Introduction

Breast cancer is one of the major causes of cancer death among women in the world. Progress in many fields of science has enabled to understand the mechanism of carcinogenesis in breast, but so far there has not been found any clear reason of breast cancer. Two main groups of breast cancer risk factors are genetic and environmental ones, among which the most important are age and breast cancer occurrence in the first and/or second step relatives. In this case, particularly important are mutations in two high penetrance genes BRCA1 and BRCA2 [1, 2], which are associated with the processes of DNA double strand breaks (DSBs) repair. A number of studies show link between hereditary as well as sporadic breast cancer and genetic instability caused by DSBs [3–6]. DSBs can be repaired by nonhomologous end-joining (NHEJ) [7], homologous recombination repair (HRR) [8] and single-strand annealing (SSA) [9]. However, HRR has been found to be a key pathway in human cells for the repair of DSBs [10]. DNA repair by homologous recombination is particularly important during and following DNA replication, when a sister chromatid is present as a template for repair. In eukaryotes, the central homologous recombination protein is RAD51 which catalyses strand transfer between a broken sequence and its undamaged homologue to allow re-synthesis of the damaged region [10–13]. BLM helicase physically and functionally interacts with RAD51 and has been reported to displace RAD51 from the nucleoprotein filament that is responsible for homology searching and strand invasion [14, 15]. BLM activity in relation to RAD51 protein depends on the conformation of the RAD51-ssDNA (single strand DNA) nucleofilaments which are formed by RAD51 protein in presence of RPA protein and ATP. BLM can destabilize RAD51-ssDNA nucleofilaments if they are in an inactive form, that is, after joining the ADP. When the RAD51-ssDNA nucleofilaments are in an active form, after the ATP joining, BLM stimulates strand exchange carried out by RAD51 [16, 17].

Loss of RAD51 and BLM proteins may contribute to an increased number of somatic mutations and rearrangements of chromosomes. It also contributes to abnormal chromosome segregation, aneuploidy, chromosomal instability, sensitivity to DNA damaging agents and loss of heterozygosity. All the disturbances can lead to reduction of the efficiency of DNA repair processes, cause genomic instability and finally lead to the development of cancer [18]. In breast cancer patients decreasing of the amount of RAD51 protein was observed in 30 % of the cases [19] whereas loss of heterozygosity was observed in 32 % of the cases [20]. The risk of malignant transformation in cells with deficiency of BLM is a hundred times greater than in cells containing a fully functional BLM helicase [21]. On the other hand, different variants of naturally occurring polymorphisms of genes involved in DNA repair can contribute to changes in the efficiency of the repair processes, which may result in the development of cancer [5, 22–27].

The fact that RAD51 and BLM proteins are acting together during DNA repair by homologous recombination prompted us to examine a correlation between polymorphic variants (SNPs) of the RAD51 gene (c.-61 G>T, rs 1801321 and g.38922 C>G, rs 4417527) and the BLM gene (g.96267 A>C, rs 2270132 and g.85394 A>G, rs 2380165) and breast cancer risk. We also studied an association between these polymorphisms of the RAD51 and BLM genes and clinical characteristics of breast cancer patients such as lymph node status, tumor grade, hormone receptors (estrogen and progesterone receptors) and epidermal growth factor receptor (HER2) expression.

Materials and Methods

Patients

Blood samples were obtained from 304 women (mean age 60 years) with sporadic breast cancer treated at the Department of Surgical Oncology, N. Copernicus Hospital (Lodz, Poland). The clinical characteristic of breast cancer patients is presented in Table 1. Blood was collected before surgical treatment and chemotherapy. The control group (319 women) consisted of age-matched women who were not diagnosed with cancer and recruited from Commune Health Clinic in Rzgow and Institute Polish Mother’s Health Center (Lodz, Poland). The Local Ethic Committee approved the study and each patient gave a written consent.

Table 1 The clinical characteristics of breast cancer patients
Full size table

Genomic DNA Isolation

Genomic DNA was prepared from peripheral blood of breast cancer patients and healthy individuals by using of commercial Blood Genomic DNA Miniprep Kit (Axygen Biosciences, CA, USA), as recommended by the manufacturer.

Selection of Polymorphisms and Primers Design

We obtained a list of SNPs in the RAD51 and BLM genes from the public domain of the National Center for Biotechnology Information the Single Nucleotide Polymorphisms database (NCBI dbSNP) at http://www.ncbi.nlm.nih.gov/snp. For this study we chose the c.-61 G>T (rs1801321) and g.38922G>C (rs4417527) polymorphisms of the RAD51 gene and the g.96267 A>C (rs2270132) and g.85394 A>G (rs2380165) polymorphisms of the BLM gene with a minor allele frequency (MAF) 0.467, 0.109, 0.376 and 0.333 in European population, respectively (submitter population ID: HapMap-CEU for all; http://www.ncbi.nlm.nih.gov/snp). Primers were designed according to the published nucleotide sequence in ENSEMBL database (RAD51 gene ID: ENSG00000051180 and BLM gene ID: ENSG00000197299) and using Primer3 software for c.-61 G>T, g.38922 G>C, g.96267 A>C SNPs (http://frodo.wi.mit.edu/) and Web-based allele-specific primer software for g.85394 SNP (http://bioinfo.biotec.or.th/WASP).

Genotype Determination

The restriction fragment length polymorphism reaction (PCR-RFLP) was used to determine the genotypes of the c.-61 G>T and the g.38922 C>G polymorphisms of the RAD51 gene and the g.96267 A>C polymorphism of the BLM gene. The polymerase chain reaction with confronting two-pair primers (PCR-CTPP) was used to determine the genotypes of the g.85394 A>G polymorphism of the BLM gene.

PCR reaction was performed in a total reaction volume of 25 μl containing 50 ng of genomic DNA, 1 U Biotools DNA polymerase (Biotools, Madrid, Spain), 1 × reaction buffer (750 mM Tris-HCl (pH 9.0), 500 mM KCl, 200 mM (NH4)2SO4), 0.2 mM of each dNTP, 1.5 mM MgCl2 and 0.25 μM of each primer (Metabion, Martinsried, Germany). The primer sequences are presented in Table 2. PCR amplifications were conducted in DNA Engine thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA). Thermal cycling conditions were as follows: initial denaturation step at 95 °C for 5 min, 34 cycles at 95 °C for 30 s, 30 s at 65 °C annealing temperature and 60 s at 72 °C.

Table 2 Primer sequences of the c.-61 G>T and the g.38922 C>G polymorphisms of the RAD51 gene and the g.96267 A>C and the g.85394 A>G polymorphisms of the BLM gene
Full size table

The products of the c.-61 G>T and the g.38922 C>G polymorphisms of the RAD51 gene and the g.96267 A > C polymorphism of the BLM gene were digested overnight with 0.2 U of the restriction enzyme NgoMIV, NlaIII and StyI (NEB New England Biolabs, Ipswich, MA, USA), respectively. The PCR products were separated into 8 % polyacrylamide gel, stained with ethidium bromide and viewed under UV light. Figure 1 presents a representative gels from analysis of the c.-61 G>T (A) and the g.38922 C>G (B) polymorphisms of the RAD51 gene and the g.96267 A>C (C) and the g.85394 A>G (D) polymorphisms of the BLM gene.

Fig. 1
figure 1

Genotypes of the c.-61 G>T (rs1801321) (a) and the g.38922 C>G (rs4417527) (b) polymorphisms of the RAD51 gene and the g.96267 A>C (rs2270132) (c) and the g.85394 A>G (rs2380165) (d) polymorphisms of the BLM gene analyzed by a 8 % polyacrylamide gel electrophoresis stained with ethidium bromide and viewed under UV light. Lane M shows GeneRuler TM 100 bp molecular weight marker; lane K on the picture C shows a control comprising PCR product without reaction with StyI restriction enzyme; lane K on the picture D shows a control comprising reaction mixture with F1 and R1 starters; all remaining lanes present genotypes indicated in the upper part of the pictures

Full size image

Statistical Analysis

Statistical analysis was performed using STATISTICA 8.0 package (Statsoft, Tulsa, OK, USA). Distributions of genotypes and alleles between groups were tested using the χ 2 analysis. The Hardy-Weinberg equilibrium was checked using the χ 2 test to compare the observed genotype frequencies with the expected frequencies among the case and control subjects. For each SNP, the odds ratios (ORs) and 95 % confidence intervals (CIs) were calculated. A linkage between genotype, cancer and clinical parameters was assessed by the logistic regression.

Results

Genotype Analysis

Breast cancer patients and controls were divided into groups corresponding to three genotypes. The genotype distributions for the c.-61 G>T and g.38922 C>G polymorphisms of the RAD51 gene were in agreement with those predicted by the Hardy-Weinberg equilibrium (p > 0.05), except for that of the c.-61 G>T polymorphism for the patients group. It is caused by very low presence of the G/G genotype of the c.-61 G>T polymorphism of the RAD51 gene in the Polish population. In the case of genotype distributions for the g.96267 A>C and g.85394 A>G polymorphisms of the BLM gene they differed significantly from those predicted by the Hardy-Weinberg equilibrium (p < 0.05), which can be caused by lack of the G/G genotype of the g.85394 A>G polymorphism and the C/C genotype of the g.96267 A>C polymorphism of the BLM gene in the Polish population. The distribution of genotypes of polymorphic variants of the RAD51 and BLM genes and intragene and intergene genotypes combinations for cancer patients and controls is shown in Table 3.

Table 3 The genotype and allele frequency and odds ratios (OR) of the c.-61 G>T and the g.38922 C>G polymorphisms of the RAD51 gene and the g.85394 A>G and the g.96267 A>C polymorphisms of the BLM gene in breast cancer patients and controls
Full size table

Clinical Parameters of Breast Cancer Patients and BLM and RAD51 Genes Polymorphism

We checked the distribution of genotypes and frequency of alleles of the BLM and RAD51 genes polymorphisms in groups of patients with different hormone receptor status, patients with positive and negative lymph node status and patients with different tumor grade (Table 4). We did not observe any association between estrogen hormone receptor status and the distribution of genotypes and frequency of alleles for four analyzed polymorphisms (data not shown). We did not observe any association between HER2 expression and the distribution of genotypes and frequency of alleles for any polymorphism (data not shown).

Table 4 The genotype, allele frequency and odds ratios (OR) of the c.-61 G>T polymorphism of the RAD51 gene and the g.96267 A>C polymorphism of the BLM gene in subjects with breast cancer with different clinical parameters
Full size table

We did not observe any correlation between the g.38922 C>G polymorphism of the RAD51 gene and the g.85394 A>G and the g.96267 A>C polymorphisms of the BLM gene and lymph node status (data not shown).

Next of the analyzed clinical features was the tumor grade described by Bloom-Richardson grading system (Table 4). We did not find any association between the c.-61 G>T and the g.38922 C>G polymorphisms of the RAD51 gene and the g.85394 A>G polymorphism of the BLM gene and tumor grade (data not shown).

Discussion

RAD51 recombinase and BLM helicase are the elements of the protein machinery performing DNA DSBs repair by homologous recombination [15]. Homologous recombination is initiated by the MRN complex (MRE11, RAD50, NBN) resecting the DSB ends to generate 3′ overhangs flanking the break site. The resulting ssDNA ends are first protected by RPA before eventually being coated with RAD51 protein to form the critical nucleoprotein filament necessary for accurate repair [11]. BLM helicase was identified in a large DNA damage surveillance complex with BRCA1 and MRN complex [28] and has been shown to associate with many homologous recombination proteins, such as topoisomerase IIIα, BLAP75/RMI1 and RAD51 [29]. Cell and biochemical studies have led to the view that BLM has both pro- and anti-recombinogenic functions. Most notable, BLM is important in stabilizing damaged replication forks and repressing aberrant recombination events, as evidenced by the dramatic increase in levels of sister chromatid exchanges (SCEs) and loss of heterozygosity in BLM null cells. In contrast, BLM is also predicted to promote homologous recombination by facilitating exonucleolytic resection of DSBs, by stimulating synthesis-dependent strand annealing, and by promoting non-crossover resolution of Holliday junctions [30]. Recently, it was shown that SUMOylation of BLM can affect its interactions with RAD51. SUMOylation of BLM facilitates repair of damaged replication forks by homologous recombination by modulating recruitment of RAD51 or retention at repair sites [31].

Defects in some DNA repair genes are associated with rare human cancer-prone disorders, like Bloom syndrome resulting from a mutation of the BLM gene, ataxia telangiectasia and Werner syndrom. Apart from these rare syndromes, deficient DNA repair is suggested as a predisposing factor in familial breast cancer and in some sporadic breast cancers [3–6]. Advances in the understating of genetic predispositions to breast cancer have also been made by screening naturally occurring polymorphisms in DNA repair genes [5, 22–27]. These studies revealed that subtle defects in DNA repair capacity arising from low-penetrance genes or their combinations are modified by other genetically determined or environmental factors and correlate to breast cancer risk.

In the present study we correlated the genetic constitution of breast cancer patients expressed by polymorphic variants of two important homologous recombination genes, BLM and RAD51, with clinical parameters of patients, including lymph node status, tumor grade, hormone receptors (estrogen and progesterone receptors) and epidermal growth factor receptor 2 (HER2) expression. The c.-61 G>T-RAD51 polymorphism is located in the 5′-untranslated region, the regulatory promoter element of the RAD51 gene. It could affect translation efficiency and mRNA stability, leading to changes in RAD51 protein levels, which, in turn, could influence on the activity of the multiprotein DNA-repair complex that includes BRCA1, BRCA2 and RAD51, and leads to breast cancer susceptibility. The g.38922 C>G-RAD51 polymorphism is located in the intron 10 in the RAD51 gene. The intron 10 is located between the exons in the core conserved domain that includes Walker A and B motifs, which have the function of ATP binding and hydrolysis activities. Because of the intron localization the g.38922 C>G polymorphism does not affect amino acid coding and therefore probably does not directly affect protein function but it is likely that may play a role in the proper recognition of non-coding element of RNA splicing and thus it may be indirectly responsible for the correct functioning of RAD51 protein. In turn, it could influence the activity of the multiprotein DNA-repair complex that includes BRCA1, BRCA2, RAD51 and other proteins. On the other hand, the observed associations between breast cancer risk and the g.38922 C>G-RAD51 polymorphism can be interpreted as affecting the mRNA maturation process and in consequence activity of RAD51 protein.

Beyond genotypes, the expression of genes or proteins might also play a role in breast cancer risk. RAD51 overexpression was associated with a higher risk of locoregional recurrence and death. RAD51 overexpression was also significantly associated with shorter locoregional recurrence-free survival and overall survival [32]. Moreover, it was shown, that RAD51 mRNA levels were inversely associated with PR status and the highest levels were detected in ER-positive/PR-negative tumors. The analysis of microarray expression data from 295 breast cancers indicate that RAD51 increased mRNA expression is associated with higher risk of tumor relapse, distant metastases and worst overall survival [33].

The g.96267 A>C polymorphism is located in the intron 20 and the g.85394 A>G polymorphism is located in the intron 17 of the BLM gene. These introns separate exons forming the evolutionarily preserved RecQ-like domain. This domain exhibits a high degree of homology with the RecQ protein of E. coli. RecQ-like domain and its constituent sub-domains are responsible for the most important functions of BLM helicase [34]. The polymorphic variant g.96267 A>C is located in the intron 20, which separates the exons encoding the HRDC domain (helicase and RNase D-like C-terminal domain), which in RecQ family-like helicases plays the role in recognizing and binding substrates and protein-protein interactions. In the case of BLM helicase, HRDC domain is also involved in the process of movement and developing of the Holliday structure. Therefore, it is likely that the polymorphic variants of the BLM gene g.96267 A>C and g.85394 A>G may play an important role in the proper recognition and cutting out introns in the process of mRNA maturation and may be responsible for the correct functioning of BLM helicase.

Our results on the genes polymorphism and breast cancer risk are partly in agreement with the study presented by Ding et al. [26]. Comparing the results of Ding et al. carried out on population of women living in Taiwan to our results we simillarly observed a protective role of the A/C variant in the case of the g.96267 A>C polymorphism of the BLM gene (Table 3). On the other hand, our results disagree with those of Ding et al. in the case of the genotype the C/C of the g.38922 C>G polymorphism of the RAD51 gene (Table 3).

It was shown that the polymorphic variant of the BLM gene − rs2380165 and the variants of the RAD51 gene – rs4417527 and rs2412546 were associated with breast cancer risk [26]. Moreover, not only the intronic polymorphisms located within the region encoding the helicase domain of BLM but also these within the RAD51-interaction domain-encoding region showed an interaction with RAD51 polymorphisms in determining breast cancer susceptibility. Recently, it was also shown that the domains responsible for RAD51 partnership coincide with the domains responsible for the binding of ssDNA, BLM100–214 and BLM1317–1367 [29]. This suggests that the ability of BLM to suppress homologous recombination may involve both the displacement of RAD51 from the D-loop and the binding of BLM to the newly liberated ssDNA to reinstate the duplex and completely impede genetic exchange through homologous recombination.

In the present study we also correlated the genotype combinations of four polymorphic variants of the BLM and RAD51 genes with breast cancer occurrence (Table 3). Especially, in the case of analysis of genotype intergene combinations and breast cancer occurrence, we detected a high increase of cancer risk (Table 3).

Relationship between the c.-61 G>T and g.38922 C>G polymorphisms of the RAD51 gene, and the g.91266 C>A and g.85394 A>G polymorphisms of the BLM gene and clinical characteristics of breast cancer patients was also studied (Table 4). Breast cancer patients were divided into groups depending on lymph node status, tumor grade and expression of estrogen and progesterone receptors (ER and PR) and HER2 receptor status. Auxiliary lymph node status and tumor size are clinical parameters directly correlated with the survival period in breast cancer. Expression level of ER and PR are useful predictive markers allowing the assessment of the response to endocrine therapy – ER- and PR-positive tumors may have a 6- to 7-higher response rate to negatives [35].

Next of the analyzed clinical features was the presence of HER2 (C-ERB/neu), belonging to the group of oncogenes of the tyrosine kinase function, which is overexpressed in about 20–30 % of breast cancers and often is associated with the resistance to a hormone therapy. Activation of this receptor leads to the uncontrolled division of cancer cells, which may be accompanied by an increased amount of vascular endothelial growth factor (VEGF) as the main factor inducing the formation of new blood vessels within the tumor. HER2 overexpression leads to early metastases in the direction of the lymph nodes, early recurrences of the disease and is associated with shorter survival and worse clinical prognosis [36]. We did not observe any association between HER2 expression and the distribution of genotypes and frequency of alleles within any of the analyzed polymorphisms (data not shown).

We also correlated the genotypes of the BLM and RAD51 polymorphic variants with tumor grade described by the Bloom-Richardson grading system (Table 4). It is a three-step scale, according to which the tumor samples taken during biopsy are classified, and then analyzed using a microscope. To the G1 category belong tumors whose cells are small, have regular shapes, show high similarity in structural terms to normal breast cells and their mitotic index does not exceed the number of seventh. To the G2 category belong tumors whose cells exhibit visible changes compared to normal breast cells and their mitotic index ranges from 8 to 14. To the last G3 category belong tumors whose cells are large and have irregular shapes. They differ significantly from normal cells, divide rapidly − their mitotic index exceeds the number 15.

The results obtained indicate that the RAD51 and BLM genes polymorphism can be associated with breast cancer risk. An association may be underlined by a common interaction between these genes. We showed that the c.-61 G>T polymorphism of the RAD51 gene might modulate breast cancer risk and be associated with progesterone receptor expression and local metastasis. We also found an association between the g.96267 A>C-BLM polymorphism and G3 grade of tumor.

Conclusion

The variability of the RAD51 and BLM genes may play a role in breast cancer occurrence. The examined polymorphisms of the RAD51 and BLM genes cannot be an independent markers of breast cancer but our studies may be useful in building a set of clinical and molecular markers helpful for diagnosis of breast cancer.