Systematic evaluation of the association between a missense variant in XRCC3 gene splicing site and the pathogenesis of ovarian cancer Open Access

Worldwide, ovarian cancer (OC) is one of the most common gynecologic cancers, and has the highest mortality rate between them and occupy the third place in mortality, only after cervical and uterine cancer [1]. At present, the global morbidity of OC is approximately 3.4%, and is on the rise [2]. It is now clear that the occurrence of OC can be caused by a variety of physiological and environmental high-risk factors such as sustained ovulation, industrial physical and chemical pollutants, etc. However, under the same level of environmental exposure, individuals have different susceptibility to OC. This suggests that the difference of genetic background may also play a key role in the pathogenesis of OC [3].

Human genomic DNA is often damaged by internal and exogenous factors, such as metabolic free radicals, viruses, ultraviolet radiation, ionizing radiation, etc. If the damaged DNA cannot be repaired effectively in time or erroneously, it may affect the specific functions of gene execution and increase the risk of cell canceration [4]. Single-nucleotide polymorphism (SNP) is the most common genetic variation in humans, which can regulate gene expression or change the function of gene products to determine individual susceptibility to disease. The X-ray cross-complementing group 3 (XRCC3) plays a key role in repairing the DNA double-strand breaks and maintaining the functional integrity of the genome. Studies showed that the SNPs of XRCC3 gene may affect its DNA damage repair efficiency and be associated with the pathogenesis of tumors [5–7].

Single nucleotide variant rs861539 G>A is located at the 241st amino acid in the coding region of XRCC3 gene, which can cause the coding amino acid changed from Threonine to Methionine (Thr241Met), and is defined as a missense mutation. In addition, rs861539 is also a SNP of a splicing site of XRCC3 gene [8]. This suggests that rs861539 may play an important regulatory role in the post-transcriptional expression of XRCC3 gene and the functional execution of XRCC3 protein products. Growing evidence suggests that rs861539 changes the function of XRCC3 and is involved in DNA damage and subsequent susceptibility to carcinogens.

Currently, a series of studies have investigated the relationship between the rs861539 and OC pathogenesis, but the conclusions are still contradictory [9–18]. So far, the relationship between rs861539 and the pathogenesis of OC, the strength of the relationship and the underlying functional mechanism are not clear. In view of this, based on the principles and methods of evidence-based medicine, this study objectively evaluated and clarified the impact of XRCC3 rs861539 G>A on the incidence of ovarian cancer, identified the potential mechanism, and provided scientific basis for the prevention of OC.

Joint search of China national knowledge infrastructure (CNKI), Wanfang Medical Network, Web of Science and NCBI PubMed database, with ‘XRCC3’, ‘DNA repair gene’, ‘polymorphism’, ‘ovarian cancer’, ‘risk’ as the search subject words or keywords, to screen the published studies on the relationship between the XRCC3 gene rs861539 and the pathogenesis of OC. The deadline is April 10, 2023. In addition, manually retrieve the references to be included in the studies to obtain the target literature.

The studies included in the meta-analysis should meet the following criteria: (1) published case–control studies on the relationship between XRCC3 rs861539 variant and OC risk, (2) definite clinical and pathological diagnosis in the patients of OC, (3) comparable non-tumor subjects in control group, and (4) providing full genotype or allele frequency data in included studies. The exclusion criteria were as follows: (1) animal and cell experiments, (2) the studies being irrelevant to the relationship between XRCC3 rs861539 and OC genetic predisposition, (3) unavailable genotyping data, and (4) abstract, case report, review, systematic review or meta-analysis.

The quality of the included studies was assessed by two investigators (QL Liang and GC Huang) by using the Newcastle–Ottawa scale (NOS). Totally, eight items were used to evaluate the selection, comparability and exposure of the study population. Specifically, if one item is met, 1 point can be obtained (except for the item of comparability containing two points), with 0–9 points. Among them, 0–3 points are rated as low quality, 4–6 points as medium quality and 7–9 points as high quality documents.

The two investigators independently extracted necessary data from the eligible studies and then crosschecked to resolve the disagreements. The information of first author, paper publication year, country, ethnicity, age (years old), genotype frequency of cases and controls, number of case and control groups in the included studies was extracted.

The odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were used to assess the strength of association between rs861539 and the OC risk. The heterogeneity across the studies was assessed by the Q-test, and was considered significant when a P-value less than 0.1 [19]. A fixed-effect model was used to calculate the pooled OR if the heterogeneity was not significant; otherwise, the random-effect model was adopted [20]. The subgroup analysis in ethnicity was also performed to detect whether Caucasian populations and Asian populations have different susceptibility to OC. Meanwhile, publication bias was tested by Begg’s method with funnel plot to identify significant publication bias [21], and the sensitivity analyses were also done to assess the influence of individual study on the pooled ORs. All data processing was undertaken by using the Stata software, version 12.0 (Stata Corp LP, College Station, TX, U.S.A.).

In addition, the false-positive report probability (FPRP) analysis was carried out to evaluate the robustness of positive associations found in the pooled analyses. The cut-off value of 0.2 and prior probability of 0.1 was set to detect the efficacy for OR of 1.2 or 1/1.2 that is most likely. True associations would be considered when FPRP value was <0.2 [22].

Random errors may increase considerably when new emerging researches were continuously included to calculate the significance, causing a false conclusion in the meta-analysis [23]. Therefore, a trial sequential analysis (TSA) was performed to decrease the random errors in present study. In parameter settings, the Type I error (α = 5%), the Type II error (β = 20%), the relative risk reduction (RRR = 20%) and two-sided graph plots were set, respectively. TSA can help to calculate the required information size (RIS) and adjust P-value with trial sequential monitoring boundary. When the TSA monitoring boundary crosses with cumulative Z-curve before the RIS is reached, it suggests that significant evidence is confirmed and further trials are not necessary [24]. The related data processing was undertaken by using the TSA software (version 0.9.5.10 beta).

Given the fact that the rs861539 is a missense variant and located in a potential splice site of the XRCC3 gene, the SNPinfo Web Server (https://manticore.niehs.nih.gov/snpinfo/guide.html) [8], Alternative Splice Site Predictor (ASSP) tool (http://wangcomputing.com/assp/index.html) [25] and VarNote-REG (http://www.mulinlab.org/varnote/application.html#REG) [26] online tools were adopted to analyze the effect of rs861539 G>A on expression regulation of XRCC3 gene. Meanwhile, Polyphen-2 (http://genetics.bwh.harvard.edu/pph2/) [27] was taken to predict the possible classification of effect on protein′s structure and function according to different sequence and structural features of the amino acid substitutions positions. The protein ID of XRCC3 was obtained from Uniport database, and the protein structure with wild-type to mutant amino acid was analyzed by Pymol 1.4.1 [28] (The PyMOL Molecular Graphics System), which would help to visualize the changes in the amino acid. The pdb filetext was downloaded from Protein Data Bank (PDB).

After searching of China national knowledge internet (CNKI), Wanfang Medical Network, Web of Science and NCBI PubMed databases, a total of 252 studies on the relationship between XRCC3 gene and OC were preliminarily retrieved. Due to 35 duplications, 217 studies were continuously screened. The literature was retrieved strictly followed the inclusion and exclusion criteria. First of all, 163 papers such as abstracts, case reports, reviews, systematic reviews or meta-analyses and unrelated studies were excluded by checking title and abstract. And then, by reading the original text in detail, 44 studies did not involve the necessary genotyping data of rs861539 variant were excluded. Finally, 10 eligible studies involving 14 research data and 16,579 subjects (6,375 cases and 10,204 controls) were recruited in present meta-analysis. See Figure 1 for the details.

The quality of studies included in this meta-analysis was assessed by utilizing the NOS. After quality evaluation, two study data were considered as low quality, four data as medium quality and eight data as high quality. The main characteristics of included studies are shown in Table 1.

The Chi-square Q test showed that there was no significant heterogeneity in the included data under the five genetic models (dominant genetic model [GA + AA vs. GG: I² = 0.30%, P_{heterogeneity} = 0.445], the recessive model [AA vs. GG + GA: I² = 0.00%, P_{heterogeneity} = 0.734], the heterozygous model [GA vs. GG: I² = 0.00%, P_{heterogeneity} = 0.575], the homozygous model [AA vs. GG: I² = 0.00%, P_{heterogeneity} = 0.520] and the allelic model [A vs. G: I² = 0.00%, P_{heterogeneity} = 0.503]). Thus, the subsequent meta-analysis was performed using a fixed-effects model.

The results of meta-analysis showed that rs861539 was significantly associated with the genetic susceptibility of OC. Comparing with the GG genotype, GA + AA genotypes could significantly reduce the risk of OC, the ORs and their 95% CIs were 0.89 (0.83–0.95), P=0.001 in the dominant genetic model, 0.88 (0.82–0.95), P=0.001 in the heterozygous genetic model. Under the allelic model, compared with the G allele, the A allele could significantly reduce the OC risk, the OR and 95% CI was 0.94 (0.89–0.98), P=0.007. However, no significant association between rs861539 and OC risk was found under the recessive genetic model and the homozygous genetic model. As shown in Figure 2 and Table 2.

The subgroup analysis in ethnicity suggested a protective effect of rs861539 on OC risk in Caucasian populations under the dominant genetic model (OR = 0.88, 95% CI: 0.82–0.94, P<0.001), the homozygous model (OR = 0.89, 95% CI: 0.80–0.98, P=0.024), the heterozygous model (OR = 0.87, 95% CI: 0.81–0.94, P<0.001) and the allelic model (OR = 0.93, 95% CI: 0.88–0.97, P=0.003) except for the recessive model. However, there was no significant association observed in the Asian populations. As demonstrated in Figure 2.

The evaluation of publication bias by Begg's test showed that the included studies were generally symmetrical distributed on both sides of the symmetry axis. It suggests that there may not be significant publication bias in the included literature in present study. As shown in Figure 3.

In this study, the sensitivity analysis was conducted by eliminating one study at a time to detect the influence of a single study data on the overall pooled effect. The results showed that no single study significantly changed the pooled OR and its corresponding 95% CI, indicating that the results of this meta-analysis were robust and the study conclusions were relatively reliable (data was not shown).

To verify the statistically significant associations, we performed the FPRP analysis with the presetting threshold of 0.2 and prior probability of 0.1. All significant associations found in the meta-analysis were considered to be true, and FPRP values for effects of rs861539 on OC risk in the dominant model, heterozygous model and allelic model were 0.009, 0.010 and 0.059, respectively. Moreover, it was found that the FPRP values were 0.001, 0.189, 0.001 and 0.026 for the effects of rs861539 on Caucasians OC risk in the subgroup analysis stratified by ethnicity under the dominant, homozygous, heterozygous and allelic models, respectively. The surprising findings indicated that the rs861539 leading a credibly decreased OC risk among Caucasian populations. As shown in Table 3.

The results of Trial Sequential Analysis (TSA) support the conclusions drawn from the meta-analysis. As depicted in Figure 4, although the cumulative Z-score line did not reach RIS, it crossed the traditional boundary and the TSA monitoring boundary, indicating that accumulative evidence was adequate to the significant associations of rs861539 and OC risk in the dominant model, the heterozygous model and the allelic model, and no further trials were required to verify these conclusions.

Since the missense mutation rs861539 of XRCC3 gene splicing site is significantly associated with OC susceptibility, we evaluated the effect of this variant on the XRCC3 gene post-transcriptional splicing and structure or function of the coding protein of XRCC3. The analysis of Exonic Splicing Enhancer (ESE) method showed that the genetic variation might affect the efficacy of splicing factors (such as SRp 55 and SF2ASF1/2). As shown in Figure 5. The ASSP tool analysis showed that XRCC3 rs861539 G>A may give rise to changes in the activity of putative splice sites. We found that score activation was 3.715 in the putative splice sites 102bp position of examined DNA sequence with rs861539 G allele, while the score activation changed to 3.272 in the same position with rs861539 A allele. The resulting change was that the putative splice site at 102 bp position changed from defined as unclasssified 3′ splice site to Alt.isoform/crytic 3′ splice site, which meaned rs861539 might have different post-transcriptional splicing regulation performance under different alleles. See Figure 6A,B.

Meanwhile, the GWAS4D prediction results revealed that rs861539 may act as an expression Quantitative Trait Loci (eQTL), affecting the expression of functional genes such as XRCC3, MARK3, APOPT1, KLC1, etc and was associated with the risk of OC in individuals (Table 4). In addition, the Polyphen2 tool analysis showed that the point mutation was possibly damaging (0.541) for structure and function of the XRCC3 protein. See Figure 7A. And the crystallographic structure changing from native amino acid to mutant amino acid analyzed by PyMOL Molecular Graphic -s System, Version 1.4.1 was shown in Figure 7B.

As a member of DNA damage repair genes family, the XRCC3 plays a key role in maintaining the functional integrity of the human genome by participating in recombination repair after genomic DNA double-strand breaks damage [29]. It is believed that if the integrity of the XRCC3 gene sequence or the function of the coding product is disturbed, its DNA damage repair efficiency will be weakened. Experimental study had also confirmed that XRCC3-deficient cells are more sensitive to cisplatin and chemoradiotherapy [30]. Studies showed that rs861539 variant of XRCC3 gene was significant associated with the onset of multiple malignancies, such as gynecological malignancies, oesophageal cancer, prostate cancer, breast cancer, etc [5,7,31–33]. At present, there have been also a series of studies on the relationship between rs861539 and the incidence of OC, but the conclusions are contradictory.

In present meta-analysis, we confirmed that rs861539 was significantly associated with the risk of human OC under the most genetic models, particularly in Caucasians. The conclusion of the present study is similar to that of a previous meta-analysis [34] but is inconsistent with the meta-analysis conclusions of Yuan C [35] and Yan Y [36]; that is, there is no positive associations between rs861539 and OC risk observed. In comparison, most of the studies they included had relatively small sample sizes, while this study further increased the number of studies and the sample sizes. The larger sample size makes the statistical tests more efficient and the conclusions more reliable [37]. In addition, a sensitivity analysis was conducted in the meta-analysis, which showed that after excluding a certain research data, the pooled OR and its corresponding 95% CI did not fluctuate significantly. Moreover, we performed a FPRP as well as a TSA to make the findings more credible. It indicates that the conclusions that rs861539 significantly affects the risk of OC among women in this study are robust and reliable. Therefore, the early identification of risk genotypes or allele carriers of rs861539 accordingly is of great clinical and social significance for the early prevention and control of OC.

However, the biological mechanisms behind the significant associations between rs861539 and the susceptibility of women to OC, as well as the role of rs861539 in the pathogenesis of OC, have not yet been elucidated. In this study, we further utilized bioinformatics analysis to explore the possible biological functions of rs861539 and preliminarily revealed the potential molecular mechanisms by which rs861539 affects individuals' susceptibility to OC. According to the functional type and the gene region localization of rs861539, we explored its possible biological functions.

With the help of ‘SNP function prediction’ tool, it can be seen that rs861539 is located at the splicing site of XRCC3 gene [8]. It was found that different alleles of rs861539 could lead to differences in the activity scores of XRCC3 post-transcriptional related splicing sites and may produce a different splicing of the post-transcriptional mRNA sequence. This indicates that rs861539 has a significant regulatory effect on the post-transcriptional splicing of XRCC3 gene, which may be one of the mechanisms to change the genetic susceptibility of individuals to OC. Meanwhile, eQTL analyses base on the GTEx database, which incorporates 127 tissue/cell type-specific epigenome data sets [27], suggests that rs861539 could regulate the expression of series of functional genes. In addition, as a missense mutation SNP, rs861539 G>A can change the encoded threonine (Thr) to methionine (Met). The analysis of PolyPhen2 and Pymol online tools showed that rs861539 variant brought about some changes in the structure of XRCC3 protein, which may cause damage to the DNA repair ability of XRCC3. Based on the above exploration, it can be revealed that rs861539 affects the DNA damage repair efficiency of XRCC3 by regulating the expression of XRCC3 gene or changing the structure of the XRCC3 product. To sum up, rs861539 affects the post-transcriptional splicing of XRCC3 gene and the expression of functional genes by changing the activity of splice sites and types of splicing factors. This evidence may provide new clues for revealing the mechanism of OC susceptibility.

This study still has some limitations. First of all, all studies included in this study have been published, and there may be some publication bias. Secondly, because the occurrence and development of OC is the result of multiple factors, such as genetic, environmental and endocrine factors, it is difficult to eliminate the effect of confounding factors in meta-analysis, which may affect the final results and conclusions. Also, evaluation of the potential gene–gene and gene–environment interaction effects on the risk of OC was limited because of lacking the original data of the reviewed studies. Finally, the present study only used bioinformatics methods to explore the mechanism of rs861539 affecting OC susceptibility, but did not carry out cell and molecular biological experiments. This weakened the strength of the evidence to some extent. Therefore, it must be cautious in accepting the conclusion of the present study.

This study supports that XRCC3 rs861539 G>A variant is significantly associated with the risk of human OC, especially in Caucasians. Involvement in the post-transcriptional regulation of XRCC3 gene and the change of the structure of the encoded XRCC3 protein may be the biological mechanisms of rs861539 leading to individual susceptibility to OC. Our finds above provide new clues for researchers to further explore the biological mechanism of OC.

The data used to support the findings of the present study are included within the article.

The authors declare that there are no competing interests associated with the manuscript.

The study was supported by the Natural Science Foundation of Guangxi Province of China [grant number 2020GXNSFAA238025] and basic ability improvement project of young and middle-aged teachers in Guangxi colleges and universities [grant numbers 2020KY12028, KY2016LX236 and 2018glmcy 020].

Qiulian Liang: Formal analysis, Methodology, Writing—original draft. Gongchen Huang: Formal analysis, Writing—original draft. Ping Zhong: Data curation, Software. Dengting Deng: Data curation, Investigation. Lin Yang: Formal analysis, Writing—review & editing. Xiangyuan Yu: Conceptualization, Funding acquisition, Project administration.

ASSP

Alternative Splice Site Predictor

CI

confidence interval

eQTL

expression Quantitative Trait Loci

ESE

Exonic Splicing Enhancer

NOS

Newcastle–Ottawa scale

OC

ovarian cancer

OR

odds ratio

PDB

Protein Data Bank

SNP

Single nucleotide polymorphism

XRCC3

X-ray cross-complementing group 3