Rheumatol Int DOI 10.1007/s00296-014-3027-x
Original Article
Associations between functional TNFR2 196 M/R polymorphisms and susceptibility to rheumatoid arthritis: a meta‑analysis Gwan Gyu Song · Sang‑Cheol Bae · Young Ho Lee
Received: 10 March 2014 / Accepted: 16 April 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Several studies have examined the effects of tumor necrosis factor receptor (TNFR) 1 +38 A/G and TNFR2 196 M/R polymorphisms on susceptibility to RA and have reported conflicting results. The purpose of this study was to examine whether the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms are associated with RA susceptibility. We performed a literature search using the Medical Literature Analysis and Retrieval System Online and Embase citation indices, and conducted a meta-analysis to examine the association between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA. Our meta-analysis included a total of 13 studies from 11 articles, consisting of 11 studies of the TNFR2 polymorphism (2,092 cases and 1,483 controls), and two studies of the TNFR1 polymorphism (672 cases and 288 controls). The meta-analysis revealed a significant association between the TNFR2 196 RR genotype and RA risk (OR 1.737, 95 % CI 1.275–2.367, P = 4.6 × 10−5). Stratification by ethnicity indicated an association between the TNFR2 196 RR genotype and RA in Europeans (OR 2.054, 95 % CI 1.305– 3.232, P = 0.002), but not in East Asians (OR 1.596, 95 % CI 0.642–3.971, P = 0.314). Analysis using a homozygote contrast model showed the same pattern for the TNFR2 196 RR genotype in a European and East Asian population. However, no association was found between the TNFR1
G. G. Song · Y. H. Lee (*) Division of Rheumatology, Department of Internal Medicine, Korea University Anam Hospital, Korea University College of Medicine, 126‑1, Anam‑dong 5‑ga, Seongbuk‑gu, Seoul 136‑705, Korea e-mail: [email protected] S.-C. Bae The Hospital for Rheumatic Diseases, Hanyang University Medical Center, Seoul, Korea
+36 A/G polymorphism and RA in a European population. Our meta-analysis demonstrated that the functional TNFR2 196 M/R polymorphism is associated with susceptibility to RA in the European population. Keywords TNFR · Meta-analysis · Polymorphism · RA
Introduction Rheumatoid arthritis (RA) is a chronic inflammatory disease that predominantly involves synovial joints and affects up to 1 % of adults worldwide. Although the etiology of RA remains undetermined, it has been established that susceptibility to RA has a genetic component [1, 2]. Tumor necrosis factor-α (TNF-α) is a potent pro-inflammatory cytokine that plays important roles in inflammatory and immune responses, including those observed in RA [3, 4]. TNF-α stimulates cytokine production, enhancing the expression of adhesion molecules and increasing neutrophil activation. The effects of TNF-α are mediated through two receptors, TNFR1 (TNFRSF1A) and TNFR2 (TNFRSF1B). Both TNFR1 and TNFR2 are expressed in the synovial tissue, cartilage, and pannus of patients with RA [5]. Several single nucleotide polymorphisms have been identified in the TNFR1 and TNFR2 genes. The two that have been studied the most intensively are an A-to-G polymorphism at position +36 in exon 1 of the TNFR1 gene (rs767455) and a T/G polymorphism at position +196 (rs1061622) in exon 6 of TNFR2, which results in a nonconservative amino acid substitution (methionine (M) to arginine (R)) within the fourth extracellular domain of TNFR2 [6]. The TNFR2 196 M/R polymorphism may affect receptor function, and the TNFR2 196 R allele transmits a stronger TNF-α signal than the TNFR2 196 M allele [7].
13
Several studies have examined the potential effects of the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms on susceptibility to RA [8–18]. However, these studies have shown mixed results due to small sample sizes and low statistical power. Meta-analysis is a powerful method for overcoming the problem of small sample sizes and inadequate statistical power in genetic studies of complex traits [19– 21]. Therefore, we conducted a meta-analysis to explore whether the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms are associated with susceptibility to RA.
Materials and methods Identification of eligible studies and data extraction We conducted a literature search for studies that examined the association between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA. We used the Medical Literature Analysis and Retrieval System Online (MEDLINE) and Embase citation indices to identify articles published through February 2014 in which the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms were identified in RA patients and controls. Additionally, all references in the identified articles were reviewed to identify studies not indexed by MEDLINE and Embase. The following key words and subject terms were used: “tumor necrosis factor receptor” “TNFR,” “TNFR1,” “TNFR2,” and “RA.” Studies were included in the analysis if they (1) were case– control studies, (2) contained genotype data, and (3) contained sufficient data to calculate odds ratios (ORs). No language restrictions were applied. We excluded the following: (1) studies containing overlapping data; (2) studies in which the number of genotypes or alleles could not be ascertained; and (3) studies in which family members had been studied, as these analyses were based on linkage considerations. The following information was extracted from each identified study: author, year of publication, ethnicity of study population, demographics, number of cases and controls, and genotype and allele frequencies of the TNFR1 and TNFR2 polymorphisms.
Rheumatol Int
which represents the proportion of between-study variability attributable to heterogeneity rather than chance [22]. I2 values of 25, 50, and 75 % were nominally assigned as low, moderate, and high estimates, respectively. The fixed-effects model assumes that a genetic factor has a similar effect on disease susceptibility across all studies investigated and that observed variations among studies are caused by chance alone [23]. The random-effects model assumes that different studies show substantial diversity and assess both withinstudy sampling error and between-study variance [24]. When study groups are homogeneous, the two models are similar. If the study groups lack homogeneity, the random-effects model usually provides wider CIs than the fixed-effects model. The random-effects model is most appropriate when there is significant between-study heterogeneity [24]. Statistical manipulations were performed using Comprehensive Meta-Analysis software (Biostat, Englewood, NJ, USA). The power of each study was computed as the probability of detecting an association between the polymorphism and RA, using a significance level of 0.05, and assuming an OR of 1.5 (small effect size). Power analysis was performed using the statistical program G*Power (http://www. psycho.uni-duesseldorf.de/aap/projects/gpower). Evaluation of sensitivity test and publication bias The Chi-squared test was used to determine whether the observed genotype frequencies in controls conformed to Hardy–Weinberg equilibrium (HWE). We performed a subgroup analysis by excluding studies in which genotype distribution in controls was inconsistent with HWE, because deviation from HWE among controls suggests the possibility of bias during control selection or genotyping errors. Sensitivity analysis was also performed to assess the influence of each individual study on the pooled or by omitting each study individually. Publication bias is typically detected using funnel plots, but because they require a range of studies of varying sizes and subjective judgment, we used Egger’s linear regression test [25], which measures funnel plot asymmetry on a natural logarithmic scale of ORs.
Evaluation of statistical associations
Results
We performed meta-analyses using (1) allelic contrast, (2) the recessive model, (3) the dominant model, and (4) homozygote contrast. Point estimates of risks, ORs, and 95 % confidence intervals (CIs) were estimated for each study. Additionally, within- and between-study variations and heterogeneities were assessed using Cochran’s Q test. Cochran’s Q test assesses the null hypothesis that all studies evaluated have the same effect. The effect of heterogeneity was quantified using I2, with a range between 0 and 100 %,
Studies included in the meta‑analysis
13
Seventy-eight studies were identified using electronic and manual searches, and 15 of these were selected for full-text review based on their title and abstract details [8–18, 26– 29]. Four studies were excluded because they contained no extractable polymorphism data [26–29]. Thus, a total of 11 studies met our inclusion criteria [8–18] (Fig. 1). One of the eligible studies contained data from two different RA groups
Rheumatol Int
Fig. 1 Flowchart representing the study design
[16] and was thus treated independently [30–34]. Thus, a total of 13 separate comparisons were considered in our meta-analysis, which were 11 studies of the TNFR2 polymorphism, consisting of 2,092 cases and 1,483 controls, including 5 European, 2 East Asian, and single South Asian, African American, Latin American, and Arab populations; and 2 studies of the TNFR1 polymorphism, consisting of 672 cases and 288 controls, including 2 European populations [15, 18]. Thus, the ethnicity-specific meta-analysis was restricted to European and East Asian populations. Selected details of the individual studies are summarized in Table 1. The statistical power of these studies ranged from 25.6 to 81.2 %. One of the studies had a statistical power exceeding 80 %. Frequency of the R allele of the TNFR2 196 M/R polymorphism by ethnicity The mean frequency of the R allele of the TNFR2 196 M/R polymorphism was 23.5 % among all healthy controls. East
Asians had a lower R allele prevalence than the other ethnic groups (16.3 %). Among healthy controls, the frequency of the TNFR2 196 R allele in the African American, Latin American, European, Arab, and South Asian populations was 16.9, 21.3, 23.6, 28.1, and 42.1 %, respectively (Table 2). Meta‑analysis of the association between the TNFR2 196 M/R polymorphism and RA A meta-analysis of all RA patients and of each ethnic patient was performed (Table 3). The meta-analysis revealed no association between the TNFR2 196 R allele and RA (OR 1.162, 95 % CI 0.996–1.428, P = 0.153) (Table 3). Stratification by ethnicity indicated no association between the TNFR2 196 R allele and RA in European and East Asian populations (OR 1.250, 95 % CI 0.942–1.660, P = 0.112; OR 1.026, 95 % CI 0.593–1.775, P = 0.927) (Table 3). Analysis using the dominant model
13
13 European
The Netherlands UK
Barton [18]
291
172 127 127 50 177 114 257 175 108 545 240 RA 381 143
160 140 127 120 130 75 183 84 62 265 137 Control 145 95
56 42 46 37 113 55 158 107 70 404 104 AA 122
MM
RA
Control
RA
Numbers
Allele contrast, b Assuming an odds ratio of 1.5 (small effect size) at a significance level of 0.05
a
Ref reference, HWE Hardy–Weinberg equilibrium, UK United Kingdom
European
Arab European South Asian Latin American European East Asian European European African American East Asian European
Ethnicity
Egypt UK UK Mexico UK Taiwan Sweden Italy USA Japan UK
Ethnicity
TNFR2 196 M/R polymorphism Hussein [8] Ghelani-1 [16] Ghelani-2 [16] Oregon-Romero [9] Potter [17] Yen [10] Dahlqvist [11] Fabris [12] Bridges [13] Shibue [14] Barton [18] TNFR1 +36 A/G polymorphism Barley [15]
Author (Ref)
Table 1 Characteristics of the studies included in the meta-analysis
152
88 67 59 13 57 54 83 60 33 124 107 AG 208
MR
44
28 18 22 0 7 5 16 8 5 17 29 GG 51
RR
44
80 68 43 77 71 28 114 57 41 208 74 AA 50
MM
Control
76
70 65 61 35 54 43 61 26 21 54 56 AG 83
MR
23
10 7 23 8 5 4 8 1 0 3 7 GG 12
RR
0.301
0.815
0.298 0.083 0.866 0.159 0.173 0.016 0.964 0.294 0.108 0.808 0.383
HWE P value
0.690
0.235
0.000 0.003 0.719 0.079 0.178 0.221 0.636 0.181 0.500 0.012 0.080
Association P valuea
54.9
63.0
44.5 37.2 35.7 25.6 41.7 27.9 55.4 36.3 25.6 81.2 49.2
Power (%)b
Rheumatol Int
Rheumatol Int Table 2 Prevalence of the R allele of the TNFR2 196 M/R polymorphism Population
European East Asian South Asian African American Latin American Arab Overall
No of studies
Numbers
M allele (%)
Case
Control
Case
Control
5 2 1 1 1 1
976 659 127 108 50 172
674 340 127 62 120 160
27.2 16.8 40.6 19.9 13.0 41.9
23.6 16.3 45.1 16.9 21.3 28.1
11
2,092
1,483
25.2
23.5
TNFR2 196 RR genotype and RA (OR 1.737, 95 % CI 1.275–2.367, P = 4.6 × 10−5) (Table 3; Fig. 2). Stratification by ethnicity indicated an association between the TNFR2 196 RR genotype and RA in the European population (OR 2.054, 95 % CI 1.305–3.232, P = 0.002), but not in the East Asian population (OR 1.596, 95 % CI 0.642– 3.971, P = 0.314) (Table 3; Fig. 2). Analysis using the homozygote contrast model showed the same pattern for the TNFR2 196 RR genotype in European and East Asian populations (Table 3). Meta‑analysis of the association between the TNFR1 +36 A/G polymorphism and RA
showed the same pattern for the TNFR2 196 R allele in Europeans and East Asians (Table 3). However, the metaanalysis revealed a significant association between the
The meta-analysis revealed no association between the TNFR1 +36 G allele and RA in the European population (OR 0.893, 95 % CI 0.737–1.083, P = 0.051) (Table 4;
Table 3 Meta-analysis of the TNFR2 196 M/R polymorphism and RA Polymorphism
R versus M allele
RR versus RM + MM (recessive)
RR + RM versus MM (dominant)
RR versus MM
Population
Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab
Number of studies
Numbers
Test of association
Test of heterogeneity
RA
Control
OR
95 % CI
P value
Model
P value
I2
11 5 2 1 1 1 1 11 5 2 1 1 1 1 11 5 2 1 1 1 1 11 5 2 1 1 1
2,092 976 659 127 108 50 172 2,092 976 659 127 108 50 172 2,092 976 659 127 108 50 172 2,092 976 659 127 108 50
1,483 674 340 127 62 120 160 1,483 674 340 127 62 120 160 1,483 674 340 127 62 120 160 1,483 674 340 127 62 120
1.162 1.250 1.026 0.937 1.219 0.554 1.840 1.737 2.054 1.596 0.947 6.143 0.131 2.917 1.135 1.217 0.942 0.901 1.060 0.629 2.071 1.857 2.217 1.488 0.894 6.475 0.122
0.946–1.428 0.942–1.660 0.593–1.775 0.658–1.334 0.685–2.169 0.286–1.071 1.330–2.545 1.275–2.367 1.305–3.232 0.642–3.971 0.497–1.805 0.361–122.1 0.007–2.314 1.368–6.220 0.896–1.438 0.863–1.715 0.482–1.842 0.538–1.510 0.549–2.046 0.302–1.311 1.328–3.231 1.120–3.081 1.392–3.521 0.590–3.753 0.436–1.832 0.349–120.1 0.007–2.163
0.153 0.112 0.927 0.719 0.500 0.079 0.500 4.6 × 10−5 0.002 0.314 0.869 0.203 0.165 0.006 0.293 0.262 0.862 0.693 0.862 0.216 0.001 0.016 0.001 0.400 0.760 0.210 0.151
R R R NA NA NA NA F F F NA NA NA NA R R R NA NA NA NA R F F NA NA NA
0.001 0.020 0.044 NA NA NA NA 0.137 0.491 0.184 NA NA NA NA 0.004 0.025 0.050 NA NA NA NA 0.023 0.223 0.109 NA NA NA
66.9 65.7 75.3 NA NA NA NA 32.6 0 43.2 NA NA NA NA 60.9 64.1 73.9 NA NA NA NA 51.7 29.7 61.0 NA NA NA
1
172
160
4.000
1.801–8.891
0.001
NA
NA
NA
OR odds ratio, CI confidence interval, F fixed-effects model, R random-effects model, NA not available
13
Rheumatol Int
Fig. 2 ORs and 95 % CIs of individual studies and pooled data for the association between the RR genotype of the TNFR2 196 M/R polymorphism and RA in the overall group (a) and in each ethnic group (b)
Table 4 Meta-analysis of the TNFR1 +36 A/G and RA Polymorphism
Population
Number of studies
Numbers
Test of association
Test of heterogeneity
RA
Control
OR
95 % CI
P value
Model
P value
I2
0.614 0.323 0.982
0 0 0
G versus A allele GG versus GA + AA (recessive) GG + GA versus AA (dominant)
European European European
2 2 2
672 672 672
288 288 288
0.893 0.755 0.920
0.737–1.083 0.525–1.087 0.687–1.232
0.051 0.131 0.576
F F F
GG versus AA
European
2
672
288
0.748
0.495–1.128
0.166
F
0
OR odds ratio, CI confidence interval, F fixed-effects model, R random-effects model
Fig. 3). Analyses using the dominant and recessive models and homozygote contrast showed no association between the TNFR1 +36 A/G polymorphism and RA risk (Table 4).
13
Heterogeneity, sensitivity test, and publication bias The meta-analysis of the TNFR2 196 R allele across all groups and within each ethnic group showed between-study
Rheumatol Int
Fig. 3 ORs and 95 % CIs of individual studies and pooled data for the association between the G allele (a) and the GG genotype (b) of the TNFR1 +36 A/G polymorphism and RA in the European population
heterogeneity (Table 3). However, no heterogeneity was identified in the meta-analysis of European and East Asian populations using either the recessive model or homozygote contrast. There was no overall heterogeneity in the meta-analyses of the TNFR1 +36 A/G polymorphism (Table 4). Sensitivity analysis showed that no individual study significantly affected the pooled OR. Publication bias results in a disproportionate number of positive studies and poses a problem for meta-analyses. However, Egger’s regression test showed no evidence of publication bias (P value > 0.1).
Discussion Although the multifactorial nature of RA is well known, genetic factors are considered to be strong determinants of these diseases, thus encouraging researchers to search for the genes responsible. Studies of the associations between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA have reported conflicting results [8–18], which is not surprising. Although association studies provide a powerful means of identifying genetic factors underlying autoimmune diseases, most reported association studies lack sufficient statistical power. All studies, with one exception, showed a statistical power less than 80 % to detect an association between the two polymorphisms and RA. Thus, thousands of study subjects are required for adequate
statistical power. In such cases, meta-analysis can be used as an alternative. In this meta-analysis, we combined data from published studies to evaluate genetic associations between the most commonly studied polymorphisms of the TNFR1 and TNFR2 genes, namely the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms, and RA. Our results showed a significant association between the TNFR2 196 RR genotype and RA risk in European, but not in East Asian population. These findings may be partially explained by ethnic differences. European populations showed a higher frequency of the TNFR2 196 R allele (28.1 %), whereas East Asian populations showed the lowest frequency of the A allele (16.3 %). TNFR2 has been considered to be a candidate gene for RA because of its chromosomal location and functional significance [5, 6]. TNFR2 is located on chromosome 1p36, which is a susceptibility locus for RA [35], and the TNFR gene has been shown to be involved in autoimmune disease in an animal model [36]. The TNFR2 196 R allele has been found to be associated with susceptibility to autoimmune diseases, including RA. TNFR2 is predominantly expressed in cells of myeloid origin, especially in stimulated T and B cells, and it has a higher affinity for TNF-α [37]. TNFR2 is the major TNF receptor found on T cells and mediates apoptosis in CD8+ T cells [38]. A previous meta-analysis has shown that the 196 R allele of the functional M196R polymorphism of TNFR2 is a risk factor for SLE in the Asian population, but not in Europeans [39]. Thus, the association reported is concordant with the present results, but the ethnic association between Europeans and Asian populations differed. Nonetheless, both studies support the hypothesis that TNFR2 plays an important role in the pathogenesis of autoimmune diseases such as RA and SLE, but the difference in ethnic associations between RA and SLE requires further investigation. In recent years, susceptibility genes of RA have been studied intensively by genome-wide association studies (GWAS). Recent data could not show importance of TNFR polymorphisms with RA susceptibility. A possible reason on difference in association results between recent GWAS and our study data needs to be explained. GWAS studies compare risk allele with control allele, but do not analyze additional genetic models. In contrast, we examined the association using various genetic models such as allelic contrast, recessive model, dominant model, and homozygote contrast. Our study using allelic contrast did not show any association between the TNFR2 196 R allele and RA, in agreement with recent RA GWAS data. However, the meta-analysis under the recessive model and the homozygote contrast model revealed a significant association between the TNFR2 196 M/R polymorphism and RA.
13
TNFR1 is located on chromosome 12p13 and has 10 exons, and it is expressed in all nucleated cells, especially cells susceptible to the cytotoxic action of TNF [13]. The TNFR1 +38 A/G polymorphism has no functional significance for TNFR1, but this polymorphism may be in linkage disequilibrium with a nearby causal variant. Our results did not show an association between the TNFR1 +36 A/G polymorphism and RA in Europeans. However, this result should be interpreted with caution, because the meta-analysis of the TNFR1 +38 A/G polymorphism and RA included only two studies, which is insufficient to provide conclusive evidence. Thus, further studies using larger numbers of subjects are required. The present study has some limitations that should be considered. First, heterogeneity and confounding factors may have distorted the analysis. Second, RA varies in severity, and in the studies included here, the RA activity level was unclear. Further research is required to examine whether an association exists between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and the activity or clinical features of RA. Third, because we included data from European and East Asian patients in our ethnicityspecific meta-analysis, our ethnicity-associated results are applicable only to these two groups. In conclusion, this meta-analysis demonstrates that the functional TNFR2 196 M/R polymorphism is associated with susceptibility to RA in Europeans. Accordingly, our findings support a role of the TNFR2 196 M/R polymorphism in the pathogenesis of RA. Large-scale studies in populations with different ethnicities are required to further explore the relationship between polymorphisms in the TNFR genes and the pathogenesis of RA. Acknowledgments This study was supported in part by a grant of the Korea Healthcare technology R&D project, Ministry for Health & Welfare, Republic of Korea (HI12C1834). Conflict of interest The authors have no financial or nonfinancial conflict of interest to declare.
References 1. Choi SJ, Rho YH, Ji JD, Song GG, Lee YH (2006) Genome scan meta-analysis of rheumatoid arthritis. Rheumatology (Oxford) 45:166–170. doi:10.1093/rheumatology/kei128 2. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG (2007) PADI4 polymorphisms and rheumatoid arthritis susceptibility: a meta-analysis. Rheumatol Int 27:827–833. doi:10.1007/s00296-007-0320-y 3. Rooney M, David J, Symons J, Di Giovine F, Varsani H, Woo P (1995) Inflammatory cytokine responses in juvenile chronic arthritis. Br J Rheumatol 34:454–460 4. Kim HR, Heo YM, Jeong KI, Kim YM, Jang HL, Lee KY, Yeo CY, Kim SH, Lee HK, Kim SR, Kim EG, Choi JK (2012) FGF-2 inhibits TNF-alpha mediated apoptosis through upregulation of Bcl2-A1 and Bcl-xL in ATDC5 cells. BMB Rep 45:287–292
13
Rheumatol Int 5. Constantin A, Dieude P, Lauwers-Cances V, Jamard B, Mazieres B, Cambon-Thomsen A, Cornelis F, Cantagrel A (2004) Tumor necrosis factor receptor II gene polymorphism and severity of rheumatoid arthritis. Arthritis Rheum 50:742–747. doi:10.1002/ art.20113 6. Kemper O, Derre J, Cherif D, Engelmann H, Wallach D, Berger R (1991) The gene for the type II (p75) tumor necrosis factor receptor (TNF-RII) is localized on band 1p36.2-p36.3. Hum Genet 87:623–624 7. Till A, Rosenstiel P, Krippner-Heidenreich A, MascherettiCroucher S, Croucher PJ, Schafer H, Scheurich P, Seegert D, Schreiber S (2005) The Met-196 - > Arg variation of human tumor necrosis factor receptor 2 (TNFR2) affects TNF-alphainduced apoptosis by impaired NF-kappaB signaling and target gene expression. J Biol Chem 280:5994–6004. doi:10.1074/jbc. M411541200 8. Hussein YM, Mohamed RH, Pasha HF, El-Shahawy EE, Alzahrani SS (2011) Association of tumor necrosis factor alpha and its receptor polymorphisms with rheumatoid arthritis in female patients. Cell Immunol 271:192–196. doi:10.1016/j.cellimm.2011.06.023 9. Oregon-Romero E, Vazquez-Del Mercado M, Navarro-Hernandez RE, Torres-Carrillo N, Martinez-Bonilla G, Estrada-Garcia I, Rangel-Villalobos H, Munoz-Valle JF (2006) Tumor necrosis factor receptor 2 M196R polymorphism in rheumatoid arthritis and osteoarthritis: relationship with sTNFR2 levels and clinical features. Rheumatol Int 27:53–59. doi:10.1007/s00296-006-0159-7 10. Yen JH, Tsai WC, Chen CJ, Ou TT, Lin CH, Hu CJ, Liu HW (2003) Tumor necrosis factor receptor 2 microsatellite and exon 6 polymorphisms in rheumatoid arthritis in Taiwan. J Rheumatol 30:438–442 11. Dahlqvist SR, Arlestig L, Sikstrom C, Linghult S (2002) Tumor necrosis factor receptor type II (exon 6) and interleukin-6 (-174) gene polymorphisms are not associated with family history but tumor necrosis factor receptor type II is associated with hypertension in patients with rheumatoid arthritis from northern Sweden. Arthritis Rheum 46:3096–3098. doi:10.1002/art.10592 12. Fabris M, Tolusso B, Di Poi E, Assaloni R, Sinigaglia L, Ferraccioli G (2002) Tumor necrosis factor-alpha receptor II polymorphism in patients from southern Europe with mild-moderate and severe rheumatoid arthritis. J Rheumatol 29:1847–1850 13. Bridges SL Jr, Jenq G, Moran M, Kuffner T, Whitworth WC, McNicholl J (2002) Single-nucleotide polymorphisms in tumor necrosis factor receptor genes: definition of novel haplotypes and racial/ethnic differences. Arthritis Rheum 46:2045–2050. doi:10.1002/art.10463 14. Shibue T, Tsuchiya N, Komata T, Matsushita M, Shiota M, Ohashi J, Wakui M, Matsuta K, Tokunaga K (2000) Tumor necrosis factor alpha 5′-flanking region, tumor necrosis factor receptor II, and HLA-DRB1 polymorphisms in Japanese patients with rheumatoid arthritis. Arthritis Rheum 43:753–757. doi:10.1002/15290131(200004)43:43.0.CO;2-O 15. Bayley JP, Bakker AM, Kaijzel EL, Huizinga TW, Verweij CL (2003) Association of polymorphisms of the tumour necrosis factor receptors I and II and rheumatoid arthritis. Rheumatology (Oxford) 42:969–971. doi:10.1093/rheumatology/keg267 16. Ghelani AM, Samanta A, Jones AC, Mastana SS (2011) Association analysis of TNFR2, VDR, A2 M, GSTT1, GSTM1, and ACE genes with rheumatoid arthritis in South Asians and Caucasians of East Midlands in the United Kingdom. Rheumatol Int 31:1355–1361 17. Potter C, Worthington J, Silman A, Barton A (2005) TNFR2 is not associated with rheumatoid arthritis susceptibility in a Caucasian population [3]. Arthritis Rheum 52:2579–2581 18. Barton A, John S, Ollier WE, Silman A, Worthington J (2001) Association between rheumatoid arthritis and polymorphism
Rheumatol Int of tumor necrosis factor receptor II, but not tumor necrosis factor receptor I, in Caucasians. Arthritis Rheum 44:61–65. doi:10.1002/1529-0131(200101)44:13.0.CO;2Q 19. Lee YH, Bae SC, Choi SJ, Ji JD, Song GG (2011) Associations between vitamin D receptor polymorphisms and susceptibility to rheumatoid arthritis and systemic lupus erythematosus a meta-analysis. Mol Biol Rep 38:3643–3651. doi:10.1007/ s11033-010-0477-4 20. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG (2006) Association of TNF-alpha -308 G/A polymorphism with responsiveness to TNFalpha-blockers in rheumatoid arthritis: a meta-analysis. Rheumatol Int 27:157–161. doi:10.1007/s00296-006-0175-7 21. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG, Nath SK, Harley JB (2007) The PTPN22 C1858T functional polymorphism and autoimmune diseases–a meta-analysis. Rheumatology (Oxford) 46:49–56. doi:10.1093/rheumatology/kel170 22. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21:1539–1558. doi:10.1002/sim.1186 23. Egger M, Smith GD, Phillips AN (1997) Meta-analysis: principles and procedures. BMJ 315:1533–1537 24. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188 25. Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629–634 26. You CG, Li XJ, Li YM, Wang LP, Li FF, Guo XL, Gao LN (2013) Association analysis of single nucleotide polymorphisms of proinflammatory cytokine and their receptors genes with rheumatoid arthritis in northwest Chinese Han population. Cytokine 61:133– 138. doi:10.1016/j.cyto.2012.09.007 27. Goeb V, Dieude P, Vittecoq O, Mejjad O, Menard JF, Thomas M, Gilbert D, Boumier P, Pouplin S, Daragon A, Fardellone P, Tron F, Cornelis F, Le Loet X (2005) Association between the TNFRII 196R allele and diagnosis of rheumatoid arthritis. Arthritis Res Ther 7:R1056–R1062. doi:10.1186/ar1777 28. Dieude P, Petit E, Cailleau-Moindrault S, Osorio J, Pierlot C, Martinez M, Faure S, Alibert O, Lasbleiz S, De Toma C, Bardin T, Prum B, Cornelis F, European Consortium on Rheumatoid Arthritis F (2002) Association between tumor necrosis factor receptor II and familial, but not sporadic, rheumatoid arthritis: evidence for genetic heterogeneity. Arthritis Rheum 46:2039– 2044. doi:10.1002/art.10101 29. Glossop JR, Dawes PT, Mattey DL (2011) Antinuclear antibodies are associated with tumor necrosis factor receptor I gene polymorphism in patients with rheumatoid arthritis. Clin Exp Rheumatol 29:609–615 30. Mrazek F, Holla LI, Hutyrova B, Znojil V, Vasku A, Kolek V, Welsh KI, Vacha J, du Bois RM, Petrek M (2005) Association
of tumour necrosis factor-alpha, lymphotoxin-alpha and HLADRB1 gene polymorphisms with Lofgren’s syndrome in Czech patients with sarcoidosis. Tissue Antigens 65:163–171. doi:10.1111/j.1399-0039.2005.00370.x 31. Pandey JP, Frederick M, Sarcoidosis ARGACCESo (2002) TNFalpha, IL1-beta, and immunoglobulin (GM and KM) gene polymorphisms in sarcoidosis. Hum Immunol 63:485–491 32. Yamaguchi E, Itoh A, Hizawa N, Kawakami Y (2001) The gene polymorphism of tumor necrosis factor-beta, but not that of tumor necrosis factor-alpha, is associated with the prognosis of sarcoidosis. Chest 119:753–761 33. Somoskovi A, Zissel G, Seitzer U, Gerdes J, Schlaak M, Muller Quernheim J (1999) Polymorphisms at position -308 in the promoter region of the TNF-alpha and in the first intron of the TNF-beta genes and spontaneous and lipopolysaccharide-induced TNF-alpha release in sarcoidosis. Cytokine 11:882–887. doi:10.1 006/cyto.1999.0498 34. Seitzer U, Swider C, Stuber F, Suchnicki K, Lange A, Richter E, Zabel P, Muller-Quernheim J, Flad HD, Gerdes J (1997) Tumour necrosis factor alpha promoter gene polymorphism in sarcoidosis. Cytokine 9:787–790. doi:10.1006/cyto.1997.0224 35. Cornelis F, Faure S, Martinez M, Prud’homme JF, Fritz P, Dib C, Alves H, Barrera P, de Vries N, Balsa A, Pascual-Salcedo D, Maenaut K, Westhovens R, Migliorini P, Tran TH, Delaye A, Prince N, Lefevre C, Thomas G, Poirier M, Soubigou S, Alibert O, Lasbleiz S, Fouix S, Bouchier C, Liote F, Loste MN, Lepage V, Charron D, Gyapay G, Lopes-Vaz A, Kuntz D, Bardin T, Weissenbach J, Ecraf S (1998) New susceptibility locus for rheumatoid arthritis suggested by a genome-wide linkage study. Proc Natl Acad Sci USA 95:10746–10750 36. Mori L, Iselin S, De Libero G, Lesslauer W (1996) Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and TNFR1-deficient mice. J Immunol 157:3178–3182 37. Heilig B, Mapara M, Brockhaus M, Krauth K, Dorken B (1991) Two types of TNF receptors are expressed on human normal and malignant B lymphocytes. Clin Immunol Immunopathol 61:260–267 38. Abe Y, Yamauchi K, Kimura S (1995) 75- but not 55-kDa tumor necrosis factor receptor is active in the homotypic aggregation and proliferation of human lymphokine-activated T killer (T-LAK) cells in vitro. J Leukoc Biol 57:462–468 39. Horiuchi T, Kiyohara C, Tsukamoto H, Sawabe T, Furugo I, Yoshizawa S, Ueda A, Tada Y, Nakamura T, Kimoto Y, Mitoma H, Harashima S, Yoshizawa S, Shimoda T, Okamura S, Nagasawa K, Harada M (2007) A functional M196R polymorphism of tumour necrosis factor receptor type 2 is associated with systemic lupus erythematosus: a case-control study and a meta-analysis. Ann Rheum Dis 66:320–324. doi:10.1136/ard.2006.058917
13
Original Article
Associations between functional TNFR2 196 M/R polymorphisms and susceptibility to rheumatoid arthritis: a meta‑analysis Gwan Gyu Song · Sang‑Cheol Bae · Young Ho Lee
Received: 10 March 2014 / Accepted: 16 April 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Several studies have examined the effects of tumor necrosis factor receptor (TNFR) 1 +38 A/G and TNFR2 196 M/R polymorphisms on susceptibility to RA and have reported conflicting results. The purpose of this study was to examine whether the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms are associated with RA susceptibility. We performed a literature search using the Medical Literature Analysis and Retrieval System Online and Embase citation indices, and conducted a meta-analysis to examine the association between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA. Our meta-analysis included a total of 13 studies from 11 articles, consisting of 11 studies of the TNFR2 polymorphism (2,092 cases and 1,483 controls), and two studies of the TNFR1 polymorphism (672 cases and 288 controls). The meta-analysis revealed a significant association between the TNFR2 196 RR genotype and RA risk (OR 1.737, 95 % CI 1.275–2.367, P = 4.6 × 10−5). Stratification by ethnicity indicated an association between the TNFR2 196 RR genotype and RA in Europeans (OR 2.054, 95 % CI 1.305– 3.232, P = 0.002), but not in East Asians (OR 1.596, 95 % CI 0.642–3.971, P = 0.314). Analysis using a homozygote contrast model showed the same pattern for the TNFR2 196 RR genotype in a European and East Asian population. However, no association was found between the TNFR1
G. G. Song · Y. H. Lee (*) Division of Rheumatology, Department of Internal Medicine, Korea University Anam Hospital, Korea University College of Medicine, 126‑1, Anam‑dong 5‑ga, Seongbuk‑gu, Seoul 136‑705, Korea e-mail: [email protected] S.-C. Bae The Hospital for Rheumatic Diseases, Hanyang University Medical Center, Seoul, Korea
+36 A/G polymorphism and RA in a European population. Our meta-analysis demonstrated that the functional TNFR2 196 M/R polymorphism is associated with susceptibility to RA in the European population. Keywords TNFR · Meta-analysis · Polymorphism · RA
Introduction Rheumatoid arthritis (RA) is a chronic inflammatory disease that predominantly involves synovial joints and affects up to 1 % of adults worldwide. Although the etiology of RA remains undetermined, it has been established that susceptibility to RA has a genetic component [1, 2]. Tumor necrosis factor-α (TNF-α) is a potent pro-inflammatory cytokine that plays important roles in inflammatory and immune responses, including those observed in RA [3, 4]. TNF-α stimulates cytokine production, enhancing the expression of adhesion molecules and increasing neutrophil activation. The effects of TNF-α are mediated through two receptors, TNFR1 (TNFRSF1A) and TNFR2 (TNFRSF1B). Both TNFR1 and TNFR2 are expressed in the synovial tissue, cartilage, and pannus of patients with RA [5]. Several single nucleotide polymorphisms have been identified in the TNFR1 and TNFR2 genes. The two that have been studied the most intensively are an A-to-G polymorphism at position +36 in exon 1 of the TNFR1 gene (rs767455) and a T/G polymorphism at position +196 (rs1061622) in exon 6 of TNFR2, which results in a nonconservative amino acid substitution (methionine (M) to arginine (R)) within the fourth extracellular domain of TNFR2 [6]. The TNFR2 196 M/R polymorphism may affect receptor function, and the TNFR2 196 R allele transmits a stronger TNF-α signal than the TNFR2 196 M allele [7].
13
Several studies have examined the potential effects of the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms on susceptibility to RA [8–18]. However, these studies have shown mixed results due to small sample sizes and low statistical power. Meta-analysis is a powerful method for overcoming the problem of small sample sizes and inadequate statistical power in genetic studies of complex traits [19– 21]. Therefore, we conducted a meta-analysis to explore whether the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms are associated with susceptibility to RA.
Materials and methods Identification of eligible studies and data extraction We conducted a literature search for studies that examined the association between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA. We used the Medical Literature Analysis and Retrieval System Online (MEDLINE) and Embase citation indices to identify articles published through February 2014 in which the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms were identified in RA patients and controls. Additionally, all references in the identified articles were reviewed to identify studies not indexed by MEDLINE and Embase. The following key words and subject terms were used: “tumor necrosis factor receptor” “TNFR,” “TNFR1,” “TNFR2,” and “RA.” Studies were included in the analysis if they (1) were case– control studies, (2) contained genotype data, and (3) contained sufficient data to calculate odds ratios (ORs). No language restrictions were applied. We excluded the following: (1) studies containing overlapping data; (2) studies in which the number of genotypes or alleles could not be ascertained; and (3) studies in which family members had been studied, as these analyses were based on linkage considerations. The following information was extracted from each identified study: author, year of publication, ethnicity of study population, demographics, number of cases and controls, and genotype and allele frequencies of the TNFR1 and TNFR2 polymorphisms.
Rheumatol Int
which represents the proportion of between-study variability attributable to heterogeneity rather than chance [22]. I2 values of 25, 50, and 75 % were nominally assigned as low, moderate, and high estimates, respectively. The fixed-effects model assumes that a genetic factor has a similar effect on disease susceptibility across all studies investigated and that observed variations among studies are caused by chance alone [23]. The random-effects model assumes that different studies show substantial diversity and assess both withinstudy sampling error and between-study variance [24]. When study groups are homogeneous, the two models are similar. If the study groups lack homogeneity, the random-effects model usually provides wider CIs than the fixed-effects model. The random-effects model is most appropriate when there is significant between-study heterogeneity [24]. Statistical manipulations were performed using Comprehensive Meta-Analysis software (Biostat, Englewood, NJ, USA). The power of each study was computed as the probability of detecting an association between the polymorphism and RA, using a significance level of 0.05, and assuming an OR of 1.5 (small effect size). Power analysis was performed using the statistical program G*Power (http://www. psycho.uni-duesseldorf.de/aap/projects/gpower). Evaluation of sensitivity test and publication bias The Chi-squared test was used to determine whether the observed genotype frequencies in controls conformed to Hardy–Weinberg equilibrium (HWE). We performed a subgroup analysis by excluding studies in which genotype distribution in controls was inconsistent with HWE, because deviation from HWE among controls suggests the possibility of bias during control selection or genotyping errors. Sensitivity analysis was also performed to assess the influence of each individual study on the pooled or by omitting each study individually. Publication bias is typically detected using funnel plots, but because they require a range of studies of varying sizes and subjective judgment, we used Egger’s linear regression test [25], which measures funnel plot asymmetry on a natural logarithmic scale of ORs.
Evaluation of statistical associations
Results
We performed meta-analyses using (1) allelic contrast, (2) the recessive model, (3) the dominant model, and (4) homozygote contrast. Point estimates of risks, ORs, and 95 % confidence intervals (CIs) were estimated for each study. Additionally, within- and between-study variations and heterogeneities were assessed using Cochran’s Q test. Cochran’s Q test assesses the null hypothesis that all studies evaluated have the same effect. The effect of heterogeneity was quantified using I2, with a range between 0 and 100 %,
Studies included in the meta‑analysis
13
Seventy-eight studies were identified using electronic and manual searches, and 15 of these were selected for full-text review based on their title and abstract details [8–18, 26– 29]. Four studies were excluded because they contained no extractable polymorphism data [26–29]. Thus, a total of 11 studies met our inclusion criteria [8–18] (Fig. 1). One of the eligible studies contained data from two different RA groups
Rheumatol Int
Fig. 1 Flowchart representing the study design
[16] and was thus treated independently [30–34]. Thus, a total of 13 separate comparisons were considered in our meta-analysis, which were 11 studies of the TNFR2 polymorphism, consisting of 2,092 cases and 1,483 controls, including 5 European, 2 East Asian, and single South Asian, African American, Latin American, and Arab populations; and 2 studies of the TNFR1 polymorphism, consisting of 672 cases and 288 controls, including 2 European populations [15, 18]. Thus, the ethnicity-specific meta-analysis was restricted to European and East Asian populations. Selected details of the individual studies are summarized in Table 1. The statistical power of these studies ranged from 25.6 to 81.2 %. One of the studies had a statistical power exceeding 80 %. Frequency of the R allele of the TNFR2 196 M/R polymorphism by ethnicity The mean frequency of the R allele of the TNFR2 196 M/R polymorphism was 23.5 % among all healthy controls. East
Asians had a lower R allele prevalence than the other ethnic groups (16.3 %). Among healthy controls, the frequency of the TNFR2 196 R allele in the African American, Latin American, European, Arab, and South Asian populations was 16.9, 21.3, 23.6, 28.1, and 42.1 %, respectively (Table 2). Meta‑analysis of the association between the TNFR2 196 M/R polymorphism and RA A meta-analysis of all RA patients and of each ethnic patient was performed (Table 3). The meta-analysis revealed no association between the TNFR2 196 R allele and RA (OR 1.162, 95 % CI 0.996–1.428, P = 0.153) (Table 3). Stratification by ethnicity indicated no association between the TNFR2 196 R allele and RA in European and East Asian populations (OR 1.250, 95 % CI 0.942–1.660, P = 0.112; OR 1.026, 95 % CI 0.593–1.775, P = 0.927) (Table 3). Analysis using the dominant model
13
13 European
The Netherlands UK
Barton [18]
291
172 127 127 50 177 114 257 175 108 545 240 RA 381 143
160 140 127 120 130 75 183 84 62 265 137 Control 145 95
56 42 46 37 113 55 158 107 70 404 104 AA 122
MM
RA
Control
RA
Numbers
Allele contrast, b Assuming an odds ratio of 1.5 (small effect size) at a significance level of 0.05
a
Ref reference, HWE Hardy–Weinberg equilibrium, UK United Kingdom
European
Arab European South Asian Latin American European East Asian European European African American East Asian European
Ethnicity
Egypt UK UK Mexico UK Taiwan Sweden Italy USA Japan UK
Ethnicity
TNFR2 196 M/R polymorphism Hussein [8] Ghelani-1 [16] Ghelani-2 [16] Oregon-Romero [9] Potter [17] Yen [10] Dahlqvist [11] Fabris [12] Bridges [13] Shibue [14] Barton [18] TNFR1 +36 A/G polymorphism Barley [15]
Author (Ref)
Table 1 Characteristics of the studies included in the meta-analysis
152
88 67 59 13 57 54 83 60 33 124 107 AG 208
MR
44
28 18 22 0 7 5 16 8 5 17 29 GG 51
RR
44
80 68 43 77 71 28 114 57 41 208 74 AA 50
MM
Control
76
70 65 61 35 54 43 61 26 21 54 56 AG 83
MR
23
10 7 23 8 5 4 8 1 0 3 7 GG 12
RR
0.301
0.815
0.298 0.083 0.866 0.159 0.173 0.016 0.964 0.294 0.108 0.808 0.383
HWE P value
0.690
0.235
0.000 0.003 0.719 0.079 0.178 0.221 0.636 0.181 0.500 0.012 0.080
Association P valuea
54.9
63.0
44.5 37.2 35.7 25.6 41.7 27.9 55.4 36.3 25.6 81.2 49.2
Power (%)b
Rheumatol Int
Rheumatol Int Table 2 Prevalence of the R allele of the TNFR2 196 M/R polymorphism Population
European East Asian South Asian African American Latin American Arab Overall
No of studies
Numbers
M allele (%)
Case
Control
Case
Control
5 2 1 1 1 1
976 659 127 108 50 172
674 340 127 62 120 160
27.2 16.8 40.6 19.9 13.0 41.9
23.6 16.3 45.1 16.9 21.3 28.1
11
2,092
1,483
25.2
23.5
TNFR2 196 RR genotype and RA (OR 1.737, 95 % CI 1.275–2.367, P = 4.6 × 10−5) (Table 3; Fig. 2). Stratification by ethnicity indicated an association between the TNFR2 196 RR genotype and RA in the European population (OR 2.054, 95 % CI 1.305–3.232, P = 0.002), but not in the East Asian population (OR 1.596, 95 % CI 0.642– 3.971, P = 0.314) (Table 3; Fig. 2). Analysis using the homozygote contrast model showed the same pattern for the TNFR2 196 RR genotype in European and East Asian populations (Table 3). Meta‑analysis of the association between the TNFR1 +36 A/G polymorphism and RA
showed the same pattern for the TNFR2 196 R allele in Europeans and East Asians (Table 3). However, the metaanalysis revealed a significant association between the
The meta-analysis revealed no association between the TNFR1 +36 G allele and RA in the European population (OR 0.893, 95 % CI 0.737–1.083, P = 0.051) (Table 4;
Table 3 Meta-analysis of the TNFR2 196 M/R polymorphism and RA Polymorphism
R versus M allele
RR versus RM + MM (recessive)
RR + RM versus MM (dominant)
RR versus MM
Population
Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab Overall European East Asian South Asian African American Latin American Arab
Number of studies
Numbers
Test of association
Test of heterogeneity
RA
Control
OR
95 % CI
P value
Model
P value
I2
11 5 2 1 1 1 1 11 5 2 1 1 1 1 11 5 2 1 1 1 1 11 5 2 1 1 1
2,092 976 659 127 108 50 172 2,092 976 659 127 108 50 172 2,092 976 659 127 108 50 172 2,092 976 659 127 108 50
1,483 674 340 127 62 120 160 1,483 674 340 127 62 120 160 1,483 674 340 127 62 120 160 1,483 674 340 127 62 120
1.162 1.250 1.026 0.937 1.219 0.554 1.840 1.737 2.054 1.596 0.947 6.143 0.131 2.917 1.135 1.217 0.942 0.901 1.060 0.629 2.071 1.857 2.217 1.488 0.894 6.475 0.122
0.946–1.428 0.942–1.660 0.593–1.775 0.658–1.334 0.685–2.169 0.286–1.071 1.330–2.545 1.275–2.367 1.305–3.232 0.642–3.971 0.497–1.805 0.361–122.1 0.007–2.314 1.368–6.220 0.896–1.438 0.863–1.715 0.482–1.842 0.538–1.510 0.549–2.046 0.302–1.311 1.328–3.231 1.120–3.081 1.392–3.521 0.590–3.753 0.436–1.832 0.349–120.1 0.007–2.163
0.153 0.112 0.927 0.719 0.500 0.079 0.500 4.6 × 10−5 0.002 0.314 0.869 0.203 0.165 0.006 0.293 0.262 0.862 0.693 0.862 0.216 0.001 0.016 0.001 0.400 0.760 0.210 0.151
R R R NA NA NA NA F F F NA NA NA NA R R R NA NA NA NA R F F NA NA NA
0.001 0.020 0.044 NA NA NA NA 0.137 0.491 0.184 NA NA NA NA 0.004 0.025 0.050 NA NA NA NA 0.023 0.223 0.109 NA NA NA
66.9 65.7 75.3 NA NA NA NA 32.6 0 43.2 NA NA NA NA 60.9 64.1 73.9 NA NA NA NA 51.7 29.7 61.0 NA NA NA
1
172
160
4.000
1.801–8.891
0.001
NA
NA
NA
OR odds ratio, CI confidence interval, F fixed-effects model, R random-effects model, NA not available
13
Rheumatol Int
Fig. 2 ORs and 95 % CIs of individual studies and pooled data for the association between the RR genotype of the TNFR2 196 M/R polymorphism and RA in the overall group (a) and in each ethnic group (b)
Table 4 Meta-analysis of the TNFR1 +36 A/G and RA Polymorphism
Population
Number of studies
Numbers
Test of association
Test of heterogeneity
RA
Control
OR
95 % CI
P value
Model
P value
I2
0.614 0.323 0.982
0 0 0
G versus A allele GG versus GA + AA (recessive) GG + GA versus AA (dominant)
European European European
2 2 2
672 672 672
288 288 288
0.893 0.755 0.920
0.737–1.083 0.525–1.087 0.687–1.232
0.051 0.131 0.576
F F F
GG versus AA
European
2
672
288
0.748
0.495–1.128
0.166
F
0
OR odds ratio, CI confidence interval, F fixed-effects model, R random-effects model
Fig. 3). Analyses using the dominant and recessive models and homozygote contrast showed no association between the TNFR1 +36 A/G polymorphism and RA risk (Table 4).
13
Heterogeneity, sensitivity test, and publication bias The meta-analysis of the TNFR2 196 R allele across all groups and within each ethnic group showed between-study
Rheumatol Int
Fig. 3 ORs and 95 % CIs of individual studies and pooled data for the association between the G allele (a) and the GG genotype (b) of the TNFR1 +36 A/G polymorphism and RA in the European population
heterogeneity (Table 3). However, no heterogeneity was identified in the meta-analysis of European and East Asian populations using either the recessive model or homozygote contrast. There was no overall heterogeneity in the meta-analyses of the TNFR1 +36 A/G polymorphism (Table 4). Sensitivity analysis showed that no individual study significantly affected the pooled OR. Publication bias results in a disproportionate number of positive studies and poses a problem for meta-analyses. However, Egger’s regression test showed no evidence of publication bias (P value > 0.1).
Discussion Although the multifactorial nature of RA is well known, genetic factors are considered to be strong determinants of these diseases, thus encouraging researchers to search for the genes responsible. Studies of the associations between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and RA have reported conflicting results [8–18], which is not surprising. Although association studies provide a powerful means of identifying genetic factors underlying autoimmune diseases, most reported association studies lack sufficient statistical power. All studies, with one exception, showed a statistical power less than 80 % to detect an association between the two polymorphisms and RA. Thus, thousands of study subjects are required for adequate
statistical power. In such cases, meta-analysis can be used as an alternative. In this meta-analysis, we combined data from published studies to evaluate genetic associations between the most commonly studied polymorphisms of the TNFR1 and TNFR2 genes, namely the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms, and RA. Our results showed a significant association between the TNFR2 196 RR genotype and RA risk in European, but not in East Asian population. These findings may be partially explained by ethnic differences. European populations showed a higher frequency of the TNFR2 196 R allele (28.1 %), whereas East Asian populations showed the lowest frequency of the A allele (16.3 %). TNFR2 has been considered to be a candidate gene for RA because of its chromosomal location and functional significance [5, 6]. TNFR2 is located on chromosome 1p36, which is a susceptibility locus for RA [35], and the TNFR gene has been shown to be involved in autoimmune disease in an animal model [36]. The TNFR2 196 R allele has been found to be associated with susceptibility to autoimmune diseases, including RA. TNFR2 is predominantly expressed in cells of myeloid origin, especially in stimulated T and B cells, and it has a higher affinity for TNF-α [37]. TNFR2 is the major TNF receptor found on T cells and mediates apoptosis in CD8+ T cells [38]. A previous meta-analysis has shown that the 196 R allele of the functional M196R polymorphism of TNFR2 is a risk factor for SLE in the Asian population, but not in Europeans [39]. Thus, the association reported is concordant with the present results, but the ethnic association between Europeans and Asian populations differed. Nonetheless, both studies support the hypothesis that TNFR2 plays an important role in the pathogenesis of autoimmune diseases such as RA and SLE, but the difference in ethnic associations between RA and SLE requires further investigation. In recent years, susceptibility genes of RA have been studied intensively by genome-wide association studies (GWAS). Recent data could not show importance of TNFR polymorphisms with RA susceptibility. A possible reason on difference in association results between recent GWAS and our study data needs to be explained. GWAS studies compare risk allele with control allele, but do not analyze additional genetic models. In contrast, we examined the association using various genetic models such as allelic contrast, recessive model, dominant model, and homozygote contrast. Our study using allelic contrast did not show any association between the TNFR2 196 R allele and RA, in agreement with recent RA GWAS data. However, the meta-analysis under the recessive model and the homozygote contrast model revealed a significant association between the TNFR2 196 M/R polymorphism and RA.
13
TNFR1 is located on chromosome 12p13 and has 10 exons, and it is expressed in all nucleated cells, especially cells susceptible to the cytotoxic action of TNF [13]. The TNFR1 +38 A/G polymorphism has no functional significance for TNFR1, but this polymorphism may be in linkage disequilibrium with a nearby causal variant. Our results did not show an association between the TNFR1 +36 A/G polymorphism and RA in Europeans. However, this result should be interpreted with caution, because the meta-analysis of the TNFR1 +38 A/G polymorphism and RA included only two studies, which is insufficient to provide conclusive evidence. Thus, further studies using larger numbers of subjects are required. The present study has some limitations that should be considered. First, heterogeneity and confounding factors may have distorted the analysis. Second, RA varies in severity, and in the studies included here, the RA activity level was unclear. Further research is required to examine whether an association exists between the TNFR1 +38 A/G and TNFR2 196 M/R polymorphisms and the activity or clinical features of RA. Third, because we included data from European and East Asian patients in our ethnicityspecific meta-analysis, our ethnicity-associated results are applicable only to these two groups. In conclusion, this meta-analysis demonstrates that the functional TNFR2 196 M/R polymorphism is associated with susceptibility to RA in Europeans. Accordingly, our findings support a role of the TNFR2 196 M/R polymorphism in the pathogenesis of RA. Large-scale studies in populations with different ethnicities are required to further explore the relationship between polymorphisms in the TNFR genes and the pathogenesis of RA. Acknowledgments This study was supported in part by a grant of the Korea Healthcare technology R&D project, Ministry for Health & Welfare, Republic of Korea (HI12C1834). Conflict of interest The authors have no financial or nonfinancial conflict of interest to declare.
References 1. Choi SJ, Rho YH, Ji JD, Song GG, Lee YH (2006) Genome scan meta-analysis of rheumatoid arthritis. Rheumatology (Oxford) 45:166–170. doi:10.1093/rheumatology/kei128 2. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG (2007) PADI4 polymorphisms and rheumatoid arthritis susceptibility: a meta-analysis. Rheumatol Int 27:827–833. doi:10.1007/s00296-007-0320-y 3. Rooney M, David J, Symons J, Di Giovine F, Varsani H, Woo P (1995) Inflammatory cytokine responses in juvenile chronic arthritis. Br J Rheumatol 34:454–460 4. Kim HR, Heo YM, Jeong KI, Kim YM, Jang HL, Lee KY, Yeo CY, Kim SH, Lee HK, Kim SR, Kim EG, Choi JK (2012) FGF-2 inhibits TNF-alpha mediated apoptosis through upregulation of Bcl2-A1 and Bcl-xL in ATDC5 cells. BMB Rep 45:287–292
13
Rheumatol Int 5. Constantin A, Dieude P, Lauwers-Cances V, Jamard B, Mazieres B, Cambon-Thomsen A, Cornelis F, Cantagrel A (2004) Tumor necrosis factor receptor II gene polymorphism and severity of rheumatoid arthritis. Arthritis Rheum 50:742–747. doi:10.1002/ art.20113 6. Kemper O, Derre J, Cherif D, Engelmann H, Wallach D, Berger R (1991) The gene for the type II (p75) tumor necrosis factor receptor (TNF-RII) is localized on band 1p36.2-p36.3. Hum Genet 87:623–624 7. Till A, Rosenstiel P, Krippner-Heidenreich A, MascherettiCroucher S, Croucher PJ, Schafer H, Scheurich P, Seegert D, Schreiber S (2005) The Met-196 - > Arg variation of human tumor necrosis factor receptor 2 (TNFR2) affects TNF-alphainduced apoptosis by impaired NF-kappaB signaling and target gene expression. J Biol Chem 280:5994–6004. doi:10.1074/jbc. M411541200 8. Hussein YM, Mohamed RH, Pasha HF, El-Shahawy EE, Alzahrani SS (2011) Association of tumor necrosis factor alpha and its receptor polymorphisms with rheumatoid arthritis in female patients. Cell Immunol 271:192–196. doi:10.1016/j.cellimm.2011.06.023 9. Oregon-Romero E, Vazquez-Del Mercado M, Navarro-Hernandez RE, Torres-Carrillo N, Martinez-Bonilla G, Estrada-Garcia I, Rangel-Villalobos H, Munoz-Valle JF (2006) Tumor necrosis factor receptor 2 M196R polymorphism in rheumatoid arthritis and osteoarthritis: relationship with sTNFR2 levels and clinical features. Rheumatol Int 27:53–59. doi:10.1007/s00296-006-0159-7 10. Yen JH, Tsai WC, Chen CJ, Ou TT, Lin CH, Hu CJ, Liu HW (2003) Tumor necrosis factor receptor 2 microsatellite and exon 6 polymorphisms in rheumatoid arthritis in Taiwan. J Rheumatol 30:438–442 11. Dahlqvist SR, Arlestig L, Sikstrom C, Linghult S (2002) Tumor necrosis factor receptor type II (exon 6) and interleukin-6 (-174) gene polymorphisms are not associated with family history but tumor necrosis factor receptor type II is associated with hypertension in patients with rheumatoid arthritis from northern Sweden. Arthritis Rheum 46:3096–3098. doi:10.1002/art.10592 12. Fabris M, Tolusso B, Di Poi E, Assaloni R, Sinigaglia L, Ferraccioli G (2002) Tumor necrosis factor-alpha receptor II polymorphism in patients from southern Europe with mild-moderate and severe rheumatoid arthritis. J Rheumatol 29:1847–1850 13. Bridges SL Jr, Jenq G, Moran M, Kuffner T, Whitworth WC, McNicholl J (2002) Single-nucleotide polymorphisms in tumor necrosis factor receptor genes: definition of novel haplotypes and racial/ethnic differences. Arthritis Rheum 46:2045–2050. doi:10.1002/art.10463 14. Shibue T, Tsuchiya N, Komata T, Matsushita M, Shiota M, Ohashi J, Wakui M, Matsuta K, Tokunaga K (2000) Tumor necrosis factor alpha 5′-flanking region, tumor necrosis factor receptor II, and HLA-DRB1 polymorphisms in Japanese patients with rheumatoid arthritis. Arthritis Rheum 43:753–757. doi:10.1002/15290131(200004)43:43.0.CO;2-O 15. Bayley JP, Bakker AM, Kaijzel EL, Huizinga TW, Verweij CL (2003) Association of polymorphisms of the tumour necrosis factor receptors I and II and rheumatoid arthritis. Rheumatology (Oxford) 42:969–971. doi:10.1093/rheumatology/keg267 16. Ghelani AM, Samanta A, Jones AC, Mastana SS (2011) Association analysis of TNFR2, VDR, A2 M, GSTT1, GSTM1, and ACE genes with rheumatoid arthritis in South Asians and Caucasians of East Midlands in the United Kingdom. Rheumatol Int 31:1355–1361 17. Potter C, Worthington J, Silman A, Barton A (2005) TNFR2 is not associated with rheumatoid arthritis susceptibility in a Caucasian population [3]. Arthritis Rheum 52:2579–2581 18. Barton A, John S, Ollier WE, Silman A, Worthington J (2001) Association between rheumatoid arthritis and polymorphism
Rheumatol Int of tumor necrosis factor receptor II, but not tumor necrosis factor receptor I, in Caucasians. Arthritis Rheum 44:61–65. doi:10.1002/1529-0131(200101)44:13.0.CO;2Q 19. Lee YH, Bae SC, Choi SJ, Ji JD, Song GG (2011) Associations between vitamin D receptor polymorphisms and susceptibility to rheumatoid arthritis and systemic lupus erythematosus a meta-analysis. Mol Biol Rep 38:3643–3651. doi:10.1007/ s11033-010-0477-4 20. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG (2006) Association of TNF-alpha -308 G/A polymorphism with responsiveness to TNFalpha-blockers in rheumatoid arthritis: a meta-analysis. Rheumatol Int 27:157–161. doi:10.1007/s00296-006-0175-7 21. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG, Nath SK, Harley JB (2007) The PTPN22 C1858T functional polymorphism and autoimmune diseases–a meta-analysis. Rheumatology (Oxford) 46:49–56. doi:10.1093/rheumatology/kel170 22. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21:1539–1558. doi:10.1002/sim.1186 23. Egger M, Smith GD, Phillips AN (1997) Meta-analysis: principles and procedures. BMJ 315:1533–1537 24. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188 25. Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629–634 26. You CG, Li XJ, Li YM, Wang LP, Li FF, Guo XL, Gao LN (2013) Association analysis of single nucleotide polymorphisms of proinflammatory cytokine and their receptors genes with rheumatoid arthritis in northwest Chinese Han population. Cytokine 61:133– 138. doi:10.1016/j.cyto.2012.09.007 27. Goeb V, Dieude P, Vittecoq O, Mejjad O, Menard JF, Thomas M, Gilbert D, Boumier P, Pouplin S, Daragon A, Fardellone P, Tron F, Cornelis F, Le Loet X (2005) Association between the TNFRII 196R allele and diagnosis of rheumatoid arthritis. Arthritis Res Ther 7:R1056–R1062. doi:10.1186/ar1777 28. Dieude P, Petit E, Cailleau-Moindrault S, Osorio J, Pierlot C, Martinez M, Faure S, Alibert O, Lasbleiz S, De Toma C, Bardin T, Prum B, Cornelis F, European Consortium on Rheumatoid Arthritis F (2002) Association between tumor necrosis factor receptor II and familial, but not sporadic, rheumatoid arthritis: evidence for genetic heterogeneity. Arthritis Rheum 46:2039– 2044. doi:10.1002/art.10101 29. Glossop JR, Dawes PT, Mattey DL (2011) Antinuclear antibodies are associated with tumor necrosis factor receptor I gene polymorphism in patients with rheumatoid arthritis. Clin Exp Rheumatol 29:609–615 30. Mrazek F, Holla LI, Hutyrova B, Znojil V, Vasku A, Kolek V, Welsh KI, Vacha J, du Bois RM, Petrek M (2005) Association
of tumour necrosis factor-alpha, lymphotoxin-alpha and HLADRB1 gene polymorphisms with Lofgren’s syndrome in Czech patients with sarcoidosis. Tissue Antigens 65:163–171. doi:10.1111/j.1399-0039.2005.00370.x 31. Pandey JP, Frederick M, Sarcoidosis ARGACCESo (2002) TNFalpha, IL1-beta, and immunoglobulin (GM and KM) gene polymorphisms in sarcoidosis. Hum Immunol 63:485–491 32. Yamaguchi E, Itoh A, Hizawa N, Kawakami Y (2001) The gene polymorphism of tumor necrosis factor-beta, but not that of tumor necrosis factor-alpha, is associated with the prognosis of sarcoidosis. Chest 119:753–761 33. Somoskovi A, Zissel G, Seitzer U, Gerdes J, Schlaak M, Muller Quernheim J (1999) Polymorphisms at position -308 in the promoter region of the TNF-alpha and in the first intron of the TNF-beta genes and spontaneous and lipopolysaccharide-induced TNF-alpha release in sarcoidosis. Cytokine 11:882–887. doi:10.1 006/cyto.1999.0498 34. Seitzer U, Swider C, Stuber F, Suchnicki K, Lange A, Richter E, Zabel P, Muller-Quernheim J, Flad HD, Gerdes J (1997) Tumour necrosis factor alpha promoter gene polymorphism in sarcoidosis. Cytokine 9:787–790. doi:10.1006/cyto.1997.0224 35. Cornelis F, Faure S, Martinez M, Prud’homme JF, Fritz P, Dib C, Alves H, Barrera P, de Vries N, Balsa A, Pascual-Salcedo D, Maenaut K, Westhovens R, Migliorini P, Tran TH, Delaye A, Prince N, Lefevre C, Thomas G, Poirier M, Soubigou S, Alibert O, Lasbleiz S, Fouix S, Bouchier C, Liote F, Loste MN, Lepage V, Charron D, Gyapay G, Lopes-Vaz A, Kuntz D, Bardin T, Weissenbach J, Ecraf S (1998) New susceptibility locus for rheumatoid arthritis suggested by a genome-wide linkage study. Proc Natl Acad Sci USA 95:10746–10750 36. Mori L, Iselin S, De Libero G, Lesslauer W (1996) Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and TNFR1-deficient mice. J Immunol 157:3178–3182 37. Heilig B, Mapara M, Brockhaus M, Krauth K, Dorken B (1991) Two types of TNF receptors are expressed on human normal and malignant B lymphocytes. Clin Immunol Immunopathol 61:260–267 38. Abe Y, Yamauchi K, Kimura S (1995) 75- but not 55-kDa tumor necrosis factor receptor is active in the homotypic aggregation and proliferation of human lymphokine-activated T killer (T-LAK) cells in vitro. J Leukoc Biol 57:462–468 39. Horiuchi T, Kiyohara C, Tsukamoto H, Sawabe T, Furugo I, Yoshizawa S, Ueda A, Tada Y, Nakamura T, Kimoto Y, Mitoma H, Harashima S, Yoshizawa S, Shimoda T, Okamura S, Nagasawa K, Harada M (2007) A functional M196R polymorphism of tumour necrosis factor receptor type 2 is associated with systemic lupus erythematosus: a case-control study and a meta-analysis. Ann Rheum Dis 66:320–324. doi:10.1136/ard.2006.058917
13
