Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Arthritis Rheum. Author manuscript; available in PMC 2011 October 1.
Published in final edited form as:
PMCID: PMC2944013

A Functional RANKL Polymorphism Associated with Younger Age at Onset of Rheumatoid Arthritis



We previously reported association of co-occurrence of HLA-DRB1 shared epitope (SE) and RANKL SNPs with younger age of RA onset in 182 rheumatoid factor positive (RF) European American (EA) early RA patients. Here, we fine-mapped the 48 kb RANKL region in the extended 210 EA RF-positive early RA cohort, sought replication of RA-associated SNPs in additional 501 EA and 298 African-Americans (AA) RA cohorts, and explored functional consequences of RA-associated SNPs.


SNP genotyping was conducted using pyrosequencing or TaqMan PCR assays. Associations of rs7984870 with RANKL expression in plasma, PBMC and isolated T cells were quantified using ELISA and RT-PCR. Site-directed mutagenesis of rs7984870 within the 2kb RANKL promoter was performed to drive the luciferase reporter gene in osteoblast and stromal cell lines. Interaction of DNA and protein was determined by electrophoretic mobility shift assay.


A single promoter SNP rs7984870 was consistently significantly associated with earlier age of RA onset in 3 independent seropositive (RF or anti-cyclic citrullinated peptide antibody positive) RA cohorts but not in seronegative RA patients. The risk C allele of rs7984870 conferred 2-fold higher plasma RANKL levels in RF-positive RA patients, significantly elevated RANKL mRNA expression in activated normal T cells, and increased promoter activity after stimulation in vitro via differential binding to transcription factor SOX5.


The RANKL promoter allele that increased transcriptional levels upon stimulation might promote interaction between activated T cells and dendritic cells, predisposing to younger RA onset in seropositive EA and/or AA individuals.

Rheumatoid arthritis (RA) is characterized by chronic inflammation and destruction of cartilage and bone. The etiology of RA is not clearly defined, but it is presumed that environmental factors trigger the disease in genetically predisposed individuals. The strongest genetic contribution to RA susceptibility is the HLA - class II alleles that contain a common amino acid motif called the “shared epitope” (SE), which is also a risk factor for disease progression and development of extraarticular manifestations (1). In addition to the SE alleles, multiple gene variants, including PTPN22, PADI4, FCRL3, CTLA4, STAT4, TRAF1/C5, CD40, IL2RB, IL21, PRKCQ, KIF5A, AFF3, REL, TNFAIP3, CD28, PRDM1 and CD2/CD58 have been associated with susceptibility to RA; however, causal variants of many of these loci remain elusive (25). Anti-cyclic citrullinated peptide antibody (ACPA) is highly specific for RA and can be detected years before the first clinical manifestations as a putative predictor for developing RA. ACPA production is associated with HLA-DRB1, PTPN22, TRAF1/C5, CD40, FCGR3A, STAT4, REL and PADI4 alleles (611), as well as with environmental factors, primarily smoking (12), suggesting different pathophysiologic mechanisms underlying seropositive and seronegative RA.

Receptor activator of nuclear factor κB ligand (RANKL [OMIM 602642]), a type II membrane protein of the tumor necrosis factor family (TNFSF11), is expressed on osteoblasts, stromal cells, activated T cells, B cells (13) and megakaryocytes (14). The RANKL, located on chromosome 13q14, encodes three isoforms: RANKL1 and RANKL2 are both transmembrane proteins with the latter lacking an intracellular domain; RANKL3 is a soluble protein containing neither an intracellular nor a transmembrane domain (15). A deficiency of RANKL, the key molecule mediating osteoclast development, activity, and survival (16), leads to osteoporosis and severe skeletal abnormalities in the murine model (17). Apart from its importance in osteoclastogenesis, RANKL participates in immune regulation by mediating differentiation of T and B lymphocytes, prolonging survival of mature dendritic cells (DCs), promoting activation of monocytes, and participating in the development of lymph nodes and medullary thymic epithelial cells (18). In RA, local and systemic increased RANKL levels are associated with bone resorption, suggesting a pivotal role in mediating bone erosion (19). Importantly, co-localization of RANKL and RANK expression at the site of DC and T cell interaction in rheumatoid synovia and lymph nodes may represent a direct link between cell-mediated immunity and bone destruction in RA (20).

We have previously observed the association of younger age of RA onset with the presence of HLA-DRB1*04 and RANKL polymorphisms in 182 RF+ EA RA patients (21). An interaction between SE and RANKL polymorphisms was associated with a mean 18 years younger age at RA onset in this cohort. Here, we used the enlarged EA RF+ RA cohort as a discovery panel to fine map RANKL variants predisposing to early RA onset, replicated in independent EA and AA early RA cohorts and tested this genetic association in EA RF+ polyarticular juvenile idiopathic arthritis (JIA) patients. Finally, we characterized functional properties of the identified and confirmed RA-associated SNP in three seropositive RA cohorts.


Study populations

The present study included 4 independent cohorts (3 RA and 1 JIA) (Table 1). The three RA cohorts were early RA long-term observational studies including 1) Discovery panel, 210 EA RF+ early RA patients were from the Western Consortium of Practicing Rheumatologists (WCPR cohort) (22), which had been enlarged from the 182 EA patients reported (21); 2) Replication panel 1, 501 independent EA early RA patients were from the Brigham Rheumatoid Arthritis Sequential Study registry (BRASS cohort); 3) Replication panel 2, 298 patients who were participating in the Consortium for the Evaluation of AAs with Early Rheumatoid Arthritis Registry (CLEAR cohort) (23). We also tested the genetic association in 80 EA RF+ polyarticular JIA patients recruited from Cincinnati's Children Hospital Medical Center, OH. All RA patients fulfilled the American College of Rheumatology 1987 revised classification criteria for RA (24). Early RA was defined as those patients who enrolled within 24 months after symptom onset and before disease-modifying antirheumatic drugs (DMARDs) therapy. This study was approved by appropriate institutional review boards.

Table 1
Clinical characteristics of the RA and JIA patients

Genotyping methods

SNPs spanning a 48 kb region from 7.3 kb upstream to 6.7 kb downstream of RANKL transcribed sequence were surveyed in the NCBI-dbSNP build 36 database. A total of 16 SNPs were genotyped in this study, including 5 promoter, 7 intronic, one 3' untranslated region (UTR), and three 3'downstream RANKL polymorphisms (Figure 1B). Among these 16 SNPs, 9 tag EA haplotypes throughout dbSNP database 124, 125 and 126, 1 promoter and 3 intronic SNPs create putative transcription factor binding sites, 1 SNP has been associated with low bone mineral density in men (25) and 2 SNPs were tested in our previous study (21). Genomic DNA was isolated from blood mononuclear cells by the standard protocol. Genotyping of RANKL SNPs was determined using PCR pyrosequencing as described (20), except rs5803141, which was genotyped using fluorescence-based PCR. In the BRASS cohort, SNPs rs7984870, rs9525641 and rs1054016 were genotyped by TaqMan real-time PCR assays (Applied Biosystems). HLA–DRB1 genotyping were determined by DNA-based sequence-specific oligonucleotide probe assay, AlleleSEQR HLA–DRB1 reagent kit or allele-specific polymerase chain reaction as described elsewhere (2628).

Figure 1
RANKL structure with tested SNPs and haplotype blocks in EA and AA patients. A. The −log10 base of genotypic p value for each SNP (comparing those homozygous for the minor allele vs. those homozygous for the major allele plus heterozygote) is ...


Concentrations of soluble RANKL levels in plasma of 23 RA patients at baseline and 24 normal controls were quantified using a commercial ELISA kit from ALPCO Diagnostics (Salem, NH).

Real-Time PCR

Two isoforms in RANKL transcript levels (isoform1: NM003701.2 and total RANKL transcript: NM033012.2) in PBMC from 38 EA normal controls were measured using ABI TaqMan assay. Relative RANKL expression in PBMCs and T cells was normalized using the expression level of CD4 and ribosomal protein (RPLPO), respectively and calculated by the 2−Δ Δ Ct method.

Cells cultures

Primary T cells were purified from peripheral blood of 8 healthy volunteers (4 CC and 4 GG genotype for rs7984870) using a T-rosette purification kit (StemCell Technologies), which were 93% CD3+ by flow cytometry. 4 × 105 T cells were plated onto 12 well plates in the presence or absence of 20ng/ml IL-2 (R&D Systems) for 24h or 48h, and then harvested to perform Real-time PCR assays.

Site-directed mutagenesis

The RANKL promoter region (−1 bp to −2 kb relative to the transcription start site) was previously cloned into the luciferase reporter pGL2 Basic vector (a kind gift from Dr. Sakamuri V. Reddy) (29). The plasmids for −1816 G/G in the RANKL promoter region were mutated into −1816 C/C using the QuikChange Lightning Site-Directed Mutagenesis kit (Stratagene). Accuracy of mutagenesis was confirmed by sequence analysis.

Cell transfection

Five microgrammes of RANKL promoter/luciferase fusion plasmid DNA or empty vector (pGL4) were transfected into human stromal cells SAKA-T and osteoblast cells hFOB 1.19 using Lipofect-AMINE kit (Invitrogen) and 100ng of pRL-SV40 control vector (Renilla luciferase) was cotransfected as an internal control for transfection efficiency. After cells treated with or without b-FGF (4ng/ml) and TNFα (10ng/ml) (R&D Systems) for 48h, cell lysates were prepared and assayed for firefly and control Renilla luciferase using the Dual-Luciferase reporter assay (Promega).

EMSA and supershift assays

Electrophoretic mobility shift assay (EMSA) and supershift assays were performed as previously described (30). After Jurkat cells were stimulated with TNFα for 24h, nuclear extract from Jurkat cells was prepared by NE-PER nuclear extraction reagent (Thermo scientific). Biotin-labeled oligonucleotides were used to perform EMSA: 5'ACCCATTAAAAAC/GAATGGAAGACACTAT3' (forward) and 5'ATAGTGTCTTCCATTG/CTTTTTAATGGGT3' (reverse). EMSAs were performed with the LightShift Chemiluminescent EMSA Kit (Pierce). The antibody used in the supershift reactions was 2μl of polyclonal rabbit anti human SOX5 (Anti-SOX5) (Santa Cruz biotechnology).

Statistical analysis

Comparisons between patient groups were made by 1-way analysis of variance (ANOVA) for three groups (3 genotypes), and unpaired t-tests for two groups (dominant or recessive models) for age of RA symptom onset. For comparison of mean values, Student's t-test was performed. The strength of pair-wise linkage disequilibrium (LD) (r2) between two genetic polymorphisms was assessed using Haploview 4.0. Statistical analyses were performed using software package Prism 4 (GraphPad Software). P values < 0.05 were considered to indicate statistical significance.


Fine-mapping RANKL SNPs for association with younger age of RA onset in EA patients (Extended RF+ early RA WCPR cohort)

We previously observed a strong association between the combined presence of HLA-DRB1*04 and RANKL polymorphisms (rs9525641 in the promoter region and rs922996 in intron 4) with a younger age of RA onset in 182 RF+ EA early RA WCPR patients (21). Because these two RA-associated SNPs showed no potential biologic functions based on bioinformatic searches, we reasoned the observed association signals are likely to reflect potential functional variants within the RANKL locus. To test this hypothesis, we assessed 16 SNPs located in the 48 kb RANKL genomic region using the enlarged discovery panel of the 210 EA RF+ early RA WCPR patients.

The minor allele of 4 RANKL SNPs, mainly in promoter (rs5803141, rs7984870 and rs9525641 in promoter, and rs1054016 in 3'UTR) were associated with earlier age at RA onset in EA patients (p < 0.05) (Figure 1A and B). These 4 SNPs are in strong linkage disequilibrium (LD) (r2 = 0.75 ~ 0.94) with each other but showed modest LD (r2 = 0.57~ 0.65) with the formerly observed RA-associated RANKL rs922996, the association of which was not confirmed in this study.

The similarities in manifestations and genetic background between RF+ adult RA and RF+ polyarticular JIA (31) prompted us to test the potential association of these 4 SNPs with juvenile arthritis in 80 EA RF+ polyarticular JIA patients. No significant association between the minor allele of these SNPs and age of disease onset was observed in this cohort, which might be attributed to the modest sample size and narrow range in age of disease onset (median = 11.1 yrs; 25% to 75% percentile = 8.8 to 12.8 yrs).

Confirmation of the association of RANKL SNPs with younger age of RA onset in an additional EA cohort (BRASS cohort) and AA cohort (CLEAR cohort)

To seek replication of the significant association of RANKL genotypes with age of RA onset, we genotyped these four SNPs in two independent RA cohorts. One replication cohort was 501 independent EA early RA patients from BRASS. The strong LD of these 4 SNPs (r2 = 0.76 − 0.93) observed in BRASS cohort was similar to the LD values in WCPR EA cohort, indicating BRASS is a well-matched independent cohort for replication study. Only the rs7984870 CC allele was associated with earlier RA onset in 344 ACPA+ EA patients (37.6 ± 12.8 yrs) compared to non-CC carriers (41.4 ± 14.6 yrs; p = 0.04) and a similar trend of association in RF+ patients (CC vs. non-CC carriers, 37.9 ± 13.0 yrs vs. 41.4 ± 14.5 yrs; p = 0.08) (Figure 1A, Table 2).

Table 2
Association of RANKL rs7984870 risk allele with younger age at RA onset in WCPR, CLEAR and BRASS cohorts

The second independent replication panel included 298 AA RA patients from CLEAR cohort. Genotypic association with younger age of RA onset was observed in two promoter SNPs (rs5803141 and rs7984870) in an ACPA+ subset (n = 172) and a trend of this association in RF+ AA RA patients (n = 211) (Figure 1A, Table 2), implicating a role of RANKL promoter SNPs in development of RA in multiple ethnic groups.

Confirmation that the interaction between HLA-DRB1 SE and RANKL promoter contributes to earlier age of RA onset in WCPR, BRASS and CLEAR RA cohorts

We previously reported a significant association of the co-occurrence of RANKL polymorphisms and HLA-DRB1*04 with younger age at onset of RA in the original WCPR cohort (21). In the present study, we defined the SE as containing any one of the following HLA-DRB1 alleles: *0101, *0102, *0401, *0404, *0405, *0408, *0410, *1001, *1402, *1406. The mean age at disease onset in RA patients harboring SE was an average of 4.7 and 4.5 years younger than that in patients without SE in seropositive (RF+ and/or ACPA+) EA WCPR and BRASS RA cohort (p = 0.03 and 0.01, respectively), but this difference was not observed in AA CLEAR RA cohort (Figure 2A). Among RF+ RA patients, co-occurrence of SE and the CC allele of rs7984870 was associated with younger age at disease onset with an average of 8.2 yrs (WCPR cohort), 4.7 yrs (BRASS cohort) and 6.7 yrs (CLEAR cohort) younger than non-CC SE positive carriers (p = 0.004, 0.01 and 0.01, respectively) (Figure 2 B–D). A similar trend of younger age at onset also was seen in ACPA+ patients in BRASS and CLEAR cohorts (mean 5.1 and 5.7 yrs earlier; p =0.01 and 0.02) (Figure 2 C–D), suggesting co-occurrence of SE and RANKL risk alleles had a stronger impact on the earlier age at disease onset than each gene variant individually. However, these genetic associations were not observed in seronegative RA patients. Stratifying RA patients into 0, 1 and 2 copies of SE subgroups, only ACPA+ BRASS EA RA showed a trend toward a younger RA onset (supplementary Table 1). Of note, RANKL CC allele was not associated with early age at RA onset in 0 copy SE RA patients, which might be attributed to the small sample size (n= 8~26 patients in each CC plus SE 0 copy subgroup) and/ or in part because the effect of rs7984870 CC on age of RA onset might depend on HLA-DRB1 (supplementary Table 1).

Figure 2
Association of shared epitope (SE) and RANKL rs7984870 CC allele with age of RA onset. A. Association of SE with age of RA onset in seropositive WCPR, BRASS and CLEAR early RA cohort. B–D. Co-occurrence rs7984870 CC allele and SE negative, SE ...

The HLA– DRB1*04 alleles represent a considerably stronger susceptibility factor than other SE alleles in RA (32). When SE was defined as only HLA-DRB1*04 alleles (included *0401, *0404, *0405, *0408, and *0410), the combined presence of SE-containing DRB1*04 and rs7984870 CC genotype displayed an even earlier age of RA onset in RF+ or ACPA+ patients (mean 13.3 yrs earlier in WCPR cohort, P =0.0003, mean 5.8 and 6.3 yrs earlier in BRASS, P = 0.01 and 0.007, and mean 9 and 9.4 yrs earlier in CLEAR, P =0.01, respectively) (Figure 2 B–D). Multiple regression analysis testing the interaction on age at RA onset with the presence of SE and risk allele in rs7984870, confirmed a significant interaction only in WCPR and BRASS EA RA patients (p < 0.05, respectively), indicating that the interaction only existed in EA patients.

Our data indicated among the 4 SNPs identified in the initial discovery WCPR cohort, only rs7984870 showed genetic association with younger age onset in 2 independent early RA replication cohorts of EA and AA ancestries. The consistent association of rs7984870 with RA led us to investigate its potential functional effect.

RANKL promoter SNP rs7984870 might affect RANKL expression level

Plasma levels of soluble RANKL (derived from membrane forms and measured by ELISA) were compared to genotypes. Frozen plasma samples had been collected from the baseline evaluation of WCPR RF+ EA RA patients who were still within 1 year of symptom onset and had not yet been treated with DMARDs. Similar to a previous report (33), our study found that soluble RANKL levels in plasma were significantly higher in RA patients (0.64 ± 0.61 pmol/L) compared to healthy controls (HC) (0.26 ± 0.18 pmol/L; p = 0.004). No significant allelic-specific difference in RANKL levels was observed in HC. However, soluble RANKL levels were approximately 2 - fold higher in RA patients carrying rs7984870 CC genotype (0.87 ± 0.53 pmol/L) than in those carrying GG genotype (0.46 ± 0.67 pmol/L; p = 0.002; Figure 3A).

Figure 3
Association of RANKL genotypes with expression levels of mRNA and secreted protein. A. Soluble RANKL levels from serum samples of healthy controls (HC) who were homozygote for either the minor C allele or the major G allele of RANKL promoter SNP rs7984870 ...

We subsequently tested whether rs7984870 variant would impact on RANKL transcription. The relative expression of RANKL isoform 1 transcript (encodes RANKL1 and RANKL2 by alternative splicing) and total RANKL transcript (isoform 1 plus isoform 2 that encodes RANKL3) in PBMC were specifically measured using real-time PCR, and normalized using the expression level of CD4 because RANKL is mainly expressed in activated T helper cells in peripheral blood. Compared to isoform 1, isoform 2 has an upstream transcription start site, suggesting it may be regulated differently. Transcript levels of RANKL isoform 2 were generated from transcripts levels of total expression minus isoform 1 in this study. We did not observe any allele-specific difference in mRNA expression of the two isoforms in PBMC from HC (Figure 3B). RANKL mRNA expression in 22 RA patients (11 CC and 11 GG genotype for rs7984870) was also examined, but the allelic difference was not conclusive, perhaps owing to a small sample size and/or disease activity (data not shown).

As shown in Figure 3B, higher expression of isoform 1 than isoform 2 was observed in PBMC, similar to a previous report (15), consistent with isoform 1 being the major RANKL transcript in blood circulation. We next tested the allelic expression levels of the major RANKL isoform in T cells in vitro. The primary T cells from 8 normal controls (4 CC and 4 GG genotype for rs7984870) were treated with or without IL-2 20ng/ml for 24 or 48 h and harvested to perform real-time PCR assays. RANKL expression was normalized using the expression level of RPLPO. RANKL transcript levels were significantly increased after in vitro activation of T cells by co-culturing with IL-2 for 48h. Interestingly, T cells carrying the rs798470 CC genotype conferred significantly higher RANKL transcript levels than those carrying GG genotype after activation by IL-2 (p = 0.04, Figure 3C).

RANKL promoter SNP rs7984870 may affect the promoter activity

2-kb RANKL promoter luciferase constructs expressing either the major G allele or the minor C allele at position −1816 were transiently transfected into the human stromal cell line SAKA-T and osteoblast cell line hFOB1.19. Basic fibroblast growth factor (b-FGF) and TNFα have been reported to upregulate RANKL expression (29;34). Transfection of the rs7984870 CC RANKL promoter displayed a 40% increaseing transcription activity in stromal cell lines and a 30% increase in osteoblast cell lines compared with GG genotype after stimulation with b-FGF; similarly, the promoter activity was increased significantly more after stimulation with TNFα when these two cell lines were transfected with rs7984870 CC construct compared with GG construct (p < 0.05, Figure 4 A–B).

Figure 4
Allelic differences in transcriptional and DNA-binding properties of rs7984870 variants. A–B. SAKA-T (left) and hFOB 1.19 (right) cell lines were transfected with a luciferase (Luc) reporter gene driven by RANKL promoter sequences (−2000/−1) ...

rs7984870 CC genotype may create a binding site to transcription factor SOX5

To investigate the effect of the polymorphism on transcription factor binding, EMSA assay was performed. The 28-bp oligonucleotide containing the homozygous minor C allele or the major G allele of rs7984870 was incubated with nuclear extracts prepared from Jurkat cells after stimulation with TNFα for 24h. Binding of nuclear proteins was observed with the oligonucleotide probe containing either the GG or the CC allele (Figure 4C, lane 2 and 5). In silico analysis indicated that transcription factor SOX5 might bind to the rs7984870 CC allele but not to the GG allele. The addition of commercial polyclonal rabbit IgG antibodies to SOX5 produced a supershift band only in CC allele (Figure 4C lane 8) but not in GG allele (Figure 4C, lane 10), suggesting only rs7984870 CC allele could bind to the SOX5 transcription factor. The addition of rabbit IgG failed to produce a supershift band (Figure 4C, lanes 7 and 9). Binding to each oligonucleotide probe was competed by a 200-fold excess of unlabeled specific oligonucleotide (Figure 4C, lanes 3 and 6).


Although the role of RANKL in RA has been well established, little is known about the contribution of its genetic variations to RA onset. We previously observed a strong association between the combined presence of HLA-DRB1*04 and RANKL polymorphisms with a younger age of RA onset in 182 RF+ EA early RA WCPR patients (21). In this extended study of 210 RF+ RA WCPR patients, we fine mapped the RANKL locus and observed association of the minor allele of 4 SNPs (rs5803141, rs7984870, rs9525641 and rs1054016) with younger age at RA onset. Subsequent replication studies of these 4 SNPs showed association of a single SNP in the promoter, the minor C allele of rs7984870, with younger age of RA onset in two independent cohorts (ACPA+ EA BRASS and AA CLEAR). These two independent RA cohorts were early RA (<24 months after symptom onset) long-term observational cohorts, and have an accurately documented date of RA onset similar to WCPR cohort.

In both the previous (21) and extended WCPR cohort, the association between rs9525641 and younger age of RA onset was present; however, the formerly observed RA-associated rs922996 was not confirmed. Recently, a RANKL intron 1 rs2277438 has been associated with RA in European population (534 cases and 516 controls) (35). This SNP was also tested in our WCPR EA cohort, but had weak LD with promoter SNP rs7984870 (r2 = 0.17) and did not show the association with age of disease onset. Possible explanations for the inability to replicate RANKL SNPs (rs5803141, rs9525641, rs922996, rs1054016 and rs2277438) in some of the studied cohorts include genetic, ethnic, and clinical heterogeneity across study populations as well as the probability that these SNPs may be linked to the actual RANKL causal allele. Several genome-wide association studies (GWAS) in RA have identified and replicated more than 20 risk loci (2;5), but have not included the RA-associated RANKL SNP described here. Given that the association with younger age of RA onset was only observed in CC homozygotes, we used a recessive genetic model to estimate the power of GWAS to capture this risk allele under the assumption of relative risk of 1.3 (36). A sample size of 10,000 cases and 10,000 controls could provide 65% power to reach a p value of 10−7, which might explain why this SNP was not identified in previous GWAS in RA (2;5;36;37). However, the consistent presence of the observed genetic association in 3 independent panels and the functional properties of the rs7984870 allele in transcription regulation of RANKL in our study make this promoter polymorphism an excellent candidate for a causal variant contributing to younger age of RA onset.

SE is a recognized genetic risk predisposing to RA onset and serves as a binding site for arthritogenic peptides, allowing their presentation to CD4+ T cells. Similar to our previous report, the co-occurrence of rs7984870 risk allele and SE displayed the earliest age of RA onset in RF+ or ACPA+ RA patients. One possible hypothesis is that co-occurrence of RANKL polymorphisms and SE contributed to younger RA onset in ACPA+/RF+ patients by causing abnormally stimulated RANKL expression, which dysregulates dendritic/T cell communication and breaks the immune tolerance to self-components in the joint, thereby enhancing T-cell reactivity towards arthritogenic antigens.

A growing number of genetic variants have been reported to contribute to ACPA+ RA (611). In our replication studies performed in 298 AA and 501 additional EA patients, the significant genetic association of rs7984870 with earlier RA onset was only shown in the ACPA+ patients but not in ACPA− RA, suggesting that the RANKL allele might be a novel disease predisposing variant in ACPA+ RA subset. Alternatively, the lack of a significant association might be attributed to the modest sample size of seronegative RA patients used in these studies.

Our study provides evidence that risk C allele of rs7984870 conferred a 2-fold higher plasma soluble RANKL level in RF+ RA patients and significantly elevated RANKL isoform 1 mRNAs expression by activated control T cells, as well as increased promoter activity after cytokine stimulation in vitro. Over expression of RANKL results in functional alteration of epidermal DCs (38) and enhances survival of DCs (13). Transfer of RANKL-stimulated MRL/lpr DCs pulsed with collagen into MRL/lpr mice have been shown acceleration autoreactivity of T and B cells and enhance the productions of pathogenic cytokines in the recipients (13). The minor C allele of rs7984870 contributed to higher RANKL expression and protein level upon activation and thus, might affect initiation of immune responses between DCs and T cells. In addition, immunoreceptor tyrosine-based activation motif (ITAM) associated receptor networks have been implicated in providing co-stimulation for RANKL signaling (39); thus, another possible explanation is that genetic variants causing differences in RANKL expression could therefore result in abnormal cross-talk in signaling networks and alteration of cellular responses to various extracellular stimuli.

An important issue is how this RANKL variant affects the expression of RANKL. Our data suggests that the minor C allele of rs7984870 creates a binding site for transcription factor SOX5, a member of the SOX family, which might explain observed alterations in RANKL transcription. SOX family have been primarily characterized as important transcriptional regulators involved in determination of cell fate and tissue specification during a number of developmental processes (40). Elevated SOX5 transcripts have been related to invasive growth and reduced apoptosis in tumors and have been associated with progression of certain cancers (4143). Recently, SOX6, another member of the SOX family, which shares 91% amino acid homology in the leucine-zipper motif with SOX5, was identified as a low bone mineral density locus in a large-scale meta-analysis of GWAS studies (44). Taken together, the differential binding of the minor C allele of rs7984870 to SOX5 transcriptional regulator might influence bone metabolism, and facilitate invasive growth and resistance to apoptosis in the synovium, leading to RA manifestations.

In summary, the association of CC genotype of rs7984870 with younger RA onset was identified and replicated in several independent cohorts in this study. This novel genetic risk factor, RANKL rs7984870 CC genotype, conferred elevated promoter activity after stimulation by cytokines, potentially via binding to transcription factor SOX5. Elevated inducible RANKL mRNA and protein levels might result in enhanced interaction between activated T cells presenting arthritogenic peptides to DCs, predisposing to earlier development of seropositive (RF+ or ACPA+) RA in EA and AA patients. Our evidence suggests that the combination of CC genotype of rs7984870 RANKL polymorphism, SE, and abnormal humoral immune responses interacts to permit an earlier onset of RA.

Supplementary Material


The Western Consortium of Practicing Rheumatologists (WCPR) registry includes the following investigators: Robert Shapiro, Maria W. Greenwald, H. Walter Emori, Fredrica E. Smith, Craig W. Wiesenhutter, Charles Boniske, Max Lundberg, Anne MacGuire, Jeffry Carlin, Robert Ettlinger, Michael H. Weisman, Elizabeth Tindall, Karen Kolba, George Krick, Melvin Britton, Rudy Greene, Ghislaine Bernard Medina, Raymond T. Mirise, Daniel E. Furst, Kenneth B. Wiesner, Robert F. Willkens, Kenneth Wilske, Karen Basin, Robert Gerber, Gerald Schoepflin, Marcia J. Sparling, George Young, Philip J. Mease, Ina Oppliger, Douglas Roberts, J. Javier Orozco Alcala, John Seaman, Martin Berry, Ken J. Bulpitt, Grant Cannon, Gregory Gardner, Allen Sawitzke, Andrew Lun Wong, Daniel O. Clegg, Timothy Spiegel, Wayne Jack Wallis, Mark Wener, Robert Fox.

The Consortium for the Longitudinal Evaluation of AAs with Early Rheumatoid Arthritis (CLEAR) includes the following investigators: University of Alabama at Birmingham: S. Louis Bridges, Jr., MD, PhD (Director); Larry W. Moreland, MD (Co-Director); George Howard, DrPH (Co-Director); Graciela S. Alarcón, MD, MPH; Emory University: Doyt L. Conn, MD; University of North Carolina: Beth L. Jonas, MD; Leigh F. Callahan, PhD; Medical University of South Carolina: Edwin A. Smith, MD; Gary S. Gilkeson, MD. Washington University: Richard Brasington, MD.

The Brigham Rheumatoid Arthritis Sequential Study registry (BRASS) includes the following investigators: David M. Lee MD, PhD; Nancy A. Shadick, MD, MPH; Jing Cui, PhD; Christine Iannaccone, MPH; Michelle Frits.

We thank all patients for enrolling in the WCPR, CLEAR and BRASS Registries, and physicians at the academic centers for referring patients. We gratefully acknowledge the following physicians, who have also enrolled patients for CLEAR registry: Adahli Estrada Massey, MD, Auburn, AL; Runas Powers, MD, Alexander City, AL; Ben Wang, MD, Memphis, TN; Jacob Aelion, MD, Jackson, TN; Sohrab Fallahi, MD, Montgomery, AL; Richard Jones, PhD, MD, Tuscaloosa, AL; Donna Paul, MD, Montgomery, AL; William Shergy, MD, Huntsville, AL.

We gratefully acknowledge CLEAR and BRASS staff and coordinators at the following sites: University of Alabama at Birmingham: Cynthia Irwin, RN, MPH; Eugene Oliver, BS; Andrew O. Westfall, MS; Selena Luckett, RN, CRNC; Laticia Woodruff, RN, MSN; Stephanie McLean, BS; Sondra Beck; Emory University: Joyce Carlone, RN, RNP; Karla Caylor, BSN, RN; Meri Eger, RN; University of North Carolina: Pat Cummins, RN; Medical University of South Carolina: Trisha Sturgill.

We gratefully acknowledge Dr. Sakamuri V. Reddy provide us stromal cell line and RANKL luciferase reporter plasmid constructs.

We thank Ornella J Rullo MD and John FitzGerarld MD at UCLA for their comments.

This work is supported by the NIAMS R21 AR55147 (BPT) and the Meyer Young Investigator Award, the Southern California Chapter of Arthritis Foundation (WT).

Reference List

(1) van der Helm-van Mil AH, Wesoly JZ, Huizinga TW. Understanding the genetic contribution to rheumatoid arthritis. Curr Opin Rheumatol. 2005;17(3):299–304. [PubMed]
(2) Barton A, Worthington J. Genetic susceptibility to rheumatoid arthritis: an emerging picture. Arthritis Rheum. 2009;61(10):1441–6. [PubMed]
(3) Eyre S, Hinks A, Flynn E, Martin P, Wilson AG, Maxwell JR, et al. Confirmation of association of the REL locus with rheumatoid arthritis susceptibility in the UK population. Ann Rheum Dis. 2009 Epub ahead of print. [PMC free article] [PubMed]
(4) Kochi Y, Yamada R, Suzuki A, Harley JB, Shirasawa S, Sawada T, et al. A functional variant in FCRL3, encoding Fc receptor-like 3, is associated with rheumatoid arthritis and several autoimmunities. Nat Genet. 2005;37(5):478–85. [PMC free article] [PubMed]
(5) Raychaudhuri S, Thomson BP, Remmers EF, Eyre S, Hinks A, Guiducci C, et al. Genetic variants at CD28, PRDM1 and CD2/CD58 are associated with rheumatoid arthritis risk. Nat Genet. 2009;41(12):1313–8. [PMC free article] [PubMed]
(6) van der Linden MP, Feitsma AL, le CS, Kern M, Olsson LM, Raychaudhuri S, et al. Association of a single-nucleotide polymorphism in CD40 with the rate of joint destruction in rheumatoid arthritis. Arthritis Rheum. 2009;60(8):2242–7. [PMC free article] [PubMed]
(7) Plenge RM, Seielstad M, Padyukov L, Lee AT, Remmers EF, Ding B, et al. TRAF1-C5 as a risk locus for rheumatoid arthritis--a genomewide study. N Engl J Med. 2007;357(12):1199–209. [PMC free article] [PubMed]
(8) Remmers EF, Plenge RM, Lee AT, Graham RR, Hom G, Behrens TW, et al. STAT4 and the risk of rheumatoid arthritis and systemic lupus erythematosus. N Engl J Med. 2007;357(10):977–86. [PMC free article] [PubMed]
(9) Thabet MM, Huizinga TW, Marques RB, Stoeken-Rijsbergen G, Bakker AM, Kurreeman FA, et al. Contribution of Fcgamma receptor IIIA gene 158V/F polymorphism and copy number variation to the risk of ACPA-positive rheumatoid arthritis. Ann Rheum Dis. 2009;68(11):1775–80. [PubMed]
(10) Cha S, Choi CB, Han TU, Kang CP, Kang C, Bae SC. Association of anti-cyclic citrullinated peptide antibody levels with PADI4 haplotypes in early rheumatoid arthritis and with shared epitope alleles in very late rheumatoid arthritis. Arthritis Rheum. 2007;56(5):1454–63. [PubMed]
(11) Szodoray P, Szabo Z, Kapitany A, Gyetvai A, Lakos G, Szanto S, et al. Anti-citrullinated protein/peptide autoantibodies in association with genetic and environmental factors as indicators of disease outcome in rheumatoid arthritis. Autoimmun Rev. 2010;9(3):140–3. [PubMed]
(12) Mahdi H, Fisher BA, Kallberg H, Plant D, Malmstrom V, Ronnelid J, et al. Specific interaction between genotype, smoking and autoimmunity to citrullinated alpha-enolase in the etiology of rheumatoid arthritis. Nat Genet. 2009;41(12):1319–24. [PubMed]
(13) Izawa T, Ishimaru N, Moriyama K, Kohashi M, Arakaki R, Hayashi Y. Crosstalk between RANKL and Fas signaling in dendritic cells controls immune tolerance. Blood. 2007;110(1):242–50. [PubMed]
(14) Bord S, Frith E, Ireland DC, Scott MA, Craig JI, Compston JE. Synthesis of osteoprotegerin and RANKL by megakaryocytes is modulated by oestrogen. Br J Haematol. 2004;126(2):244–51. [PubMed]
(15) Suzuki J, Ikeda T, Kuroyama H, Seki S, Kasai M, Utsuyama M, et al. Regulation of osteoclastogenesis by three human RANKL isoforms expressed in NIH3T3 cells. Biochem Biophys Res Commun. 2004;314(4):1021–7. [PubMed]
(16) Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T, et al. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell. 1998;93(2):165–76. [PubMed]
(17) Kong YY, Feige U, Sarosi I, Bolon B, Tafuri A, Morony S, et al. Activated T cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature. 1999;402(6759):304–9. [PubMed]
(18) Leibbrandt A, Penninger JM. RANK/RANKL: regulators of immune responses and bone physiology. Ann N Y Acad Sci. 2008;1143:123–50. [PubMed]
(19) Choi Y, Arron JR, Townsend MJ. Promising bone-related therapeutic targets for rheumatoid arthritis. Nat Rev Rheumatol. 2009;5(10):543–8. [PMC free article] [PubMed]
(20) Page G, Miossec P. RANK and RANKL expression as markers of dendritic cell-T cell interactions in paired samples of rheumatoid synovium and lymph nodes. Arthritis Rheum. 2005;52(8):2307–12. [PubMed]
(21) Wu H, Khanna D, Park G, Gersuk V, Nepom GT, Wong WK, et al. Interaction between RANKL and HLA-DRB1 genotypes may contribute to younger age at onset of seropositive rheumatoid arthritis in an inception cohort. Arthritis Rheum. 2004;50(10):3093–103. [PubMed]
(22) Paulus HE, Bulpitt KJ, Ramos B, Park G, Wong WK. Relative contributions of the components of the American College of Rheumatology 20% criteria for improvement to responder status in patients with early seropositive rheumatoid arthritis. Arthritis Rheum. 2000;43(12):2743–50. [PubMed]
(23) Bridges SL, Jr., Hughes LB, Mikuls TR, Howard G, Tiwari HK, Alarcon GS, et al. Early rheumatoid arthritis in African-Americans: the CLEAR Registry. Clin Exp Rheumatol. 2003;21(5 Suppl 31):S138–S145. [PubMed]
(24) Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988;31(3):315–24. [PubMed]
(25) Dong SS, Liu XG, Chen Y, Guo Y, Wang L, Zhao J, et al. Association analyses of RANKL/RANK/OPG gene polymorphisms with femoral neck compression strength index variation in Caucasians. Calcif Tissue Int. 2009;85(2):104–12. [PMC free article] [PubMed]
(26) O'Dell JR, Nepom BS, Haire C, Gersuk VH, Gaur L, Moore GF, et al. HLA-DRB1 typing in rheumatoid arthritis: predicting response to specific treatments. Ann Rheum Dis. 1998;57(4):209–13. [PMC free article] [PubMed]
(27) Hughes LB, Morrison D, Kelley JM, Padilla MA, Vaughan LK, Westfall AO, et al. The HLA-DRB1 shared epitope is associated with susceptibility to rheumatoid arthritis in African Americans through European genetic admixture. Arthritis Rheum. 2008;58(2):349–58. [PMC free article] [PubMed]
(28) Ferucci ED, Majka DS, Parrish LA, Moroldo MB, Ryan M, Passo M, et al. Antibodies against cyclic citrullinated peptide are associated with HLA-DR4 in simplex and multiplex polyarticular-onset juvenile rheumatoid arthritis. Arthritis Rheum. 2005;52(1):239–46. [PubMed]
(29) Roccisana JL, Kawanabe N, Kajiya H, Koide M, Roodman GD, Reddy SV. Functional role for heat shock factors in the transcriptional regulation of human RANK ligand gene expression in stromal/osteoblast cells. J Biol Chem. 2004;279(11):10500–7. [PubMed]
(30) Leopold JA, Dam A, Maron BA, Scribner AW, Liao R, Handy DE, et al. Aldosterone impairs vascular reactivity by decreasing glucose-6-phosphate dehydrogenase activity. Nat Med. 2007;13(2):189–97. [PMC free article] [PubMed]
(31) Phelan JD, Thompson SD, Glass DN. Susceptibility to JRA/JIA: complementing general autoimmune and arthritis traits. Genes Immun. 2006;7(1):1–10. [PubMed]
(32) Weyand CM, Hicok KC, Conn DL, Goronzy JJ. The influence of HLA-DRB1 genes on disease severity in rheumatoid arthritis. Ann Intern Med. 1992;117(10):801–6. [PubMed]
(33) Hein GE, Meister M, Oelzner P, Franke S. sRANKL and OPG in serum and synovial fluid of patients with rheumatoid arthritis in comparison to non-destructive chronic arthritis. Rheumatol Int. 2008;28(8):765–9. [PubMed]
(34) Tunyogi-Csapo M, Kis-Toth K, Radacs M, Farkas B, Jacobs JJ, Finnegan A, et al. Cytokine-controlled RANKL and osteoprotegerin expression by human and mouse synovial fibroblasts: fibroblast-mediated pathologic bone resorption. Arthritis Rheum. 2008;58(8):2397–408. [PubMed]
(35) Assmann G, Koenig J, Pfreundschuh M, Epplen JT, Kekow J, Roemer K, et al. Genetic Variations in Genes Encoding RANK, RANKL, and OPG in Rheumatoid Arthritis: A Case-Control Study. J Rheumatol. 2010 [PubMed]
(36) Skol AD, Scott LJ, Abecasis GR, Boehnke M. Joint analysis is more efficient than replication-based analysis for two-stage genome-wide association studies. Nat Genet. 2006;38(2):209–13. [PubMed]
(37) Harley JB. IL-7Ralpha and multiple sclerosis risk. Nat Genet. 2007;39(9):1053–4. [PubMed]
(38) Loser K, Mehling A, Loeser S, Apelt J, Kuhn A, Grabbe S, et al. Epidermal RANKL controls regulatory T-cell numbers via activation of dendritic cells. Nat Med. 2006;12(12):1372–9. [PubMed]
(39) Ivashkiv LB. Cross-regulation of signaling by ITAM-associated receptors. Nat Immunol. 2009;10(4):340–7. [PMC free article] [PubMed]
(40) Lefebvre V, Dumitriu B, Penzo-Mendez A, Han Y, Pallavi B. Control of cell fate and differentiation by Sry-related high-mobility-group box (Sox) transcription factors. Int J Biochem Cell Biol. 2007;39(12):2195–214. [PMC free article] [PubMed]
(41) Huang DY, Lin YT, Jan PS, Hwang YC, Liang ST, Peng Y, et al. Transcription factor SOX-5 enhances nasopharyngeal carcinoma progression by down-regulating SPARC gene expression. J Pathol. 2008;214(4):445–55. [PubMed]
(42) Ueda R, Yoshida K, Kawase T, Kawakami Y, Toda M. Preferential expression and frequent IgG responses of a tumor antigen, SOX5, in glioma patients. Int J Cancer. 2007;120(8):1704–11. [PubMed]
(43) Zafarana G, Gillis AJ, van Gurp RJ, Olsson PG, Elstrodt F, Stoop H, et al. Coamplification of DAD-R, SOX5, and EKI1 in human testicular seminomas, with specific overexpression of DAD-R, correlates with reduced levels of apoptosis and earlier clinical manifestation. Cancer Res. 2002;62(6):1822–31. [PubMed]
(44) Rivadeneira F, Styrkarsdottir U, Estrada K, Halldorsson BV, Hsu YH, Richards JB, et al. Twenty bone-mineral-density loci identified by large-scale meta-analysis of genome-wide association studies. Nat Genet. 2009;41(11):1199–206. [PMC free article] [PubMed]