Immunohistochemical Expression of DNA Repair Proteins in Familial Breast Cancer Differentiate BRCA2-Associated Tumors
http://www.100md.com
《临床肿瘤学》
the Human Genetics Department, Breast and Gynecological Cancer Laboratory, Immunohistochemical Unit, Centro Nacional de Investigaciones Oncológicas, Madrid, and the Genetics Department, Pathology Department, Hospital de la Santa Creu i Sant Pau, Barcelona, and the Department of Pathology, Fundación Jiménez Díaz, Universidad Autónoma, Madrid, Spain
Laboratory of Cancer Genetics, Tampere University Hospital and Institute of Medical Technology, University of Tampere, Tampere, Finland
Genetic Pathology Evaluation Centre, British Columbia Cancer Agency, University of British Columbia, Vancouver, BC, Canada
Department of Pathology, Department of Obstetrics and Gynecology, Helsinki University Central Hospital, Helsinki, Finland
Program in Cancer Genetics, Departments of Oncology and Human Genetics, McGill University, Montreal, Quebec, Canada
ABSTRACT
PURPOSE: Morphologic and immunohistochemical studies of familial breast cancers have identified specific characteristics associated with BRCA1 mutation-associated tumors when compared with BRCA2 and non-BRCA1/2 tumors, but have not identified differences between BRCA2 and non-BRCA1/2 tumors. Because BRCA1 and BRCA2 genes participate in the DNA repair pathway, we have performed an immunohistochemical study with markers related to this pathway to establish the profile of the three groups.
MATERIALS AND METHODS: We have studied two tissue microarrays that include 103 familial and 104 sporadic breast tumors, with a panel of DNA repair markers including ATM, CHEK2, RAD51, RAD50, XRCC3, and proliferating cell nuclear antigen.
RESULTS: We found more frequent expression of CHEK2 in BRCA1 and BRCA2 tumors than in non-BRCA1/2 and sporadic tumors. We found absence of nuclear expression and presence of cytoplasmic expression of RAD51 in BRCA2 tumors that differentiate them from other familial tumors. We validated these results with a new series of patient cases. The final study with 253 familial patient cases (74 BRCA1, 71 BRCA2, 108 non-BRCA1/2), and 288 sporadic patient cases, has allowed us to confirm our preliminary results. Because BRCA2 tumors present a specific immunohistochemical profile for RAD51 and CHEK2 markers that is different from non-BRCA1/2 tumors, we have built a multivariate model with these markers that distinguish both tumors with an estimated probability of at least 76%.
CONCLUSION: Our results suggest that BRCA2 tumors demonstrate more cytoplasmic and less nuclear RAD51 staining, and increased CHEK2 staining. This pattern may distinguish BRCA2 from familial non-BRCA1/2 tumors.
INTRODUCTION
Approximately 5% to 10% of all breast cancers are hereditary and may be attributable to mutations in several high penetrance susceptibility genes, of which only two have been identified: BRCA1 and BRCA2.1,2
Although earlier estimates suggested that BRCA1 and BRCA2 mutations were responsible for 75% of site-specific breast cancer families and the majority of breast and ovarian cancer families,3,4 recent data indicate that these percentages may have been overestimated. In fact, the percentage of high-risk families associated with mutations in these genes seems to be around 25%5-9 in all groups tested, including the Spanish population.10,11
Because genetic testing for BRCA1 and BRCA2 is expensive and time-consuming because of the large size of both genes and the low percentage of altered cases, it is important to find clinical or pathologic features that could suggest or exclude the presence of BRCA1 or BRCA2 mutations in a given patient. In this respect, recent studies analyzing pathologic and immunohistochemical characteristics have shown that BRCA1 tumors tend to be of a higher grade than BRCA2 tumors. BRCA1 tumors also tend to be negative for hormone receptors and p53 positive, while the majority of BRCA2 tumors are hormone receptor-positive and p53 negative.12-14 In addition, we recently analyzed markers for cell cycle, apoptosis, and adhesion molecules in BRCA1 and BRCA2 tumors. We found striking differences between both types of tumors in type D cyclins and their associated CDK4 and CDKIs (p16, p21, and p27), which were downregulated in BRCA1 with respect to BRCA2 carcinomas.15 Basal markers were also shown to be specific of a subset of BRCA1 carcinomas.16,17 In contrast, non-BRCA1/2 tumors present similar immunohistologic anomalies to those of BRCA2 tumors with regard to proliferation, hormone receptors, cell cycle, and apoptosis when compared with BRCA1 tumors. They also cannot be distinguished from BRCA2 tumors using such markers, confirming the similarities between both types of tumors pointed out previously.14,18
One of the main functions of BRCA1 and BRCA2 is DNA double-strand break (DSB) repair by homologous recombination (HR).19-21 BRCA1 is activated by the protein kinase ATM, that initiates cell cycle changes after DNA damage.22 ATM phosphorylates numerous downstream targets, including p53, MDM2, CHEK2, NBS1, RAD9, and BRCA1.23 CHEK2 phosphorylates BRCA1, and this phosphorylation is required for the release of BRCA1 from CHEK2 and is important for cellular response.24 The CHEK2 1100delC mutation, is associated with familial breast cancer. Tumors carrying this mutation often show reduced or absent CHEK2 protein expression.25,26
BRCA1 interacts with the MRE11/RAD50/NBS1 protein complex,27,28 which is involved in both nonhomologous end joining (NHEJ) and HR in yeast and vertebrate cells.29 BRCA1 also promotes RAD51 homologous repair at a nuclear level through BRCA2. Exposure of cells to agents that induce DNA damage and replication arrest causes BRCA2 to colocalize with RAD51 at intranuclear sites that also contain the replication protein proliferating cell nuclear antigen (PCNA).30,31
The formation of RAD51 foci in response to DNA damage is dependent on BRCA2 and a series of proteins known as the RAD51 paralogs (RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3), a multiprotein complex that forms a heterodimer with each of the gene products playing a different role in DSB repair by HR.32 Some of these genes have already been proposed as breast cancer susceptibility genes.33
Currently, we do not know the functions of other susceptibility genes responsible for non-BRCA1/2 cases. To define possible differences between BRCA-associated tumors, and especially between BRCA2 and familial non-BRCA1/2 tumors that could be associated to the DNA repair pathway, we assessed the expression of some of the proteins covering several steps in this pathway by immunohistochemistry. Our results show differences in the expression of some of these markers that can help to distinguish different types of familial tumors and specifically BRCA2 tumors.
MATERIALS AND METHODS
Study Patients
Patients were obtained from three centers in Spain: the Spanish National Cancer Center, the Fundación Jimenez Díaz in Madrid, and the Hospital Sant Pau in Barcelona. They were selected from high-risk breast cancer families with at least three women affected with breast cancer and/or ovarian cancer, one of them younger than 50 years of age, or a male with breast cancer. The index case of each family was screened for mutations in the BRCA1 and BRCA2 genes by a combination of SSCP, CSGE, and sequencing techniques. Some of these results have been reported previously.10,11,34
Immunohistochemical analysis was performed on tumors distributed across two tissue microarrays (TMAs). Data from one of the TMAs were recently published.14 The second TMA was built following the same methodology. Tumors included were 33 invasive ductal carcinomas (IDC) from patients with mutations in BRCA1 (mean age, 40 years), 24 tumors with mutations in BRCA2 (19 IDC, one in situ ductal carcinoma, and four invasive lobular carcinomas; mean age, 46 years), and 46 familial non-BRCA1/2 tumors (41 IDC, three ductal carcinoma-in-situ, and two invasive lobular carcinomas; mean age, 47 years). All tumors were duplicated in the TMA. We also included normal breast and tonsil tissues as an internal control.
In addition to familial patient cases, a consecutive series of women with breast cancer was studied: 104 sporadic IDC with an age distribution similar to that of familial patient cases (mean age, 45 years).
We have validated the results obtained for CHEK2, RAD51, and RAD50 with a new independent series of patients with breast cancer from other countries. These include 150 familial patient cases (41 BRCA1, 47 BRCA2, 62 non-BRCA1/2), and 184 sporadic breast tumors, the majority of them previously published.35-37 The distribution and the origin of all patient cases are summarized in Table 1.
We have genotyped all the familial patient cases for BRCA1 and BRCA2 genes but not the sporadic patient cases. The consecutive series of sporadic patient cases was not tested because the frequency of mutations in BRCA1 and BRCA2 in these women is low (around 2%).38,39
Immunohistochemical Analysis
Immunohistochemical staining was performed by the DAKO EnVision system (DAKO, Copenhagen, Denmark) with a heat-induced, antigen retrieval step. Sections from the tissue microarray were immersed in boiling 10 mmol/L sodium citrate at pH 6.5 for 2 minutes in a pressure cooker, adding EDTA for RAD51. For ATM, proteinase K was added at a dilution of 1:40 for 10 minutes at room temperature. Antibodies, dilutions, and suppliers are listed in Table 2.
Morphological Studies and Immunohistochemical Scoring
Before tissue microarray construction, total sections of each hematoxylin-eosin stained tumor were evaluated and classified according to the WHO classification.
Two pathologists (E.H. and J.P.) simultaneously evaluated the immunohistochemical staining to avoid observer subjectivity as much as possible.
Taking into account that all immunohistochemical markers used are related with DNA repair and have to be active in the nucleus, we only considered the percentage of stained nuclei in each tumor, independent of the intensity since the variation in intensity was low. In addition, we used cutoff points to establish dichotomous variables to clarify results. For nuclear expression we used the mean (among all patient cases) of the percentage of stained cells as the cutoff point. When the percentage of stained cells was 10% or higher, we considered the tumor as positive for ATM, RAD51, and XRCC3. We considered 80% or higher as positive for RAD50 and PCNA because positive patient cases had stained 100% or near 100% of cells. The mean was 95% (SD = 8%) for positive patient cases and 70% (SD = 14%) for negative patient cases. For CHEK2 we used the cutoff of 60% defined by Kilpivaara et al.40 The mean was 80% (SD = 8%) for positive patient cases and 15% (SD = 16%) for negative patient cases. For RAD50 we did not find any patient case without staining. There was variation in nuclear staining intensity with respect to normal tissue, and we considered a lower intensity than normal tissue as negative, even if the percentage was higher than 80%.
Concerning the expression in normal breast and tonsil control tissues, expression of CHEK2 was seen in 60% of normal ductal cells, and of RAD50 and PCNA in 80% of cells. There was no staining for RAD51, ATM, nor XRCC3. For tonsil, RAD50 and PCNA staining was observed in 80%, and XRCC3, CHEK2, and RAD51 staining in 60% of germinal center lymphocytes and the basal layers of the squamous epithelium. ATM staining was seen in 80% of B and T lymphocytes outside the germinal center. For all markers the staining was in the nucleus.
For RAD51 and XRCC3, cytoplasmic staining was also considered because it was present in many tumors. Intensity was scored as either negative (non- and low-staining) or positive (high staining) (Fig 1).
Samples from other series (Table 1) were studied for RAD51, CHEK2, and RAD50 markers, and analyzed by the same pathologists (E.H. and J.P.) who were blinded to the mutation status as in the first set of tumors.
Tumors with poor correlation between duplicates (1% to 2%) were excluded from the analyses. Cores were missing for 5% to 15% of patient cases. Some cores (around 1%) were stromal tissue without breast tumor in both cores, because of errors in the selection of the tumor area.
Statistical Analysis
The Pearson's 2 test was performed to determine differences between groups using SPSS for Windows (SPSS Inc, Chicago, IL). Multivariate analyses of associations between tumor type (BRCA2 v non-BRCA1/2) and markers were carried out by unconditional logistic regression using Stata 8 (StataCorp, College Station, TX). Interactions were assessed using the log-likelihood ratio significance test. Statistical significance was declared for nominal P values less than .05.
RESULTS
Table 3 summarizes the results obtained with the group of DNA repair antibodies. For familial tumors, ATM presented similar expression in all groups, while the expression of CHEK2 was significantly more frequent among BRCA1 tumors (60.7%) and BRCA2 tumors (56.5%) than in non-BRCA1/2 tumors (21.6%) (P < .006).
Nuclear staining for RAD50 was significantly less frequent in BRCA1-associated tumors (23.3%) than in BRCA2 tumors (55.6%) and familial non-BRCA1/2 tumors (60.0%), while, for the PCNA, a greater percentage of positive tumors was found in BRCA1 tumors.
We did not observe nuclear expression of RAD51 in BRCA2 tumors, but cytoplasmic expression was highly positive in 57.1% of the patient cases. For BRCA1 tumors and non-BRCA1/2 tumors, the frequencies of nuclear and cytoplasmic expression were similar, and significantly different from BRCA2 tumors 22.6% and 30.4% versus 0% (P < .02) for nuclear staining and 29.0% and 23.9% versus 57.1% (P < .04) for high cytoplasmic staining.
No significant differences were found in the expression of XRCC3.
We analyzed these markers in our sample of sporadic cases (Table 3) and found similar frequencies of expression as obtained for non-BRCA1/2 tumors except a lower frequency of RAD51 nuclear and cytoplasmic staining and a lower frequency of XRCC3 nuclear staining.
We validated the results obtained for RAD51, CHEK2, and RAD50 with a new independent series of hereditary and sporadic tumor samples. Tumors from 41 BRCA1 patients, 47 BRCA2 patients, 62 non-BRCA1/2 patients, and 184 sporadic patient cases were analyzed for these markers (Table 4). We obtained results similar to those previously found: for CHEK2 we observed a higher percentage of BRCA1 tumors and BRCA2 tumors with expression than in non-BRCA1/2 tumors and sporadic tumors (P < .005 for all comparisons). An association was also observed for the expression of RAD50 that was significantly less frequent in BRCA1-associated tumors than in BRCA2 tumors, non-BRCA1/2 tumors, and sporadic tumors (P < .05). Finally, a low percentage of BRCA2 tumors with nuclear staining (15%) for RAD51, and a high percentage with high cytoplasmic staining (47%) was found. Although, in the first case the differences were significant compared with all other sub-types studied (P < .003 for all comparisons, Table 4), for cytoplasmic staining there was only a nonsignificant tendency when compared with non-BRCA1/2 tumors. This was as a result of the fact that some non-BRCA1/2 patient cases presented not only cytoplasmic but also nuclear expression (data not shown).
Table 5 presents results from analyses of the combined data from the original series and the validation set (253 familial tumors and 288 sporadic tumors in total). We considered all four combinations of staining for cytoplasmic and nuclear RAD51. Among those with negative nuclear staining, high cytoplasmic expression of RAD51 was found more frequently in BRCA2 tumors (48.5%) versus BRCA1 tumors (16.7%), non-BRCA1/2 tumors (16.7%), and sporadic tumors (14.9%; P < .001). Consistent with previous findings, the percentage of tumors with RAD51 nuclear staining, with or without cytoplasmic staining, was significantly higher in BRCA1 and non-BRCA1/2 mutation carriers, than in BRCA2 tumors. In addition, CHEK2 was more often expressed in BRCA1 tumors (66.0%) and BRCA2 tumors (56.4%), than in non-BRCA1/2 tumors and sporadic tumors (23.9% and 25.8%, respectively; P < .001). Finally, the number of positive tumors for RAD50 was, in all cases, significantly lower in BRCA1 tumors (39.0%) than in BRCA2 tumors (72.1%), non-BRCA1/2 tumors (68.9%), and sporadic tumors (72.4%; P < .001). We investigated associations between markers and found that of those that were correlated, only CHEK2 and RAD51 nuclear staining (positively associated, P = .01) were both associated with BRCA2 mutation status. This potential confounding of the association of each with BRCA2 versus non-BRCA1/2 tumors was considered in the multivariate analysis, which included both these markers.
Among 92 non-BRCA1/2 tumors and 54 BRCA2 tumors with results for all markers mutation status was modeled using multivariate unconditional logistic regression to differentiate the two types of tumors (Table 6). For this analysis we included RAD51 nuclear and cytoplasmic staining and CHEK2 expression, since these markers significantly differentiated BRCA2 tumors and non-BRCA1/2 tumors in univariate analyses (data not shown). An interaction between RAD51 nuclear and cytoplasmic staining was also included and was statistically significant (P = .05), consistent with the results from Table 5. The logistic model estimated a probability of 81% (95% CI, 63% to 91%) of being a BRCA2 tumor if RAD51 nuclear expression was negative and the expression of cytoplasmic RAD51 and CHEK2 were both positive. Furthermore, the model estimated a probability of 76% (95% CI, 63% to 86%) of being a non-BRCA1/2 tumor if nuclear expression of RAD51, cytoplasmic expression of RAD51, and CHEK2 expression were all negative, and a probability of over 87% if all three were positive or RAD51 nuclear expression was positive and CHEK2 expression was negative (the latter independent of RAD51 cytoplasmic expression, Table 6).
DISCUSSION
In previous analyses of different markers related to, among others, proliferation, hormone receptors, cell cycle, apoptosis, cell adhesion, and basal markers, we found differences between BRCA1 tumors and BRCA2 tumors and between BRCA1 tumors and non-BRCA1/2 tumors. However, BRCA2 tumors and non-BRCA1/2 tumors always presented a similar immunohistochemical pattern.14,15,17,41 These findings suggested that the differences between BRCA2 tumors and non-BRCA1/2 tumors were caused by other mechanisms. For this reason, we investigated markers for the DNA repair pathway, in which BRCA1 and BRCA2 are involved. By using two tissue microarrays that include 33 BRCA1, 24 BRCA2, 46 non-BRCA1/2, and 104 sporadic patient cases, we analyzed a number of markers related to DSB repair by HR to establish the immunohistochemical profile of familial breast tumors and to better understand the biology of these cases. Our results show important differences in some markers involved in this pathway that have been validated with a new independent series of 150 familial tumors and 184 breast cancer tumors from other countries. The most important differences are that BRCA2 tumors demonstrate more cytoplasmic and less nuclear RAD51 staining, and increased CHEK2 staining. This result can help to differentiate BRCA2 tumors from non-BRCA1/2 tumors
Concerning CHEK2, we have found that this marker is expressed in the majority of BRCA1 tumors (66%) and BRCA2 tumors (56.5%), while this is only observed in approximately 25% of non-BRCA1/2 tumors and sporadic tumors (P < .001; Table 5). The frequency of positive cases among non-BRCA1/2 tumors and sporadic tumors observed in this study is much lower than reported in previous studies25,26,40 and could be because of the different dilution (1:25 v 1:2000) and detection of the primary antibody (DAKO EnVision system v VECTASTAIN Elite ABC kit [Vector Laboratories, Inc, Burlingame, CA]) that we have used. CHEK2 activates BRCA1 directly and BRCA2 indirectly in response to DNA damage. So, the increase in the number of positive patient cases for CHEK2 in the BRCA1 and BRCA2 defective tumors is consistent with CHEK2 acting upstream of BRCA1 and BRCA2 in response to DNA DSBs. This association is not observed in non-BRCA1/2 tumors and sporadic tumors, suggesting that CHEK2 level could be determined by the feedback from the levels of BRCA1 and BRCA2. So, if one of these targets were altered, they would not respond to CHEK2 signal activation.
One of the most significant findings was the relationship between RAD51 and BRCA2-associated tumors. We found a small percentage of BRCA2 tumors with RAD51 nuclear expression (10%), and a high percentage of cases (50%) with high cytoplasmic expression when compared with the other groups of tumors studied (P < .001; Table 5).
This lack of expression of RAD51 in the nucleus was expected, since BRCA2 directly interacts with RAD51 and plays a critical role in controlling the actions of RAD51 in the DNA repair process at both the focus formation and the DNA binding level.42 The role of BRCA2 in reparation is to transport RAD51 to the nuclear site where its action is required.20 The vast majority of BRCA2 mutants are nonfunctional and they do not translocate to the nucleus. Therefore, the truncated BRCA2 protein could retain its ability to bind RAD51 if the mutation does not affect the site of RAD51 interaction, but RAD51 would remain in the cytoplasm. This would explain why 50% of our BRCA2 tumors present high cytoplasmic expression levels, while in all other breast tumor sub-types that we studied, this percentage is around 15% (P < .001 for all comparisons). So, using a total of 74 BRCA1, 71 BRCA2, 108 non-BRCA1/2, and 288 sporadic tumors, we confirmed that RAD51 is a marker associated with the presence of a germ-line BRCA2 mutation. However, in approximately 10% of our BRCA2 tumors, RAD51 was found in the nucleus, indicating the existence of a BRCA2-independent RAD51 transport route to the nucleus, as previously suggested.43 We analyzed whether, in these patient cases, any correlation exists with the type of mutation, and we observed that there is a random distribution of mutations along the BRCA2 gene, discarding any relation with the site of RAD51 interaction (data not shown).
The findings of RAD51 expression in BRCA2 tumors are significantly different from those in the rest of familial tumors and sporadic tumors that show presence of nuclear expression and absence of cytoplasmic expression in 15% to 30% versus 9% in BRCA2 tumors. Using RAD51 and CHEK2 we applied a multivariate logistic model that can predict the status of BRCA2 and non-BRCA1/2 mutation carriers, with an accuracy of at least 76% in approximately 60% of cases. This model has great importance since at presents a number of different markers such as ER, PR, Ki-67, p53, and CK5/6 allows us to differentiate BRCA1 tumors from BRCA2 tumors and non-BRCA1/2 tumors, but until now it was not possible to differentiate BRCA2 tumors from non-BRCA1/2 tumors. Furthermore, we have established an immunohistochemical profile for BRCA2 tumors represented by negative RAD51 nuclear expression, positive RAD51 cytoplasm, and CHEK2 expression (Table 6). These data could help to select candidate families for genetic studies, especially those families with a male breast cancer.
Another interesting finding related to RAD51 was that, while a high frequency of BRCA1 tumors and non-BRCA1/2 tumors were positive for nuclear RAD51 (35% and 40%, respectively), sporadic patient cases presented a significantly lower percentage of nuclear expression (20%; P < .001; Table 5). Because we could consider sporadic patient cases as a model for tumors not associated with alterations in DNA repair genes, the similar pattern of frequent RAD51 positivity in BRCA1 tumors and non-BRCA1/2 tumors may reflect that this group or part of this group, present a deficient response to DNA damage.
Nuclear expression of RAD50 was lower in BRCA1-associated tumors (40%) than in the other familial patient cases and sporadic patient cases (70%; P < .001; Table 5). The BRCA1 protein foci, which appear after ionizing radiation, colocalize in a subset of the cell population with the nuclear foci formed by the MRE11/RAD50/NBS1 complex.27,28 This finding suggests that BRCA1 has a role in the cellular response to DNA DSBs through the NHEJ.44 How these foci are formed, and what draws BRCA1 to them is currently unknown. Because BRCA1 and the MRE11/RAD50/NBS1 complex are components of a super complex of DNA repair proteins,27 it is possible that BRCA1 mutations produce alterations in the expression of proteins that bind directly, like RAD50.28 So, these alterations could compromise the DNA repair through the NHEJ pathway,44,45 although there are some contradictory results.46
We have found similar levels of expression for ATM and XRCC3 proteins among the different tumors, suggesting that the expression of either gene is not affected in the majority of sporadic and familial breast cancers. Finally, with regard to the high positivity of the proliferating cell nuclear antigen PCNA in BRCA1 tumors, the data are consistent with those observed for other proliferating antigens described in these tumors, such as Ki67.14,47
In summary, we have shown by immunohistochemistry that CHEK2 is more frequently positive in the BRCA1 tumors and BRCA2 tumors, than in non-BRCA1/2 tumors and sporadic tumors. We have also found that BRCA2 may be necessary for translocating RAD51 from the cytoplasm to the nucleus. When BRCA2 is mutated, RAD51 is accumulated in the cytoplasm (high level expression) and only a few patient cases show nuclear expression (10%). This situation defines a main profile for BRCA2 tumors characterized by a high expression of RAD51 in the cytoplasm and negative expression in the nucleus that is different from non-BRCA1/2 tumors. All together, our results allow us to conclude that an immunohistochemical assay with RAD51 and CHEK2 antibodies could help to differentiate BRCA2 tumors from non-BRCA1/2 tumors. This information could assist in the selection of candidate families for genetic studies, especially for families with male breast cancer. Although more studies are necessary.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank the Spanish National Tumor Bank Network for providing samples, Ana Díez, Maria Jesús Acu?a, Raquel Pajares, and Nina Puolakka for their technical assistance and Laura Salonen, Carl Blomqvist, MD, and Louis R. Bégin for their kind help.
NOTES
Supported in part by projects from the Spanish Ministry of Health (FIS G03-179), the Spanish Ministry of Science and Technology (SAF01-075 and SAF03-02497) and the Susan G. Komen Breast Cancer Foundation (BCTR 01-97). E.H. is a recipient of a research grant from the Fondo de Investigaciones Sanitarias.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Wooster R, Bignell G, Lancaster J, et al: Identification of the breast cancer susceptibility gene BRCA2. Nature 378:789-792, 1995
Miki Y, Swensen J, Shattuck-Eidens D, et al: A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266:66-71, 1994
Ford D, Easton DF, Stratton M, et al: Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62:676-689, 1998
Easton DF, Bishop DT, Ford D, et al: Genetic linkage analysis in familial breast and ovarian cancer: Results from 214 families. The Breast Cancer Linkage Consortium. Am J Hum Genet 52:678-701, 1993
Shih HA, Couch FJ, Nathanson KL, et al: BRCA1 and BRCA2 mutation frequency in women evaluated in a breast cancer risk evaluation clinic. J Clin Oncol 20:994-999, 2002
Hakansson S, Johannsson O, Johansson U, et al: Moderate frequency of BRCA1 and BRCA2 germ-line mutations in Scandinavian familial breast cancer. Am J Hum Genet 60:1068-1078, 1997
Serova OM, Mazoyer S, Puget N, et al: Mutations in BRCA1 and BRCA2 in breast cancer families: Are there more breast cancer-susceptibility genes? Am J Hum Genet 60:486-495, 1997
Meindl A: Comprehensive analysis of 989 patients with breast or ovarian cancer provides BRCA1 and BRCA2 mutation profiles and frequencies for the German population. Int J Cancer 97:472-480, 2002
van Beers EH, van Welsem T, Wessels LF, et al: Comparative genomic hybridization profiles in human BRCA1 and BRCA2 breast tumors highlight differential sets of genomic aberrations. Cancer Res 65:822-827, 2005
Osorio A, Barroso A, Martinez B, et al: Molecular analysis of the BRCA1 and BRCA2 genes in 32 breast and/or ovarian cancer Spanish families. Br J Cancer 82:1266-1270, 2000
Diez O, Osorio A, Duran M, et al: Analysis of BRCA1 and BRCA2 genes in Spanish breast/ovarian cancer patients: A high proportion of mutations unique to Spain and evidence of founder effects. Hum Mutat 22:301-312, 2003
Lakhani SR, Easton DF, Stratton MR, et al: Pathology of familial breast cancer: Differences between breast cancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases. Breast Cancer Linkage Consortium. Lancet 349:1505-1510, 1997
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 20:2310-2318, 2002
Palacios J, Honrado E, Osorio A, et al: Immunohistochemical characteristics defined by tissue microarray of hereditary breast cancer not attributable to BRCA1 or BRCA2 mutations: Differences from breast carcinomas arising in BRCA1 and BRCA2 mutation carriers. Clin Cancer Res 9:3606-3614, 2003
Palacios J, Honrado E, Osorio A, et al: Phenotypic characterization of BRCA1 and BRCA2 tumors based in a tissue microarray study with 37 immunohistochemical markers. Breast Cancer Res Treat 90:5-14, 2005
Palacios J, Honrado E, Osorio A, et al: Re: Germline BRCA1 mutations and a basal epithelial phenotype in breast cancer. J Natl Cancer Inst 96:712-714, 2004
Foulkes WD, Stefansson IM, Chappuis PO, et al: Germline BRCA1 mutations and a basal epithelial phenotype in breast cancer. J Natl Cancer Inst 95:1482-1485, 2003
Lakhani SR, Gusterson BA, Jacquemier J, et al: The pathology of familial breast cancer: Histological features of cancers in families not attributable to mutations in BRCA1 or BRCA2. Clin Cancer Res 6:782-789, 2000
Cipollini G, Tommasi S, Paradiso A, et al: Genetic alterations in hereditary breast cancer. Ann Oncol 15 Suppl 1:I7–I13, 2004
Venkitaraman AR: Cancer susceptibility and the functions of BRCA1 and BRCA2. Cell 108:171-182, 2002
Bertwistle D, Ashworth A: Functions of the BRCA1 and BRCA2 genes. Curr Opin Genet Dev 8:14-20, 1998
Samuel T, Weber HO, Funk JO: Linking DNA damage to cell cycle checkpoints. Cell Cycle 1:162-168, 2002
Bakkenist CJ, Kastan MB: DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421:499-506, 2003
Lee JS, Collins KM, Brown AL, et al: HCds1-mediated phosphorylation of BRCA1 regulates the DNA damage response. Nature 404:201-204, 2000
Vahteristo P, Bartkova J, Eerola H, et al: A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer. Am J Hum Genet 71:432-438, 2002
Oldenburg RA, Kroeze-Jansema K, Kraan J, et al: The CHEK2*1100delC variant acts as a breast cancer risk modifier in non-BRCA1/BRCA2 multiple-case families. Cancer Res 63:8153-8157, 2003
Wang Y, Cortez D, Yazdi P, et al: BASC, a super complex of BRCA1-associated proteins involved in the recognition and repair of aberrant DNA structures. Genes Dev 14:927-939, 2000
Zhong Q, Chen CF, Li S, et al: Association of BRCA1 with the hRad50-hMre11-p95 complex and the DNA damage response. Science 285:747-750, 1999
Yamaguchi-Iwai Y, Sonoda E, Sasaki MS, et al: Mre11 is essential for the maintenance of chromosomal DNA in vertebrate cells. Embo J 18:6619-6629, 1999
Chen J, Silver DP, Walpita D, et al: Stable interaction between the products of the BRCA1 and BRCA2 tumor suppressor genes in mitotic and meiotic cells. Mol Cell 2:317-328, 1998
Scully R, Chen J, Ochs RL, et al: Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage. Cell 90:425-435, 1997
Schild D, Lio YC, Collins DW, et al: Evidence for simultaneous protein interactions between human Rad51 paralogs. J Biol Chem 275:16443-16449, 2000
Rodriguez-Lopez R, Osorio A, Ribas G, et al: The variant E233G of the RAD51D gene could be a low-penetrance allele in high-risk breast cancer families without BRCA1/2 mutations. Int J Cancer 110:845-849, 2004
Campos B, Diez O, Domenech M, et al: BRCA2 mutation analysis of 87 Spanish breast/ovarian cancer families. Ann Oncol 12:1699-1703, 2001
Eerola H, Heikkila P, Tamminen A, et al: Histopathological features of breast tumours in BRCA1, BRCA2, and in mutation negative breast cancer families. Breast Cancer Res 7:R93–R100, 2005
Syrjakoski K, Kuukasjarvi T, Waltering K, et al: BRCA2 mutations in 154 Finnish male breast cancer patients. Neoplasia 6:541-545, 2004
Goffin JR, Chappuis PO, Begin LR, et al: Impact of germline BRCA1 mutations and overexpression of p53 on prognosis and response to treatment following breast carcinoma: 10-year follow up data. Cancer 97:527-536, 2003
Prevalence and penetrance of BRCA1 and BRCA2 mutations in a population-based series of breast cancer cases. Anglican Breast Cancer Study Group. Br J Cancer 83:1301-1308, 2000
Martinez-Ferrandis JI, Vega A, Chirivella I, et al: Mutational analysis of BRCA1 and BRCA2 in Mediterranean Spanish women with early-onset breast cancer: Identification of three novel pathogenic mutations. Hum Mutat 22:417-418, 2003
Kilpivaara O, Bartkova J, Eerola H, et al: Correlation of CHEK2 protein expression and c. 1100delC mutation status with tumor characteristics among unselected breast cancer patients. Int J Cancer 113:575-580, 2005
Chappuis PO, Donato E, Goffin JR, et al: Cyclin E expression in breast cancer: Predicting germline BRCA1 mutations, prognosis and response to treatment. Ann Oncol 16:735-742, 2005
Tarsounas M, Davies AA, West SC: RAD51 localization and activation following DNA damage. Philos Trans R Soc Lond B Biol Sci 359:87-93, 2004
Tarsounas M, Davies D, West SC: BRCA2-dependent and independent formation of RAD51 nuclear foci. Oncogene 22:1115-1123, 2003
Bau DT, Fu YP, Chen ST, et al: Breast cancer risk and the DNA double-strand break end-joining capacity of nonhomologous end-joining genes are affected by BRCA1. Cancer Res 64:5013-5019, 2004
Zhong Q, Boyer TG, Chen PL, et al: Deficient nonhomologous end-joining activity in cell-free extracts from Brca1-null fibroblasts. Cancer Res 62:3966-3970, 2002
Wang H, Zeng ZC, Bui TA, et al: Nonhomologous end-joining of ionizing radiation-induced DNA double-stranded breaks in human tumor cells deficient in BRCA1 or BRCA2. Cancer Res 61:270-277, 2001
Armes JE, Trute L, White D, et al: Distinct molecular pathogeneses of early-onset breast cancers in BRCA1 and BRCA2 mutation carriers: A population-based study. Cancer Res 59:2011-2017, 1999(Emiliano Honrado, Ana Oso)
Laboratory of Cancer Genetics, Tampere University Hospital and Institute of Medical Technology, University of Tampere, Tampere, Finland
Genetic Pathology Evaluation Centre, British Columbia Cancer Agency, University of British Columbia, Vancouver, BC, Canada
Department of Pathology, Department of Obstetrics and Gynecology, Helsinki University Central Hospital, Helsinki, Finland
Program in Cancer Genetics, Departments of Oncology and Human Genetics, McGill University, Montreal, Quebec, Canada
ABSTRACT
PURPOSE: Morphologic and immunohistochemical studies of familial breast cancers have identified specific characteristics associated with BRCA1 mutation-associated tumors when compared with BRCA2 and non-BRCA1/2 tumors, but have not identified differences between BRCA2 and non-BRCA1/2 tumors. Because BRCA1 and BRCA2 genes participate in the DNA repair pathway, we have performed an immunohistochemical study with markers related to this pathway to establish the profile of the three groups.
MATERIALS AND METHODS: We have studied two tissue microarrays that include 103 familial and 104 sporadic breast tumors, with a panel of DNA repair markers including ATM, CHEK2, RAD51, RAD50, XRCC3, and proliferating cell nuclear antigen.
RESULTS: We found more frequent expression of CHEK2 in BRCA1 and BRCA2 tumors than in non-BRCA1/2 and sporadic tumors. We found absence of nuclear expression and presence of cytoplasmic expression of RAD51 in BRCA2 tumors that differentiate them from other familial tumors. We validated these results with a new series of patient cases. The final study with 253 familial patient cases (74 BRCA1, 71 BRCA2, 108 non-BRCA1/2), and 288 sporadic patient cases, has allowed us to confirm our preliminary results. Because BRCA2 tumors present a specific immunohistochemical profile for RAD51 and CHEK2 markers that is different from non-BRCA1/2 tumors, we have built a multivariate model with these markers that distinguish both tumors with an estimated probability of at least 76%.
CONCLUSION: Our results suggest that BRCA2 tumors demonstrate more cytoplasmic and less nuclear RAD51 staining, and increased CHEK2 staining. This pattern may distinguish BRCA2 from familial non-BRCA1/2 tumors.
INTRODUCTION
Approximately 5% to 10% of all breast cancers are hereditary and may be attributable to mutations in several high penetrance susceptibility genes, of which only two have been identified: BRCA1 and BRCA2.1,2
Although earlier estimates suggested that BRCA1 and BRCA2 mutations were responsible for 75% of site-specific breast cancer families and the majority of breast and ovarian cancer families,3,4 recent data indicate that these percentages may have been overestimated. In fact, the percentage of high-risk families associated with mutations in these genes seems to be around 25%5-9 in all groups tested, including the Spanish population.10,11
Because genetic testing for BRCA1 and BRCA2 is expensive and time-consuming because of the large size of both genes and the low percentage of altered cases, it is important to find clinical or pathologic features that could suggest or exclude the presence of BRCA1 or BRCA2 mutations in a given patient. In this respect, recent studies analyzing pathologic and immunohistochemical characteristics have shown that BRCA1 tumors tend to be of a higher grade than BRCA2 tumors. BRCA1 tumors also tend to be negative for hormone receptors and p53 positive, while the majority of BRCA2 tumors are hormone receptor-positive and p53 negative.12-14 In addition, we recently analyzed markers for cell cycle, apoptosis, and adhesion molecules in BRCA1 and BRCA2 tumors. We found striking differences between both types of tumors in type D cyclins and their associated CDK4 and CDKIs (p16, p21, and p27), which were downregulated in BRCA1 with respect to BRCA2 carcinomas.15 Basal markers were also shown to be specific of a subset of BRCA1 carcinomas.16,17 In contrast, non-BRCA1/2 tumors present similar immunohistologic anomalies to those of BRCA2 tumors with regard to proliferation, hormone receptors, cell cycle, and apoptosis when compared with BRCA1 tumors. They also cannot be distinguished from BRCA2 tumors using such markers, confirming the similarities between both types of tumors pointed out previously.14,18
One of the main functions of BRCA1 and BRCA2 is DNA double-strand break (DSB) repair by homologous recombination (HR).19-21 BRCA1 is activated by the protein kinase ATM, that initiates cell cycle changes after DNA damage.22 ATM phosphorylates numerous downstream targets, including p53, MDM2, CHEK2, NBS1, RAD9, and BRCA1.23 CHEK2 phosphorylates BRCA1, and this phosphorylation is required for the release of BRCA1 from CHEK2 and is important for cellular response.24 The CHEK2 1100delC mutation, is associated with familial breast cancer. Tumors carrying this mutation often show reduced or absent CHEK2 protein expression.25,26
BRCA1 interacts with the MRE11/RAD50/NBS1 protein complex,27,28 which is involved in both nonhomologous end joining (NHEJ) and HR in yeast and vertebrate cells.29 BRCA1 also promotes RAD51 homologous repair at a nuclear level through BRCA2. Exposure of cells to agents that induce DNA damage and replication arrest causes BRCA2 to colocalize with RAD51 at intranuclear sites that also contain the replication protein proliferating cell nuclear antigen (PCNA).30,31
The formation of RAD51 foci in response to DNA damage is dependent on BRCA2 and a series of proteins known as the RAD51 paralogs (RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3), a multiprotein complex that forms a heterodimer with each of the gene products playing a different role in DSB repair by HR.32 Some of these genes have already been proposed as breast cancer susceptibility genes.33
Currently, we do not know the functions of other susceptibility genes responsible for non-BRCA1/2 cases. To define possible differences between BRCA-associated tumors, and especially between BRCA2 and familial non-BRCA1/2 tumors that could be associated to the DNA repair pathway, we assessed the expression of some of the proteins covering several steps in this pathway by immunohistochemistry. Our results show differences in the expression of some of these markers that can help to distinguish different types of familial tumors and specifically BRCA2 tumors.
MATERIALS AND METHODS
Study Patients
Patients were obtained from three centers in Spain: the Spanish National Cancer Center, the Fundación Jimenez Díaz in Madrid, and the Hospital Sant Pau in Barcelona. They were selected from high-risk breast cancer families with at least three women affected with breast cancer and/or ovarian cancer, one of them younger than 50 years of age, or a male with breast cancer. The index case of each family was screened for mutations in the BRCA1 and BRCA2 genes by a combination of SSCP, CSGE, and sequencing techniques. Some of these results have been reported previously.10,11,34
Immunohistochemical analysis was performed on tumors distributed across two tissue microarrays (TMAs). Data from one of the TMAs were recently published.14 The second TMA was built following the same methodology. Tumors included were 33 invasive ductal carcinomas (IDC) from patients with mutations in BRCA1 (mean age, 40 years), 24 tumors with mutations in BRCA2 (19 IDC, one in situ ductal carcinoma, and four invasive lobular carcinomas; mean age, 46 years), and 46 familial non-BRCA1/2 tumors (41 IDC, three ductal carcinoma-in-situ, and two invasive lobular carcinomas; mean age, 47 years). All tumors were duplicated in the TMA. We also included normal breast and tonsil tissues as an internal control.
In addition to familial patient cases, a consecutive series of women with breast cancer was studied: 104 sporadic IDC with an age distribution similar to that of familial patient cases (mean age, 45 years).
We have validated the results obtained for CHEK2, RAD51, and RAD50 with a new independent series of patients with breast cancer from other countries. These include 150 familial patient cases (41 BRCA1, 47 BRCA2, 62 non-BRCA1/2), and 184 sporadic breast tumors, the majority of them previously published.35-37 The distribution and the origin of all patient cases are summarized in Table 1.
We have genotyped all the familial patient cases for BRCA1 and BRCA2 genes but not the sporadic patient cases. The consecutive series of sporadic patient cases was not tested because the frequency of mutations in BRCA1 and BRCA2 in these women is low (around 2%).38,39
Immunohistochemical Analysis
Immunohistochemical staining was performed by the DAKO EnVision system (DAKO, Copenhagen, Denmark) with a heat-induced, antigen retrieval step. Sections from the tissue microarray were immersed in boiling 10 mmol/L sodium citrate at pH 6.5 for 2 minutes in a pressure cooker, adding EDTA for RAD51. For ATM, proteinase K was added at a dilution of 1:40 for 10 minutes at room temperature. Antibodies, dilutions, and suppliers are listed in Table 2.
Morphological Studies and Immunohistochemical Scoring
Before tissue microarray construction, total sections of each hematoxylin-eosin stained tumor were evaluated and classified according to the WHO classification.
Two pathologists (E.H. and J.P.) simultaneously evaluated the immunohistochemical staining to avoid observer subjectivity as much as possible.
Taking into account that all immunohistochemical markers used are related with DNA repair and have to be active in the nucleus, we only considered the percentage of stained nuclei in each tumor, independent of the intensity since the variation in intensity was low. In addition, we used cutoff points to establish dichotomous variables to clarify results. For nuclear expression we used the mean (among all patient cases) of the percentage of stained cells as the cutoff point. When the percentage of stained cells was 10% or higher, we considered the tumor as positive for ATM, RAD51, and XRCC3. We considered 80% or higher as positive for RAD50 and PCNA because positive patient cases had stained 100% or near 100% of cells. The mean was 95% (SD = 8%) for positive patient cases and 70% (SD = 14%) for negative patient cases. For CHEK2 we used the cutoff of 60% defined by Kilpivaara et al.40 The mean was 80% (SD = 8%) for positive patient cases and 15% (SD = 16%) for negative patient cases. For RAD50 we did not find any patient case without staining. There was variation in nuclear staining intensity with respect to normal tissue, and we considered a lower intensity than normal tissue as negative, even if the percentage was higher than 80%.
Concerning the expression in normal breast and tonsil control tissues, expression of CHEK2 was seen in 60% of normal ductal cells, and of RAD50 and PCNA in 80% of cells. There was no staining for RAD51, ATM, nor XRCC3. For tonsil, RAD50 and PCNA staining was observed in 80%, and XRCC3, CHEK2, and RAD51 staining in 60% of germinal center lymphocytes and the basal layers of the squamous epithelium. ATM staining was seen in 80% of B and T lymphocytes outside the germinal center. For all markers the staining was in the nucleus.
For RAD51 and XRCC3, cytoplasmic staining was also considered because it was present in many tumors. Intensity was scored as either negative (non- and low-staining) or positive (high staining) (Fig 1).
Samples from other series (Table 1) were studied for RAD51, CHEK2, and RAD50 markers, and analyzed by the same pathologists (E.H. and J.P.) who were blinded to the mutation status as in the first set of tumors.
Tumors with poor correlation between duplicates (1% to 2%) were excluded from the analyses. Cores were missing for 5% to 15% of patient cases. Some cores (around 1%) were stromal tissue without breast tumor in both cores, because of errors in the selection of the tumor area.
Statistical Analysis
The Pearson's 2 test was performed to determine differences between groups using SPSS for Windows (SPSS Inc, Chicago, IL). Multivariate analyses of associations between tumor type (BRCA2 v non-BRCA1/2) and markers were carried out by unconditional logistic regression using Stata 8 (StataCorp, College Station, TX). Interactions were assessed using the log-likelihood ratio significance test. Statistical significance was declared for nominal P values less than .05.
RESULTS
Table 3 summarizes the results obtained with the group of DNA repair antibodies. For familial tumors, ATM presented similar expression in all groups, while the expression of CHEK2 was significantly more frequent among BRCA1 tumors (60.7%) and BRCA2 tumors (56.5%) than in non-BRCA1/2 tumors (21.6%) (P < .006).
Nuclear staining for RAD50 was significantly less frequent in BRCA1-associated tumors (23.3%) than in BRCA2 tumors (55.6%) and familial non-BRCA1/2 tumors (60.0%), while, for the PCNA, a greater percentage of positive tumors was found in BRCA1 tumors.
We did not observe nuclear expression of RAD51 in BRCA2 tumors, but cytoplasmic expression was highly positive in 57.1% of the patient cases. For BRCA1 tumors and non-BRCA1/2 tumors, the frequencies of nuclear and cytoplasmic expression were similar, and significantly different from BRCA2 tumors 22.6% and 30.4% versus 0% (P < .02) for nuclear staining and 29.0% and 23.9% versus 57.1% (P < .04) for high cytoplasmic staining.
No significant differences were found in the expression of XRCC3.
We analyzed these markers in our sample of sporadic cases (Table 3) and found similar frequencies of expression as obtained for non-BRCA1/2 tumors except a lower frequency of RAD51 nuclear and cytoplasmic staining and a lower frequency of XRCC3 nuclear staining.
We validated the results obtained for RAD51, CHEK2, and RAD50 with a new independent series of hereditary and sporadic tumor samples. Tumors from 41 BRCA1 patients, 47 BRCA2 patients, 62 non-BRCA1/2 patients, and 184 sporadic patient cases were analyzed for these markers (Table 4). We obtained results similar to those previously found: for CHEK2 we observed a higher percentage of BRCA1 tumors and BRCA2 tumors with expression than in non-BRCA1/2 tumors and sporadic tumors (P < .005 for all comparisons). An association was also observed for the expression of RAD50 that was significantly less frequent in BRCA1-associated tumors than in BRCA2 tumors, non-BRCA1/2 tumors, and sporadic tumors (P < .05). Finally, a low percentage of BRCA2 tumors with nuclear staining (15%) for RAD51, and a high percentage with high cytoplasmic staining (47%) was found. Although, in the first case the differences were significant compared with all other sub-types studied (P < .003 for all comparisons, Table 4), for cytoplasmic staining there was only a nonsignificant tendency when compared with non-BRCA1/2 tumors. This was as a result of the fact that some non-BRCA1/2 patient cases presented not only cytoplasmic but also nuclear expression (data not shown).
Table 5 presents results from analyses of the combined data from the original series and the validation set (253 familial tumors and 288 sporadic tumors in total). We considered all four combinations of staining for cytoplasmic and nuclear RAD51. Among those with negative nuclear staining, high cytoplasmic expression of RAD51 was found more frequently in BRCA2 tumors (48.5%) versus BRCA1 tumors (16.7%), non-BRCA1/2 tumors (16.7%), and sporadic tumors (14.9%; P < .001). Consistent with previous findings, the percentage of tumors with RAD51 nuclear staining, with or without cytoplasmic staining, was significantly higher in BRCA1 and non-BRCA1/2 mutation carriers, than in BRCA2 tumors. In addition, CHEK2 was more often expressed in BRCA1 tumors (66.0%) and BRCA2 tumors (56.4%), than in non-BRCA1/2 tumors and sporadic tumors (23.9% and 25.8%, respectively; P < .001). Finally, the number of positive tumors for RAD50 was, in all cases, significantly lower in BRCA1 tumors (39.0%) than in BRCA2 tumors (72.1%), non-BRCA1/2 tumors (68.9%), and sporadic tumors (72.4%; P < .001). We investigated associations between markers and found that of those that were correlated, only CHEK2 and RAD51 nuclear staining (positively associated, P = .01) were both associated with BRCA2 mutation status. This potential confounding of the association of each with BRCA2 versus non-BRCA1/2 tumors was considered in the multivariate analysis, which included both these markers.
Among 92 non-BRCA1/2 tumors and 54 BRCA2 tumors with results for all markers mutation status was modeled using multivariate unconditional logistic regression to differentiate the two types of tumors (Table 6). For this analysis we included RAD51 nuclear and cytoplasmic staining and CHEK2 expression, since these markers significantly differentiated BRCA2 tumors and non-BRCA1/2 tumors in univariate analyses (data not shown). An interaction between RAD51 nuclear and cytoplasmic staining was also included and was statistically significant (P = .05), consistent with the results from Table 5. The logistic model estimated a probability of 81% (95% CI, 63% to 91%) of being a BRCA2 tumor if RAD51 nuclear expression was negative and the expression of cytoplasmic RAD51 and CHEK2 were both positive. Furthermore, the model estimated a probability of 76% (95% CI, 63% to 86%) of being a non-BRCA1/2 tumor if nuclear expression of RAD51, cytoplasmic expression of RAD51, and CHEK2 expression were all negative, and a probability of over 87% if all three were positive or RAD51 nuclear expression was positive and CHEK2 expression was negative (the latter independent of RAD51 cytoplasmic expression, Table 6).
DISCUSSION
In previous analyses of different markers related to, among others, proliferation, hormone receptors, cell cycle, apoptosis, cell adhesion, and basal markers, we found differences between BRCA1 tumors and BRCA2 tumors and between BRCA1 tumors and non-BRCA1/2 tumors. However, BRCA2 tumors and non-BRCA1/2 tumors always presented a similar immunohistochemical pattern.14,15,17,41 These findings suggested that the differences between BRCA2 tumors and non-BRCA1/2 tumors were caused by other mechanisms. For this reason, we investigated markers for the DNA repair pathway, in which BRCA1 and BRCA2 are involved. By using two tissue microarrays that include 33 BRCA1, 24 BRCA2, 46 non-BRCA1/2, and 104 sporadic patient cases, we analyzed a number of markers related to DSB repair by HR to establish the immunohistochemical profile of familial breast tumors and to better understand the biology of these cases. Our results show important differences in some markers involved in this pathway that have been validated with a new independent series of 150 familial tumors and 184 breast cancer tumors from other countries. The most important differences are that BRCA2 tumors demonstrate more cytoplasmic and less nuclear RAD51 staining, and increased CHEK2 staining. This result can help to differentiate BRCA2 tumors from non-BRCA1/2 tumors
Concerning CHEK2, we have found that this marker is expressed in the majority of BRCA1 tumors (66%) and BRCA2 tumors (56.5%), while this is only observed in approximately 25% of non-BRCA1/2 tumors and sporadic tumors (P < .001; Table 5). The frequency of positive cases among non-BRCA1/2 tumors and sporadic tumors observed in this study is much lower than reported in previous studies25,26,40 and could be because of the different dilution (1:25 v 1:2000) and detection of the primary antibody (DAKO EnVision system v VECTASTAIN Elite ABC kit [Vector Laboratories, Inc, Burlingame, CA]) that we have used. CHEK2 activates BRCA1 directly and BRCA2 indirectly in response to DNA damage. So, the increase in the number of positive patient cases for CHEK2 in the BRCA1 and BRCA2 defective tumors is consistent with CHEK2 acting upstream of BRCA1 and BRCA2 in response to DNA DSBs. This association is not observed in non-BRCA1/2 tumors and sporadic tumors, suggesting that CHEK2 level could be determined by the feedback from the levels of BRCA1 and BRCA2. So, if one of these targets were altered, they would not respond to CHEK2 signal activation.
One of the most significant findings was the relationship between RAD51 and BRCA2-associated tumors. We found a small percentage of BRCA2 tumors with RAD51 nuclear expression (10%), and a high percentage of cases (50%) with high cytoplasmic expression when compared with the other groups of tumors studied (P < .001; Table 5).
This lack of expression of RAD51 in the nucleus was expected, since BRCA2 directly interacts with RAD51 and plays a critical role in controlling the actions of RAD51 in the DNA repair process at both the focus formation and the DNA binding level.42 The role of BRCA2 in reparation is to transport RAD51 to the nuclear site where its action is required.20 The vast majority of BRCA2 mutants are nonfunctional and they do not translocate to the nucleus. Therefore, the truncated BRCA2 protein could retain its ability to bind RAD51 if the mutation does not affect the site of RAD51 interaction, but RAD51 would remain in the cytoplasm. This would explain why 50% of our BRCA2 tumors present high cytoplasmic expression levels, while in all other breast tumor sub-types that we studied, this percentage is around 15% (P < .001 for all comparisons). So, using a total of 74 BRCA1, 71 BRCA2, 108 non-BRCA1/2, and 288 sporadic tumors, we confirmed that RAD51 is a marker associated with the presence of a germ-line BRCA2 mutation. However, in approximately 10% of our BRCA2 tumors, RAD51 was found in the nucleus, indicating the existence of a BRCA2-independent RAD51 transport route to the nucleus, as previously suggested.43 We analyzed whether, in these patient cases, any correlation exists with the type of mutation, and we observed that there is a random distribution of mutations along the BRCA2 gene, discarding any relation with the site of RAD51 interaction (data not shown).
The findings of RAD51 expression in BRCA2 tumors are significantly different from those in the rest of familial tumors and sporadic tumors that show presence of nuclear expression and absence of cytoplasmic expression in 15% to 30% versus 9% in BRCA2 tumors. Using RAD51 and CHEK2 we applied a multivariate logistic model that can predict the status of BRCA2 and non-BRCA1/2 mutation carriers, with an accuracy of at least 76% in approximately 60% of cases. This model has great importance since at presents a number of different markers such as ER, PR, Ki-67, p53, and CK5/6 allows us to differentiate BRCA1 tumors from BRCA2 tumors and non-BRCA1/2 tumors, but until now it was not possible to differentiate BRCA2 tumors from non-BRCA1/2 tumors. Furthermore, we have established an immunohistochemical profile for BRCA2 tumors represented by negative RAD51 nuclear expression, positive RAD51 cytoplasm, and CHEK2 expression (Table 6). These data could help to select candidate families for genetic studies, especially those families with a male breast cancer.
Another interesting finding related to RAD51 was that, while a high frequency of BRCA1 tumors and non-BRCA1/2 tumors were positive for nuclear RAD51 (35% and 40%, respectively), sporadic patient cases presented a significantly lower percentage of nuclear expression (20%; P < .001; Table 5). Because we could consider sporadic patient cases as a model for tumors not associated with alterations in DNA repair genes, the similar pattern of frequent RAD51 positivity in BRCA1 tumors and non-BRCA1/2 tumors may reflect that this group or part of this group, present a deficient response to DNA damage.
Nuclear expression of RAD50 was lower in BRCA1-associated tumors (40%) than in the other familial patient cases and sporadic patient cases (70%; P < .001; Table 5). The BRCA1 protein foci, which appear after ionizing radiation, colocalize in a subset of the cell population with the nuclear foci formed by the MRE11/RAD50/NBS1 complex.27,28 This finding suggests that BRCA1 has a role in the cellular response to DNA DSBs through the NHEJ.44 How these foci are formed, and what draws BRCA1 to them is currently unknown. Because BRCA1 and the MRE11/RAD50/NBS1 complex are components of a super complex of DNA repair proteins,27 it is possible that BRCA1 mutations produce alterations in the expression of proteins that bind directly, like RAD50.28 So, these alterations could compromise the DNA repair through the NHEJ pathway,44,45 although there are some contradictory results.46
We have found similar levels of expression for ATM and XRCC3 proteins among the different tumors, suggesting that the expression of either gene is not affected in the majority of sporadic and familial breast cancers. Finally, with regard to the high positivity of the proliferating cell nuclear antigen PCNA in BRCA1 tumors, the data are consistent with those observed for other proliferating antigens described in these tumors, such as Ki67.14,47
In summary, we have shown by immunohistochemistry that CHEK2 is more frequently positive in the BRCA1 tumors and BRCA2 tumors, than in non-BRCA1/2 tumors and sporadic tumors. We have also found that BRCA2 may be necessary for translocating RAD51 from the cytoplasm to the nucleus. When BRCA2 is mutated, RAD51 is accumulated in the cytoplasm (high level expression) and only a few patient cases show nuclear expression (10%). This situation defines a main profile for BRCA2 tumors characterized by a high expression of RAD51 in the cytoplasm and negative expression in the nucleus that is different from non-BRCA1/2 tumors. All together, our results allow us to conclude that an immunohistochemical assay with RAD51 and CHEK2 antibodies could help to differentiate BRCA2 tumors from non-BRCA1/2 tumors. This information could assist in the selection of candidate families for genetic studies, especially for families with male breast cancer. Although more studies are necessary.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank the Spanish National Tumor Bank Network for providing samples, Ana Díez, Maria Jesús Acu?a, Raquel Pajares, and Nina Puolakka for their technical assistance and Laura Salonen, Carl Blomqvist, MD, and Louis R. Bégin for their kind help.
NOTES
Supported in part by projects from the Spanish Ministry of Health (FIS G03-179), the Spanish Ministry of Science and Technology (SAF01-075 and SAF03-02497) and the Susan G. Komen Breast Cancer Foundation (BCTR 01-97). E.H. is a recipient of a research grant from the Fondo de Investigaciones Sanitarias.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Wooster R, Bignell G, Lancaster J, et al: Identification of the breast cancer susceptibility gene BRCA2. Nature 378:789-792, 1995
Miki Y, Swensen J, Shattuck-Eidens D, et al: A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266:66-71, 1994
Ford D, Easton DF, Stratton M, et al: Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62:676-689, 1998
Easton DF, Bishop DT, Ford D, et al: Genetic linkage analysis in familial breast and ovarian cancer: Results from 214 families. The Breast Cancer Linkage Consortium. Am J Hum Genet 52:678-701, 1993
Shih HA, Couch FJ, Nathanson KL, et al: BRCA1 and BRCA2 mutation frequency in women evaluated in a breast cancer risk evaluation clinic. J Clin Oncol 20:994-999, 2002
Hakansson S, Johannsson O, Johansson U, et al: Moderate frequency of BRCA1 and BRCA2 germ-line mutations in Scandinavian familial breast cancer. Am J Hum Genet 60:1068-1078, 1997
Serova OM, Mazoyer S, Puget N, et al: Mutations in BRCA1 and BRCA2 in breast cancer families: Are there more breast cancer-susceptibility genes? Am J Hum Genet 60:486-495, 1997
Meindl A: Comprehensive analysis of 989 patients with breast or ovarian cancer provides BRCA1 and BRCA2 mutation profiles and frequencies for the German population. Int J Cancer 97:472-480, 2002
van Beers EH, van Welsem T, Wessels LF, et al: Comparative genomic hybridization profiles in human BRCA1 and BRCA2 breast tumors highlight differential sets of genomic aberrations. Cancer Res 65:822-827, 2005
Osorio A, Barroso A, Martinez B, et al: Molecular analysis of the BRCA1 and BRCA2 genes in 32 breast and/or ovarian cancer Spanish families. Br J Cancer 82:1266-1270, 2000
Diez O, Osorio A, Duran M, et al: Analysis of BRCA1 and BRCA2 genes in Spanish breast/ovarian cancer patients: A high proportion of mutations unique to Spain and evidence of founder effects. Hum Mutat 22:301-312, 2003
Lakhani SR, Easton DF, Stratton MR, et al: Pathology of familial breast cancer: Differences between breast cancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases. Breast Cancer Linkage Consortium. Lancet 349:1505-1510, 1997
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 20:2310-2318, 2002
Palacios J, Honrado E, Osorio A, et al: Immunohistochemical characteristics defined by tissue microarray of hereditary breast cancer not attributable to BRCA1 or BRCA2 mutations: Differences from breast carcinomas arising in BRCA1 and BRCA2 mutation carriers. Clin Cancer Res 9:3606-3614, 2003
Palacios J, Honrado E, Osorio A, et al: Phenotypic characterization of BRCA1 and BRCA2 tumors based in a tissue microarray study with 37 immunohistochemical markers. Breast Cancer Res Treat 90:5-14, 2005
Palacios J, Honrado E, Osorio A, et al: Re: Germline BRCA1 mutations and a basal epithelial phenotype in breast cancer. J Natl Cancer Inst 96:712-714, 2004
Foulkes WD, Stefansson IM, Chappuis PO, et al: Germline BRCA1 mutations and a basal epithelial phenotype in breast cancer. J Natl Cancer Inst 95:1482-1485, 2003
Lakhani SR, Gusterson BA, Jacquemier J, et al: The pathology of familial breast cancer: Histological features of cancers in families not attributable to mutations in BRCA1 or BRCA2. Clin Cancer Res 6:782-789, 2000
Cipollini G, Tommasi S, Paradiso A, et al: Genetic alterations in hereditary breast cancer. Ann Oncol 15 Suppl 1:I7–I13, 2004
Venkitaraman AR: Cancer susceptibility and the functions of BRCA1 and BRCA2. Cell 108:171-182, 2002
Bertwistle D, Ashworth A: Functions of the BRCA1 and BRCA2 genes. Curr Opin Genet Dev 8:14-20, 1998
Samuel T, Weber HO, Funk JO: Linking DNA damage to cell cycle checkpoints. Cell Cycle 1:162-168, 2002
Bakkenist CJ, Kastan MB: DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421:499-506, 2003
Lee JS, Collins KM, Brown AL, et al: HCds1-mediated phosphorylation of BRCA1 regulates the DNA damage response. Nature 404:201-204, 2000
Vahteristo P, Bartkova J, Eerola H, et al: A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer. Am J Hum Genet 71:432-438, 2002
Oldenburg RA, Kroeze-Jansema K, Kraan J, et al: The CHEK2*1100delC variant acts as a breast cancer risk modifier in non-BRCA1/BRCA2 multiple-case families. Cancer Res 63:8153-8157, 2003
Wang Y, Cortez D, Yazdi P, et al: BASC, a super complex of BRCA1-associated proteins involved in the recognition and repair of aberrant DNA structures. Genes Dev 14:927-939, 2000
Zhong Q, Chen CF, Li S, et al: Association of BRCA1 with the hRad50-hMre11-p95 complex and the DNA damage response. Science 285:747-750, 1999
Yamaguchi-Iwai Y, Sonoda E, Sasaki MS, et al: Mre11 is essential for the maintenance of chromosomal DNA in vertebrate cells. Embo J 18:6619-6629, 1999
Chen J, Silver DP, Walpita D, et al: Stable interaction between the products of the BRCA1 and BRCA2 tumor suppressor genes in mitotic and meiotic cells. Mol Cell 2:317-328, 1998
Scully R, Chen J, Ochs RL, et al: Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage. Cell 90:425-435, 1997
Schild D, Lio YC, Collins DW, et al: Evidence for simultaneous protein interactions between human Rad51 paralogs. J Biol Chem 275:16443-16449, 2000
Rodriguez-Lopez R, Osorio A, Ribas G, et al: The variant E233G of the RAD51D gene could be a low-penetrance allele in high-risk breast cancer families without BRCA1/2 mutations. Int J Cancer 110:845-849, 2004
Campos B, Diez O, Domenech M, et al: BRCA2 mutation analysis of 87 Spanish breast/ovarian cancer families. Ann Oncol 12:1699-1703, 2001
Eerola H, Heikkila P, Tamminen A, et al: Histopathological features of breast tumours in BRCA1, BRCA2, and in mutation negative breast cancer families. Breast Cancer Res 7:R93–R100, 2005
Syrjakoski K, Kuukasjarvi T, Waltering K, et al: BRCA2 mutations in 154 Finnish male breast cancer patients. Neoplasia 6:541-545, 2004
Goffin JR, Chappuis PO, Begin LR, et al: Impact of germline BRCA1 mutations and overexpression of p53 on prognosis and response to treatment following breast carcinoma: 10-year follow up data. Cancer 97:527-536, 2003
Prevalence and penetrance of BRCA1 and BRCA2 mutations in a population-based series of breast cancer cases. Anglican Breast Cancer Study Group. Br J Cancer 83:1301-1308, 2000
Martinez-Ferrandis JI, Vega A, Chirivella I, et al: Mutational analysis of BRCA1 and BRCA2 in Mediterranean Spanish women with early-onset breast cancer: Identification of three novel pathogenic mutations. Hum Mutat 22:417-418, 2003
Kilpivaara O, Bartkova J, Eerola H, et al: Correlation of CHEK2 protein expression and c. 1100delC mutation status with tumor characteristics among unselected breast cancer patients. Int J Cancer 113:575-580, 2005
Chappuis PO, Donato E, Goffin JR, et al: Cyclin E expression in breast cancer: Predicting germline BRCA1 mutations, prognosis and response to treatment. Ann Oncol 16:735-742, 2005
Tarsounas M, Davies AA, West SC: RAD51 localization and activation following DNA damage. Philos Trans R Soc Lond B Biol Sci 359:87-93, 2004
Tarsounas M, Davies D, West SC: BRCA2-dependent and independent formation of RAD51 nuclear foci. Oncogene 22:1115-1123, 2003
Bau DT, Fu YP, Chen ST, et al: Breast cancer risk and the DNA double-strand break end-joining capacity of nonhomologous end-joining genes are affected by BRCA1. Cancer Res 64:5013-5019, 2004
Zhong Q, Boyer TG, Chen PL, et al: Deficient nonhomologous end-joining activity in cell-free extracts from Brca1-null fibroblasts. Cancer Res 62:3966-3970, 2002
Wang H, Zeng ZC, Bui TA, et al: Nonhomologous end-joining of ionizing radiation-induced DNA double-stranded breaks in human tumor cells deficient in BRCA1 or BRCA2. Cancer Res 61:270-277, 2001
Armes JE, Trute L, White D, et al: Distinct molecular pathogeneses of early-onset breast cancers in BRCA1 and BRCA2 mutation carriers: A population-based study. Cancer Res 59:2011-2017, 1999(Emiliano Honrado, Ana Oso)