Association of N-acetyltransferases 1 and 2 Polymorphisms with Susceptibility to Head and Neck Cancers—A Meta-Analysis, Meta-Regression, and Trial Sequential Analysis

Background and objective:N-acetyltransferases 1 and 2 (NAT1 and NAT2) genes have polymorphisms in accordance with slow and rapid acetylator phenotypes with a role in the development of head and neck cancers (HNCs). Herein, we aimed to evaluate the association of NAT1 and NAT2 polymorphisms with susceptibility to HNCs in an updated meta-analysis. Materials and methods: A search was comprehensively performed in four databases (Web of Science, Scopus, PubMed/Medline, and Cochrane Library until 8 July 2021). The effect sizes, odds ratio (OR) along with 95% confidence interval (CI) were computed. Trial sequential analysis (TSA), publication bias and sensitivity analysis were conducted. Results: Twenty-eight articles including eight studies reporting NAT1 polymorphism and twenty-five studies reporting NAT2 polymorphism were involved in the meta-analysis. The results showed that individuals with slow acetylators of NAT2 polymorphism are at higher risk for HNC OR: 1.22 (95% CI: 1.02, 1.46; p = 0.03). On subgroup analysis, ethnicity, control source, and genotyping methods were found to be significant factors in the association of NAT2 polymorphism with the HNC risk. TSA identified that the amount of information was not large enough and that more studies are needed to establish associations. Conclusions: Slow acetylators in NAT2 polymorphism were related to a high risk of HNC. However, there was no relationship between NAT1 polymorphism and the risk of HNC.


Introduction
Cellular inflammation and immunity can play a significant role in various stages of carcinogenesis [1] such as head and neck cancers (HNCs). HNC mortality rates are elevating and disproportionately affect people in low-and middle-income countries and areas with restricted resources [2]. Global Burden of Disease Study (GBD) in 2016 estimated 512,492 deaths due to HNC (a minimum of 15,018 deaths in North Africa and the Middle East to a maximum of 199,280 in South Asia) and predicted the death count to reach 705,901 in 2030 [3,4]. HNC involves a series of tumors originating in the oropharynx, hypopharynx, oral cavity, lip, larynx, or nasopharynx [5]. Smoking, alcohol consumption, and high-risk human papillomaviruses have been related to HNC [5][6][7]. In connection with the role of genetics in HNC, several recent meta-analyses have reported the association of polymorphisms with the risk of HNCs [8][9][10][11].
A number of heterocyclic and aromatic amines are the main carcinogenic compounds of tobacco smoke [12,13] that their metabolism in humans is complex and includes acetylation as a main pathway for DNA mutation and the onset of carcinogenesis [14]. In particular, two N-acetyltransferases, NAT1 and NAT2 perform a role in catalyzing the deactivation and activation of several carcinogenic amines through N-and O-acetylation, respectively [14,15]. Both NAT genes (NAT1 and NAT2) have polymorphisms in humans and in accordance with slow and rapid acetylator phenotypes [16]. The NAT2 metabolized gene is located in region 10 of chromosome 8p21, which contains two exons with a long intron of about 8.6 kb [17]. Exon 1 is very short (100 bp) and the entire protein-coding region in Exon 2 is 870 bp [18]. Also, the NAT1 gene is located on the short arm of chromosome 8 (8p21) [19,20]. NAT1 accelerates acetylation specifically for arylamine receptor structures such as p-aminosalicylic and p-aminobenzoic acids [21] and NAT2 acetylates other arylamine-acceptor structures, such as isoniazid, sulfasalazine, procainamide, and caffeine [19].
Evidence from the published articles on the relationship between NAT1 and NAT2 polymorphisms and HNC susceptibility is conflicting [22,23]. The association between the polymorphisms (NAT1 and NAT2) and the HNC risk has been evaluated by one [24] and four [25][26][27][28] meta-analyses, respectively. However, these studies were published several years ago with the most recent one being published in 2015. Therefore, through this metaanalysis, we intend to update the evidence on the association between the polymorphisms and the HNC risk by including more studies. In addition, we aim to conduct trial sequential analysis (TSA) and meta-regression.

Study Design
The present meta-analysis follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocols [29]. The PI/ECO (population, intervention/exposure, comparison, and outcome) question was: Are polymorphisms of NAT1 and NAT2 associated with the risk of HNC?

Identification of Articles
A search was comprehensively performed by one author (M.S.) in four databases of Web of Science, Scopus, PubMed/Medline, and Cochrane Library until 8 July 2021, without any restrictions in language, publication year, age, and sex to retrieve the relevant articles ( Figure 1). The titles and abstracts of the relevant articles were assessed by the same author (M.S.); subsequently, the full-texts of the articles found to be relevant based on the eligibility criteria were downloaded. The search strategy included: ("N-acetyl transferases" or "Nacetyltransferase" or "NAT2" or "NAT1") and ("mouth" or "OSCC" or "oral" or "tongue" or "head and neck" or "HNSCC" or "nasopharyngeal" or "nasopharynx" or "oropharyngeal" or "salivary gland" or "laryngeal" or "larynx" or "hypopharyngeal" or "pharyngeal" or "pharynx" or "oral cavity" or "hypopharynx") and ("tumor" or "carcinoma" or "cancer" or "neoplasm") and ("allele" or "variant" or "polymorphism" or "genotype" or "gene"). The reference lists of the retrieved articles were reviewed to ensure that no important study was missed. Another author (H.M.) re-checked the process of searching and article selection. A lack of agreement between both authors was resolved by another author (J.T.).

Eligibility Criteria
The inclusion criteria were: (1) case-control studies reporting slow and rapid acetylators of NAT1 and NAT2 polymorphisms in HNC patients and controls, (2) HNC patients were diagnosed clinically and pathologically, and (3) HNC patients had no other systemic diseases and controls were healthy or free of tumors. On the contrary, meta-analyses, review studies, articles with incomplete data, studies without a control group, animal studies, conference papers, book chapters, and comment papers were excluded.

Data Summary
The data of the articles involved in the meta-analysis were separately retrieved by two authors (M.S. and S.B.). Extracted data included names of the authors, publication year, study country, ethnicity, number of cases, tumor type, source of controls, genotyping method, quality score, age, and gender distribution.

Quality Evaluation
The quality scoring was completed by one author (M.S.) based on the Newcastle-Ottawa Scale (NOS) scale [30] that a study is judged on three broad perspectives: the selection (4 scores); the comparability (2 scores); and the outcome (3 scores) for nonrandomized studies, respectively. The maximum possible score was nine and high-quality studies were those with a score of ≥7.

Statistical Analysis
The effect sizes, odds ratios (OR) along with 95% confidence interval (CI), were calculated using the Review Manager 5.3 (RevMan 5.3; the Cochrane Collaboration, the Nordic Cochrane Centre, Copenhagen, Denmark) as well as subgroup analyses, quantifying the association between NAT1 and NAT2 polymorphisms and the HNC risk. A p-value (2-sided) < 0.05 was considered as a significant value. A random-effects model [31] was performed when I 2 statistic represented a significant heterogeneity (P heterogeneity < 0.1 or I 2 > 50%) and if the heterogeneity was insignificant, a fixed-effect model [32] was applied.
Subgroup analyses were performed based on the ethnicity of study participants, control source in the study, tumor type, sample size, and genotyping method used in a study. To adjust for the effect of sample sizes, gender, and age distribution of the subjects included in the studies, a meta-regression analysis was conducted.
Publication bias was assessed applying funnel plots, Egger's and/or Begg's tests with a p-value (2-sided) < 0.05 demonstrating the existence of publication bias. Sensitivity analyses ("one-study-removed" and "cumulative" analyses) were conducted to evaluate the stability of pooled ORs. The meta-regression, publication bias, and sensitivity analysis were analyzed using the Comprehensive Meta-Analysis version 2.0 (CMA 2.0) software (CMA 2.0; Biostat Inc., Englewood, NJ, USA).
To illustrate false-positive or negative conclusions from meta-analyses [33], trial Sequential Analysis (TSA) software (version 0.9.5.10 beta) (Copenhagen Trial Unit, Centre for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark) was used to evaluate TSA for analyses [34]. A futility threshold can be checked by the TSA to determine the effectiveness or ineffectiveness before information size is reached. The required information size (RIS) and a two-sided boundary type were computed with an alpha risk of 5% and beta risk of 20%. There were enough studies where the Z-curve reached the RIS line or the boundary line or entered the futility area. Otherwise, the amount of information was not enough and more evidence was needed.

Primer Sequences
The primer sequences of NAT1 and NAT2 are shown in the studies of Katoh et al. [35] and Chen et al. [36], respectively.

Study Selection
From the four electronic databases and manual searching, 265 records were identified. After excluding the duplicates and irrelevant records, 48 full-text articles met the eligibility criteria ( Figure 1). Then, 20 full-texts were removed (five were meta-analyses, one was an umbrella review, three were reviews, one was a book, one was an animal study, five articles had no control groups, one article had insufficient data, and three articles did not report genotypes of slow and rapid). Finally, 28 articles were used in the meta-analysis.

Pooled Analyses
The pooled OR for the association between NAT1 polymorphism and the risk of HNC from eight studies was 0.89 (95% CI: 0.77, 1.02; p = 0.09; I 2 = 48%), (Figure 2). The pooled effect estimate was not significant demonstrating no association between NAT1 polymorphism and the risk of HNC. Forest plot in Figure 3 illustrates that the pooled OR was 1.22 (95% CI: 1.02, 1.46; p = 0.03; I 2 = 74%) for the relationship between NAT2 polymorphism and the HNC risk. This indicates that slow acetylators are related to high risk of HNC.

Subgroup Analyses
When there was one study for a subgroup, we could delete it [61]. Subgroup analyses were performed based on ethnicity, sample size, control source, genotyping method, and tumor type (Table 3). With regards to NAT1 polymorphism, no subgroup differences were observed. For NAT2 polymorphism, significant subgroup effects were observed for ethnicity and the control source. Slow acetylators among Asians and also the populationbased studies could be effective factors on the pooled result of the association between NAT2 polymorphism and the HNC risk.

Meta-Regression
The meta-regression analyses assessing the effect of publication year, the sample size, and the mean age and gender distribution of cases and controls on the risk of HNC in NAT1 and NAT2 polymorphisms are shown in Table 4. Sample size, the mean age of cases, and the percentage of males in the controls were confounding factors for the pooled result of the association between NAT2 polymorphism and the HNC susceptibility. With an increase in sample size, age of the cases, and percentage of males in the controls, the OR decreased. Table 4. Meta-regression analysis of association between N-acetyltransferases 1 and 2 (NAT1 and NAT2) polymorphisms and the risk of head and neck cancer (slow vs. rapid acetylators).

Trial Sequential Analysis
TSA for both polymorphisms (NAT1 and NAT2) and the HNC risk is illustrated in Figure 4. The Z-curve (blue line) did not reach the RIS or the boundary lines or enter the futility area for either polymorphism and therefore, the amount of information was not large enough, suggesting the need for more studies.

Discussion
This meta-analysis showed a significant relationship between NAT2 polymorphisms and the HNC susceptibility with slow acetylators being at higher risk for HNC than rapid acetylators. For NAT2 polymorphism, the ethnicity, the control source, and genotyping methods could modify the association of this polymorphism and the HNC risk. In addition, TSA showed the amount of information for the association between the polymorphisms (NAT1 and NAT2) and the HNC risk was not large enough.
The findings from studies exploring the association of NAT1 polymorphism with other cancers and HNC are different. One meta-analysis [24] found NAT1 polymorphism to be related to the risk of lung, colorectal, head and neck, bladder, and gastric carcinomas, but not with prostate, breast, and pancreatic carcinomas and non-Hodgkin's lymphoma.
Varzim et al. [47] checked the association between NAT1 polymorphism and the laryngeal cancer risk and found that the association depends on tumor location. Among the eight studies included in our meta-analyses [23,35,38,39,44,47,52,60] which evaluated the association between NAT1 polymorphism and the HNC risk, just one study [35] reported a protective role of NAT1 slow acetylators in the HNC patients while the rest of the studies did not find any association.
Comparing the individual studies included in the meta-analysis, differences were observed between the studies. For example, five studies [41,46,48,55] found an elevated risk of HNC for NAT2 slow acetylators, one found a protective role of these acetylators in HNC patients, and three did not find any association between NAT2 polymorphism and the HNC risk [23,45,49].
Effective factors on the association between NAT polymorphisms and the risk of HNC were not included in our analysis due to low numbers of studies, including smoking, gene combination, and the linkage disequilibrium. One study [41] found an elevated frequency of the NAT2 slow acetylator genotypes among HNC patients who smoked less than those who smoked more frequently. Another study reported an association in cases with a smoking history ≤30 years in duration [35]. These contradictory results [35,41,46] suggest the need to evaluate the effect of NAT polymorphisms independent of the history of smoking. In addition, assessing the frequencies of gene-gene combination (NAT2 with GSTM1, XPD, and CYP1A1) between cases with laryngeal cancer and the controls, the frequency of combinations was superior to cases than in controls where the numbers of combinations had an increased risk of laryngeal cancer and the numbers of other combinations had a protective role [40]. The linkage disequilibrium between the genes of NAT1 and NAT2 has been observed in HNC [23,38,62] and other cancers [63][64][65]. Research [66] showed the highest level of carcinogen-DNA adducts formation in cases with acetylation activity of NAT1 rapid and NAT2 slow. Therefore, future studies should consider the linkage between these polymorphisms.
The limitations of the present meta-analysis were: (1) low sample size in some studies. (2) In a number of the involved studies, the controls were not well matched to the cases.
(3) Low numbers of studies entered to the analysis as shown by TSA. (4) Existence of publication bias and high heterogeneity between the analyses.

Conclusions
There was no association between NAT1 polymorphism and susceptibility to HNC, whereas an association between and NAT2 polymorphism and the HNC risk was found. Slow acetylators of NAT2 polymorphism were at greater risk for HNC than the rapid acetylators. Despite the stability of the results, the presence of high heterogeneity, publication bias, and confounding factors warrant the need for more studies to confirm the results of the present meta-analysis as well as TSA.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.