Genetic risk factors for rheumatoid arthritis differ in caucasian and korean populations

Vol. 60, No. 2, February 2009, pp 364–371 2009, American College of Rheumatology Genetic Risk Factors for Rheumatoid Arthritis Differ in Caucasian and Korean Populations Hye-Soon Lee,1 Benjamin D. Korman,2 Julie M. Le,2 Daniel L. Kastner,2 Elaine F. Remmers,2 Objective. Recent studies have identified a num-
in Caucasian patients with RA contributed significantly
ber of novel rheumatoid arthritis (RA) susceptibility
to disease in Koreans. Although tag SNPs covering the
loci in Caucasian populations. The aim of this study was
PTPN22 linkage disequilibrium block were polymor-
to determine whether the genetic variants at 4q27, 6q23,
phic, they did not reveal any disease association, and
CCL21, TRAF1/C5, and CD40 identified in Caucasians
resequencing did not identify any new common coding
are also associated with RA in a Korean case–control
region variants in this population. The 6q23 and 4q27
collection. We also comprehensively evaluated the ge-
SNPs assayed were nonpolymorphic in this population,
netic variation within PTPN22, a well-established auto-
and the TRAF1/C5, CD40, and CCL21 SNPs did not
immune disease–associated gene.
show any evidence for association with RA in this
Methods. We designed an experiment to thor-
population of Korean patients.
oughly evaluate the PTPN22 linkage disequilibrium
Conclusion. The genetic risk factors for RA are
region, using tag single-nucleotide polymorphisms
different in Caucasian and Korean patients. Although
(SNPs) and disease-associated SNPs at 5 RA-associated
patients of different ethnic groups share the HLA region
loci recently identified in Caucasians, in 1,128 Korean
as a major genetic risk locus, most other genes shown to
patients with RA and 1,022 ethnically matched control
be significantly associated with disease in Caucasians
subjects. We also resequenced the PTPN22 gene to seek
novel coding variants that might be contributing to
appear not to play a role in Korean patients with RA.
disease in this population.
Results. None of the susceptibility loci identified
Rheumatoid arthritis (RA) is a chronic auto- immune arthritis characterized by progressive joint de-struction. Both genetic and environmental factors have Supported in part by the Korea Health 21 R&D Project (grant 01-PJ3-PG6-01GN11-0002) and by the NIH (Intramural Pro- been shown to play a role in the development of RA.
gram of the National Institute of Arthritis and Musculoskeletal and Although the largest genetic risk factor predisposing to Skin Diseases). Mr. Korman’s work was supported by the NIH ClinicalResearch Training Program. Dr. Gregersen’s work was supported by RA, a common set of alleles at HLA–DRB1, has been grants from the NIH (R01-AR-44422 and N01-AI-95386).
associated with RA in populations of both Caucasian 1Hye-Soon Lee, MD, PhD: Hanyang University College of and Asian ancestry (1,2), many other risk loci have been Medicine and the Hospital for Rheumatic Diseases, Seoul, SouthKorea, and The Feinstein Institute for Medical Research, Manhasset, shown to confer RA susceptibility in only 1 ethnic group.
New York; 2Benjamin D. Korman, BS, Julie M. Le, BS, Daniel L.
For example, although PTPN22 has been consistently Kastner, MD, PhD, Elaine F. Remmers, PhD: National Institute of shown to be associated with RA in Caucasians, the 620W Arthritis and Musculoskeletal and Skin Diseases, Bethesda, Maryland;3Peter K. Gregersen, MD: Feinstein Institute for Medical Research, risk allele in this gene is not found and thus is not Manhasset, New York; 4Sang-Cheol Bae, MD, PhD, MPH: Hanyang disease-associated in Asian populations (3–7). In con- University College of Medicine and the Hospital for RheumaticDiseases, Seoul, South Korea.
trast, PADI4, SLC22A4, and FCRL3 have been associ- Dr. Gregersen has received consulting fees, speaking fees, ated with RA in studies of Asian patients, but the and/or honoraria from Roche Pharmaceuticals (less than $10,000).
associations were weak or negative in populations of Address correspondence and reprint requests to Elaine F.
Remmers, PhD, National Institutes of Health, 9 Memorial Drive, European ancestry (6,8–14). Although these results sug- 9/1W108, MSC 0908, Bethesda, MD 20892-0908. E-mail: remmerse@ gest that the genetic risk alleles that confer susceptibility to RA are heterogeneous across major ethnic groups, Submitted for publication June 10, 2008; accepted in revised recent reports have identified a common STAT4 haplo- DIFFERENT GENETIC RISK FACTORS IN CAUCASIAN AND KOREAN PATIENTS WITH RA type that confers a similar degree of risk of RA in both staging system proposed by Steinbrocker et al (23) was used as Asian and Caucasian populations (15,16).
a measure of the radiographic severity of RA.
Outside of the HLA region, the R620W variant PTPN22 resequencing. Genomic DNAs from 48 rheu-
matoid factor–positive Korean patients with RA were se- of the PTPN22 gene is the most consistently and most quenced over the regions representing the promoter region strongly disease-associated variant identified in RA as- (up to Ϫ2,000 bp), all 21 exons including exon–intron bound- sociation studies performed in Caucasian populations.
aries, and the 3Ј–untranslated region of PTPN22, using stan- Although this variant is not found in Asians, it is possible dard dideoxy DNA sequencing techniques (Polymorphic DNA that other variants within this important candidate gene SNP selection and assay design. We used the Tagger
could contribute to genetic susceptibility to RA in Asian method (ref. 24, and http://www.broad.mit.edu/mpg/tagger/) to populations. In addition to PTPN22, new RA suscepti- select SNPs polymorphic in the HapMap combined Han bility loci in Caucasian populations have been identified Chinese in Beijing (CHB) and Japanese in Tokyo (JPT) in several recent genetic studies. The TRAF1/C5 locus database (25) across the 400-kb linkage disequilibrium block has been identified both by a genome-wide association that includes PTPN22. In addition, all variants identified fromthe PTPN22 exon resequencing were selected. A total of 34 study and as a candidate gene (7,17). Two independent SNPs/variants were selected from the PTPN22 linkage disequili- followup genome-wide association studies identified a brium block. Last, we identified SNPs that have been reported susceptibility locus on 6q23 near TNFAIP3 (18,19). A as associated with RA in either a Caucasian (TRAF1/C5, 6q23, large meta-analysis using genome-wide association and 4q27, CD40, and CCL21 loci) or Asian (PADI4) population replication collections identified the CD40 and CCL21 and selected 1 disease-associated SNP per region (a total of 6 loci (20), and a 4q27 region including IL2/IL21 has been SNPs were selected). The 40 selected SNPs were used todesign multiplex genotyping assays using Sequenom RealSNP identified through a candidate gene approach after (www.realsnp.com) and eXTEND applications (Sequenom, being shown to be associated with other autoimmune diseases (21). To date, there has been no study in an Genotyping. Multiplex PCR was used to amplify DNA
Asian population of SNPs in these new regions associ- products containing up to 25 SNPs in a single reaction from ated with RA. Therefore, the aim of this study was to 5 ng genomic DNA. Synthetic oligonucleotides that bind determine whether PTPN22 variants or any of 5 disease- adjacent to the SNP site were then hybridized and extendedwith nucleotides complementary to the template SNP site, associated SNPs recently identified in Caucasians also using modified nucleotides that terminate the extension reac- tion at the interrogated SNP and generate alternate productsof mass sufficiently different to be separated by mass spec-trometry. The extended products were separated by matrix- PATIENTS AND METHODS
assisted laser desorption ionizationϪtime-of-flight mass spec- Study population. The study group comprised 1,128
trometry, and the genotypes were determined by SpectroTyper Korean patients with RA who were enrolled consecutively software (Sequenom). Calls were evaluated and edited by from the outpatient clinic of the Hospital for Rheumatic cluster analysis, which was performed with SpectroTyper soft- Diseases, Hanyang University, Seoul, South Korea and 1,022 ethnically matched control subjects for whom no history of RA Statistical analysis. Association tests for SNPs and
or other autoimmune diseases was noted in the self-reported haplotypes in patients and control subjects and the regional questionnaire at the time of enrollment. All patients with RA linkage disequilibrium structure for analyzed SNPs were de- met the American College of Rheumatology (formerly, the termined using Haploview software, version 4.0 (26). We American Rheumatism Association) 1987 revised criteria for excluded SNPs with significant deviation from Hardy- the classification of RA (22). The study was approved by the Weinberg equilibrium (P Ͻ 0.005) and SNPs with a minor institutional review board of Hanyang University Hospital.
allele frequency in controls of Ͻ0.01. We reanalyzed the Written informed consent was obtained from all participants.
association of all SNPs with specific subgroups of RA (CCP All subjects were typed for HLA–DRB1 subtypes, antibody–positive RA, severe erosive RA, or SE-positive RA) using polymerase chain reaction (PCR) and sequence-specific using multivariate logistic regression with adjustments for age oligonucleotide probe hybridization. Because the second most and sex. These subgroup analyses were performed using SAS significant risk allele for RA in the Korean population was software, version 9.1 (SAS Institute, Cary, NC).
DRB1*0901, we regarded DRB1*0901 as a member of theshared epitope (SE) group, in addition to the following alleles: DRB1*0101, *0102, *0401, *0404, *0405, *0408, *0413, *1001,and *1402. Among 1,128 patients with RA, cyclic citrullinated Characteristics of the patients and control sub-
peptide (CCP) antibody data were available for 807 patients.
jects. Among the patients with RA, the mean Ϯ SD age
These data were based on quantitative duplicate measure-ments by enzyme-linked immunosorbent assay, using the DI- was 52.4 Ϯ 11.9 years, the mean Ϯ SD age at disease ASTAT anti-CCP kit (FCCP 200; Axis-Shield, Dundee, Scot- onset was 39.9 Ϯ 11.5 years, 89.7% were female, 81.4% land, UK) according to the manufacturer’s instructions. The were positive for the SE, 17.0% showed no erosive Association tests of SNPs in the PTPN22 region in 1,128 Korean patients with rheumatoid * SNP ϭ single-nucleotide polymorphism; var. ϭ variant.
† R620W variant.
‡ Hardy-Weinberg disequilibrium.
changes (Steinbrocker stage I), and 83.0% demonstrated nlm.nih.gov/SNP/) (the ss numbers are as follows: erosions (Steinbrocker stages II, III, and IV). Among ss103510562, ss103510564, ss103510567, ss103510570, RA patients with CCP antibody data (n ϭ 807), 681 ss103510573, and ss103510576). Although all of the novel (84.4%) were anti-CCP positive. The mean Ϯ SD age of variants were rare (i.e., found in only a single DNA of control subjects was 36.8 Ϯ 12.5 years, and 86.3% were the 48 resequenced), they were included in the case– control genotyping. One of the novel variants was found PTPN22 sequence variants identified by rese-
in the promoter region, 4 were intronic, and 1 was a quencing. Nineteen sequence variants (i.e., variants
identified in 1 or more resequenced DNAs) were iden- Genotype assay results. Of the 40 SNPs included
tified in the exonic (exons and intronic splice junction in the assay design, assays for 2 SNPs failed (Table 1) regions) and the promoter regions of PTPN22 in the (additional information is presented in Supplementary resequenced Korean DNA samples. Of these, 13 had Table 1, available on the Arthritis & Rheumatism Web previously been identified, and 6 were novel. The novel site at http://www3.interscience.wiley.com/journal/ variants were submitted to the Database of Single- 76509746/home). The genotype call rate was Ͼ95% for Nucleotide Polymorphisms (dbSNP; http://www.ncbi.
the remaining markers, with the exception of rs6537798, DIFFERENT GENETIC RISK FACTORS IN CAUCASIAN AND KOREAN PATIENTS WITH RA DIFFERENT GENETIC RISK FACTORS IN CAUCASIAN AND KOREAN PATIENTS WITH RA PTPN22 haplotype association analysis in Korean patients taining the PTPN22 gene was similar in individuals in with rheumatoid arthritis and control subjects* the HapMap Centre d’Etude du Polymorphisme Hu- main Utah residents with ancestry from northern and Western Europe (CEU) and the combined CHB and JPT populations, with a 400-kb region of strong linkagedisequilibrium containing 6 genes (MAGI3, PHTF1, RSBN1, PTPN22, C1orf178, AP4B1) identified in both populations (see Supplementary Figure 1, available on the Arthritis & Rheumatism Web site at http://www3.
* Haplotypes are those identified in Table 2, with the addition of 2 interscience.wiley.com/journal/76509746/home). The 24 single-nucleotide polymorphisms, rs1217412 and rs3761935. The A PTPN22 tag SNPs that passed quality control captured allele of rs1217412 was found only on haplotype 1B, and the G alleleof rs3761935 was found only on haplotype 2. OR ϭ odds ratio; 95% (r2 Ͼ 0.8) 90% of the 185 HapMap variants within the 400-kb region that had a minor allele frequency of Ͼ5%in the CHB ϩ JPT population. Although the linkagedisequilibrium structure of the PTPN22 gene region was which had a 77.7% call rate. No patients or control similar in individuals in the HapMap CEU and CHB ϩ subjects were polymorphic at rs2476601 (R620W), while JPT populations, the frequencies of the common haplo- 1–9 minor alleles (of 4,300 alleles) were present in the 6 types in these 2 populations were very different. Eleven novel PTPN22 variants, the 6q23 SNP rs6920220, and of the 13 tag SNPs used for the haplotype association the 4q27 SNP rs6822844 (Supplementary Table 1 and analysis of the 85-kb PTPN22 gene region (see below) Table 2). Additionally, the PTPN22 tag SNP rs1018592 were genotyped in the HapMap CHB ϩ JPT and CEU was not analyzed, because of significant deviation from samples. These 11 SNPs defined only 4 common haplo- types (those with Ͼ5% frequency) in both populations.
RA association tests. No statistically significant
All 4 common haplotypes found in the Asian samples association (P Ͻ 0.05) with RA was detected between were also common haplotypes in the CEU samples, but any of the 24 polymorphic PTPN22 tag SNPs (Table 1and Supplementary Table 1) from the linkage disequili- the haplotype frequencies were strikingly different be- brium block covering the AP4B1, C1orf178, PTPN22, tween the 2 populations (Table 3). Interestingly, the RSBN1, PTHF1, and MAG13 genes. In addition, Cauca- 620W autoimmune disease–associated variant was found sian RA–associated SNPs at TRAF1/C5, CD40, and only on a portion (approximately half) of the copies of CCL21 were found to be polymorphic in the Korean haplotype 1 in the CEU samples, but was not found on population but not to be associated with disease (Table any of the haplotype 1 copies in the CHB ϩ JPT samples 2). As noted above, the 4q27 and 6q23 variants were too rare in this population to be evaluated for association Haplotype analysis. PTPN22 haplotypes were
(Table 2). The PADI4 SNP rs2240340 was noted to be constructed using the SNPs from Table 1, numbers 6 strongly disease-associated, with a P value of 1.15 ϫ through 25, including only those with a minor allele 10Ϫ8 and an odds ratio of 1.43 (95% confidence interval frequency Ͼ0.05, using Haploview software. These 13 1.26–1.62) for the presence of the disease-associated T SNPs covered the entire 85-kb PTPN22 gene region. The allele on patient chromosomes compared with control 4 common haplotypes (each with a frequency Ͼ5%) chromosomes (Table 2). In the subgroup analyses, there together accounted for 96.6% of the chromosomes of were no significant genotypic or allelic associations of the individuals genotyped. Association tests were per- any of these SNPs in subgroups of patients classified formed for each of these 4 haplotypes, and none were according to radiographic severity (Steinbrocker stage I found to be associated with disease (Table 4).
and stages II–IV), SE status (SE positive and SE nega-tive), and the CCP status (CCP antibody positive and DISCUSSION
CCP antibody negative) compared with control subjects,except for PADI4 SNP rs2240310, which was signifi- Recent whole-genome association studies in Eu- cantly associated with RA irrespective of SE status (P Ͻ ropean ancestry case–control collections have shown 0.01) or CCP antibody positivity (P Ͻ 0.0001).
that after HLA variants, SNPs within the PTPN22 PTPN22 linkage disequilibrium. The linkage dis-
genetic region are consistently the second most signifi- equilibrium structure of the broad genomic region con- cantly disease-associated variants in the genome (3,7).
However, the proposed causative variant, an SNP en- are associated with RA in Koreans. In comparison, we coding a structural variant (R620W) of the PTPN22 genotyped the PADI4 SNP rs2240340, for which we protein, is not present in the Asian population. We previously reported an association in a subset (545 therefore sought to determine whether other PTPN22 patients and 392 control subjects) of this case–control alleles are associated with RA in Asian patients. We also series (9), and found a striking association, with a P evaluated variants in 5 other loci that have strong value for association of 1.15 ϫ 10Ϫ8, in this larger evidence for association in Caucasians with RA in a large Korean RA case–control series.
These data provide convincing evidence that the SNP variants within the PTPN22 linkage disequili- genetic risk factors for RA differ substantially between brium block were evaluated using a tag SNP approach to Asian and Caucasian populations. Although the HLA interrogate the known common variants in the region.
region has been clearly demonstrated to be the most None of the 24 polymorphic SNPs covering the genomic substantial genetic risk factor in populations of both region including PTPN22 were associated with suscepti- ancestries, the other genetic contributions seem to be bility to RA in this Korean population. Of note, we have very different. PTPN22 consistently has the second confirmed that the R620W variant is nonpolymorphic in strongest genetic effect in whole-genome association this population and have shown that a promoter SNP, studies in Caucasians, yet our data indicate that not only rs2488457, which has been reported to be associated is the R620W variant not present in Koreans, but also with RA in Caucasians and with type 1 diabetes mellitus that no other common polymorphisms in the region are in Japanese (27), had no evidence for association in significant risk alleles in Koreans. The other loci associ- Korean patients with RA. The genotyped SNPs cover ated with RA in Caucasians for which we assayed SNPs 90% of the known polymorphic SNPs in the 400-kb (TRAF1/C5, 6q23, 4q27, CD40, CCL21) were derived PTPN22 linkage disequilibrium block (r2 Ͼ 0.8), and from rigorously interpreted and well-replicated studies none were associated with disease. Furthermore, none and probably represent many of the next tier of suscep- of the 4 common PTPN22 haplotypes were associated tibility loci in populations of European ancestry. The with disease. Coupled with the fact that no common fact that none of these SNPs demonstrated association PTPN22 structural variants were identified by rese- in this large Korean study reveals a large degree of quencing, these data provide strong evidence that com- genetic heterogeneity of RA across continental ancestry mon PTPN22-region genetic variants are not associated differences, especially when juxtaposed with the PADI4 with RA in this population. This is in stark contrast to locus, which has been strongly associated with RA in this the very strong association signals seen in multiple and other Asian populations but which has shown little case–control studies in Caucasians with RA.
evidence for association in Caucasians (14).
Although none of the common PTPN22 variants In conclusion, we observed no association be- were associated with disease, the novel rare variants are tween RA in Koreans and tag SNPs covering the entire more difficult to evaluate. Interestingly, 3 of the novel PTPN22 region or with individual SNPs shown to be variants identified (1 promoter and 2 intronic) were strongly associated with disease in Caucasians at the found in a single patient with RA but in no control sub- TRAF1/C5, 6q23, 4q27, CD40, and CCL21 loci. When jects. Further work will be required to determine whether contrasted with other loci such as PADI4, FCRL3, and any of the rare variants of PTPN22 might be private SLC22A4, which have been shown to be associated with mutations that contribute to disease susceptibility.
RA in Asian but not Caucasian populations, these data This study also demonstrated that SNPs in 5 indicate that most of the disease-associated variants additional regions (TRAF1/C5, 6q23, 4q27, CD40, and have arisen independently in different continental pop- CCL21) that are strongly associated with RA in Cauca- ulations, and that different points in disease-associated sians are either nonpolymorphic (6q23 and 4q27) or are pathways are influenced by genetic risk factors in the not associated with RA in Korean patients (TRAF1/C5, different populations. Therefore, one would expect that CD40, and CCL21). For these SNPs, we can say with whole-genome association studies in Asian populations certainty only that the SNP associated with disease in would reveal more and different susceptibility loci than Caucasians is not associated with disease in Koreans.
those that have been described in Caucasian patients Further investigation and fine mapping similar to what with RA. The identification of such loci should expand we have done with PTPN22 would be needed to deter- our understanding of the processes that lead to the mine whether other polymorphic SNPs in these regions DIFFERENT GENETIC RISK FACTORS IN CAUCASIAN AND KOREAN PATIENTS WITH RA AUTHOR CONTRIBUTIONS
America and Sweden: association of susceptibility with PTPN22,CTLA4, and PADI4. Am J Hum Genet 2005;77:1044–60.
Dr. Remmers had full access to all of the data in the study and 12. Suzuki A, Yamada R, Chang X, Tokuhiro S, Sawada T, Suzuki M, takes responsibility for the integrity of the data and the accuracy of the et al. Functional haplotypes of PADI4, encoding citrullinating enzyme peptidylarginine deiminase 4, are associated with rheuma- Study design. Lee, Korman, Kastner, Remmers, Gregersen, Bae.
toid arthritis. Nat Genet 2003;34:395–402.
Acquisition of data. Lee, Korman, Le, Kastner, Remmers, Gregersen,
13. Tokuhiro S, Yamada R, Chang X, Suzuki A, Kochi Y, Sawada T, et al. An intronic SNP in a RUNX1 binding site of SLC22A4, Analysis and interpretation of data. Lee, Korman, Le, Kastner,
encoding an organic cation transporter, is associated with rheu- matoid arthritis. Nat Genet 2003;35:341–8.
Manuscript preparation. Lee, Korman, Kastner, Remmers, Gregersen,
14. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG. PADI4 polymor- phisms and rheumatoid arthritis susceptibility: a meta-analysis.
Statistical analysis. Lee, Korman, Remmers.
15. Lee HS, Remmers EF, Le JM, Kastner DL, Bae SC, Gregersen PK. Association of STAT4 with rheumatoid arthritis in the Korean REFERENCES
population. Mol Med 2007;13:455–60.
16. Remmers EF, Plenge RM, Lee AT, Graham RR, Hom G, Behrens 1. Jun KR, Choi SE, Cha CH, Oh HB, Heo YS, Ahn HY, et al.
TW, et al. STAT4 and the risk of rheumatoid arthritis and systemic Meta-analysis of the association between HLA-DRB1 allele and lupus erythematosus. N Engl J Med 2007;357:977–86.
rheumatoid arthritis susceptibility in Asian populations. J Korean 17. Kurreeman FA, Padyukov L, Marques RB, Schrodi SJ, Seddighza- deh M, Stoeken-Rijsbergen G, et al. A candidate gene approach 2. Silman AJ, Pearson JE. Epidemiology and genetics of rheumatoid identifies the TRAF1/C5 region as a risk factor for rheumatoid arthritis. Arthritis Res 2002;4 Suppl 3:S265–72.
3. Wellcome Trust Case Control Consortium. Genome-wide associ- 18. Plenge RM, Cotsapas C, Davies L, Price AL, de Bakker PI, Maller ation study of 14,000 cases of seven common diseases and 3,000 J, et al. Two independent alleles at 6q23 associated with risk of shared controls. Nature 2007;447:661–78.
rheumatoid arthritis. Nat Genet 2007;39:1477–82.
4. Begovich AB, Carlton VE, Honigberg LA, Schrodi SJ, Chokkal- 19. Thomson W, Barton A, Ke X, Eyre S, Hinks A, Bowes J, et al.
ingam AP, Alexander HC, et al. A missense single-nucleotide Rheumatoid arthritis association at 6q23. Nat Genet 2007;39: polymorphism in a gene encoding a protein tyrosine phosphatase (PTPN22) is associated with rheumatoid arthritis. Am J Hum 20. Raychudhuri S, Remmers EF, Lee AT, Hackett R, Guiducci C, Burtt NP, et al. Common variants at CD40 and other loci confer 5. Gregersen PK, Lee HS, Batliwalla F, Begovich AB. PTPN22: risk of rheumatoid arthritis. Nat Genet 2008;40:1216–23.
setting thresholds for autoimmunity. Semin Immunol 2006;18: 21. Zhernakova A, Alizadeh BZ, Bevova M, van Leeuwen MA, Coenen MJ, Franke B, et al. Novel association in chromosome4q27 region with rheumatoid arthritis and confirmation of type 1 6. Mori M, Yamada R, Kobayashi K, Kawaida R, Yamamoto K.
diabetes point to a general risk locus for autoimmune diseases.
Ethnic differences in allele frequency of autoimmune-disease- associated SNPs. J Hum Genet 2005;50:264–6.
22. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, 7. Plenge RM, Seielstad M, Padyukov L, Lee AT, Remmers EF, Cooper NS, et al. The American Rheumatism Association 1987 Ding B, et al. TRAF1-C5 as a risk locus for rheumatoid arthritis: revised criteria for the classification of rheumatoid arthritis.
a genomewide study. N Engl J Med 2007;357:1199–209.
8. Hu X, Chang M, Saiki RK, Cargill MA, Begovich AB, Ardlie KG, 23. Steinbrocker O, Traeger CH, Batterman RC. Therapeutic criteria et al. The functional 169T3 C single-nucleotide polymorphism in in rheumatoid arthritis. JAMA 1949;140:659–62.
FCRL3 is not associated with rheumatoid arthritis in white North 24. De Bakker PI, Yelensky R, Pe’er I, Gabriel SB, Daly MJ, Altshuler Americans. Arthritis Rheum 2006;54:1022–4.
D. Efficiency and power in genetic association studies. Nat Genet 9. Kang CP, Lee HS, Ju H, Cho H, Kang C, Bae SC. A functional haplotype of the PADI4 gene associated with increased rheuma- 25. The International HapMap Consortium. The International Hap- toid arthritis susceptibility in Koreans. Arthritis Rheum 2006;54: Map Project. Nature 2003;426:789–96.
26. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and 10. Kochi Y, Yamada R, Suzuki A, Harley JB, Shirasawa S, Sawada T, visualization of LD and haplotype maps. Bioinformatics 2005;21: et al. A functional variant in FCRL3, encoding Fc receptor-like 3, is associated with rheumatoid arthritis and several autoimmunities.
27. Kawasaki E, Awata T, Ikegami H, Kobayashi T, Maruyama T, Nakanishi K, et al. Systematic search for single nucleotide poly- 11. Plenge RM, Padyukov L, Remmers EF, Purcell S, Lee AT, morphisms in a lymphoid tyrosine phosphatase gene (PTPN22): Karlson EW, et al. Replication of putative candidate-gene associ- association between a promoter polymorphism and type 1 diabetes ations with rheumatoid arthritis in Ͼ4,000 samples from North in Asian populations. Am J Med Genet A 2006;140:586–93.

Source: http://www.naracstudy.org/NaracStudy/documents/Lee%20Korean%20Genetics%20RA%20ARFeb%202009.pdf