Homologous recombination DNA repair gene RAD51, XRCC2 & XRCC3 polymorphisms and breast cancer risk in South Indian women

Background Homologous recombination repair (HRR) accurately repairs the DNA double-strand breaks (DSBs) and is crucial for genome stability. Genetic polymorphisms in crucial HRR pathway genes might affect genome stability and promote tumorigenesis. Up to our knowledge, the present study is the first to investigate the impact of HRR gene polymorphisms on BC development in South Indian women. The present population-based case-control study investigated the association of polymorphisms in three key HRR genes (XRCC2-Arg188His, XRCC3-Thr241Met and RAD51-G135C) with BC risk. Materials and methods Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method was used for genotyping the HRR variants in 491 BC cases and 493 healthy women. Results We observed that the XRCC3 Met allele was significantly associated with BC risk [OR:1.27 (95% CI: 1.02–1.60); p = 0.035]. In addition, the homozygous mutant (C/C) genotype of RAD51 G135C variant conferred 2.19 fold elevated risk of BC [OR: 2.19 (95% CI: 1.06–4.54); p = 0.034]. Stratified analysis of HRR variants and BC clinicopathological features revealed that the XRCC3-Thr241Met and RAD51-G135C variants are associated with BC progression. Combined SNP analysis revealed that the individuals with RAD51-C/C, XRCC2-Arg/Arg, and XRCC3-Thr/Thr genotype combination have three-fold increased BC risk. Conclusion The present study imparts additional evidence that genetic variants in crucial HRR pathway genes might play a pivotal role in modulating BC risk in South Indian women.


Introduction
Breast cancer (BC) is a complex polygenic disease that arises due to the synergistic effect of several genetic variations and environmental factors. Molecular epidemiological studies have suggested that approximately 80% of BC's inherited susceptibility is due to the combined effect of several low penetrant gene variants rather than the high penetrant gene mutations [1]. Moreover, it has been observed that most of the epidemiological studies have been conducted on participants from European ancestry. Therefore, simultaneously increasing the representation of participants from other populations is highly recommended [2]. BC possesses a significant health burden in both developed and developing countries. Surprisingly, the highest BC incidence rate was observed in the Chennai (South Indian city) registry, and the BC burden was estimated to escalate up to 233 per 1000 females by 2026 in India [3]. Furthermore, the impact of candidate gene polymorphisms on BC risk has not been completely ascertained in South Indian women.
Several endogenous or exogenous factors trigger aggressive DNA damages, such as doublestrand break (DSB) lesions, which are generally repaired by the DSB repair. DSB is repaired via two pathways: Homologous recombination repair (HRR) and non-homologous end joining (NHEJ) DSB repair. HRR genes function as genomic caretakers, and germline mutations in crucial HRR genes have been strongly associated with tumor predisposition [4,5]. In HRR, the MRN (MRE11/RAD50/NBN) complex binds to the ends of DSBs. The nucleotides from the 5' end of DSBs are excised by MRE11, thereby resulting in 3' single-strand DNA (ssDNA) overhangs [6]. RAD51 forms a nucleoprotein filament by binding to ssDNA and facilitates strand invasion into homologous DNA duplex using various mediator proteins such as XRCC2, XRCC3, and BRCA2. The newly synthesized DNA then dissociates to anneal with an opposite DNA strand, and ligation completes the HRR process [7,8]. A G>C substitution (rs1801320) at position 135, in the 5' untranslated region (UTR) of RAD51 has been reported as a modulator of RAD51 DNA repair capacity (DRC). Individuals with the C allele had the lowest DRC, thereby suggesting the RAD51 G135C variant has a functional role in modulating BC susceptibility [9].
Similarly, X-ray repair cross-complementing 2 (XRCC2), a member of the RAD51 family of proteins, possess walker motifs A and B (which are ATP binding domains) and is a crucial protein that mediates HRR [10,11]. Interestingly, XRCC2 functions as an enhancer of RAD51 activity, and loss of XRCC2 protein activity results in a critical delay in the initial RAD51 response to the DNA damage [12]. A non-synonymous variation (rs3218536) caused due to c.563G>A substitution in exon 3 of XRCC2 gene results in substitution of Arg to His amino acid at codon 188. Moreover, site-directed mutagenesis of XRCC2 revealed that non-conservative amino acid substitution at 188th amino acid position significantly affects cell's sensitivity to DNA damage [13]. Furthermore, the XRCC2 Arg188His variant was found to modify BC risk in women with reduced plasma folate levels [14]. Likewise, XRCC3, a RAD51 paralog, controls the fidelity of HR and is essential for stabilising heteroduplex DNA in HRR. Furthermore, a mutation in X-ray repair cross-complementing 3 (XRCC3) generates severe chromosomal instability [15]. A common variant (rs861539) in the XRCC3 gene is a c.722C>T substitution in exon 7, which results in Thr to Met amino acid substitution at codon 241. Additionally, individuals carrying the Met allele had increased DNA adduct levels in the lymphocyte DNA [16,17]. Moreover, in vitro studies suggested that the XRCC3-241Met variant increased an individual's cancer risk [18]. Besides, Song et al. conducted a meta-analysis and highlighted that the Thr241Met variant was significantly associated with a higher risk of radiation-induced early adverse outcomes, as well as specific detrimental effects such as mucositis and acute skin toxicity [19].
Last decade, there have been conflicting reports regarding the impact of the HRR gene polymorphisms on BC risk [20][21][22][23][24][25][26]. Besides, there is a paucity of information regarding the impact of HRR gene polymorphisms on South Indian women's BC etiology. Hence, we conducted this population-based case-control study to evaluate the impact of XRCC2-Arg188His, XRCC3-Thr241Met, and RAD51-G135C variants with BC risk in South Indian women.

Study subjects
The present study investigated 984 subjects, which included 491 histopathologically confirmed breast cancer cases and 493 healthy women from South India. The institutional ethics committee of Dr. G.V.N Cancer Institute (ECR/436/INST/TN/2013) and MMHRC (ECR/398/INST/ TN/2013/RR-16) approved the present study. The samples were collected following the tenets of the Declaration of Helsinki and its later amendments. Written informed consent was obtained from the study participants. Peripheral blood of BC patients was collected from the medical oncology department of Dr. G.V.N Cancer Institute and Meenakshi Mission Hospital & Research Centre between June 2017 and January 2020. Blood samples of healthy (cancerfree) women, who are age and ethnicity matched to cases and without a family history of cancer, were collected during the same period. The surgically resected primary tumors of the BC patients were graded according to the Scarf-Bloom-Richardson grading system and staged based on the American Joint Committee of Cancer (AJCC) system. Clinico-pathological characteristics of the BC patients such as menopausal status, age at disease onset, hormonal receptor status, tumor grade, tumor stage, and metastasis extent were noted with the help of a medical oncologist.

Genomic DNA extraction
Antecubital venepuncture was performed to draw 3-5ml of venous blood from the study participants in a commercially available sterile K2-EDTA coated vacutainer (BD Vacutainer1, Franklin Lakes, USA). The genomic DNA was extracted from the whole blood samples using the HiPurA SPP blood DNA isolation kit (HiMedia TM , Mumbai, India) following the manufacturer's protocol. The isolated DNA was quantitatively assessed for purity using the Nano-Drop 2000 TM spectrophotometer (Thermo Fischer Scientific, USA). All the genomic DNA samples were stored at -20˚C until further analysis.

Genotyping
PCR-RFLP (Polymerase Chain Reaction-Restriction Fragment Length Polymorphism) method was used for genotyping XRCC2-Arg188His, XRCC3-Thr241Met, and RAD51-G135C variants. The PCR reactions were carried out in a final volume of 25μl reaction mixture comprising 12.5μl of 2x GoTaq Green master mix (Promega, Madison, USA), 0.5μl of each primer (10μM), 100-150ng genomic DNA, and nuclease-free water. PCR reactions were carried out using T100 TM thermal cycler (Bio-Rad, CA, USA). The PCR amplicons were run on 1% ethidium bromide-stained agarose gel and visualized using Bio-Rad XR + gel documentation system (Bio-Rad, CA, USA). The list of primers, annealing temperature, and RFLP conditions utilized is given in Table 1. The investigated variants' genotypes were determined by performing electrophoresis of the digested PCR products in ethidium bromide-stained 4% agarose gel and visualized using a gel documentation system (Bio-Rad, CA, USA). To assess the genotyping quality, genotyping was repeated in random 10% of the samples, and the results were 100% concordant.

Statistical analysis
With respect to the SNPs investigated, the genotypes of the controls were assessed for their agreement with Hardy-Weinberg equilibrium (HWE) using the χ2 goodness of fit test. Odds ratio (OR) and 95% confidence interval (CI) were determined by performing unconditional logistic regression analysis using SNPStats online software (https://www.snpstats.net/). Stratified analysis was carried out between the investigated SNPs and clinicopathological features of BC cases to evaluate SNPs' role in disease progression. P-value <0.05 was considered as statistically significant. Multifactor-dimensionality reduction (MDR) analysis and interaction dendrogram was constructed using the MDR software package (MDR 3.0.2) to evaluate the impact of gene-gene interaction on BC risk. The best interaction model was selected based on the highest cross-validation consistency (CVC) and testing balance accuracy (TBA). Further, STRING software was used to visualize protein-protein interaction (https://string-db.org/).

Characteristics of the study population
The present study comprised of 491 breast cancer cases and 493 healthy women as controls. The mean age of BC onset in the patients was 52.1±10.99 yrs. The clinicopathological features of BC cases are summarized in Table 2.

Hardy-Weinberg Equilibrium (HWE) test
The observed genotype frequency for the studied polymorphic loci were in accordance with Hardy-Weinberg equilibrium in the controls. Allele and genotype distribution of XRCC2-Arg188His, XRCC3-Thr241Met, and RAD51-G135C variants XRCC2-Arg188His variant analysis. Genotype frequency distribution of the XRCC2-Arg188His variant (Arg/Arg, Arg/His and His/His genotypes) was 79.9%, 19.3% and 0.8% in controls and 76.6%, 21.6% and 1.8% in BC cases, respectively. However, no significant association was observed between the genotype and allele frequency distribution of the XRCC2-Arg188His variant in BC cases and controls (Table 3) (S1 File).

Association between HRR gene polymorphisms and BC clinicopathological characteristics
To evaluate the association between the HRR gene polymorphisms and various BC clinicopathological features, BC patients were stratified based on their genotypes and clinicopathological characteristics. Pertaining to the XRCC3 Thr241Met variant, the heterozygous (Thr/
Concerning the XRCC2 Arg188His variant and clinicopathological characteristics, we did not observe a significant association between various BC clinicopathological characteristics and Arg188His polymorphism in the breast cancer patients (S1 Table).

MDR analysis
The combined effect of HRR gene variants (XRCC2-Arg188His, XRCC3-Thr241Met and RAD51-G135C) on BC risk was evaluated using MDR analysis. MDR evaluates the effect of SNP-SNP interaction on the risk of developing a multi-factorial disease such as breast cancer. MDR divides the data into a training dataset (9/10) and an independent testing dataset (1/10). A higher TBA value indicates that the observed interaction accurately predicts the case-control status. Moreover, a TBA score greater than 0.5 indicates that the interaction combination observed is not by chance, and a score of 1.00 highlights that the observed interaction combination is the best [27]. Furthermore, the model that has the highest TBA and CVC can be identified as the best interaction model. Fig 1 represents the various models predicted by MDR.
In the present study, MDR analysis revealed that among the different models, the interaction between XRCC3-Thr241Met and RAD51-G135C variants were observed to be the best interacting model under the two-loci model (TBA: 0.538; CVC: 10/10). Further investigation of the combinatorial impact of HRR variants on BC risk showed that the XRCC2 -Arg/Arg, XRCC3 -Thr/Thr, and RAD51 -C/C genotype combination was elevated in BC cases compared to controls, thereby conferring elevated risk of BC [OR: 3.29 (95%CI: 1.17-9.23); p = 0.024] ( Table 5). Fig 2 depicts the first ten protein partners that interact with XRCC2, XRCC3, and RAD51, which includes BRCA1, BRCA2, and RAD51D proteins. The STRING database collects and integrates all the publicly available protein-protein interaction sources and aids in the visualization of both direct (physical) and indirect (functional) interactions [28].

Discussion
Screening for certain commonly occurring polymorphisms has advanced our understanding of the crucial role played by genetics in BC predisposition. Moreover, various studies have reported that subtle variation in DNA repair capacity caused by the combination of low-penetrant genes and other influential factors such as environment modulates BC risk. Also, defects in DNA DSB repair has been identified as a common denominator for mammary carcinogenesis. Moreover, the RAD51 gene family, including XRCC2, XRCC3, and RAD51, is highly polymorphic in nature [29]. Previous reports on the impact of HRR pathway gene variations from various ethnicities have yielded inconsistent results. Hence, the present study aims to clarify the role of genetic variants in crucial HRR genes (XRCC2, XRCC3, and RAD51) towards BC development.
In the present study, we found that the XRCC2 Arg188His variant was not associated with BC predisposition in South Indian women. In line with our report, various studies observed a similar lack of association in Pakistani [30], Caucasian [26], Portuguese [31], African-American and white [32] women. Similarly, various meta-analysis studies investigating the role of the XRCC2 Arg188His variant on BC development reported that the XRCC2 Arg188His variant was not directly associated with BC predisposition [33][34][35]. Interestingly, a study by Silva et al. [31] highlighted that individuals who have never-breast fed and are heterozygous (Arg/ His) for the XRCC2 rs3218536 variant had reduced risk for BC.

PLOS ONE
Investigating the role of XRCC3 Thr241Met variant and BC risk, we observed that the Met allele was associated with BC risk. However, in the present study, we found that the heterozygous (Thr/Met) and homozygous mutant (Met/Met) genotypes were not significantly associated with BC risk. A recent meta-analysis study based on 55 case-control studies on XRCC3 Thr241Met variant and BC risk concluded that the XRCC3 Thr241Met variant was associated with BC risk in Arabian and Asian populations [36]. Additionally, another metaanalysis concluded that the XRCC3 Thr241Met variant was associated with a weakly elevated BC risk [37]. Similarly, another study on Thai [38] and South American [24] women reported that the 241Met carriers were at elevated BC risk. Furthermore, a study by Santos et al. [39] additionally observed that the XRCC3 Thr241Met variant was slightly associated with an increased risk of BC in individuals with elevated chromosomal damage. However, reports from certain ethnicities have observed a lack of association between XRCC3 Thr241Met variant and BC risk [25,26,[40][41][42][43].
Interestingly, various studies reported that the RAD51 G135C variant modified BC risk in BRCA2 mutation carriers [44,45]. Antoniou et al. [44] suggested that the RAD51 G13C variant located in the 5'UTR region might also affect alternate splicing. Thus, the RAD51 135C allele might cause an overall decrease in RAD51 protein abundance. The present study investigated the role of RAD51 G135C variant and BC risk, and we observed that the homozygous mutant (C/C) genotype was associated with an elevated risk of BC in South Indian women. A similar association was observed in a study conducted on mixed ethnicity (subjects were pooled from 19 studies including 13 countries) [44], Polish [40], and European [22] women. We also observed that the mutant CC genotype elevated the risk of metastasis in individuals with the homozygous mutant genotype. In line with our report, Weigmans et al. [46] speculated that breast tumors that overexpress RAD51 might have an elevated chance of disease progression and metastasis. Additionally, they observed that RAD51 promotes the expression of pro-metastatic genes and decreases the metastasis suppressor gene expression. Several meta-analyses highlighted the RAD51 G135C variant could function as a potential candidate biomarker for various cancers, particularly breast cancer [47][48][49]. Sekhar et al. [50] observed that the homozygous mutant variant (C/C) elevated BC risk in an ethnic-specific manner. However, on the other hand, few studies suggested that RAD51 tolerates very minimal dysfunctional sequence variation, and the RAD51 G135C variant might not contribute towards BC susceptibility [23,26,51]. In contrast, another study in Polish women highlighted that the RAD51 135C allele reduced the risk of BC in BRCA1 5382insC mutation carriers [52]. Furthermore, the identification of RAD51 foci in gBRCA mutant tumors was correlated with PARP inhibitor resistance [53,54]. Overall, RAD51 could be therapeutically targeted in the future using small molecule inhibitors [55] and could be used as a marker and target in neoadjuvant endocrine treatment [56]. Up to our knowledge, the present study is the first to report the impact of genetic polymorphisms in crucial HRR gene on BC development in South Indian women. Future studies that investigate the mechanistic role of HRR gene polymorphisms on BC predisposition, and studies that evaluate the prognostic potential of these SNPs might enhance our current understanding of the precise role played by these variants towards mammary carcinogenesis. Investigating common genetic variants that predispose BC development in developing nations such as India might aid primary health care providers in formulating schemes that could enable early diagnosis and lessen disease burden.
Supporting information S1 File. Genotyping results of the study participants. (XLSX) S1