Relationship between XPA, XPB/ERCC3, XPF/ERCC4, and XPG/ERCC5 Polymorphisms and the Susceptibility to Head and Neck Carcinoma: A Systematic Review, Meta-Analysis, and Trial Sequential Analysis

Background and Objectives: Nucleotide Excision Repair (NER), the most extensively researched DNA repair mechanism, is responsible for repairing a variety of DNA damages, and Xeroderma Pigmentosum (XP) genes participate in NER. Herein, we aimed to update the previous results with a meta-analysis evaluating the association of XPA, XPB/ERCC3, XPF/ERCC4, and XPG/ERCC5 polymorphisms with the susceptibility to HNC. Materials and Methods: PubMed/Medline, Web of Science, Scopus, and Cochrane Library databases were searched without any restrictions until 18 November 2023 to find relevant studies. The Review Manager 5.3 (RevMan 5.3) software was utilized to compute the effect sizes, which were expressed as the odds ratio (OR) with a 95% confidence interval (CI). Results: Nineteen articles were involved in the systematic review and meta-analysis that included thirty-nine studies involving ten polymorphisms. The results reported that the CC genotype of rs17655 polymorphism showed a significantly decreased risk of HNC in the recessive model (OR: 0.89; 95%CI: 0.81, 0.99; p-value is 0.03). In addition, the CT genotype (OR: 0.65; 95%CI: 0.48, 0.89; p-value is 0.008) of the rs751402 polymorphism was associated with a decreased risk, and the T allele (OR: 1.28; 95%CI: 1.05, 1.57; p-value is 0.02), the TT (OR: 1.74; 95%CI: 1.10, 2.74; p-value is 0.02), and the TT + CT (OR: 2.22; 95%CI: 1.04, 4.74; p-value is 0.04) genotypes were associated with an increased risk of HNC. Conclusions: The analysis identified two polymorphisms, rs17655 and rs751402, as being significantly associated with the risk of HNC. The study underscored the influence of various factors, such as the type of cancer, ethnicity, source of control, and sample size on these associations.


Introduction
According to the most recent GLOBOCAN data (2020), head and neck cancer (HNC) is the seventh most prevalent cancer worldwide, with approximately 890,000 new cases (about 4.5% of all global cancer diagnoses) and 450,000 deaths annually (about 4.6% of all global cancer deaths) [1], and both its incidence and mortality rates are increasing [2,3].This concerning trend is largely due to avoidable risk factors such as tobacco and alcohol consumption, areca nut ingestion, and sexually transmitted HPV infections [1].HNC originates from epithelial cells and typically develops in the oral cavity, pharynx, and larynx [4].
While environmental carcinogens and cancer-causing viruses are the primary causative factors, it is clear that genetic predisposition also plays a significant role in modulating the risk of HNC [5][6][7][8][9].Nucleotide Excision Repair (NER), the most extensively researched DNA repair mechanism, is responsible for repairing a variety of DNA damage, including thymidine dimers, oxidative DNA damage, bulky DNA adducts, and cross-links [10].NER is a flexible system that detects and repairs DNA damage induced by both internal and external factors, including therapeutic drugs [11].Mutations in NER factors can impair cellular health and lead to human diseases, such as Xeroderma Pigmentosum (XP) [12].Xeroderma Pigmentosum (XP) is a rare genetic disorder that is inherited in an autosomal recessive manner.It is characterized by a deficiency in the NER mechanism due to single-nucleotide mutations [13].XP can be further classified into seven unique subgroups, referred to as complementation groups, ranging from XPA to XPG.These are typically seen as a less severe variant form of the disorder [14].Each of these complementation groups signifies the existence of a causative mutation in one of the seven XP genes that participate in NER [15].
DNA repair mechanisms are crucial in safeguarding cells from DNA damage and preserving genomic integrity [16].The NER pathway is the main method for eliminating bulky adducts from DNA, serving as a key component of the cellular defense against a wide range of structurally unrelated DNA lesions [17].The NER pathway comprises several stages: The initial stage of the NER pathway involves the identification of damage by a protein complex, which includes XPC.The subsequent stage entails the unwinding of the DNA by a complex that includes XPD and the excision of the damaged single-stranded nucleotide fragment by molecules such as XPG, ERCC1, and XPF [18][19][20].
Three meta-analyses reported the association between XP polymorphisms and the risk of HNC risk with different results [21-23], as well as original studies [24][25][26][27].Therefore, we aimed to design a meta-analysis evaluating the association of XPA, XPB/ERCC3, XPF/ERCC4, and XPG/ERCC5 polymorphisms with the susceptibility to HNC.

Study Design
The meta-analysis was carried out following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [28].The research question, framed in the context of PECO (population, exposure, comparison, and outcome), was as follows: is there a correlation between XPA, XPB, XPF, and XPG polymorphisms and the susceptibility to HNC in case-control studies?

Identification of Articles
A researcher (M.S.) conducted a thorough literature review using various databases, including PubMed/Medline, Web of Science, Scopus, and Cochrane Library, without any restrictions until 18 November 2023, to find relevant studies.Another researcher (E.S.) reviewed the titles and abstracts of the retrieved articles and obtained the full texts of those that met the inclusion criteria.The search strategy included ("xerodermapigmentosum" or "xeroderma pigmentosum" or "xeroderma pigmentosa" or "ERCC*" or "excision repair cross-complementing") and ("oral" or "OSCC" or "oral squamous cell" or "head and neck" or "HNSCC" or "salivary gland" or "nasopharyngeal" or "nasopharynx" or "tongue" or "nasopharyngeal" or "oropharyngeal" or "oropharynx" or "laryngeal" or "hypopharyngeal" or "oropharyngolaryngeal" or "hypopharynx" or "oral cavity") and ("cancer*" or "tumor*" or "carcinoma*" or "neoplasm*") and ("gene*" or "polymorphism*" or "variant*" or "genotype*" or "allele*").To ensure no significant study was missed, the reference lists of the articles were also examined.The search and selection procedures were further confirmed by another author (M.M.I.).

Selection Criteria
The exclusion criteria include review articles, meta-analyses, and systematic reviews.Articles with incomplete data or those that did not include a control group are also excluded.Studies conducted on animals, conference papers, and comment papers are not considered.Duplicate studies and book chapters are also excluded.Studies where the control group Medicina 2024, 60, 478 3 of 14 included individuals with any systemic disease or where the cases were under treatment are not considered.

Data Summary
Two authors (E.S. and M.S.) collected the data from the included studies in the metaanalysis independently.Any differences that arose were settled through joint discussion.

Quality Evaluation
The quality of the studies was evaluated by one author (M.S.) using the Newcastle-Ottawa Scale (NOS) tool, a well-established method for assessing the quality of nonrandomized studies in meta-analyses [29].The NOS tool assigns a maximum score of 9 for case-control studies, with a score of 7 or above indicating high quality.The scores were then reviewed for accuracy by another author (E.S.).Any discrepancies between the authors were resolved through discussion, ensuring a thorough and consensus-based evaluation process.

Statistical Analyses
The Review Manager 5.3 (RevMan 5.3) software was utilized to compute the effect sizes, which were expressed as the odds ratio (OR) with a 95% confidence interval (CI).This reflected the prevalence of XPA, XPB, XPF, and XPG polymorphisms in patients with HNC and control subjects.The pooled OR's significance was ascertained using the Z-test, with a two-sided p-value of less than 0.10 indicating significance.A random-effects model [30] was employed if P heterogeneity was <0.10 (I 2 > 50%), signifying significant heterogeneity.If the heterogeneity was not significant, a fixed-effect model [31] was used.
A subgroup analysis was conducted to pinpoint any significant differences in the pooled ORs within these groups.Additionally, a meta-regression analysis was performed using a random-effects model to depict a linear relationship between the auxiliary variables in the study and the effect size.The presence of publication bias was evaluated using a Begg's funnel plot and Egger's regression test.The p-values from both Egger's and Begg's tests were computed, with a two-sided p-value less than 0.10, indicating the presence of publication bias.
In terms of sensitivity analysis, two methods were employed to assess the stability and consistency of the pooled ORs: The "one-study-removed" analysis was conducted to ascertain if any single study had a disproportionate impact on the overall estimate.The "cumulative" analysis was conducted to evaluate the impact of each additional study on the overall estimate.All these analyses related to publication bias and sensitivity analyses were performed using the Comprehensive Meta-Analysis version 3.0 (CMA 3.0) software.
The STRING database version 12.0 (https://string-db.org/),used for protein-protein interaction (PPI) network analysis, was utilized to investigate the functional interactions among the genes under studyThe sources of interaction and species were restricted to "Homo sapiens", and an interation score exceeding 0.900 was used to construct the PPI networks.In these networks, the nodes represent proteins, and the edges represent the interactions between them.This method was employed to explore potential interactions among differentially expressed genes (DEGs) associated with various tissues.
To mitigate the risk of drawing false-positive or negative conclusions from metaanalyses [32], a TSA was performed.This analysis was conducted using the TSA software (version 0.9.5.10 beta) [33].The TSA allows for the setting of a futility threshold, which can identify a non-effect outcome before the necessary information size is reached.The required information size (RIS) was determined with an alpha risk of 5%, a beta risk of 20%, and a two-sided boundary type.The heterogeneity (D 2 ) was evaluated for the prevalence of the XPA and XPG polymorphisms in HNC patients and controls.If the Z-curve intersected the RIS line, it suggested that the studies included an adequate number of cases and that the conclusions were trustworthy.If not, it indicated that the available information was inadequate and more data were needed.

Study Selection
A total of 941 records were identified among the databases and the electronic sources (Figure 1).After removing irrelevant records, reading the titles/abstracts, and then excluding full-text articles with reasons, 19 articles were involved in the systematic review and meta-analysis.These articles included 39 studies involving 10 polymorphisms.
required information size (RIS) was determined with an alpha risk of 5%, a beta risk of 20%, and a two-sided boundary type.The heterogeneity (D 2 ) was evaluated for the prevalence of the XPA and XPG polymorphisms in HNC patients and controls.If the Z-curve intersected the RIS line, it suggested that the studies included an adequate number of cases and that the conclusions were trustworthy.If not, it indicated that the available information was inadequate and more data were needed.

Study Selection
A total of 941 records were identified among the databases and the electronic sources (Figure 1).After removing irrelevant records, reading the titles/abstracts, and then excluding full-text articles with reasons, 19 articles were involved in the systematic review and meta-analysis.These articles included 39 studies involving 10 polymorphisms.

Pooled Analysis
A summary of forest plot analyses is reported for the association of each polymorphism in five genetic models with the risk of HNC (Table 2).The forest plots are included in Supplementary File S1.The results reported that the CC genotype of rs17655 polymorphism showed a significantly decreased risk of HNC in the recessive model (OR: 0.89; 95%CI: 0.81, 0.99; p-value is 0.03).In addition, the CT genotype (OR: 0.65; 95%CI: 0.48, 0.89; p-value is 0.008) of the rs751402 polymorphism was associated with a decreased risk, and the T allele (OR: 1.28; 95%CI: 1.05, 1.57; p-value is 0.02), the TT (OR: 1.74; 95%CI: 1.10, 2.74; p-value is 0.02), and the TT + CT (OR: 2.22; 95%CI: 1.04, 4.74; p-value is 0.04) genotypes were associated with an increased risk of HNC.Therefore, among ten polymorphisms, just the rs17655 and rs751402 polymorphisms were associated with the HNC risk.

Subgroup Analysis
A subgroup analysis was performed on the pooled analyses of two polymorphisms (rs17655 and rs1800975) with sufficient studies (Table 3).With regards to the rs17655 polymorphism, the results suggested that in the Asian population, individuals with the C allele or the CC genotype have a decreased risk of HNC.In larger studies (sample size ≥ 400), the CC + CT genotype had an increased risk of HNC.In laryngeal cancer cases, the CC genotype had a decreased risk of HNC.In hospital-based controls, the CC genotype and, in population-based controls, the CC + CT genotype had a decreased risk of HNC.Therefore, the findings suggested that the cancer subtype and the characteristics of the study population (such as ethnicity, control source, and sample size) can influence the association of the rs17655 polymorphism with the risk of HNC.
With regards to rs1800975 polymorphism, the results suggested that in the Caucasian population, individuals with the A allele and the A allele and AA and AA + GA genotypes have a decreased risk of HNC.In hospital-based controls and individuals with oral cancer, the AA genotype had a decreased risk of HNC.Therefore, the findings suggested that the cancer subtype and the characteristics of the study population (such as ethnicity and control source) can affect the association of rs1800975 polymorphism with the risk of HNC.

Meta-Regression Analysis
Table 4 reports a random-effect meta-regression analysis for two polymorphisms with sufficient studies (rs17655 and rs1800975).The results showed that the publication year, sample size, and quality score were not confounding factors for these polymorphisms.

Sensitivity Analysis
Both "cumulative" and "one-study-removed" analyses showed the stability of the pooled results for the rs17655, rs1047768, rs6498486, and rs1800975 polymorphisms.To remove the studies with a deviation of HWE in controls showed that in contrast to the initial pooled analysis, the CC genotype of the rs17655 polymorphism did not associate with the risk of HNC, but this removal did not change the initial pooled analysis of the rs1800975 polymorphism (Table 5).Therefore, a deviation of HWE in controls can impact the association of rs17655 polymorphism with the risk of HNC but not for the rs1800975 polymorphism.

TSA
In other cases, the Z-curve not crossing the RIS line in the TSA for the rs17655 and rs1800975 polymorphisms associated with HNC risk in five genetic models suggests that the current evidence is not sufficient to conclusively determine the association.More trials may be needed to reach a definitive conclusion.Supplementary File S2 shows the TSA plots.

Publication Bias
Supplementary File S3 shows the funnel plots for the rs17655 and rs1800975 polymorphisms in five genetic models.Begg's test showed publication bias for the rs1800975 polymorphism in the allelic (p = 0.083) and the recessive (p = 0.083) models.

STRING Results
Figure 2 shows the PPI network graph and heatmap for the XPA, XPB/ERCC3, XPF/ERCC4, and XPG/ERCC5 from the STRING database.Among the interactions, there are curated and experimental interactions between all XPs together.

Discussion
The analysis revealed that only two polymorphisms, rs17655 and rs751402, were associated with HNC risk.The influence of factors such as the type of cancer, ethnicity, source of control, and sample size on these associations was highlighted in the subgroup analysis.The association of rs17655 with HNC risk was affected by deviations from HWE in controls.Further studies may be required for a conclusive result.There was a moderate publication bias for rs1800975 polymorphism.Interactions between all XPs were demonstrated in the PPI network.
NER is a well-researched pathway in the human body that repairs various forms of damage to double-helix DNA.This process involves four steps: identifying the lesion, marking and unwinding the damaged DNA segment, excising the oligonucleotide, and ligating new strands [42,[49][50][51].Variations in the core NER genes can alter the NER capa-

Discussion
The analysis revealed that only two polymorphisms, rs17655 and rs751402, were associated with HNC risk.The influence of factors such as the type of cancer, ethnicity, source of control, and sample size on these associations was highlighted in the subgroup analysis.The association of rs17655 with HNC risk was affected by deviations from HWE in controls.Further studies may be required for a conclusive result.There was a moderate publication bias for rs1800975 polymorphism.Interactions between all XPs were demonstrated in the PPI network.
NER is a well-researched pathway in the human body that repairs various forms of damage to double-helix DNA.This process involves four steps: identifying the lesion, marking and unwinding the damaged DNA segment, excising the oligonucleotide, and ligating new strands [42,[49][50][51].Variations in the core NER genes can alter the NER capability by affecting the expression and functionality of key proteins [42,52,53].Several factors, including post-translational modifications and interactions with other proteins, regulate the proteins' ability to engage in the NER pathway [54].In our meta-analysis, the PPI demonstrated that XPA, XPB, XPF, and XPG have robust interactions.
Certain XP proteins are involved in pathways that include repairing oxidative damage, removing DNA cross-links, and transcription [55,56].The study of XP genetics primarily centered on changes in expression levels and polymorphisms as they pertain to a range of physiological responses.These include susceptibility to specific types of cancer, reactions to DNA-damaging chemotherapy drugs, and aging [54].The binding of XPA to damaged DNA is significantly enhanced by its interaction with other components of the NER [57][58][59].DNA damage and repair can influence several cellular processes, such as replication and transcription, mutagenesis, and apoptosis.Therefore, they may play crucial roles in an organism's development and pathology, including cancer [60].
A meta-analysis [61] indicated that XPA is a minor risk factor for the development of cancer.Another meta-analysis [62] evaluated XPG polymorphisms and found that rs1047768 polymorphism was linked to an increased risk of lung cancer, rs2227869 polymorphism was associated with a decreased risk of cancer in population-based studies, and rs751402 and rs873601 polymorphisms were connected to the risk of gastric cancer.A metaanalysis by Jiang et al. [21] suggested that the rs17655 polymorphism was a risk factor for HNC susceptibility, particularly in laryngeal cancer and in the Asian population.However, another meta-analysis [22] suggested that the rs17655 polymorphism may not be associated with the genetic susceptibility of HNC overall but might contribute to HNC susceptibility in the European population.The meta-analysis by Wu et al. [23] indicated that the XPA rs1800975 polymorphism may not be associated with overall HNC susceptibility but with oral carcinoma susceptibility.The differences between results for subtypes of HNC can show that cancers originating from different sites in the HNC may have different tumor biology [63].
Our meta-analysis revealed that some XP polymorphisms are associated with HNC risk, and a subgroup analysis reported that factors such as the type of cancer, ethnicity, source of control, and sample size have an impact on this association.
This meta-analysis had many limitations: (1) The limited number of published studies on eight polymorphisms prevented us from conducting subgroup analysis or metaregression analysis.(2) High heterogeneity was observed among several analyses, possibly due to the small number of studies.(3) There was a lack of adequate participants in the analyses based on TSA.(4) Many studies deviated from HWE in controls.On the other hand, the study had three strengths: (1) Most of the studies scored high in terms of quality.
(2) Most analyses showed no publication bias.(3) The results were stable.

Conclusions
The analysis identified two polymorphisms, rs17655 and rs751402, as being significantly associated with the risk of HNC.Specifically, the CC genotype of rs17655 and the CT genotype of rs751402 were linked to a decreased risk of HNC, while the T allele and TT and TT + CT genotypes of rs751402 were connected to an increased risk.The study underscored the influence of various factors, such as the type of cancer, ethnicity, source of control, and sample size on these associations.However, the association of rs17655 with HNC risk was found to be influenced by deviations from HWE in controls.
The findings could potentially guide the development of personalized treatment strategies for HNC based on a patient's genetic profile.However, due to the moderate

Figure 1 .
Figure 1.Flowchart of study selection.Figure 1. Flowchart of study selection.

Figure 1 .
Figure 1.Flowchart of study selection.Figure 1. Flowchart of study selection.

Figure 2 .
Figure 2. Protein-protein interaction (PPI) network of the examined genes.The nodes indicate proteins, and the edges indicate the number of interactions.Different colors represent different levels of evidence of a connection between proteins.This analysis had an average confidence score of more than 0.900.The PPI enrichment had a p-value less than 5.49 × 10 −11 , a total of 4 nodes and 6 edges, and an average node degree of 3.

Figure 2 .
Figure 2. Protein-protein interaction (PPI) network of the examined genes.The nodes indicate proteins, and the edges indicate the number of interactions.Different colors represent different levels of evidence of a connection between proteins.This analysis had an average confidence score of more than 0.900.The PPI enrichment had a p-value less than 5.49 × 10 −11 , a total of 4 nodes and 6 edges, and an average node degree of 3.

Table 2 .
Summary of forest plot analyses.

Table 3 .
Subgroup analysis of two polymorphisms with sufficient studies.
N: Number of studies.

Table 5 .
Pooled analysis for the polymorphisms to remove the studies with a deviation of Hardy-Weinberg equilibrium in controls.