Male breast cancer (MBC) is a rare disease, representing <1% of all breast cancers and <1% of all cancers in men.1, 2 All histological types of breast cancer have been identified in men, with infiltrating ductal carcinoma representing about 80% of cases. Similarly to FBC, hormonal, genetic, and environmental factors are involved in the etiology of MBC.3 Conditions associated with increased estrogen or decreased androgen levels, such as Klinefelter syndrome and testicular disorders, predispose to MBC.1 Particularly, the Klinefelter syndrome, characterized by 47,XXY karyotype, is associated with a 50-fold increase in breast cancer risk.4 The genetic factor most clearly associated with MBC is BRCA2 germline mutation and its frequency in men with breast cancer varies considerably (4–40%) depending on the population investigated.3, 5 Germline BRCA1 mutations are also found in male breast cancer patients.5
The possible role and clinical utility of oncogene alterations have been widely investigated in FBC. The most extensively studied is the ERBB2 oncogene that belongs to the epidermal growth factor receptor family. ERBB2 is amplified and overexpressed in 10–34% of FBC and has been shown to have prognostic and predictive value.6 Targeted therapy against ERBB2 is currently undergoing clinical trials and shows potential benefit in the treatment of metastatic disease.7 In addition to ERBB2, the MYC and CCND1 oncogenes have also been implicated in FBC. MYC is amplified in 15% of FBC and the amplification has been associated with poor prognosis and development of lymph node metastases.8CCND1, a critical factor involved in cell cycle progression, is amplified in 5–23% of FBC and associates with increased risk of relapse and decreased survival.9 Besides these known oncogene loci, genome-wide screening by comparative genomic hybridization has revealed several chromosomal regions that frequently show increased copy number in FBC. Of these, the 17q23 and 20q13 regions are most frequently affected and subsequent studies have highlighted several candidate oncogenes, such as ZNF217 and PPM1D, within these regions.10, 11, 12
To date, only a few studies have investigated the possible involvement of genes, that are known to be commonly amplified in FBC, in breast tumors from male patients. Such comparisons between the genetic composition of FBC and MBC are likely to provide new knowledge on the pathogenesis of MBC and might also provide useful information for the clinical management of MBC. Previous immunohistochemical analyses have shown overexpression of CCND1 and MYC in 58% and 82% of cases, respectively.13, 14 The frequency of ERBB2 protein overexpression has varied considerably from one study to another with a third of tumors (range = 0–95%) showing elevated expression.1ERBB2 amplification has been assessed in a single study and no amplifications were observed.15 The aim of our present study was to evaluate the potential involvement of genes commonly amplified in FBC in a large series of MBC using the TMA technology. Amplification frequencies of ERBB2, MYC, CCND1, ZNF217 and PPM1D were investigated in 128 MBC specimens using FISH and the protein expression levels of ER, PR, CCND1 and ERBB2 were evaluated using IHC.
MATERIAL AND METHODS
We identified all MBC patients (n = 237) diagnosed in Finland between years 1967–1996 from the nation-wide Finnish Cancer Registry. The patients were originally treated in hospitals throughout Finland and all 128 patients, from whom we were able to obtain a formalin-fixed paraffin-embedded tumor specimen, were included in our study. Patient medical records and follow-up data were not available for our study. All tumors were reviewed by one pathologist (T.K.) and included 111 invasive ductal carcinomas, 12 lobular carcinomas, 4 DCIS, and 1 Paget disease. Thirty percent of the tumors were histological Grade 3, 56% Grade 2, and 14% Grade 1. The use of the Finnish Cancer Registry data was approved by the Ministry of Social Affairs and Health and the use of the tissue specimens by the Ethics Committee of the Pirkanmaa Hospital District.
The TMA was constructed as described.16 Briefly, a tissue arraying instrument (Beecher Instruments, Silver Spring, MD) was used to create holes in the recipient paraffin block, to obtain cylindrical tissue cores biopsies with a diameter of 0.6 mm from histologically representative areas of the donor blocks, and to transfer these biopsies to the recipient block. A single core biopsy was obtained from each tumor block. After construction of the TMA block, multiple 5-μm sections were cut, transferred to glass slides, and dried overnight at 60°C.
Copy number analysis by FISH
Spectrum Orange-labeled ERBB2, CCND1, and MYC probes as well as corresponding FITC-labeled centromeric probes for two-color FISH were obtained from Vysis (Downers Grove, IL). Separate one-color hybridizations were carried out with Spectrum Orange-labeled ZNF217 and chromosome 20 centromeric probes (Vysis). Two overlapping BAC clones (RP11-634F5/AC025515 and RP11-1081E4/AC079202) specific for PPM1D were identified from genomic sequence databases using the blastn program. The specificity of the BAC clones was confirmed by PCR. The BAC clones were labeled with Spectrum Orange-dUTP using random priming and hybridized with FITC-labeled chromosome 17 centromeric probe.
FISH on TMA was carried out as described previously.17 Briefly, TMA sections were treated according to the Paraffin Pretreatment Reagent kit protocol (Vysis), denaturated at 94°C for 5 min in Tth-buffer [10 mmol/l Tris-Hcl, pH 8.9 (25°C), 0.1 mol/l KCl, 1.5 mmol/l MgCl, 50 μg/ml BSA, 0.05% Tween 20 (v/v)], treated with Proteinase K (10 μg/ml in PBS) at 37°C for 10 min, dehydrated, and air-dried. After overnight hybridization at 37°C, slides were washed and counterstained with 0.2 μM 4′,6′-diamidino-2-phenylindole.
FISH signals were evaluated using a Zeiss fluorescence microscope (Jena, Germany). The specimens containing tight clusters of signals or ≥3-fold increase in the number of gene specific probe signals in relation to the centromeric signals in at least 10% of the tumor cells were considered to be amplified. A minimum of 50 tumor cells were evaluated per specimen.
The TMA sections were stained using an automated immunostaining system (TechMate 500 Plus, DAKO, Glostrup, Denmark). Slides were de-waxed, washed thoroughly with xylene and alcohol, and stained with antibodies to ER, PR, CCND1 and ERBB2 using avidin-biotin enhanced immunoperoxidase technique (ChemMate EnVision Detection Kit, DAKO). The following primary antibodies and dilutions were used: ER (6F11, Novocastra Laboratories, Newcastle, United Kingdom) at 1:80, PR (PgR636, DAKO) at 1:80, CCND1 (P2D11F11, Novocastra Laboratories) at 1:10, and ERBB2 (CB11, Novocastra Laboratories) at 1:100. The slides were counterstained with hematoxylin and embedded. Evaluation of immunohistochemistry was done using a light microscope equipped with a 20× objective. Normal breast samples were used as positive controls for ER and PR staining and FISH validated positive tumor specimens for CCND1 and ERBB2 staining. The immunohistochemical staining results for ER and PR were scored by using estimates of relative nuclear staining intensity and the percentage of positively stained carcinoma cells (histoscore). A histoscore ≥100 was considered as positive.18 CCND1 was considered positive when at least 5% of the cells exhibited intensive nuclear staining. The ERBB2 staining was evaluated according to the HercepTest scoring system.19
The correlation between amplification/overexpression and tumor grade was examined using Kruskal-Wallis test. Fisher's exact test was used to evaluate possible associations between ER and CCND1 expression levels.
We used FISH on TMA to evaluate the amplification frequencies of 5 genes (ERBB2, CCND1, MYC, PPM1D and ZNF217) in 128 primary breast tumors from male patients (Table I). Successful hybridizations were obtained in 73% of specimens on the TMA. The CCND1 oncogene was most frequently amplified in 12% (13 of 109) of cases evaluated (Fig. 1). Amplification of ERBB2 and MYC was observed in only one case each (of 91 and 88 evaluated samples, respectively). ZNF217 and PPM1D amplifications were equally rare and were found in 2 primary tumors each (of 98 and 80 evaluated tumors, respectively). The overall amplification frequency was low with 3.7% of the tumors showing amplification of at least one of the genes tested. One primary tumor showed amplification of both CCND1 and ZNF217, whereas other amplifications were independent occurring in different tumors.
Table I. Amplification and Overexpression Frequencies in MBC1
ND not determined
The protein expression of ER, PR, HER and CCND1 were studied by using IHC on TMA. Seventy-two percent of the primary tumors expressed ER and 70% were PR positive. Five primary tumors (4%) showed ERBB2 overexpression whereas the remaining 124 tumors did not exhibit positive staining for ERBB2 (Table I). The single case with ERBB2 amplification by FISH was also overexpressing the ERBB2 protein. Sixty-three percent of the primary tumors, including all cases with CCND1 amplification, showed positive staining for CCND1 protein in tumor cells (Table I, Fig. 1). Sixty-nine of these (88%) were also ER positive and a statistically significant association was found between positive staining for ER and CCND1 (p < 0.0001) (Table II). No statistical correlation was found between amplification or overexpression of any of the genes examined and the histopathological tumor grade.
Table II. Relationship Between ER and CCND1 Protein Expression in 123 MBC
Low or no expression
Activation of oncogenes through various mechanisms, such as DNA amplification, is known to contribute to the development and progression of cancer.20 Oncogene amplification is especially common in solid tumors and in FBC such amplifications have clinical value as prognostic or predictive markers.6, 8, 9 Moreover, the ERBB2 oncogene has recently received attention as a target for antibody-based therapy.7 MBC is a rare disease representing only 1% of all breast cancer cases, but does have clinical importance especially in families with BRCA2 mutations where 11% of breast carcinomas occur in men.21 The current knowledge on the frequency and extent of oncogene amplifications in MBC is very limited. A previous study has illustrated that the most common DNA copy number changes in MBC resemble those observed in FBC.22 The possible involvement of specific oncogenes implicated in FBC has not been systematically evaluated in MBC although such information might have important clinical implications. The main aim of our study was to determine the possible relevance of genes frequently amplified in FBC in the pathogenesis of MBC. Specifically, we hypothesized that comparison between amplification frequencies in FBC and MBC would shed light on the genetic basis of MBC and might provide new information that would be useful for the clinical management of MBC patients.
We examined amplification frequencies of the ERBB2, CCND1, MYC, PPM1D and ZNF217 genes in 128 primary tumors from male breast cancer patients using high-throughput TMA technology. Our FISH analyses on TMA showed an overall success rate of 73% with most of the failures due to missing or unrepresentative samples or poor hybridization quality. This result is in good concordance with previous TMA studies,23, 24 even though our samples were exclusively breast tumors that are difficult to evaluate due to the melting of the fat during the hybridization procedure.17 In addition, many of the samples were derived from very old archival tissue with variable fixation that is also known to hamper the FISH analysis.
A low overall frequency of amplification was observed with only 3.7% of male breast tumors showing amplification. ERBB2 amplification was detected in a single case confirming the results from a previous study where no amplification was detected among 58 samples.15 Although only a single tumor with ERBB2 amplification was found, we decided to evaluate the ERBB2 protein expression levels because of the clinical significance of this gene in FBC.6, 7 The number of male breast tumors with ERBB2 overexpression was also low (4%). Several previous studies have assessed ERBB2 overexpression by IHC in MBC with variable results.1 Combined data from these studies indicate overexpression in an average of 37% (range = 0–95%) of tumors. These highly variable results might be due to different antigen retrieval techniques, fixatives and antibodies, as well as to differences in the scoring systems.1 Recently, Bloom et al.15 applied strict criteria to define positive ERBB2 staining and found overexpression in only 1.7% of specimens, which is consistent with our results. The low ERBB2 amplification and overexpression frequencies indicate that MBC is clearly distinct from FBC, where ERBB2 alterations are seen in 10–34% of tumors.6 ERBB2 overexpression, however, is also known to be less common in women with hormone receptor-positive breast cancer25 and therefore the infrequent involvement of ERBB2 might reflect the high frequency of ER and PR positivity (72% and 70%, respectively) observed in our material. Overall, our results indicate that the well-known clinical utility of ERBB2 as a prognostic or predictive indicator in breast cancer is not valid for male patients and similarly the possible clinical utility of ERBB2-targeted therapies seems to be negligible.
Amplifications of MYC, PPM1D and ZNF217 were observed in <2% of MBC cases. These frequencies are also clearly lower than those reported in FBC and therefore the contribution of these genes to the pathogenesis of MBC seems less important. These genes, however, are located at chromosomal regions where the amplicons in FBC are complex and contain multiple genes.10, 11, 26, 27, 28 It is therefore possible that other genes from these regions are amplified more frequently in MBC. Even if this would be the case, the results obtained here clearly highlight distinct differences in the amplification frequencies at these loci between MBC and FBC, indicating dissimilar genetic backgrounds for these disease entities.
CCND1 amplification was detected in 12% of MBCs and overexpression in 63% of tumors, thus validating results from a previous study where CCND1 overexpression was detected in 58% of MBC cases.13 There was a good correlation between amplification and overexpression with all CCND1 amplified tumors showing overexpression. In contrast with the other genes examined in our study, the CCND1 alteration frequencies are in excellent agreement with reports on FBC.9 This result indicates that with regard to this specific genetic alteration the breast tumors from male and female patients are alike. We also found a statistically significant correlation between CCND1 and ER expression (p < 0.0001) again confirming previous findings on FBC29 and strengthening the observation on genetic similarity involving the CCND1 associated pathway. This link between CCND1 and ER is interesting because CCND1 contributes to the estrogen-induced mitogenesis in breast cancer cells.30 Overall, previous studies have indicated that CCND1 plays a specific role in normal mammary gland and mammary carcinogenesis. Analyses of preneoplastic lesions suggest that CCND1 overexpression is an early event in mammary carcinogenesis.31, 32 Transgenic mice overexpressing CCND1 in the mammary gland exhibit precocious lobuloalveolar breast development similar to that observed in pregnancy and eventually develop mammary adenocarcinomas.33 Moreover, CCND1 deficient mice are resistant to breast cancer induced by ERBB2 and RAS oncogenes.34 These findings point to a central role of CCND1 in breast cancer pathogenesis and our results suggest that this is true also for MBC.
Hormone receptors and their clinical value have been extensively characterized in FBC. ER and PR positivity correlate with better survival and response to estrogen antagonists, such as tamoxifen.35 Endocrine therapy is effective in approximately one-third of FBCs and up to 80% of tumors expressing both ER and PR. The vast majority of MBCs are hormone receptor positive (81% in previous studies, 71% in our series), and therefore tamoxifen has become a mainstay of therapy.1 Randomized clinical trials on tamoxifen therapy in MBC are missing and retrospective studies have shown contradictory results. Some studies have found improved survival in male patients treated with tamoxifen36, 37, 38 whereas others suggested that ER or PR expression in MBC had no association with treatment response or prognosis.39, 40, 41 Interestingly, CCND1 was recently shown to be constitutively expressed in tamoxifen-resistant cells42 and elevated CCND1 mRNA levels were associated with shorter clinical response to tamoxifen.43 These results suggest that CCND1 overexpression in ER-positive tumors could contribute to tamoxifen-resistance.
In conclusions, our findings on oncogene amplification frequencies in MBC indicate substantial differences in the molecular pathogenesis between male and female breast cancer. Most strikingly, the ERBB2 oncogene amplification does not seem to contribute to MBC whereas CCND1 alterations were detected at frequencies similar to FBC. The lack of ERBB2 involvement indicates that the utility of ERBB2 as a prognostic or predictive marker or as a therapeutic target in MBC is insignificant as compared to that in FBC. The high frequency of expression of hormone receptors and CCND1 as well as the significant association found for ER and CCND1 overexpression imply that hormonal regulation plays a major role also in the development of MBC.
The authors thank Ms. K. Rouhento, Ms. R. Randen and J. Rauta, MSc, for their skillful assistance.