The Association between PON1 (Q192R and L55M) Gene Polymorphisms and Risk of Cancer: A Meta-Analysis Based on 43 Studies

Q192R and L55M polymorphism were considered to be associated with the development of multiple cancers. Nevertheless, the results of these researches were inconclusive and controversial. Therefore, we conducted a meta-analysis of all eligible case-control studies to assess the association between PON1 (Q192R and L55M) gene polymorphisms and risk of cancer. With the STATA 14.0 software, we evaluated the strength of the association by using the odds ratios (ORs) and 95% confidence intervals (CIs). A total of 43 case-control publications 19887 cases and 23842 controls were employed in our study. In all genetic models, a significant association between PON1-L55M polymorphisms and overall cancer risk was observed. Moreover, in the stratified analyses by cancer type, polymorphism of PON1-L55M played a risk factor in the occurrence of breast cancer, hematologic cancer, and prostate cancer. Similarly, an increased risk was observed in the Caucasian and Asian population as well as hospital-based group and population-based group. For PON1-Q192R polymorphisms, in the stratified analyses by cancer type, PON1-Q192R allele was associated with reduced cancer risks in breast cancer. Furthermore, for racial stratification, there was a reduced risk of cancer in recession model in Caucasian population. Similarly, in the stratification analysis of control source, the overall risk of cancer was reduced in the heterozygote comparison and dominant model in the population-based group. In conclusion, PON1-Q192R allele decreased the cancer risk especially breast cancer; there was an association between PON1-L55M allele and increased overall cancer risk. However, we need a larger sample size, well-designed in future and at protein levels to confirm these findings.


Introduction
Cancer is one of the diseases caused by a combination of genetic and environmental factors [1]. The PON1 gene, located on the long arm of chromosome 7q21.3, is an antioxidant enzyme that has strong lipophilic antioxidant properties, which can maintain the balance of antioxidant-oxidant [2,3]. Simultaneously, PON1 is also an esterase involved in scavenging reactive oxygen species by binding to highdensity lipoprotein (HDL). Studies have shown that oxidative stress may participate in the process of cell proliferation and malignant transformation and damage DNA as well as other biological molecules, resulting in the occurrence of tumors [4]. The ability of PON1 detoxification of carcinogenic oxidative stress products makes it possible for researchers to predict PON1 gene polymorphism in cancer susceptibility [5].
At present, with the deep development of genetic studies of PON1, studies have found that PON1-Q192R and PON1-L55M, the two most common functional genetic polymorphisms in PON1, were identified at positions 192 and 55 [6]. PON1-Q192R polymorphism (rs662A > G) was caused by the glutamine (Q genotype) substituted for the arginine (R genotype) 192 of the gene 6 exon of the PON 1 gene [7]. PON1-L55M (rs854560) was originated from the replacement of 55 leucines (L genotype) by methionine (M genotype) at 2 BioMed Research International third exon 55 [8]. In addition, it has been shown that the two functional SNP, Q192R and L55M, were associated with the risk of multiple tumors [9,10], such as oral cancer [11], lung cancer [12], and embryonal tumors [13].
According to the important role of PON1 in the development of tumor and the correlation between genotype and phenotype, we speculate that the variation of PON1 gene Q192 R and L55M may be related to tumor susceptibility. However, the data of many studies are contradictory and uncertain at present. Therefore, a comprehensive meta-analysis should be conducted to determine the relationship between Q192R and L55M polymorphism and cancer risk.

Search Strategy.
We conducted a systematic literature search in the PubMed, Embase, and Web of Science for all related studies before June 10, 2019 via utilizing the following terms: "polymorphism OR paraoxonase 1 OR PON1" AND "tumor OR malignancy OR cancer OR carcinoma OR neoplasm". In addition, we extracted the reference of the original articles on this issue to carry out a hand search for extra studies. The results deduced from these articles were limited to humans. When the publication referred to more than one cancer type or ethnicity, we deled with data respectively. Besides, if different authors published articles based on the same population or one author used similar data in an article, we picked out the report with the latest study and largest sample size.

Inclusion Criteria and Exclusion
Criteria. The enrolled studies must contain the following inclusion criteria: (1) publication that evaluated the association between PON1-L55M, or PON1-Q192R polymorphism and the risk of cancer. (2) The genotype frequency may be obtainable from cases and controls, or we could gain it via computing. In addition, studies were excluded when they would meet these exclusion criteria: (1) reviews, case reports, or case-only studies; (2) studies with deficient genotype frequency date; (3) animals reports; and (4) replicate studies.

Data Extraction.
The authors were able to excerpt relevant data from these qualified studies independently, and the following information would be seized: first author's last name, publishing year, the ethnicity of each population, the genotyping methods, the control of source, cancer types, number of cases and controls, and P value of Hardy-Weinberg equilibrium. When encountering divergences, we analyzed the report and reached a consistent agreement lastly. Report included in this meta-analysis (n=43) 43 for PON1-Q192R, 28 for PON1-L55M and dominant (ML+MM versus LL; RR+RQ versus QQ). Besides, stratified analyses were conducted via ethnicity, cancer type, control source, and genotyping method. However, when any cancer type is less than two studies, we would segment it into the "other cancers" group. In addition, 2test-based Q-statistic test [14] was taken to assess the research heterogeneity while 2 values and P values [15] were used for quantifying. When I 2 < 50% and P>0.10, it indicates that there was no significant heterogeneity, and ORs could be pooled by a fixed-effects model. Otherwise, the random effects model would be adopted [16]. Furthermore, sensitivity analysis, from the qualified removing a single research study and revealing the individual data set to merge OR influence, was applied to estimate the stability of these data. (P<0.05 was regarded as statistically significant [17].) Finally, potential publication bias was estimated by symmetry of funnel plot of Begg's test as well as Egger's test [15,18], and being statistically significant was considered when P<0.05. All statistical tests were performed with STATA Software (version 14.0, state Corp), and P<0.05 for any genetic models or tests was identified as statistically significant.

Association between PON1-L55M and Cancer Susceptibility.
Our study had uncovered that the PON1-L55M polymorphism was significantly associated with an increased risk of the overall cancers under all the genetic models ( ). In addition, we identified an increased risk by stratified analysis using genotyping method.

Publication Bias and Sensitivity Analysis.
A sensitivity analysis was carried out to detect the impact of individual papers on whole data by getting rid of one report at a time from the pooled analysis. And no individual report has been significantly affected by the pooled OR. Figure 4 showed the plot of the sensitivity analysis for evaluating the association between PON1-Q192R and cancer risk (RR versus QQ). Besides, we perform Egger's test and Begg's funnel plot to evaluate publication bias ( Figure 5

Discussion
Several studies have indicated that PON1, which is one of xenobiotic metabolising enzymes, plays a crucial role in the detoxification of carcinogenic compounds and decreases oxidative stress. Genetic polymorphisms can influence the enzyme and modify its activity, resulting in an impact on individual sensitivity to certain pathologies [61]. Indeed, a great deal of researches have showed that polymorphisms     Notes: * statistically significant (P<0.05); P value a : P value of Q test for heterogeneity test; I 2 : 0%-25% means no heterogeneity, 25%-50% means modest heterogeneity, and 50% means high heterogeneity.
Abbreviations: CI, confidence interval; OR, odds ratio; PCR-RFLP, polymerase chain reaction-restriction fragment length polymorphism.     encoding the gene of these enzymes have been linked to the progression of cancer [49,62]. Furthermore, several variants of PON1, including Q192R and L55M, have been found to be a biologically reasonable candidate which has an obvious influence on cancer. PON1 (Q192R and L55M) gene polymorphisms were related to many types of cancer, such as breast, prostate, and hepatocellular carcinoma [20,50,63]. For instance, PON1-L55M polymorphism may increase the risk in multiple cancer types, such as prostate and breast cancers but decrease renal cell carcinoma and ovarian cancer risk. As for PON1-Q192R, it has been revealed to suppress expression in lung [64] and pancreatic cancer [65] and reduce the risk of breast and prostate cancers. And the results of these researches were inconclusive and controversial. In our work, in all genetic models we have identified the significant association between PON1-L55M polymorphism  and overall cancer risk, while PON1-Q192R allele was not associated with reduced overall cancer risks. In the stratified analysis, we observed an increased risk in the Caucasian population and the Asian population, as well as the hospitalbased group and population-based group under all the five genetic models in the PON1-L55M polymorphism. Similarly, a significantly increased risk of the overall cancers under the homozygote, allele contrast, recessive, and dominant models was uncovered in hematological tumor in the PON1-L55M polymorphism. Nevertheless, in the PON1-Q192R polymorphism, we also observe a reduced risk of the overall cancers in the allele contrast and dominant models. Meanwhile, we could obtain an interesting phenomenon that PON1-L55M polymorphism acts as a risk factor in all the five genetic models and there was an association between Q192R polymorphism and a reduced risk for cancer progression (except recessive model) after stratified analyses by cancer type, especially breast cancer. Thus, we can obtain that PON1 (Q192R and L55M) gene polymorphisms play a vital role in the development of breast cancer, whose mechanism maybe as follows: there was a critical association between L allele and higher PON1 serum concentrations while M variant decreased the stability of this enzyme. Therefore, the blood concentration of PON1 was reduced in this way; then, the activity of the enzyme was influenced, which may increase the vulnerability to genomic damage by reducing the inflammatory oxidant and the detoxifying ability of dietary carcinogens, thereby increasing the risk of breast cancer [5]. Furthermore, breast cancer becomes more susceptible to genomic damage as a result of lower levels of PON1 which could decrease the ability to detoxify inflammatory oxidants and dietary carcinogens [5]. Similarly, the exchange of Q and R could produce an enzyme which has a higher detoxification activity when there were potential carcinogenic products of oxidative stress and lipid peroxidation [66,67]. In addition, not only genetic factors but also other contributors including nutrition and lifestyle can significantly affect PON1 enzyme activity, thereby reducing the risk of breast cancer [68]. To sum up, PON1, as a member of lipid peroxidation scavenging systems, may have an impact on malignant transformation and cell proliferation in the progression of breast cancer [69]. In the ethnographic analysis, we found ethnic groups having different results, which may be due to ethnic living habits, living environment, and genetic factors.
Previous meta-analysis also reported the association of PON1 polymorphism with cancer risk [10,70]. As far as we know, we are the first of the typical functional polymorphism of the PON1 gene including all the published and defined case-control studies that have been conducted in a comprehensive meta-analysis. Compared with previous researches, our report was more persuasive and we have carried out a more detailed analysis to demonstrate our results. First and most obviously, the data we collected in our study was up-to-date, and we could keep up with the research front. Secondly, we included more qualified studies and larger sample size, which indicates that we are relatively more accurate in assessing that association between the PON1 gene SNPs and the risk of cancer.
Despite the association between PON1 (Q192R and L55M) gene polymorphism and cancer risk which has been studied in detail, we should note some limitations at the same time. First of all, the quantity of publications collected in our study was limited and there was a relatively small sample size of the report. What is more, Caucasian accounted for the most of the registered publications and there were no Africans. Furthermore, some of publications would only publish positive results, which could make the meta-analysis less credible. Lastly, our results were based on the estimates of single-factor, which could lead to serious confusion and bias due to the lack of raw data, and there is a need to adjust the effect size with possible confounders related to lifestyle risk factors, such as age, obesity, alcohol consumption, and smoking.
In conclusion, our study has demonstrated that PON1-Q192R can significantly reduce the risk of cancer and the polymorphism of PON1-L55M is a risk factor leading to cancer, especially breast cancer. Next, we need a larger sample size at protein levels to confirm whether PON1 polymorphisms may be potential genetic markers of tumor prognosis and identify its role in the risk of women developing breast cancer.