Introduction

Mutations in the BRCA1 and the BRCA2 genes have been linked with the susceptibility to breast and ovarian cancer (Miki et al. 1994; Wooster et al. 1995; Tavtigian et al. 1996). Mutation carriers of these genes are at high risk of breast and ovarian cancer (Narod et al. 1995; Ford et al. 1998; Thorlacius et al. 1998; Neuhausen 1999; Rebbeck 1999; Struewing et al. 1997; Anglian Breast Cancer Study Group 2000). The two genes have large coding sequences consisting of 48 exons in total, and a large number of mutations and SNPs are reported in the Breast Cancer Information Core (BIC) database (http://research.nhgri.nih.gov/bic/) and the dbSNP database (http://www.ncbi.nlm.nih.gov/SNP/index.html). The majority of mutations described are protein-truncating mutations containing frame-shift mutations and nonsense mutations. In the BIC database, 55.9% of genetic variants are reported as pathogenic mutations containing mainly protein-truncating mutations and a small fraction of pathogenic missense mutations, and 39.3% of variants are categorized as “unclassified variants.” These variants contain coding SNPs (cSNPs) that result in amino-acid substitutions or SNPs located at exon–intron boundaries. We have also reported five protein-truncating mutations and 12 “unclassified variants” that have been found only once in 24 breast cancer families (Sakayori et al. 2003). To elucidate whether detected SNPs are pathogenic mutations or not, it is important to accumulate SNP information (both the type and allele frequency) in patients and the general population.

In this study, we extended our study on the BRCA1 and BRCA2 sequencing project to 55 breast cancer patients from 50 Japanese breast cancer families. We evaluated the detected SNPs by comparing the allele frequency of the SNPs in healthy volunteers.

Subjects and methods

Fifty-five enrolled patients (including 50 probands and five relatives) with a history of breast cancer from 50 unrelated high-risk breast cancer families were selected according to the criteria defined by the Tohoku Familial Cancer Society (Sakayori et al. 2003). An additional 28 Japanese volunteers with no breast cancer family history were also enrolled to analyze the specific BRCA1 and BRCA2 variations detected in the breast cancer patients. We obtained informed consent from all patients and volunteers, and the independent studies were approved by the Familial Cancer Society and the Ethical Committee of Tohoku University Graduate School of Medicine. To identify the sequence variations in the BRCA1 and BRCA2 genes of the familial breast cancer patients, we sequenced approximately 23 kb genomic regions of the BRCA1 (8.4 kb) and the BRCA2 (14.6 kb) containing all coding exons and their franking introns using a method described previously (Sakayori et al. 2003). The identified SNPs found in the familial breast cancer patients were also examined in the genomic DNA from the healthy volunteers by PCR-RFLP analysis using PCR primers, cycle conditions and restriction enzymes for PCR-RFLP analysis (http://www.idac.tohoku.ac.jp/dep/co/data/saka/brca02.htm) or by DNA sequence analysis. We used Genbank (U14680 and U61268 for BRCA1, U43746 and X95152-77 for BRCA2) and the BIC database as the reference sequences of BRCA1 and BRCA2 genes.

Results and discussion

By DNA sequencing analysis of the BRCA1 and BRCA2 genes for 55 patients from 50 unrelated breast cancer families; we detected 55 SNPs (21 SNPs in BRCA1 and 34 SNPs in BRCA2). Among these SNPs, we found nine protein-truncating mutations (four in BRCA1 and five in BRCA2) in ten patients containing six novel mutations that were not found in the BIC database (http://research.nhgri.nih.gov/bic/) (Table 1). The existence of mutations were also confirmed by the stop codon assay in yeast (Ishioka et al. 1997; Sakayori et al. 2003). The percentage of protein-truncating mutations in the examined families was 20% (ten of 50), comparable with several reports describing the frequency of protein-truncating mutations of the two genes in Japanese breast cancer families (Inoue et al. 1995; Takano et al. 1997; Inoue et al. 1997; Ikeda et al. 2001). Although the frequency was also similar to results from Western populations, the number of patients studied in our and other Japanese populations was smaller than that in Western countries. Therefore, it is necessary to study a larger number of patients to clarify the mutation frequency in Japanese familial breast cancer.

Table 1 Sequence variations detected in the BRCA1 and the BRCA2 genes
Full size table

Among the remaining 46 SNPs (17 SNPs in BRCA1 and 29 SNPs in BRCA2), 18 and 31 have been reported in the dbSNP database and in the BIC database, respectively, (Table 1). To evaluate whether the SNPs are also found in Japanese healthy volunteers, all but six SNPs were further examined by PCR-RFLP analysis or by DNA sequence analysis. Twenty-six SNPs were found in the volunteers at least once with different allele frequencies (Table 2). In these 40 SNPs, we predicted that 27 SNPs were common polymorphisms and probably played no direct role in the tumorigenesis of breast cancer. The remaining 13 SNPs were quite low in allele frequency (<0.05) and were not found in the healthy volunteers or the dbSNP database. In addition, there were six SNPs that failed to examine in the volunteer group. In these six SNPs, one was predicted to be common and five to be rare from the allele frequencies in the patients. Overall, we predicted the 18 rare SNPs are candidates of pathogenic mutations and that the remaining 28 were common polymorphisms and probably have no direct roll in tumorigenesis of breast cancer. In the 18 rare SNPs, five (BRCA1-13, BRCA1-17, BRCA1-18, BRCA2-2, BRCA2-3) were located at the exon–intron boundaries and two (BRCA2-23, BRCA2-27) were located in introns far from exon–intron boundaries. Although SNPs at the exon–intron boundaries may affect normal RNA splicing, our RT-PCR analysis showed negative data for splicing abnormalities. In the remaining 11 rare cSNPs, seven (BRCA1-4, BRCA1-20, BRCA2-12, BRCA2-19, BRCA2-25, BRCA2-30 and BRCA2-34) were nonsynonymous substitution resulting in the amino-acid substitutions, and four (BRCA2-8, BRCA2-17, BRCA2-31 and BRCA2-33) were synonymous (silent) changes. These SNPs may directly affect the functions of BRCA1 and BRCA2 proteins or affect normal splicing by acting as possible cryptic splicing sites. Unfortunately, we have failed to clarify this issue mainly because there are no reliable functional assays of either BRCA1 or BRCA2 protein for many nonsynonymous changes.

Table 2 Allele frequencies of SNPs in the BRCA1 and the BRCA2 genes
Full size table

To confirm the pathogenic effect of rare SNPs, both the development of functional assays for these gene products and more intensive SNP analysis including an investigation into whether these SNPs cosegregate with breast cancer onset in families are necessary.