adhesion that has been shown to be expressed in basal-like carcinomas,4 or p53, a protein important in maintaining the integrity of the genome; this gene is frequently mutated and the protein accumulated in BRCA1-associated breast cancers13 (Table 3).
We also analyzed our biomarker data restricting the analysis to women younger than age 50 years to reduce the possibility that some of the cancers among women with BRCA1 germline mutations were sporadic rather than mutation related (Table 4). Although both basal cytokeratin and EGFR expression were somewhat more frequent among mutation carriers than among noncarriers in this subgroup, the differences were not statistically significant. This analysis is limited by the small sample size, and should be reexamined in other cohorts.
Several earlier studies have evaluated the role of expression of basal cytokeratins and other markers to help identify women with breast cancer who might harbor BRCA1 germline mutations. Lakhani et al12 reported improved sensitivity for detection of BRCA1 mutation carriers in their cohort by adding basal markers to their model. They found that ER-negative, basal cytokeratin-positive (CK14 and/or CK5/6) tumors constituted about 70% of the cancers among BRCA1 mutation carriers but only 9% of cancers among controls. Our study differs from that of Lakhani et al in that our population was restricted to women with triple-negative breast cancers. Eerola and colleagues7 evaluated basal cytokeratin expression in BRCA1 mutation carriers and compared the frequency with that seen in sporadic controls. These authors found that the proportion of cases expressing basal cytokeratins was higher among BRCA1 mutation carriers. However, as with our study, basal cytokeratin expression did not independently predict BRCA1 mutation status.7 Among 17 BRCA1-related ER-negative, HER2-negative breast cancers, Foulkes et al9 noted that 15 (88%) expressed CK5/6 whereas CK5/6 expression was seen in only 25 of 55 (45%) non-BRCA1–related ER/HER2-negative tumors. They concluded that among ER-negative/HER2-negative cancers expression of CK5/6 was associated with a 9-fold risk of having a BRCA1 mutation.9 However, in that study there were 16 tumors reported as grade 1 which were ER negative. In our experience, grade 1 invasive ductal carcinomas are virtually always ER positive. Thus, the accuracy of the ER results in these 16 cases is questionable. If one eliminates these 16 cases from their analysis, the total number of BRCA1-negative cases with CK5/6 positivity increases to 64% (25/39) from the 45% shown in their Table 1. Given this, the difference in the prevalence of CK5/6 positivity among women with and without BRCA1 mutations is not significantly different (88% vs. 64%, respectively).
One potential limitation of this study is the relatively small size of the patient population. Although this is one of the largest studies to evaluate BRCA1 mutation prevalence in triple-negative breast cancers to date, there were only 20 women with such mutations. It is possible that with a larger sample size the results might differ, particularly among women younger than 50 years of age where the addition of basal cytokeratins and EGFR might help to predict BRCA1 mutation carrier status.
In conclusion, the results of this study confirm the high prevalence of expression of basal cytokeratins and/or EGFR in women with sporadic triple-negative breast cancers. These markers are also commonly expressed in the triple-negative tumors of women with germline BRCA1 mutations. Therefore, in this dataset, expression of basal-like markers do not help to predict which women with triple-negative breast cancers are likely to harbor BRCA1 germline mutations.
1. Abd El-Rehim DM, Ball G, Pinder SE, et al. High-throughput protein expression analysis using tissue microarray technology of a large well-characterised series identifies biologically distinct classes of breast cancer
confirming recent cDNA expression analyses. Int J Cancer. 2005;116:340–350.
2. Abd El-Rehim DM, Pinder SE, Paish CE, et al. Expression of luminal and basal cytokeratins
in human breast carcinoma. J Pathol. 2004;203:661–671.
3. Arnes JB, Brunet JS, Stefansson I, et al. Placental cadherin and the basal
epithelial phenotype of BRCA1-related breast cancer
. Clin Cancer Res. 2005;11:4003–4011.
4. Armes JE, Egan AJ, Southey MC, et al. The histologic phenotypes of breast carcinoma occurring before age 40 years in women with and without BRCA1 or BRCA2 germline mutations: a population-based study. Cancer. 1998;83:2335–2345.
5. Brenton JD, Carey LA, Ahmed AA, et al. Molecular classification and molecular forecasting of breast cancer
: ready for clinical application? J Clin Oncol. 2005;23:7350–7360.
6. Cheang MC, Voduc D, Bajdik C, et al. Basal
-like breast cancer
defined by five biomarkers has superior prognostic value than triple-negative phenotype. Clin Cancer Res. 2008;14:1368–1376.
7. Eerola H, Heinonen M, Heikkila P, et al. Basal cytokeratins
in breast tumours among BRCA1, BRCA2 and mutation-negative breast cancer
families. Breast Cancer
8. Farshid G, Balleine RL, Cummings M, et al. Morphology of breast cancer
as a means of triage of patients for BRCA1 genetic testing. Am J Surg Pathol. 2006;30:1357–1366.
9. Foulkes WD, Stefansson IM, Chappuis PO, et al. Germline BRCA1 mutations and a basal
epithelial phenotype in breast cancer
. J Natl Cancer Inst. 2003;95:1482–1485.
10. Korkko J, Annunen S, Pihlajamaa T, et al. Conformation sensitive gel electrophoresis for simple and accurate detection of mutations: comparison with denaturing gradient gel electrophoresis and nucleotide sequencing. Proc Natl Acad Sci USA. 1998;95:1681–1685.
11. Laakso M, Loman N, Borg A, et al. Cytokeratin 5/14-positive breast cancer
: true basal
phenotype confined to BRCA1 tumors. Mod Pathol. 2005;18:1321–1328.
12. Lakhani SR, Reis-Filho JS, Fulford L, et al. Prediction of BRCA1 status in patients with breast cancer
using estrogen receptor and basal
phenotype. Clin Cancer Res. 2005;11:5175–5180.
13. Lakhani SR, Van De Vijver MJ, Jacquemier J, et al. The pathology of familial breast cancer
: predictive value of immunohistochemical markers estrogen receptor, progesterone receptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2. J Clin Oncol. 2002;20:2310–2318.
14. Lynch BJ, Holden JA, Buys SS, et al. Pathobiologic characteristics of hereditary breast cancer
. Hum Pathol. 1998;29:1140–1144.
15. Matros E, Wang ZC, Lodeiro G, et al. BRCA1 promoter methylation in sporadic breast tumors: relationship to gene expression profiles. Breast Cancer
Res Treat. 2005;91:179–186.
16. Nielsen TO, Hsu FD, Jensen K, et al. Immunohistochemical and clinical characterization of the basal
-like subtype of invasive breast carcinoma. Clin Cancer Res. 2004;10:5367–5374.
17. Tai YC, Chen S, Parmigiani G, et al. Incorporating tumor immunohistochemical markers in BRCA1 and BRCA2 carrier prediction. Breast Cancer