A comprehensive meta-analysis of the association between three IL 1 B polymorphisms and rheumatoid arthritis

Rheumatoid arthritis (RA) is an immune-mediated chronic inflammatory disease that causes huge destruction to human body. IL1B encodes key mediator IL-1β protein, which plays an important role in the pathogenesis of inflammatory syndromes. The aim of this study was to evaluate the association between IL1B polymorphisms and RA. A meta-analysis was performed on the association between three IL1B polymorphisms (IL1B-31: rs1143627; IL1B-511: rs16944; IL1B + 3954: rs1143634) and RA. A trend of significant association was observed between IL1B + 3954 and RA (p = 0.06, odd ratio (OR) = 1.19, 95% confidential interval (CI) = 1.00 1.42). A significant association was found in Europeans under the dominant model between IL1B-511T and RA (p = 0.03, OR = 0.89, 95% CI = 0.81 0.99). Our meta-analysis indicated that IL1B − 511-T played a protective role against RA in Europeans, and that IL1B + 3954-T had the potential to increase the risk of RA. Future large-scale studies should be considered to confirm the association between IL1B polymorphisms and RA.


INTRODUCTION
Rheumatoid arthritis (RA) is an immune-mediated chro-nic inflammatory disease [1] that can lead to low bone mineral density [2], depression [3], obstructive lung disease [4] and Cardiovascular diseases [5], causing huge destruction to human body.Twin studies estimated that heritability of RA liability was up to 60% [6].Familybased studies demonstrated that genetic factors played a more important role in the development of RA than environmental factors did [7,8].
Discrepancy among the association studies might be due to the different ethnic background, inefficient sample size [30], or the uncorrected physiological status among the association studies [31].Meta-analysis is often used to enhance statistical power and to draw a more convincing conclusion by pooling up the research data from individual association study [32].The goals of our metaanalyses were to find out the causes of the above inconsistent findings among various case-control association studies, and to evaluate the contribution of IL1B polymorphisms to RA.

Data Collection
A systematic literature searching was performed in PubMed/MEDLINE without language restriction, using the keywords "rheumatoid arthritis IL1Bassociation" and "rheumatoid arthritis IL1B polymorphism" to identify available articles.We also checked Chinese databases (WanFang, WeiPu and CNKI) using the same keywords.The inclusion criteria of the literatures for the metaanalyses comprise the following items: (1) It was an original case-control study with an assessment of the association between IL1B polymorphisms and RA risk in humans; (2) It contains sufficient information to infer the odd ratios (ORs) and 95% confidential intervals (95% CI); (3) Genotype distribution of each polymorphism in controls met Hardy-Weinberg equilibrium (HWE).All of the association studies between IL1B polymorphisms and RA were fully considered and carefully selected in July 2013.We extracted or calculated the following information from each study: Genetic locus, first author's name, year of publication, ethnicity, numbers of cases and controls, control source, HWE for controls, the result of individual studies about the association of IL1B − 31, IL1B − 511 and IL1B + 3954 with RA and power analysis for each of the involved studies.

Statistical Analysis
Arlequin program was used to test HWE [33].Power and Sample Size Calculation program was applied to calculate the power of each study [34].Review Manager 5 was used for the meta-analysis [35].Statistical heterogeneity was tested using Cochran's Q statistic and I² test [36] to decide the type of analysis to be used in the metaanalysis.For the studies with minimal to moderate heterogeneity (I 2 < 50%), the fixed-effect model would be used for the meta-analysis.For the studies with significant heterogeneity (I 2 > = 50%), the random-effect model would be used.Funnel plots are also drawn to observe the potential publication bias.

DISCUSSION
In the current meta-analyses, we summarized the associations of three IL1B variants with RA from 21 studies (22 stages) among 5888 cases and 5760 controls.Our results showed a trend of association between IL1B + 3954-T and RA (Table 2 and Figure 2) and a significant association under the dominant model between IL1B-511-T and RA in Europeans (Table 2 and Figure 2).Single nucleotide polymorphisms (SNPs) occur in a high frequency in the human genome, which may affect the function of genes [44].IL1B + 3954 in the exon 5 and IL1B − 511 in the promoter are two key polymorphisms of IL1B that play an important role in inflammatory diseases [9].Previous studies proved that IL1B-511-T in- creased LPS-induced IL-1β production by 2 -3 folds and showed higher levels of IL-1Ra [45].Different conclusions were shown for IL1B + 3954-T.Some researches indicated that it might increase plasma levels of IL-1β [10,46], but some others found it had no influence or reduced IL-1 levels [16,20,47].IL1B − 511 and IL1B + 3954 showed a wide association with diseases like gastric cancer [48,49], breast cancer [50], aspirin-tolerant asthma [51], left ventricular systolic dysfunction [52], hip osteoarthritis [53] and RA [10,[21][22][23][24]29].
Several other RA association studies observed dominant effect among a handful of SNPs such as −607A/C polymorphism of IL-18 gene [54], −670A/G polymorphism of FAS gene [55], rs1343151 and rs10489629 of IL-23R gene [56] and −173G/C polymorphism MIF gene [57].The significant association of IL1B-511-T polymorphism under the dominant model may provide a new hint in the pathogenesis of RA.
Significant heterogeneity showed in the overall analysis (I 2 = 65%, Table 2) and dominant model (I² = 57%, Table 2) of IL1B + 3954.A subgroup study by ethnicity (Table 2) showed that significant heterogeneity was only found in Asians (I 2 = 56% in dominant model of IL1B − 511 in Europeans, I 2 = 84% in overall analysis and dominant model of IL1B + 3954 in Europeans).Frequency of IL1B-511-C and in Asians (Hapmap-HCB) is 0.547 that is lower to 0.642 in Europeans (Hapmap-CEU).And the allele frequency of IL1B+3954-C and in Asians (Hapmap-HCB) is 0.988 that is much higher to 0.792 in Europeans (Hapmap-CEU).A further analysis of the two polymorphisms showed a non-significant ethnic difference between Asians and Europeans (IL1B − 511: Fst = 0.0094; IL1B + 3954: Fst = 0.0988).A further power analysis suggested there was a lack of power for the subgroup meta-analysis in Asians (power = 0.701 in IL1B-511, power = 0.214 in IL1B + 3954, Table 2), suggesting that the non-significant association in Asians might be due to the small sample size in the existing case-control association studies in Asian population.In contrast, the power in the meta-analysis in European populations for IL1B − 511 and IL1B + 3954 polymorphisms are 0.998 and 0.937 (Table 2).
Compared with the previous two meta-analysis studies [24,58] about the polymorphisms of IL1B and RA, our meta-analyses included 13 and 8 more case-control studies than the studies by P. Harrison et al. [24] and Young LEE et al. [58], respectively.Our research showed that IL1B − 511 was significantly associated with RA in Europeans under the dominant model, and a trend association of IL1B + 3954 with RA.Moreover, our study grouped Turkish population into Caucasians instead of Asians in the subgroup meta-analysis according to the fact that the ancestors of major Turkish population were from Europe [59][60][61].We performed HWE test for the controls in all the involved studies, and excluded one [20] that was included in previous meta-analysis [58].With an enhanced power and stricter selection criteria, our meta-analyses produced a more reliable conclusion than the previous meta-analysis studies.
However, our study presented several limitations that needed to be carefully considered.Firstly, there are only a limited number of associations in non-Caucasian populations.A lack of power in the non-Caucasian studies suggested that non-significant results in Asians and other population needed to be taken with caution.Future studies with larger samples size are required to establish the association of IL1B polymorphisms with RA.Secondly, RA is a complex disease that different physiological status of RA may exist in the cases.All the existing casecontrol studies didn't perform a stratified analysis by the RA disease stage.This may partially explain the discrepancies in the current case-control studies.Thirdly, genetic heterogeneity may exist in IL1B since there are 826 known IL1B polymorphisms.Our meta-analyses only focused on three IL1B SNPs that might not fully represent the overall contribution of IL1B variations.Other IL1B polymorphisms needed to be analyzed for their contribution to RA in the future.Fourthly, the positive findings of current study might not reach a very precise statistical significance by the certain extent multiple testing in our analyses.
In conclusion, our meta-analysis observed a trend association of IL1B + 3954-T with RA and a significant association under the dominant model between IL1B − 511-T and RA in Europeans.Further researches are required to confirm our findings and to discover the underlying mechanisms of other polymorphisms of IL1B that might contribute to the risk of RA.

Figure 1 .
Figure 1.Flowchart of selection process for meta-analyses.

Figure 2 .
Figure 2. Forest plots of the three SNPs with RA.A) Forest plot of IL1B − 31 in overall analysis; B) Forest plot of IL1B − 511 in overall analysis; C) Forest plot of Dominant model of IL1B − 511 in Europeans; D) Forest plot of IL1B + 3954 in overall analysis.

Figure 3 .
Figure 3. Funnel plots of three SNPs with RA.A) Funnel plot of IL1B − 31 in overall analysis; B) Funnel plot of IL1B − 511 in overall analysis; C) Funnel plot of Dominant model of IL1B − 511 in Europeans; D) Funnel plot of IL1B + 3954 in overall analysis.
* : Amount of Stages; NA: Not applicable.