Meta-analysis of HLA-DRB1 polymorphism in Latin American patients with rheumatoid arthritis

Abstract

Objectives

To estimate the common effect size of HLA-DRB1 alleles on rheumatoid arthritis (RA) susceptibility across Latin America populations through a meta-analysis combining the results of published data.

Methods

Case-control studies on HLA-DRB1 association with RA in Latin America were searched up to October 2006. Genotype frequencies were extracted according to both shared epitope (SE) and HLA-DR4 positive or negative alleles. The effect summary odds ratio (OR) and 95% confidence intervals was obtained. Heterogeneity and publication bias were assessed.

Results

Eight studies containing 684 cases and 1015 controls were included. Under the random effects model, the common OR was 3.28 (1.93, 5.60) (p < 0.0001) and 3.54 (2.47, 5.05) (p = 4.22 × 10^− 12) for HLA-DR4 and SE, respectively. There was no evidence of publication bias according to Funnel plot and Egger's regression test (p = 0,445 for DR4 and p = 0,464 for SE meta-analysis). Significant heterogeneity was observed for HLA-DR4 (I² = 81.06%, Q = 36.96, p = 0.000005) but not for the SE meta-analysis.

Conclusions

HLA-DR4 and SE positive HLA-DRB1 alleles (mainly HLA-DRB1⁎0404) are associated with RA in Latin Americans. Heterogeneity is expected owing to the diverse degree of admixture between the examined populations. Our findings support the HLA as a major susceptibility locus for RA and validate the SE hypothesis in Latin America.

Introduction

Rheumatoid arthritis (RA) is a chronic and systemic autoimmune disease characterized by symmetric polyarthritis with progressive damage of diarthrodial joints leading to disability, increased comorbidity and premature mortality [1]. RA is a complex and heterogeneous genetic disease in which multiple genetic and non-genetic factors interact over the time [1]. Latin America (i.e. the geographical area defined by Mexico, Central America, South America, and the islands of the Caribbean) is a growing and foremost region with a population of about 515 million people. The RA prevalence reported in Latin America is considered to be less than 0.5% [2], [3], [4].

Evidence of familial clustering was the first indication of a genetic susceptibility to RA. Using variance modelling, twin studies have determined that the RA heritability is about 60% based on the higher monozygotic twins concordance rates (12–15%) compared to dizygotic twins (2–4%) [5]. Familial aggregation for RA, measured by the sibling recurrent risk ratio, vary from 2 to 17 depending upon the disease prevalence in the population used as reference [6]. Segregation analysis showed that the RA pattern of inheritance likely follows a mixed model, with a major gene (probably at the MHC) operating against a polygenic background [7]. The above findings indicate an important genetic component in RA.

Several whole-genome scans for RA have been performed with multi-locus nonparametric linkage analysis to identify susceptibility loci (Fig. 1) [8], [9], [10], [11], [12], [13], [14], [15], [16]. The HLA region have been the only one with significant evidence of linkage (LOD score > 3.6 or P < 3 × 10^− 5) across all the genome scans except the Japanese study [8]. In addition, consistent evidence gathered from association studies indicates that HLA-DRB1 is the strongest susceptibility locus for RA, accounting for about 30% of the total genetic component [6]. Hence, other non-HLA loci also contribute to disease susceptibility even though most of the non-HLA loci identified in individual studies have not been replicated [17]. This can be explained by insufficient sample size, low statistical power, and genetic or clinical heterogeneity. Genome-wide meta-analysis for RA have confirmed the HLA loci as the greatest susceptibility factor and provided greater statistical power to identify non-HLA loci at 1p, 2q, 5p, 6, 8p, 12, 16 y 18q [18], [19], [20]. A 2-locus linkage analysis detected improved evidence of linkage for loci on 6q16.3 and 16p12.3 after conditioning for HLA, and suggested an epistatic interaction between these loci and HLA-DRB1[21].

Thirty years have elapsed since Peter Stastny found RA to be associated with the mixed leukocyte culture type Dw4 (HLA-DRB1^⁎0401) [22]. Further investigations have established more RA-associated alleles, mainly HLA-DRB1^⁎0401, ^⁎0404 and ^⁎0408 in Caucasians; HLA-DRB1^⁎0405 in Spaniards, Japanese and Jews, HLA-DRB1^⁎0101/2 in Israelis, HLA-DRB1^⁎1402 in some Native Americans such as Pima and Yakima Indians, and HLA-DRB1^⁎1001 in Greeks [Reviewed in [23], [24]. By comparing amino acid sequences encoded by the disease-associated HLA-DRB1 alleles listed above, Gregresen et al. demonstrated a conserved motif (L-L-E-[Q/R]-[R/K]-R-A-A) comprising residues 67–74 in the third hypervariable region of the DRβ1 chain and named it the shared epitope[25]. These residues constitute an α-helical domain forming one side of the antigen binding site, a site likely to affect antigen presentation. Thus, the SE might selectively bind an arthritogenic peptide which could favour an autoimmune response. Negatively charged anchor residues best fit into this positive binding pocket 4, as in the case of the Collagen II peptide (261–263), an epitope in human-transgenic mouse models of RA [17]. HLA alleles with a negatively charged amino acid at any one of these positions are not associated with the disease. These alleles usually contain an aspartic acid (D) at residue 70 or the common motif D⁷⁰-E-R-A-A⁷⁴ (HLA-DRB1⁎0103, ⁎0402, ⁎1102, ⁎1103, ⁎1301, ⁎1302, and ⁎1304), and may protect against RA or favours a less erosive disease [26].

Current evidence indicates that the SE association with RA is highly complex and varying among populations [17], [27]. Moreover, the genotype penetrance is small because each susceptibility SE-allele is present in 5–15% of the normal population and the absolute risk for developing RA in SE-encoding allele carrying individuals is relatively low [24]. These observations indicate that SE is neither necessary nor sufficient for disease to occur, and successive investigations have suggested that the primary role of the SE could be related to severity of disease rather than to susceptibility [28].

HLA studies in Latin American patients with RA are scarce and results diverse. Such inconsistencies might be attributable to variations in genetic background and to insufficient sample size. Review and meta-analysis are both tools to synthesize and combining data across studies, the latter being a more objective statistical method to estimate the common effect across several studies improving statistical power. Thus, to estimate the genetic contribution of SE positive HLA-DRB1 alleles to disease susceptibility across different Latin American populations, we undertook a systematic review along with a meta-analysis of all case-control studies on HLA-DRB1 polymorphisms associated with RA conducted to date.