Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune disease with an estimated prevalence of ~1 % in the general population. RA is characterized by a chronic inflammation of the synovial membrane and the progressive destruction of the joint cartilage and subchondral bone. RA is a complex disease in which the interplay between multiple genetic and environmental factors determines not only its onset but also its evolution to more severe forms [
1].
The extent of joint damage is a clinically relevant and highly variable feature among RA patients. Recent studies have shown that there is a genetic predisposition to develop higher levels of joint destruction in RA. The heritability of this clinically relevant trait has been estimated to range from 45 to 60 % [
2]. The identification of the genes and genetic pathways that contribute to increase the risk of joint damage in RA will clearly be of high value for the future development of prognostic tests as well as for the design of more efficient therapeutic approaches.
The interleukin (IL)-6 signaling pathway is strongly associated with RA pathophysiology [
3]. IL-6 cytokine is highly expressed in the inflamed synovial tissue of RA patients and influences the functionality of multiple cell types including macrophages and T and B lymphocytes and osteoclasts [
4]. These effects are mediated by the binding of the cytokine to the IL-6 receptor protein (IL6R, CD126), which is expressed both on the surface of the cell membranes as well as in a soluble form. Targeting the cell-bound and soluble IL-6R using monoclonal antibodies has proven to be an efficacious therapy in RA, significantly reducing structural damage [
5]. Consequently, the gene encoding IL-6R is a strong candidate for association with the level of joint destruction in RA.
We performed a candidate gene study of the association of IL-6 receptor gene (IL6R) with the level of joint damage in RA. To test this hypothesis we identified single nucleotide polymorphisms (SNPs) tagging the IL6R locus and analyzed their association in a discovery cohort of RA patients from Spain. Using an independent cohort of patients, we have subsequently validated the observed associations.
Discussion
The key measure of severity in RA is the presence of joint erosion as determined by radiological examination. The IL-6 signaling pathway is essential in the proinflammatory network that contributes to tissue destruction in RA. Blocking a key member of this network, the IL-6 receptor has proven highly efficacious in the inhibition of joint damage [
5]. We have performed a candidate-gene association study to test the association of genetic variation at
IL6R with the level of erosions in RA.
To date, three main mRNA isoforms of
IL6RA have been described, with 10, 9 and 7 exons (mRNA isoforms 1, 2 and 3, respectively). We have found a strong association between a SNP in the first intron common to all isoforms and joint damage (SNP rs4845618,
P = 0.00011), and a moderate association of a SNP in the sixth intron of isoforms 1 and 2 (rs4845374,
P = 0.0021). After combining both variants rs4845618 and rs4845374 as a haplotype, the association with the level of joint damage in RA was found to be stronger (
P = 6.45 × 10
−5). This result suggests that neither of the two associated variants are the causal polymorphism that influences joint destruction, and instead they are proxies for another polymorphism. The LD pattern associated with the most strongly associated SNP, rs4845618, suggests that the causal polymorphism lies within the transcribed sequence of
IL6R (Fig.
1b). In this region lies SNP rs2228145, a variant that was recently found to have profound functional implications in
IL6R expression [
14]. Importantly, this variant has also been associated with RA risk [
14,
15]. However, the LD between rs4845374 and the SNP associated with joint damage rs4845618 is moderate in the Caucasian European population, like our Spanish RA patient cohorts (
r
2
= 0.44; 1KG data). Furthermore, the marker tagging rs2228145, SNP rs4453032 (
r
2
= 0.91), showed no evidence of association with the level of joint erosions in RA (
P = 0.51; Table
1). These results therefore suggest that the genetic variant associated with joint erosions in RA is not the same variant associated with disease risk. Future studies using next-generation sequencing analysis will be valuable to identify the precise variation at the IL6R locus that increases the risk of developing joint erosions at the
IL6R locus.
In most human complex traits, the vast majority of genetic associations are being identified outside coding regions [
16], and therefore the functional link is generally difficult to identify. In the present study, the SNP showing the strongest association to joint damage, rs4845618, is an intronic SNP of
IL6R. Evaluating publicly available results on tissue expression quantitative loci (eQTL) analysis [
17], there is evidence supporting rs4845618 being a highly significant cis-eQTL for
IL6R expression in whole blood (
P = 1.1 × 10
−16; data not shown). Also, analyzing the most recent regulatory and epigenetic evidence generated from large international consortia such as the Roadmap Epigenomics project [
18], there is strong regulatory evidence associated with this SNP (Table S3 in Additional file
1). In particular, the rs4845618 SNP region is associated with regulatory activity in more than 50 different human cell types, including CD4
+ T cells. Together these results support the present genetic association with joint damage being likely to influence
IL6R regulatory activity.
Tocilizumab is a monoclonal antibody against IL6R, prevents downstream IL-6 signaling and has been shown to significantly reduce the signs and symptoms of RA. In particular, there is increasing evidence that IL6R blockade prevents structural joint damage [
19]. IL-6 is a pleiotropic inflammatory cytokine and therefore many biological mechanisms can explain the efficacy of this treatment [
3]. From these mechanisms, the inhibition of inflammatory osteoclastogenesis after IL6R blocking [
20] is clearly the most direct biological process influencing the level of joint destruction in RA. There is recent suggestive evidence that variants at
IL6R are associated with the clinical response to tocilizumab [
21]. While these results are still preliminary, they provide additional support for the importance of variation at this gene with disease proinflammatory activity. Future studies analyzing the association between SNPs in the present study with the response to tocilizumab are possible.
In the present study we prioritized the selection of the best tagging SNPs of
IL6R variation over previously published variants, in order to capture most of the common variation in this gene. For example,
IL6R SNP rs8192284 (now merged into SNP rs2228145) – which had been previously associated with disease activity in RA [
22] – was not selected for genotyping. Importantly, however, rs8192284 is in very high LD with the genotyped SNP rs4453032 (
r
2
= 0.91, 1000 Genomes Project data on a Caucasian European population). As shown in the discovery stage, there is no evidence for association between SNP rs4453032 and joint damage in RA (
P = 0.51; Table
1). The high LD between rs8192284 and rs4453032 therefore supports the lack of association of this variant with joint damage in RA. This result confirms that the objective selection of the most informative markers is a powerful approach to identify the relevant genetic variation associated with disease severity in RA.
One potential limitation of this study is that using a simplified score might have reduced the statistical power of the association. It is possible that using a more standard score such as the SHS we would have better assessed joint damage in the RA cohorts, and therefore found a more significant association with the genetic variation at the IL6R locus. Nonetheless, the significant association of IL6R found both in the discovery cohort and in the validation cohort strongly supports the usefulness of the S-score to capture the essential features of this trait in RA. While the loss of information from not using a reference score like SHS would be unacceptable in clinical studies such as treatment efficacy analysis, this simplified erosion score could be a practical tool in those instances like genetic studies in which a large number of patient data must be collected from multiple centers and there is no possibility to perform more complex scores of joint damage.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MLL, IGA, and SM designed the study. MLL, AJ, AA, IGA, and SM analyzed the data. NP and RT performed molecular genetic studies. MLL, JM, BF-G, IUG, FJB, JDC, MA-L, AO, HC, JT, AE, MA, AO, GA, LR, IGA, and SM performed acquisition of the data and critically revised the manuscript. MLL, IGA, AJ, SM, and JM wrote the manuscript. All authors read and approved the final manuscript.