Arg462Gln and Asp541Glu polymorphisms in ribonuclease L and prostate cancer risk: a meta-analysis

Objective The association between ribonuclease L (RNASEL) gene polymorphisms and prostate cancer risk has been widely reported, but the results of these studies remained controversial and underpowered. We performed a meta-analysis of 28 studies to evaluate the association between Arg462Gln and Asp541Glu polymorphisms in the RNASEL gene and prostate cancer risk. Methods Odds ratios (ORs) with 95% confidence intervals (CIs) were estimated to assess the association between RNASEL polymorphisms and prostate cancer risk. Results A significantly increased prostate cancer risk was found for the Arg462Gln polymorphism in Africans (Gln/Gln vs Arg/Arg: OR = 2.50, 95%CI = 1.28-4.87; Gln/Gln vs Gln/Arg + Arg/Arg: OR = 2.54, 95%CI = 1.30-4.95), but not in Europeans and Asians. Additionally, the Asp541Glu polymorphism was associated with increased total prostate cancer risk (Glu-allele vs Asp-allele: OR = 1.04, 95%CI = 1.01-1.07; Glu/Glu vs Asp/Asp: OR = 1.22, 95%CI = 1.03-1.46; Glu/Glu vs Glu/Asp + Asp/Asp: OR = 1.09, 95%CI = 1.02-1.16). In the stratified analysis for the Asp541Glu polymorphism, there was a significantly increased prostate cancer risk in Africans and Europeans, and in hospital-based prostate cancer cases. Conclusion The meta-analysis results showed evidence that RNASEL Arg462Gln and Asp541Glu polymorphisms are associated with prostate cancer risk and could be low-penetrance prostate cancer susceptibility biomarkers.


INTRODUCTION
Prostate cancer is the most common non-skin cancer and the second leading cause of cancer-relateddeath of men in the USA, with an estimated 192,000 new cases and 27,000 deaths annually [1] . The under-lying etiology of prostate cancer remains poorly understood, with both genetic predisposition and environmental factors likely to play a role. A recent study showed that prostate cancer patients with good and poor survival tend to aggregate in families, providing evidence of heritability in the prognosis of prostate cancer [2] . Despite this strong evidence for a genetic component in prostate cancer, little progress has been made in identifying a major gene or genes [3] .
There are different polymorphism sites in the RNASEL gene. Sequence analysis of RNASEL identified two common missense mutations: Arg462Gln (rs486907) and Asp541Glu (rs627928). The Arg-462Gln polymorphism reduces the ability of the cell to cause apoptosis in response to activation by 2-5 (A) and the enzymatic activity is also three times less than normal [14] , whereas the Asp541Glu variant has no known effects on RNASEL function [15] . Previously, Li et al. [16] performed a meta-analysis of the association of these two polymorphisms of RNASEL and prostate cancer risk, and found that the Asp541Glu polymorphism was associated with an increased risk of prostate cancer in Caucasians. However, they failed to find a significant association between the Arg462Gln polymorphism and prostate cancer risk in Africans. The main reason to explain this conclusion may be that they did not include any case-control studies with the African population.

Publication search
We attempted to include all the case-control studies published to date on the association between RNA-SEL gene polymorphisms and prostate cancer risk. Eligible studies were identified by searching the electronic literature in PubMed for relevant reports (the last search update was 7 May 2010, using the search terms "RNASEL", "polymorphism" and "prostate cancer"). A total of 64 articles were retrieved, of which 13 articles met the inclusion criteria for studies examining the association between the RNASEL Arg462Gln and/or Asp541Glu polymorphism and prostate cancer risk.

Inclusion and exclusion criteria
The following inclusion criteria were used for literature selection: 1) studies evaluating the RNASEL Arg462Gln and/or Asp541Glu polymorphism and prostate cancer risk; 2) case-control studies; 3) studies that contained information about available genotype frequency that could help evaluate the results in the papers; 4) studies that were published in English; 5) studies with full-text manuscripts. Major exclusion criteria were: 1) studies with no control population; 2) studies with no available data on genotype frequency; 3) studies with a clear bias of accrual.

Data extraction
Data was extracted from each study by two investigators independently according to the prespecified inclusion criteria. Data extracted from these articles included: first author, year of publication, origin, ethnicity, source of cases (population-based and hospitalbased), sample size (case/control), frequency of genotype in cases and controls, and Hardy-Weinberg equilibrium of controls. Different ethnic descents were categorized as European, Asian and African.

Statistical analysis
Odds ratios (ORs) with 95% confidence intervals (CIs) were used to measure the strength of the association between the RNASEL Arg462Gln or Asp541Glu polymorphism and prostate cancer risk based on the genotype frequencies in cases and controls. We explored the association between allele Gln of RNASEL Arg462Gln or allele Glu of Asp541Glu and prostate cancer risk, as well as the heterozygote comparison Gln/Arg vs Arg/Arg or Glu/Asp vs Asp/Asp, the homozygote comparison Gln/Gln vs Arg/Arg or Glu/Glu vs Asp/Asp and the risk of (Gln/Gln+Gln/Arg) vs Arg/ Arg or (Glu/Glu + Glu/Asp) vs Asp/Asp and Gln/Gln vs (Gln/Arg +Arg/Arg) or Glu/Glu vs (Glu/Asp + Asp/ Asp) in dominant and recessive models, respectively. Subgroup analysis was performed by ethnicity and source of cases (population-based and hospital-based) for the Arg462Gln or Asp541Glu polymorphism. The fixed effects model and the random effects model were used to calculate the pooled OR. Heterogeneity assumption was evaluated with a chi-squarebased q test among the studies. A P value of more than 0.05 for the q-test indicated a lack of heterogeneity among the studies. A random effects model (Der-Simonian and Laird method [30] ) was used when the P value for the heterogeneity test was <0.05; otherwise, a fixed effects model (Mantel-Haenszel method [31] ) was used. The departure of frequencies of RNASEL polymorphisms from expectation under Hardy-Weinberg equilibrium was assessed by χ 2 test in controls, and P < 0.05 was considered statistically significant. Publication bias was assessed with the Egger test [32] , and P < 0.05 was considered statistically significant. All statistical tests for this meta-analysis were performed with the Stata software (Version 10.0, Stata-Corp LP, USA).
With the exception of one study [17] , the distribution of genotypes in the controls was consistent with the Hardy-Weinberg equilibrium.

Test of heterogeneity of the RNASEL Arg-462Gln polymorphism
There was significant heterogeneity for the Gln allele vs the Arg allele (P heterogeneity = 0.028) in the homozygote comparison (P heterogeneity =0.001) and the recessive models (P heterogeneity = 0.001), but not in the heterozygote comparison (P heterogeneity = 0.997) and the dominant models (P heterogeneity = 0.740).

Sensitivity analysis
Sensitivity analysis was used to determine whether modification of the inclusion criteria of the metaanalysis affected the final results. These were carried out by limiting the meta-analysis to the studies conforming to Hardy-Weinburg equilibrium and altering corresponding statistical variables and analysis models. Moreover, no other single study influenced the summary OR qualitatively as indicated by sensitivity analysis.

Bias diagnosis
The Begg's funnel plot and Egger's test were performed to assess the publication bias of the literature. The shape of the funnel plots did not reveal any evidence of obvious asymmetry in the homozygote comparison of the Arg462Gln and the Glu-allele vs the The squares and horizontal lines correspond to the study-specific OR and 95% CI. The area of the squares reflects the weight (inverse of the variance). The diamond represents the summary OR and 95% CI. Study African Robbins [18] Shook [19] Shea [29] Subtotal Asian Nakzato [25] Subtotal European Beuten [17] Shook [19] Shook [19] Cybulski [20] Rokman [27] Wang [26] Wiklund [24] Maier [23] Noonan-Wheeler [28] Subtotal Overall Asp-allele of Asp541Glu. Then, Egger's test was used to provide statistical evidence of funnel plot symmetry. The results still did not reveal any evidence of publication bias (Gln/Gln vs Arg/Arg, t = 1.91, P = 0.079; the Glu-allele vs the Asp-allele, t = 1.08, P = 0.305. Fig. 3).

DISCUSSION
RNASEL encodes the RNASEL protein, a constitutively expressed latent endoribonclease that mediates the IFN-inducible 2-5A system. The protein functions partly as a tumor suppressor, acting to induce apoptosis of cancer cells, is associated with the activity of interferons. Casey et al. [14] and Xiang et al. [15] found that deficiency in the Arg462Gln variant is correlated with a decrease in enzyme dimerization to the active form and a reduction in the ability to cause apoptosis. However, the Asp541Glu variant has no effect on RNASEL function. The rapid growth of RNASEL genetic studies provides numerous opportunities for studiny disease association. The present meta-analysis involved 7,461 cases and 6,963 controls concerning the Arg462Gln polymorphism in the protein kinase region of RNASEL, and 5,353 cases and 4,321 controls concerning the Asp541Glu polymorphism in the ribonuclease domain of RNASEL. We then explored the association between these two potentially functional polymorphisms of RNASEL and prostate cancer risk.
Previously, Li et al. [16] performed a meta-analysis of the association of these two variants in RNASEL and prostate cancer risk, which included 11 casecontrol studies (each study contains the frequency of genotypes), and suggested that the Asp541Glu polymorphism was associated with an increased risk of prostate cancer in Caucasians. In this article, the authors divided prostate cancer into familial/hereditary prostate cancer and familial/sporadic prostate cancer.
In our meta-analysis, novel case-control studies in the last 4 y were included, and the familial/hereditary and familial/sporadic prostate cancer subjects were used together. The only Asian study [29] found decreased familial prostate cancer risk in Japanese with the genotype Gln/Gln. Larger studies involving a wider spectrum of Asian populations are needed for a more definitive evaluation of the relationship between the Arg462Gln polymorphism and prostate cancer risk in Asians.
The true role of the RNASEL Arg462Gln and As-p541Glu polymorphism and their influence on prostate cancer risk are similarly controversial. Genotype frequency of various polymorphic loci may exhibit racial differences. For our purposes, we divided the world's population into three broad racial groups, Asian, African and European [33] .
The RNASEL Arg462Gln polymorphism was implicated in up to 13% of prostate cancer cases, with the resultant enzyme 3 times less active than the wildtype enzyme, and an association was found among Europeans and Africans between the RNASEL Arg462Gln polymorphism and sporadic prostate cancer risk [14] . Our results suggested that the RNASEL Arg462Gln polymorphism was associated with prostate cancer risk in the African population rather than in European and Asian population, which confirmed the hypothesis described above. Moreover, our analysis suggested that the Asp541Glu polymorphism was associated with increased prostate cancer risk among Europeans and Africans but not among Asians in all genetic mod-els, although Asp541Glu produced similar levels of RNASEL activity to those of the wildtype enzyme [15] . It may be a reflection of the differences in genetic backgrounds and gene-environment interactions in the etiology. Other factors such as time-lag bias and publication bias may also have played a role. In time-lag bias, studies with 'negative' results take longer time to be published, whereas studies with 'positive' results are published much more quickly [34] . In publication bias, small studies with 'negative' results are never published, whereas equally small studies with similar quality but 'positive' results would appear in the literature [35][36] . We examined these possibilities and found that 'positive' studies were reported more in 462Gln and 541Glu allele carriers' studies (especially in the European and African studies).
In the source of cases, interestingly, we also found evidence for the association between the Arg462Gln polymorphism and prostate cancer risk using hospitalbased cases, the same as the Asp541Glu polymorphism.
Some limitations of this meta-analysis should be mentioned. First of all, the number of published studies included in our meta-analysis was not sufficiently large for a comprehensive analysis, particularly for the Asian population. Further studies should be carried out to confirm such an effect in Asians. Second, publication bias might have occurred, and our Egger's test results may have a substantial risk of being affected by such a bias. Third, the interactions between genegene, gene-environment and even different polymorphic loci of the same gene may modulate prostate cancer risk. Fourth, in some RNASEL polymorphism studies, a small number of cases and controls were included. Fifth, in some original studies, the authors divided the prostate cancer into two groups: familial/ hereditary prostate and familial/sporadic prostate cancer, whereas, in our meta-analysis, we grouped them together, which may have influenced the results. Sixth, our meta-analysis was based on unadjusted estimates. A more precise analysis should be conducted if individual information, including other covariates, such as age, gender and prostate cancer stage, becomes available. In spite of these issues, our meta-analysis also had two advantages. First, substantial numbers of cases and controls were pooled from different studies, which significantly increased the statistical power of the analysis. Second, the quality of case-control studies included in the current meta-analysis was satisfactory based on our selection criteria.
In summary, the present meta-analysis found novel evidence that the RNASEL Arg462Gln polymorphism is associated with increased prostate cancer risk in Africans; moreover, the RNASEL Asp541Glu polymorphism is associated with prostate cancer risk in Europeans and Africans, confirming the results of previous meta-analyses. These two polymorphisms may be low-penetrance susceptibility biomarkers for prostate cancer. Further prospective studies with a larger population of participants worldwide are expected to examine associations between these two polymorphisms in RNASEL and prostate cancer.