Somatic SNPs of the BRCA2 gene at the fragments encoding RAD51 binding sites of canine mammary tumors
Abstract
Mammary tumors are the most common tumor type both in women and in female dogs. In women, heritable breast cancers have been linked mutations in the breast cancer susceptibility gene BRCA2 and it contains eight BRC repeats in exon 11 that bind to RAD51. In this study, we investigated the sequence variations of BRC1-BRC8 and C-terminus of canine BRCA2 gene. From a total of 64 canine patients with mammary tumors, 31 mammary tumors with benign and malign carcinomas and the 3 normal mammary glands were used for the study. In this study, 19 SNPs of exon 11 of BRCA2 in canine mammary tumors were detected for the first time. The c.2383A>C (T1425P) SNP was found to be the most probable disease-associated nsSNP. Our findings suggest that T1425P variation in BRC3 to be the most probable disease-associated nsSNP and may affect RAD51 binding strength.
Introduction
Tumors of the mammary glands are the most common tumors to affect entire female dogs representing between 50 and 70% of all tumors types,1 which is three times higher rate of incidence than humans. No other animal species has such high probability of onset of mammary tumors.2 In humans, heritable breast cancers have been linked with mutations in the breast cancer susceptibility gene BRCA2. Genetic analysis, including detection of deleterious mutations and splicing variants, to identify BRCA2 mutation carriers is strongly advocated, as the lifetime risk of breast cancer is high (81–88%) for females carrying a BRCA2 mutations.3 Canine mammary tumor (CMT) shows many similarities with human breast cancer in histology, genetics and biology. The spontaneous occurrence and clinical similarity of canine to human cancers makes the dog a valuable animal model to study genetic predisposition to cancer.4
The primary function of BRCA2 is homologous recombination (HR), and its mediates the recruitment of recombinase RAD51 to DNA double-strand breaks (DSBs); RAD51 recruitment is not only essential for HR but also responsible for the tumor-suppressive function of this repair process.5
It has been reported that BRCA2 mutations are associated with the development of mammary tumors in dogs.6-8 The canine BRCA2 cDNA is 11 kb in length and encodes 3471 amino acids.9 The BRCA2 gene contains 8 BRC repeats (BRC1–BRC8), encoding 39 amino acid sequence, which play a crucial role in binding to RAD51. The C terminus of BRCA2, which is corresponding to exon 27, contains a cyclin-dependent kinase (CDK) phosphorylation site that also binds RAD51.5 Sequence homologies of canine and human BRCA2 and RAD51 proteins are 68 and 99%, respectively.9 Although the overall sequence similarity among BRCA2 orthologous is limited, the BRC repeats in different species are well conserved, suggesting that they are essentials to the biological functions of BRCA2.10
Based on the above considerations, the objective of this study is to investigate the genetic variations in BRCA2 gene RAD51 binding sites (exons 11 and 27) in canine mammary tumors (CMTs).
Materials and methods
Tissue samples, histopathology and DNA isolation
Tumoral masses and/or biopsies taken from the mammary gland regions of 64 canine patients were examined histopathologically and a total of 31 mammary tumors (benign n = 10 and malign n = 21) were used for the study. Radiological examination of the thorax was performed to reveal detectable pulmonary metastases. Mammary gland tumor samples were resected surgically at the Uludag, Istanbul and Firat Universities, Faculties of Veterinary Medicine, Department of Obstetrics and Gynecology. For histopathological examination, a representative part of the tumor samples were fixed in 10% buffered formalin for 48 h then they were processed according to routine tissue dehydration and clearing procedures and embedded in paraffin. Serial tissue sections at a thickness of 5 µm were stained with hematoxylin-eosin and they were evaluated histopathologically according to the current canine mammary tumors grading and classification guide.11
Clinical and demographic information of the dogs were recorded at the time of diagnosis for the following criteria, age, breed, clinical stage, macroscopic appearance of the tumoral masses, ovariohysterectomy history, hormonal treatments, prior chemotherapy. Besides all those information, the characteristics and clinicopathological findings has been shown in Table S1, Supporting Information. None of the animals had undergone chemotherapy.
Genomic DNA was extracted from the 31 CMTs and 3 normal mammary glands tissue samples using the DNeasy blood and tissue isolation kit (Qiagen Cat No: 69504) according to the manufacturer's protocol.
PCR and DNA sequencing
For PCR amplification to BRCA2 gene exon 11 and exon 27 regions were amplified as described Yoshikawa et al.3 Primers and product sizes has been given in (Table S2). PCR products were analyzed by 2% agarose gel electrophoresis. Then, PCR samples were sequenced from both directions, following the purification of PCR products. Direct sequencing was performed on 3100 ABI PRISM sequencer (Applied Biosystems, Foster City, USA), and sequenced by commercial services. Sequences were obtained with the same primers used for PCR amplification.
Data analysis
Sequences were analyzed using the BIOEDIT ver 7.2.5 software12 for sequence alignment. The DnaSP software 5.10.0113 was used to calculate polymorphic sites, average number of nucleotide differences (k), number of haplotypes (h), nucleotide diversity (π) haplotype diversity (Hd); Watterson's theta estimator for the studied species separately using a haplotype sequence was obtained.
In silico analysis
In this study, 21 SNPs of BRCA2 exon 11 has been exposed to in silico analysis using 7 different algorithms; SIFT, PROVEAN, PolyPhen, MutPred, I-Mutant, Consurf and Swiss Model.
The impact of amino acid variants on protein structure via analysis of multiple sequence alignments was performed using SIFT (Sorting Intolerant From Tolerant), which uses sequence homology to predict whether an amino acid substitution will affect protein function and hence, potentially alter the phenotype. It gives a normalized probability score value that the amino acid change is tolerated. If the score value less than 0.05, the amino acid change is predicted deleterious.14
PROVEAN (Protein Variation Effect Analyzer) is a sequence-based predictor that estimates the effect of protein sequence variation on protein function. The variant is considered deleterious when the PROVEAN score is less than −2.5 and is predicted to be ‘Neutral’ if the score is above the threshold.15 The PolyPhen server was used to study the functional consequences of nsSNPs. The classification of the nsSNPs results in Possibly Damaging and Probably Damaging (PSIC > 0.5) or Benign (PSIC < 0.5).16 MutPred is a web-based tool to predict the molecular cause of disease-related amino acid substitution.17 It determines the changes at atomic and molecular level induced by the amino acid substitution.18 I-Mutant 2.019 is a support vector machine (SVM) tool for the prediction of protein stability free-energy change (ΔΔG or DDG) on a specific nsSNP. It predicts the free energy changes starting from either the protein structure or the protein sequence. A negative DDG value means that the mutation decreases the stability of the protein, while a positive DDG value indicates an increase in stability.18
ConSurf was used for high-throughput characterization of the functional regions in the protein.20 The conserved regions were predicted via conservation scores and coloring scheme and further divided into distinct scales of nine grades.21
We used to SWISS-MODEL to generate a 3D structural model for canine BRCA2. SWISS-MODEL is a platform for automated protein structure prediction based on homology modeling.22 There is a brief outline of the method, which is divided into four general stages: fold assignment, alignment of target and template sequences, model building based on the alignment with a selected template, and structure validation.21
Results
In accordance with the literature information, our sequence results have been compared with Z7566423 and FJ464397-FJ464411 GenBank sequences.9 Among them Z75664 has been used as a reface sequence for BRCA2 gene exon 11 region.
According to the sequence analysis of the BRCA2 gene exon 11 region, 29 different haplotypes (28 SNPs) were obtained. The results of healthy mammary gland sequences were shown three haplotypes, which are haplotype1 (H1), H2 and H4. Compared with the sequence in the healthy mammary gland haplotypes, 26 haplotypes has been detected in canine mammary tumors. Haplotype 20 (KX090067) was excluded from the analysis, because of the lack of pathology result (Table 1). After radiological examination seven individuals (samples 3, 9, 12, 14, 18–20) were shown to pulmonary metastases. Among those samples, two samples (samples 9 and 18) were found to belonged H12.
Ten of these 26 haplotypes were found only in benign tumor samples, whereas the rest of 21 haplotypes were found in malign tumor samples. In total, we identified 28 SNPs in BRCA2 exon 11; among those genetic variations, 21 were non-synonymous SNPs (nsSNPs) (75%) and 7 were synonymous SNPs (25%).
A total of 19 variations (11 malign, 4 benign, 4 both malign and benign) of BRCA2 in canine mammary tumors in this study were detected for the first time when compared with available nucleotides in GenBank. The BRCA2 gene exon 11 haplotypes has been deposited in GenBank database under the access number: KX090048-KX090076.
The variations c.1828A>C and c.2383A>C, c.2414G>A are located in the BRC repeat 1 and BRC repeat 3, respectively. When we focus on those BRC repeat domain variations, including the high frequency of one SNP (c.2414G>A) occurred sporadically in all tumor types. In the malignant tumor samples, the c.2414G>A variation occurred in 14 of 21 dogs, but not in osteosarcoma.
For the exon 27 analyses, ‘ENSCAFE00000069954’ has been used as a reference sequence. Only one indel polymorphism was determined in exon 27 of BRCA2 (c.10204ins/delAAA) and 10204insAA polymorphism changes the amino acid sequence (M3332IK) (Fig. 1). The deletion (10204delAAA) was only found in four samples for CMT (samples 8 and 15 for malign tumors and samples 22 and 24 for benign tumors), and only in one individual from the normal mammary gland sample. The presence of AAA insertion observed in the rest of the dogs. The BRCA2 gene exon 27 haplotypes has been deposited in GenBank database under the access number: KX090077.
A total of 21 nsSNPs for BRCA2 gene exon 11 region were analyzed to identify the deleterious mutations. Of these, 9 were found to be damaging (score < 0.005) by SIFT and 11 nsSNPs were predicted to be deleterious with a PROVEAN. It was observed that the variants Q1072P, S1078W, A1108G, K1179Q, T1425P, T1559P and L1601I are probably damaging in Polyphen analysis. A total of four nsNSPs were predicted to be damaging by SIFT and deleterious by PROVEAN and probably damaging by Polyphen. The DDG (ΔΔG) predicted by I-Mutant classified 18 nsSNPs as decreasing the stability of mutated protein and 3 as increasing it. These results were summarized in Table S3.
Finally, the c.2383A>C (T1425P) SNP was predicted to be the most probable disease associated nsSNP by MutPred (Table S4) with confident hypothesis prediction (P = 0.045).
Conserved regions of BRCA2 protein were analyzed using the ConSurf server. It was observed that variants T1425P have a conservation scale of 8 (Fig. 2).
The homologous model generated using a template (1n0w, Homo sapiens) covered a total of 3446 amino acid residues of BRCA2 protein (Fig. 3). The ConSurf and SWISS MODEL results indicate that T1425P variant region have well conserved. Further, I-Mutant results showed that this genomic variant is likely to decrease the stability of canine BRCA2 protein. [Colour figure can be viewed at wileyonlinelibrary.com]
Discussion
The BRCA2 protein plays an important role in the HR repair of DSBs and promotes the assembly of RAD51 recombinase onto single-stranded DNA. To initiate DSB repair by HR, BRCA2 interacts with RAD51 at the highly conserved BRC repeats of BRCA2, which consist of 13 conserved amino acid residues.8 Until now, there is limited information available about the genetic variation in BRCA2 gene RAD51 binding/interaction sites in dogs.8, 9, 24, 25 Studying the interaction between canine BRC repeats and RAD51 may extend our knowledge of the DNA repair and tumor suppressor roles of BRCA2.26
In the present study, we focus on BRCA2 gene RAD51 binding sites (exon 11 and exon 27) in dogs with mammary tumors and 26 haplotypes (28 SNPs) in exon 11 and one indel polymorphism (c. 10204ins/delAAA) in exon 27 were identified.
In this study, 19 variants (c.783T>C, c.866T>C, c.1270A>G, c.1325>A>C, c.1343C>G, c.1344T>G, c.1390C>G, c.1397T>C, c.1433C>G, c.1443A>C, c.1452T>C, c.1642A>C, c.1659A>C, c.1692A>C, c.1828A>C, c.2066A>C, c.2785A>C, c.2911C>A, c.3620C>G) were novel SNPs when compared with available nucleotides in GenBank (Z75664 and FJ464397-FJ464411).
Among 28 SNPs in this study, c.1828A>C was found in the BRC1 region, whereas c.2383A>C and c.2414G>A had been determined in the BRC3 region. Among eight BRC repeats in BRCA2 gene (BRC1–8), BRC3 and BRC4 are two best-characterized ones exhibiting strong interaction with RAD51.8
There were 5 of 21 (23%) malignant mammary tumors (carsinosarcoma, adenosquamous carcinoma, anaplastic carcinoma, tubular adenocarcinoma and solid carcinoma) containing base change at 2383 position (p. 1425T>P) and for three of these five individuals were seen only Rottweiler dog breeds of CMT. However, the number of dogs in this study is too limited for exploring any differences between high and low risk breeds.
Hsu et al.9 reported that SNPs at codon 511 and 2414 in canine BRCA2 gene to appear frequently in dogs with mammary tumors, which is the similar to results presented in this study. However, in this study 2414 variants have been found to be tolerated/neutral by SIFT, PROVEAN and PolyPhen tools; whereas 511 variants (K801Q) was found to be damaging. Loss of methylation (P = 0.007) and loss of ubiquitination (P = 0.0087) were predicted by MutPred at this site (c.511A>C). This implied that 511A>C variants may account for potential structural and functional alteration of BRCA2.
According to in silico analysis results four amino acid substitutions (S1078W, A1108G, T1425P, T1559P) predicted to have damaging/deleterious effect by SIFT, PROVEAN and PolyPhen tools. Among these amino acid substitutions, only T1425P was located at the BRC3 domain.
When results of all the tools were used to detect high-risk variations the functional SNPs at positions S1078W, A1108G, T1425P and T1559P predicted to have damaging/deleterious effect by SIFT, PROVEAN and PolyPhen servers and were used for further structural analysis. Conserved regions of BRCA2 protein were predicted using the ConSurf tool and T1425P has a conservation scale of 8, thereby showing more chances to alter the protein structure. A genetic variant located in strongly conserved regions of the gene is more likely to be disease-associated than polymorphisms occurring in regions with a lower degree of conservation.25 Finally, the T1425P was found to be the most probable disease-associated variant by MutPred with confident hypothesis prediction (P = 0.045).
Ochiai et al.8 investigated polymorphic alleles of BRC3 and BRC4 in multiple dog breeds and to explore the influence of canine BRC3 polymorphisms on the RAD51 interaction, they performed the mammalian two-hybrid assay (MTH). Researchers were not found polymorphisms except for the previously identified T1425P and K1435R and they found that T1425P polymorphisms in BRC3 that reduced the binding strength for RAD51 might affect the full-length BRC functions, which are the similar results presented in the study.
In summary, among 21 SNPs only T1425P substitution was found to be the most probable disease-associated variant for exon 11.
Exon 27 is located mainly downstream of the BRCA2 gene. This exon contains sequence that code for the C-terminus of BRCA2 and 3′-UTR. The C-terminal sequence is highly conserved and in addition to RAD51 interaction function, it plays some important roles in mammals. For example, the nuclear localization signals (NLSs) at the C-terminus are essential for the nuclear localization of human BRCA2. Deletion of the C-terminal NLSs prevents BRCA2 localization to the nucleus, which results in dysfunction.24
Yoshikawa et al.24 and Enginler et al.7 found a higher allele frequency for the 10204insAAA than for the 10204delAAA, which are the similar results presented in this study. Also, according to Yoshikawa et al.24 10204insAAA was associated with significantly higher mammary tumor morbidity rate in dogs than the 10204delAAA. In this study, 10204delAAA was only found in four samples for CMT (samples 8 and 15 for malign tumors and samples 22 and 24 for benign tumors), and only in one individual from the normal mammary gland sample, but the number of dogs in this study is too limited for exploring any differences between high and low risk breeds as to this polymorphism.
Breast cancer is a complex multifactorial disease, which results from an interplay of environmental, reproductive, lifestyle, and genetic risk factors.27 Canine cancers occur spontaneously, and have similar clinical presentation and pathophysiology to equivalent human cancers. Identification of risk for developing cancer is a key factor in health management of the individual and the population. Once constitutional changes to the canine genome have been determined, these data offer a powerful approach to screen and stratify populations by risk. Within purebred lines, knowledge of risk will play a key role in selecting more informed breeding programs designed to reduce carefully the frequency of deleterious alleles in the population. The inherited risk of a cancer may not be directly related to the genome dysregulation associated with initiation and/or propagation of tumor cells, and so be of little to no value in making decisions about patient care once a cancer develops. Genome-wide assessment of changes at the somatic level is therefore another key area of active research in comparative oncology.28
In conclusion, we have identified 28 somatic SNPs in canine BRCA2 exon 11 region, among them 19 SNPs were detected for the first time in this study. A total of 21 were nsNSPs that might cause altered protein function, four of which were predicted to have damaging effect on the protein. According to in silico analysis results four amino acid substitutions predicted to have damaging effect. Among these amino acid substitutions, only T1425P was located in well-conserved regions. Furthermore, Swiss model result has been confirmed the MutPred detection, which is loss of sheet at T1425P position. When these results were analyzed together, we speculate that T1425P variation in BRC3 to be the most probable disease associated nsSNP and may affect RAD51 binding strength. Although, our results suggest that T1425P may most probable disease associated variation, its association with RAD51 requires further validation on large number of animals with a sample from healthy tissue of the same individual, in order to distinguish germline mutations from somatic mutations.
Acknowledgements
The work was supported by Scientific Research Projects Council of Firat University, project number VF1311.
Conflict of Interest
The authors declare that they have no conflict of interest.
Ethical approval
The study was approved by the ‘Firat University Animal Researches and Ethic Committee’ (Verdict number: 2012/06/63).
