- Email: [email protected]
Association of PTPN22 haplotypes with type 1 diabetes in the Japanese population
Human Immunology 71 (2010) 795–798
Contents lists available at ScienceDirect
Association of PTPN22 haplotypes with type 1 diabetes in the Japanese population Matsuo Taniyama a, Taro Maruyama b, Teruaki Tozaki c, Yasuko Nakano c, Yoshiyuki Ban d,* a
Division of Endocrinology and Metabolism, Department of Internal Medicine, Showa University, Fujigaoka Hospital, Yokohama, Japan Department of Medicine, Saitama Social Insurance Hospital, Saitama, Japan c Department of Pharmacogenomics, Showa University School of Pharmacy, Tokyo, Japan d Division of Diabetes, Metabolism and Endocrinology, Department of Internal Medicine, Showa University School of Medicine, Tokyo, Japan b
A R T I C L E
I N F O
Article history: Received 21 January 2010 Accepted 19 May 2010 Available online 25 May 2010
Keywords: Type 1 diabetes PTPN22 Haplotype Genetic susceptibility
A B S T R A C T
The R620W polymorphism in the protein-tyrosine-phosphatase nonreceptor type 22 gene (PTPN22) confers susceptibility to type 1 diabetes (T1D) and other autoimmune diseases. This polymorphism is reportedly nonpolymorphic in the Asian population. Additional polymorphisms and specific haplotypes have also been associated with T1D, rheumatoid arthritis (RA) and Graves’ disease in Caucasians. We examined whether PTPN22 single nucleotide polymorphisms (SNPs) other than R620W and haplotypes are associated with T1D in the Japanese population. We compared the allele frequencies of five haplotype-tagging SNPs in the PTPN22 gene, 2 of which are reportedly associated with RA in Caucasians (rs3789604 and rs1310182), and compared haplotype distributions between 184 Japanese T1D patients and 179 healthy controls. rs3789604 was not associated with T1D in our Japanese subjects. The frequency of the C allele of rs1310182 differed significantly between T1D patients and controls. Permutation analysis revealed the distribution of this haplotype to differ significantly between T1D patients and controls. One rare haplotype that included the susceptibility allele of rs1310182 was more frequent, while another rare haplotype that included the protective allele of rs1310182 was absent, in T1D patients. This significant haplotype distribution difference suggests that polymorphisms in the PTPN22 gene other than R620W are involved in either predisposition to or protection from T1D in the Japanese population. 䉷 2010 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics.
1. Introduction Type 1 diabetes (T1D) is a multifactorial disease caused by T-cell–mediated destruction of pancreatic  cells [1]. Both genetic predisposition and environmental factors are known to play important roles in the development of this disease. Until recently, only six gene regions were commonly accepted as being associated with T1D: human leukocyte antigen (HLA) class II, insulin (INS), cytotoxic T-lymphocyte antigen-4 (CTLA4), protein-tyrosine-phosphatase nonreceptor type 22 (PTPN22), interleukin-2 receptor-␣ (IL-2RA), and interferon-induced helicase-1(IFIH1) [2–5]. The pace of discovery of genetic associations with T1D has now accelerated. Owing to large-scale genome-wide association studies (GWAS), as well as candidate gene studies, the total number of confirmed loci, including the MHC region, is now 43 [6]. However, much remains to be elucidated regarding these associations and the established T1D loci. Studies with larger sample numbers and more markers are needed. Recently, 257 single-nucleotide polymorphisms (SNPs) were genotyped in 19 candidate genes, including INS, CTLA4,
* Corresponding author. E-mail address: [email protected] (Y. Ban).
PTPN22, IL-2RA, and IFIH1, in 2300 affected sib-pair (ASP) families and were tested for associations with T1D as part of the Type I Diabetes Genetics Consortium’s candidate gene study [7]. The data from these families provided no consistent evidence for associations of the other 14 candidate genes tested with T1D [7]. The possibility remains that loci such as the small ubiquitin-related modifiers– 4 gene (SUMO4) could be associated with T1D in non-European populations, which might provide an example of population-specific genetic regulation of T1D [8]. Very recently, all GWAS data published to date for seven common types of autoimmune disease (AID) were compiled, allowing a network-based analysis of shared susceptibility genes at different levels of significance [9]. Although involvement of the MHC region in chromosome 6p21 is not in question for most types of AID, the complex genetic architecture of this locus poses a significant analytical challenge [9]. Statistically significant excess sharing of nonMHC genes was found between T1D and all other AIDs studied, a result also seen for rheumatoid arthritis (RA) [9]. A smaller but significant degree of sharing was observed for multiple sclerosis, celiac disease, and Crohn’s disease [9]. The availability of GWAS data allows systematic analysis of similarities and differences among several types of AID [9].
0198-8859/10/$32.00 - see front matter 䉷 2010 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics. doi:10.1016/j.humimm.2010.05.016
796
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
The PTPN22 gene encodes lymphoid-specific protein-tyrosinephosphatase (Lyt), and was thereby identified as the fourth genetic locus associated with T1D. The R620W polymorphism (1858T variant, dbSNP no. rs2476601) in PTPN22, is reportedly associated with T1D in Caucasians [10]. This polymorphism also shows associations with other types of AID, such as autoimmune thyroid disease, RA, and systemic lupus erythematosus (SLE). The PTPN22 locus is thus regarded as a general autoimmunity locus [11]. In Asian populations, the R620W polymorphism was originally reported to be nonpolymorphic [12,13]. Recently, Kawasaki et al. described heterozygosity (molecular heterosis) of the polymorphism rs2488457 in the promoter region of the PTPN22 gene (⫺1123 G/C) as being associated with acute-onset T1D in the Japanese population [14]. In Caucasians, this promoter polymorphism is strongly linked to R620W, and several studies have revealed the association of ⫺1123 G/C and AID to be a result of linkage to R620W, and that the real susceptibility allele is 1858T (620W) [15,16]. Carlton et al. identified 37 SNPs in the PTPN22 gene and reported that several, including rs1310182 (designated as SNP27) and rs3789604 (designated as SNP37), were also associated with RA independently of R620W in the Caucasian population [17]. They also studied the haplotypes and found that some not containing 620 W (1858T) were associated with RA. However, these SNPs were not associated with Japanese RA [13]. Because the PTPN22 gene is associated with several types of AID [9], the existence of disease susceptibility alleles other than 1858T in the Japanese population seems likely. In the present study, we examined whether the SNPs associated with RA in Caucasians were also associated with T1D in the Japanese population. We also analyzed haplotypes and found that the PTPN22 gene plays a role in either susceptibility to or the protection from T1D in the Japanese population.
labeled probe methods, and a LightScanner (Idaho Technology, Salt Lake City, UT) according to the manufacturer’s instructions. Details of the primer sequences used for typing the five SNPs are available in Table 1.
2. Subjects and methods
Because of the strong LD among the five variants, haplotype analysis was undertaken using SNPAlyze. We identified seven haplotypes, four of which correlated with haplotypes 1, 3, 4 and 5, as previously reported by Carlton et al. (Table 3). Four haplotypes (1– 4) were found to be relatively common, whereas the remaining three haplotypes (5–7) were rare. In addition, haplotype distributions differed significantly between T1D and control subjects according to the permutation procedure (p ⫽ 0.0001). Haplotype 7, which contained the rs1310182 risk allele, was found to be positively associated with T1D (p ⫽ 0.0013). In contrast, haplotype 5, which did not contain the rs1310182 risk allele, was found to be protective (permutation p ⬍ 0.0001).
2.1. Subjects A total of 184 Japanese patients (67 male and 117 female) with T1D were included in this study. All study participants lived in the Tokyo metropolitan area. T1D was diagnosed on the basis of sudden onset of severe symptoms or rapid progression to overt diabetes and dependence on exogenous insulin because of absolute insulin deficiency. The T1D diagnosis was based on the 1997 Committee of the American Diabetes Association criteria [18]. All subjects were insulin dependent at the time of the study, although their -cell reserves, estimated by fasting serum C-peptide measures, varied. Patients with slowly progressive T1D were not included in this study. The median patient age at diagnosis was 28.7 years. A total of 179 age-matched healthy female subjects exhibiting normal glucose tolerance based on normal fasting plasma glucose and HbA1c levels, with no family history of T1D or any other types of AID, served as controls. All patients and controls were informed of the purpose of the study, and their consent was obtained before starting the experiments. The study was preapproved by the ethics committee of each participating institution. 2.2. Genotyping of SNPs DNA was extracted from whole blood using a Puregene kit (Gentra Systems, Minneapolis, MN) and the following five SNPs in the PTPN22 gene were analyzed: rs12760457 (SNP18) in intron 11, rs2797415 (SNP25) in intron 15, rs1310182 (SNP27) in intron 16, rs2476599 (SNP33) in intron 19 and rs3789604 (SNP37) in putative transcription factor-binding sites (TRBS) downstream from the PTPN22 gene. rs1310182 and rs3789604 were analyzed using the SNAPshot system (PE Applied Biosystems, Foster City, CA) based on the manufacturer’s protocol. rs12760457, rs2797415, and rs2476599 were genotyped using high-resolution melting and un-
2.3. Statistical analysis Case-control analysis and testing for Hardy–Weinberg equilibrium of the SNPs were performed using SNPAlyze version 7.0 (Dynacom, Yokohama, Japan) [19]. Hardy–Weinberg equilibrium was tested separately for all loci in subjects and controls. These tests in the subject and control groups showed no significant deviations from Hardy–Weinberg equilibrium for any of the SNPs. Linkage disequilibrium (LD) between SNPs was evaluated by D= and r2 of pairwise LD calculated using the SNPAlyze computer program. Haplotype frequencies for multiple loci were estimated using phase estimation and the expectation-maximization algorithm. Permutation p values were calculated by comparing haplotype frequencies between patients and controls based on 10,000 replications using the SNPAlyze program. 3. Results 3.1. Genotyping analysis Frequency analysis revealed that rs3789604 (SNP37) was not associated with T1D. Among the five SNPs studied, only rs1310182 (SNP27) differed in frequency between T1D patients and controls (Table 2). However, the risk allele in our Japanese subjects was found to be C, whereas that in Caucasian RA subjects was found to be T [17]. We also examined rs2488457 (⫺1123 G/C) in our population, but found no association between T1D and this SNP (data not shown). 3.2. Haplotype analysis
4. Discussion T1D incidences vary considerably among ethnic groups. Scandinavian individuals have an incidence as high as 35 in 100,000 per year, whereas Japanese individuals have a very low incidence (1.7 in 100,000 per year) [1]. The low incidence of T1D in the Japanese population is probably attributable to the difference in genetic predisposition to T1D, with Asians having a low prevalence of susceptibility haplotypes, whereas these genetic factors are relatively common in Europeans. The HLA class II loci are known to be strongly associated with T1D in all ethnic groups [3], and the haplotypes DRB1*0301-DQA1*0501-DQB1*0201, and DRB1*0401 DQA1*0301-DQB1*0302 are strongly associated with T1D in Caucasians. Genotypes heterogeneous for the DRB1*0401/DRB1*0301 alleles represent an extreme risk for T1D development [20]. These two alleles do not exist in the Japanese population, whereas DRB1*0405 and DRB1*0901 have been identified as risk variants. These differences in HLA genotypes may explain, in part, the low incidence of T1D in the Japanese population [21]. In recent years, the PTPN22 gene has emerged as an important genetic locus for several AIDs, including T1D [2]. The risk variant
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
797
Table 1 Primers used for typing the five SNPs dbSNP ID
Forward primer
Reverse primer
Interrogator primer or 3=-amino–linked probes
rs12760457 rs2797415 rs1310182 rs2476599 rs3789604
TATCACTGTTGCTCAGACTGG GCTGAGCAACATTTGATGTC ATGGACATATTTTCCCATGATG TCAGGTACCTCCTCGAAAAG CAGCATGAGGTTTCTCCTTC
CAGTAACTGAAGCAAAATAGATG GGATTTGGGAACATTTGGTG AAATGCCTACTGTATGCCAG CACTGTGTGAACTGCATATGC CAATAGAAGCTCGGTTGGAG
TTTCCGTACCCTCTATTCTGC ATACTATTCCTAAGAACCGAAC GATCGATCGATCGATCGTACAAGAGGTTCAC CCTAGTATTACGTATGCATGTGC CCACAAACACACATTTAAATGGCCCGACCGCCCCCCCTCCGCGCA
SNP, single nucleotide polymorphism.
PTPN22 620W (1858T) has been validated as a gain-of-function variant and demonstrates increased catalytic activity of Lyp [22]. The 1858T variant is virtually absent in Japanese individuals [12,13], which may largely explain the low incidence of T1D in this population [21]. In the present study, we focused on whether the proposed risk variants of PTPN22 for Caucasian RA patients are also risk variants for T1D in the Japanese population, independently of 1858T. The marker rs3789604, which was originally designated SNP37 in the Carlton et al. report [17], was not found to be associated with T1D in our Japanese cohort. The contrasting results of these two studies may be attributable to a difference in disease ontology or ethnic differences. The marker rs3811021, which is in almost absolute LD with rs3789604, was not associated with T1D in Caucasians [23,24]. In contrast, Ichimura et al. reported that rs3789604 was associated with Graves’ disease in the Japanese population, although the risk allele differed from that associated with RA in Caucasians [26]. Thus, rs3789604C does not appear to be a consistent risk variant for AID. In the present study, the frequency of rs1310182 (SNP27) differed significantly between patients and controls. In the study by Carlton et al., the association between rs1310182 and RA was attributed to linkage to rs3789604. Furthermore, in the present study, the risk allele differed from that identified by Carlton et al. When the haplotype data described below are considered, it seems unlikely that the rs1310182 variant itself is responsible for T1D risk in the Japanese population. In fact, allele T of rs1310182 was not associated with T1D in a Sardinian population [24]. With the exception of the haplotype that contains 1858T, the associations of PTPN22 haplotypes with AID have not been consistently demonstrated by previous studies. The haplotype that contains rs3789604C was associated with Caucasian RA patients in one study [17]; however, this association was not seen in RA [27] or Graves’ disease [25] examined in other studies of Caucasian populations. Onengut-Gumuscu et al. reported that two haplotypes, one containing rs3789604C, were over-transmitted to affected offspring when haplotype association tests were conducted on T1D patients homozygous for the C1858 nonrisk allele [23]. In addition,
Carlton et al. reported that one of these haplotypes confers protection against RA [17]. This haplotype was found to be transmitted to a lesser degree in a family-based association analysis of T1D [23]. In contrast, the haplotype termed Carlton’s haplotype 3 appeared to provide protection against Graves’ disease in Caucasians [25]. Although Japanese do not have the 1858T variant, the haplotype distribution differed significantly between T1D patients and controls, as evidenced by the permutation procedure used in this study. One rare haplotype, termed haplotype 7, was found to be significantly more frequent in T1D, whereas another rare haplotype (haplotype 5) was absent in T1D. These two haplotypes were not described in previous reports on Caucasian populations. Haplotype 7 contains the rs1310182 risk allele, whereas haplotype 5 does not. By contrast, haplotypes 3 and 4 containing the risk allele of rs1310182 did not appear to confer susceptibility to T1D. In addition, haplotypes 1, 2, and 6 do not contain the rs1310182 risk allele, and they did not appear to be protective. These results support the hypothesis that the rs1310182 variant itself is not responsible for susceptibility to or protection from T1D, and that the frequency differences in the above two haplotypes may partially explain the significant differences in the frequency of the rs1310182 allele. As both haplotypes described as being associated with T1D are rare, the association may not be clinically relevant. Our results suggest an association of the C allele of rs1310182 with T1D in our population. A haplotype analysis, including the five SNPs showed over-representation and under-representation, respectively, of two rare haplotypes in Japanese T1D patients. When combined with the results of previous studies, our data indicate the associations between haplotypes of the PTPN22 gene and AID to be complicated. Identification of precise risk or protective alleles in the region of the PTPN22 gene linked to the associated haplotypes requires further investigation. Recently, a loss-of-function mutation in the PTPN22 gene that does not exist in Asian populations was shown to confer protection against human SLE [28]. Finally, it has also been proposed that significant LD may extend for several hundred kilobases surrounding PTPN22, and thus potentially encompass additional genes, including RSBN1 and DCLRE1B [23].
Table 2 Genotype and allele frequencies of five SNPs in T1D patients and controls db SNP ID
Allele
T1D
(%)
Control
(%)
p Value
rs12760457
C T
339 29
92.1 7.9
335 23
93.6 6.4
NS
T C
208 160
56.5 43.5
218 140
60.9 39.1
T C
272 96
73.9 26.1
291 67
81.3 18.7
G A
306 62
83.1 16.8
315 43
88.0 12.0
A C
287 81
78.0 22.0
295 63
82.4 17.6
rs2797415
rs1310182
rs2476599
rs3789604
SNP, single nucleotide polymorphism; T1D, type 1 diabetes; NS, not significant.
NS
0.017
NS
NS
Genotype
T1D
(%)
Control
(%)
C/C C/T T/T T/T T/C C/C T/T T/C C/C G/G G/A A/A A/A A/C C/C
155 29 0 65 78 41 103 66 15 129 48 7 115 57 12
84.2 15.8 0.0 35.3 42.4 22.3 56.0 35.9 8.2 70.1 26.1 3.8 62.5 31.0 6.5
157 21 1 71 76 32 120 51 8 138 39 2 124 47 8
87.7 11.7 0.6 40.0 42.5 17.9 67.0 28.5 4.5 77.1 21.8 1.1 69.3 26.3 4.5
p Value
NS
NS
NS
NS
NS
798
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
Table 3 Frequencies of five SNP haplotypes Haplotype no.
SNP 18-25-27-33-37
DM
Control
1 2 3 4 5 6 7
C-T-T-G-A C-C-T-G-C C-C-C-A-A T-C-C-G-A C-C-T-G-A C-T-T-G-C C-C-C-G-A
0.536 0.195 0.139 0.071 0.000 0.021 0.039
0.592 0.170 0.113 0.061 0.050 0.012 0.003
Permutation p value
⬍0.0001 0.0013
SNP numbers are those designated by Carlton et al. [11]. 18: rs12760457, 25: rs2797415, 27: rs1310182, 33: rs2476599, 37: rs3789604.
Therefore, the possibility that a gene other than PTPN22 is responsible for the association with T1D cannot be ruled out. Acknowledgments This work was supported in part by a Showa University Grantin-aid for Innovative Collaborative Research Projects (to Y.B.), the Showa University Medical Foundation (to Y.B.), a grant from the Showa University School of Medicine Alumni Association (to Y.B.), and a grant from the Yamaguchi Endocrine Research Association (to Y.B.). References [1] Notokins L, Lernmark Å. Autoimmune type 1 diabetes: Resolved and unresolved. J Clin Invest 2001;108:1247–52. [2] Pugliese A, Eisenbarth GS. Type 1 diabetes mellitus of man: genetic susceptibility and resistance. Adv Exp Med Biol 2004;552:170 –203. [3] She JX. Susceptibility to type 1 diabetes: HLA-DQ and DR revisited. Immunol Today 1996;17:323–9. [4] Wellcome Trust. Case Control Consortium Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 2007;447:661–78. [5] Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K, Plagnol V, et al. Robust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 2007;39:857– 64. [6] T1D Base. Available at: http://www.T1DBase.org. Last accessed 2010. [7] Howson JMM, Walker NM, Smyth DJ, Todd JA, and the Type I Diabetes Genetics Consortium. Analysis of 19 genes for association with type I diabetes in the Type I Diabetes Genetics Consortium families. Genes Immun 2009;10:S74 – 84. [8] Podolsky R, Prasad Linga-Reddy MV, She J-X, and the Type I Diabetes Genetics Consortium. Analyses of multiple single-nucleotide polymorphisms in the SUMO4/IDDM5 region in affected sib-pair families with type I diabetes. Genes Immun 2009;10:S16 –20. [9] Baranzini SE. The genetics of autoimmune diseases: A networked perspective. Curr Opin Immunol 2009;21:596 – 605.
[10] Bottini N, Musumeci L, Alonso A, Rahmouni S, Nika K, Rostamkhani M, et al. A functional variant of lymphoid tyrosine phosphatase is associated with type I diabetes. Nat Genet 2004;36:337– 8. [11] Bottini N, Vang T, Cucca F, Mustelin T. Role of PTPN22 in type 1 diabetes and other autoimmune diseases. Semin Immunol 2006;18:207–13. [12] Ban Y, Tozaki T, Taniyama M, Tomita M, Ban Y. The codon 620 single nucleotide polymorphism of the protein tyrosine phosphatase-22 gene does not contribute to autoimmune thyroid disease susceptibility in the Japanese. Thyroidology 2005;15:1115– 8. [13] Ikari K, Momohara S, Inoue E, Tomatsu T, Hara M, Yamanaka H, Kamatani N. Haplotype analysis revealed no association between the PTPN22 gene and RA in a Japanese population. Rheumatology 2006;45:1345– 8. [14] Kawasaki E, Awata T, Ikegami H, Kobayashi T, Maruyama T, Nakanishi K, et al. Systematic search for single nucleotide polymorphisms in a lymphoid tyrosine phosphatase gene (PTPN22): Association between a promoter polymorphism and type 1 diabetes in Asian populations. Am J Med Genet 2006;140:586 –93. [15] Cinek O, Hradsky O, Ahmedov G, Slavcev A, Kolouskova S, Kulich M, Sumnik Z. No independent role of the -1123 G⬎C and⫹2740 A⬎G variants in the association of PTPN22 with type 1 diabetes and juvenile idiopathic arthritis in two Caucasian populations. Diabetes Res Clin Pract 2007;76:297–303. [16] Viken MK, Olsson M, FlÇm ST, Førre O, Kvien TK, Thorsby E, Lie BA. The PTPN22 promoter polymorphism -1123G⬎C association cannot be distinguished from the 1858C⬎T association in a Norwegian rheumatoid arthritis material. Tissue Antigens 2007;70:190 –7. [17] Carlton VE, Hu X, Chokkalingam AP, Schrodi SJ, Brandon R, Alexander HC, et al. PTPN22 genetic variation: Evidence for multiple variants associated with rheumatoid arthritis. Am J Hum Genet 2005;77:567– 81. [18] Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997. [19] SNPAlyze. Available at: http://www.dynacom.co.jp/e/products/package/ snpalyze/. Last accessed 2010. [20] Baschal EE, Eisenbarth GS. Extreme genetic risk for type 1A diabetes in the post-genome era. J Autoimmun 2008;31:1– 6. [21] Ikegami H, Kawabata Y, Noso S, Fujisawa T, Ogihara T. Genetics of type 1 diabetes in Asian and Caucasian populations. Diabetes Res Clin Pract 2007; 77(Suppl 1):S116 –21. [22] Vang T, Congia M, Macis MD, Musumeci L, OrrÛ V, Zavattari P, et al. Autoimmuneassociated lymphoid tyrosine phosphatase is a gain-of-function variant. Nat Genet 2005;37:1317–19. [23] Onengut-Gumuscu S, Buckner JH, Concannon P. A haplotype-based analysis of the PTPN22 locus in type 1 diabetes. Diabetes 2006;55:2883–9. [24] Zoledziewska M, Perra C, OrrÚ V, Moi L, Frongia P, Congia M, et al. Further evidence of a primary, causal association of the PTPN22 620W variant with type 1 diabetes. Diabetes 2008;57:229 –34. [25] Heward JM, Brand OJ, Barrett JC, Carr-Smith JD, Franklyn JA, Gough SC. Association of PTPN22 haplotypes with Graves’ disease. J Clin Endocrinol Metab 2007;92:685–90. [26] Ichimura M, Kaku H, Fukutani T, Koga H, Mukai T, Miyake I, et al. Associations of protein tyrosine phosphatase nonreceptor 22 (PTPN22) gene polymorphisms with susceptibility to Graves’ disease in a Japanese population. Thyroidology:18:625–30. [27] Hinks A, Eyre S, Barton A, Thomson W, Worthington J. Investigation of genetic variation across the protein tyrosine phosphatase gene in patients with rheumatoid arthritis in the UK. Ann Rheum Dis 2007;66:683– 6. [28] OrrÛ V, Tsai SJ, Rueda B, Fiorillo E, Stanford SM, Dasgupta J, et al. A loss-offunction variant of PTPN22 is associated with reduced risk of systemic lupus erythematosus. Hum Mol Genet 2009;18:569 –79.
Contents lists available at ScienceDirect
Association of PTPN22 haplotypes with type 1 diabetes in the Japanese population Matsuo Taniyama a, Taro Maruyama b, Teruaki Tozaki c, Yasuko Nakano c, Yoshiyuki Ban d,* a
Division of Endocrinology and Metabolism, Department of Internal Medicine, Showa University, Fujigaoka Hospital, Yokohama, Japan Department of Medicine, Saitama Social Insurance Hospital, Saitama, Japan c Department of Pharmacogenomics, Showa University School of Pharmacy, Tokyo, Japan d Division of Diabetes, Metabolism and Endocrinology, Department of Internal Medicine, Showa University School of Medicine, Tokyo, Japan b
A R T I C L E
I N F O
Article history: Received 21 January 2010 Accepted 19 May 2010 Available online 25 May 2010
Keywords: Type 1 diabetes PTPN22 Haplotype Genetic susceptibility
A B S T R A C T
The R620W polymorphism in the protein-tyrosine-phosphatase nonreceptor type 22 gene (PTPN22) confers susceptibility to type 1 diabetes (T1D) and other autoimmune diseases. This polymorphism is reportedly nonpolymorphic in the Asian population. Additional polymorphisms and specific haplotypes have also been associated with T1D, rheumatoid arthritis (RA) and Graves’ disease in Caucasians. We examined whether PTPN22 single nucleotide polymorphisms (SNPs) other than R620W and haplotypes are associated with T1D in the Japanese population. We compared the allele frequencies of five haplotype-tagging SNPs in the PTPN22 gene, 2 of which are reportedly associated with RA in Caucasians (rs3789604 and rs1310182), and compared haplotype distributions between 184 Japanese T1D patients and 179 healthy controls. rs3789604 was not associated with T1D in our Japanese subjects. The frequency of the C allele of rs1310182 differed significantly between T1D patients and controls. Permutation analysis revealed the distribution of this haplotype to differ significantly between T1D patients and controls. One rare haplotype that included the susceptibility allele of rs1310182 was more frequent, while another rare haplotype that included the protective allele of rs1310182 was absent, in T1D patients. This significant haplotype distribution difference suggests that polymorphisms in the PTPN22 gene other than R620W are involved in either predisposition to or protection from T1D in the Japanese population. 䉷 2010 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics.
1. Introduction Type 1 diabetes (T1D) is a multifactorial disease caused by T-cell–mediated destruction of pancreatic  cells [1]. Both genetic predisposition and environmental factors are known to play important roles in the development of this disease. Until recently, only six gene regions were commonly accepted as being associated with T1D: human leukocyte antigen (HLA) class II, insulin (INS), cytotoxic T-lymphocyte antigen-4 (CTLA4), protein-tyrosine-phosphatase nonreceptor type 22 (PTPN22), interleukin-2 receptor-␣ (IL-2RA), and interferon-induced helicase-1(IFIH1) [2–5]. The pace of discovery of genetic associations with T1D has now accelerated. Owing to large-scale genome-wide association studies (GWAS), as well as candidate gene studies, the total number of confirmed loci, including the MHC region, is now 43 [6]. However, much remains to be elucidated regarding these associations and the established T1D loci. Studies with larger sample numbers and more markers are needed. Recently, 257 single-nucleotide polymorphisms (SNPs) were genotyped in 19 candidate genes, including INS, CTLA4,
* Corresponding author. E-mail address: [email protected] (Y. Ban).
PTPN22, IL-2RA, and IFIH1, in 2300 affected sib-pair (ASP) families and were tested for associations with T1D as part of the Type I Diabetes Genetics Consortium’s candidate gene study [7]. The data from these families provided no consistent evidence for associations of the other 14 candidate genes tested with T1D [7]. The possibility remains that loci such as the small ubiquitin-related modifiers– 4 gene (SUMO4) could be associated with T1D in non-European populations, which might provide an example of population-specific genetic regulation of T1D [8]. Very recently, all GWAS data published to date for seven common types of autoimmune disease (AID) were compiled, allowing a network-based analysis of shared susceptibility genes at different levels of significance [9]. Although involvement of the MHC region in chromosome 6p21 is not in question for most types of AID, the complex genetic architecture of this locus poses a significant analytical challenge [9]. Statistically significant excess sharing of nonMHC genes was found between T1D and all other AIDs studied, a result also seen for rheumatoid arthritis (RA) [9]. A smaller but significant degree of sharing was observed for multiple sclerosis, celiac disease, and Crohn’s disease [9]. The availability of GWAS data allows systematic analysis of similarities and differences among several types of AID [9].
0198-8859/10/$32.00 - see front matter 䉷 2010 Published by Elsevier Inc. on behalf of American Society for Histocompatibility and Immunogenetics. doi:10.1016/j.humimm.2010.05.016
796
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
The PTPN22 gene encodes lymphoid-specific protein-tyrosinephosphatase (Lyt), and was thereby identified as the fourth genetic locus associated with T1D. The R620W polymorphism (1858T variant, dbSNP no. rs2476601) in PTPN22, is reportedly associated with T1D in Caucasians [10]. This polymorphism also shows associations with other types of AID, such as autoimmune thyroid disease, RA, and systemic lupus erythematosus (SLE). The PTPN22 locus is thus regarded as a general autoimmunity locus [11]. In Asian populations, the R620W polymorphism was originally reported to be nonpolymorphic [12,13]. Recently, Kawasaki et al. described heterozygosity (molecular heterosis) of the polymorphism rs2488457 in the promoter region of the PTPN22 gene (⫺1123 G/C) as being associated with acute-onset T1D in the Japanese population [14]. In Caucasians, this promoter polymorphism is strongly linked to R620W, and several studies have revealed the association of ⫺1123 G/C and AID to be a result of linkage to R620W, and that the real susceptibility allele is 1858T (620W) [15,16]. Carlton et al. identified 37 SNPs in the PTPN22 gene and reported that several, including rs1310182 (designated as SNP27) and rs3789604 (designated as SNP37), were also associated with RA independently of R620W in the Caucasian population [17]. They also studied the haplotypes and found that some not containing 620 W (1858T) were associated with RA. However, these SNPs were not associated with Japanese RA [13]. Because the PTPN22 gene is associated with several types of AID [9], the existence of disease susceptibility alleles other than 1858T in the Japanese population seems likely. In the present study, we examined whether the SNPs associated with RA in Caucasians were also associated with T1D in the Japanese population. We also analyzed haplotypes and found that the PTPN22 gene plays a role in either susceptibility to or the protection from T1D in the Japanese population.
labeled probe methods, and a LightScanner (Idaho Technology, Salt Lake City, UT) according to the manufacturer’s instructions. Details of the primer sequences used for typing the five SNPs are available in Table 1.
2. Subjects and methods
Because of the strong LD among the five variants, haplotype analysis was undertaken using SNPAlyze. We identified seven haplotypes, four of which correlated with haplotypes 1, 3, 4 and 5, as previously reported by Carlton et al. (Table 3). Four haplotypes (1– 4) were found to be relatively common, whereas the remaining three haplotypes (5–7) were rare. In addition, haplotype distributions differed significantly between T1D and control subjects according to the permutation procedure (p ⫽ 0.0001). Haplotype 7, which contained the rs1310182 risk allele, was found to be positively associated with T1D (p ⫽ 0.0013). In contrast, haplotype 5, which did not contain the rs1310182 risk allele, was found to be protective (permutation p ⬍ 0.0001).
2.1. Subjects A total of 184 Japanese patients (67 male and 117 female) with T1D were included in this study. All study participants lived in the Tokyo metropolitan area. T1D was diagnosed on the basis of sudden onset of severe symptoms or rapid progression to overt diabetes and dependence on exogenous insulin because of absolute insulin deficiency. The T1D diagnosis was based on the 1997 Committee of the American Diabetes Association criteria [18]. All subjects were insulin dependent at the time of the study, although their -cell reserves, estimated by fasting serum C-peptide measures, varied. Patients with slowly progressive T1D were not included in this study. The median patient age at diagnosis was 28.7 years. A total of 179 age-matched healthy female subjects exhibiting normal glucose tolerance based on normal fasting plasma glucose and HbA1c levels, with no family history of T1D or any other types of AID, served as controls. All patients and controls were informed of the purpose of the study, and their consent was obtained before starting the experiments. The study was preapproved by the ethics committee of each participating institution. 2.2. Genotyping of SNPs DNA was extracted from whole blood using a Puregene kit (Gentra Systems, Minneapolis, MN) and the following five SNPs in the PTPN22 gene were analyzed: rs12760457 (SNP18) in intron 11, rs2797415 (SNP25) in intron 15, rs1310182 (SNP27) in intron 16, rs2476599 (SNP33) in intron 19 and rs3789604 (SNP37) in putative transcription factor-binding sites (TRBS) downstream from the PTPN22 gene. rs1310182 and rs3789604 were analyzed using the SNAPshot system (PE Applied Biosystems, Foster City, CA) based on the manufacturer’s protocol. rs12760457, rs2797415, and rs2476599 were genotyped using high-resolution melting and un-
2.3. Statistical analysis Case-control analysis and testing for Hardy–Weinberg equilibrium of the SNPs were performed using SNPAlyze version 7.0 (Dynacom, Yokohama, Japan) [19]. Hardy–Weinberg equilibrium was tested separately for all loci in subjects and controls. These tests in the subject and control groups showed no significant deviations from Hardy–Weinberg equilibrium for any of the SNPs. Linkage disequilibrium (LD) between SNPs was evaluated by D= and r2 of pairwise LD calculated using the SNPAlyze computer program. Haplotype frequencies for multiple loci were estimated using phase estimation and the expectation-maximization algorithm. Permutation p values were calculated by comparing haplotype frequencies between patients and controls based on 10,000 replications using the SNPAlyze program. 3. Results 3.1. Genotyping analysis Frequency analysis revealed that rs3789604 (SNP37) was not associated with T1D. Among the five SNPs studied, only rs1310182 (SNP27) differed in frequency between T1D patients and controls (Table 2). However, the risk allele in our Japanese subjects was found to be C, whereas that in Caucasian RA subjects was found to be T [17]. We also examined rs2488457 (⫺1123 G/C) in our population, but found no association between T1D and this SNP (data not shown). 3.2. Haplotype analysis
4. Discussion T1D incidences vary considerably among ethnic groups. Scandinavian individuals have an incidence as high as 35 in 100,000 per year, whereas Japanese individuals have a very low incidence (1.7 in 100,000 per year) [1]. The low incidence of T1D in the Japanese population is probably attributable to the difference in genetic predisposition to T1D, with Asians having a low prevalence of susceptibility haplotypes, whereas these genetic factors are relatively common in Europeans. The HLA class II loci are known to be strongly associated with T1D in all ethnic groups [3], and the haplotypes DRB1*0301-DQA1*0501-DQB1*0201, and DRB1*0401 DQA1*0301-DQB1*0302 are strongly associated with T1D in Caucasians. Genotypes heterogeneous for the DRB1*0401/DRB1*0301 alleles represent an extreme risk for T1D development [20]. These two alleles do not exist in the Japanese population, whereas DRB1*0405 and DRB1*0901 have been identified as risk variants. These differences in HLA genotypes may explain, in part, the low incidence of T1D in the Japanese population [21]. In recent years, the PTPN22 gene has emerged as an important genetic locus for several AIDs, including T1D [2]. The risk variant
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
797
Table 1 Primers used for typing the five SNPs dbSNP ID
Forward primer
Reverse primer
Interrogator primer or 3=-amino–linked probes
rs12760457 rs2797415 rs1310182 rs2476599 rs3789604
TATCACTGTTGCTCAGACTGG GCTGAGCAACATTTGATGTC ATGGACATATTTTCCCATGATG TCAGGTACCTCCTCGAAAAG CAGCATGAGGTTTCTCCTTC
CAGTAACTGAAGCAAAATAGATG GGATTTGGGAACATTTGGTG AAATGCCTACTGTATGCCAG CACTGTGTGAACTGCATATGC CAATAGAAGCTCGGTTGGAG
TTTCCGTACCCTCTATTCTGC ATACTATTCCTAAGAACCGAAC GATCGATCGATCGATCGTACAAGAGGTTCAC CCTAGTATTACGTATGCATGTGC CCACAAACACACATTTAAATGGCCCGACCGCCCCCCCTCCGCGCA
SNP, single nucleotide polymorphism.
PTPN22 620W (1858T) has been validated as a gain-of-function variant and demonstrates increased catalytic activity of Lyp [22]. The 1858T variant is virtually absent in Japanese individuals [12,13], which may largely explain the low incidence of T1D in this population [21]. In the present study, we focused on whether the proposed risk variants of PTPN22 for Caucasian RA patients are also risk variants for T1D in the Japanese population, independently of 1858T. The marker rs3789604, which was originally designated SNP37 in the Carlton et al. report [17], was not found to be associated with T1D in our Japanese cohort. The contrasting results of these two studies may be attributable to a difference in disease ontology or ethnic differences. The marker rs3811021, which is in almost absolute LD with rs3789604, was not associated with T1D in Caucasians [23,24]. In contrast, Ichimura et al. reported that rs3789604 was associated with Graves’ disease in the Japanese population, although the risk allele differed from that associated with RA in Caucasians [26]. Thus, rs3789604C does not appear to be a consistent risk variant for AID. In the present study, the frequency of rs1310182 (SNP27) differed significantly between patients and controls. In the study by Carlton et al., the association between rs1310182 and RA was attributed to linkage to rs3789604. Furthermore, in the present study, the risk allele differed from that identified by Carlton et al. When the haplotype data described below are considered, it seems unlikely that the rs1310182 variant itself is responsible for T1D risk in the Japanese population. In fact, allele T of rs1310182 was not associated with T1D in a Sardinian population [24]. With the exception of the haplotype that contains 1858T, the associations of PTPN22 haplotypes with AID have not been consistently demonstrated by previous studies. The haplotype that contains rs3789604C was associated with Caucasian RA patients in one study [17]; however, this association was not seen in RA [27] or Graves’ disease [25] examined in other studies of Caucasian populations. Onengut-Gumuscu et al. reported that two haplotypes, one containing rs3789604C, were over-transmitted to affected offspring when haplotype association tests were conducted on T1D patients homozygous for the C1858 nonrisk allele [23]. In addition,
Carlton et al. reported that one of these haplotypes confers protection against RA [17]. This haplotype was found to be transmitted to a lesser degree in a family-based association analysis of T1D [23]. In contrast, the haplotype termed Carlton’s haplotype 3 appeared to provide protection against Graves’ disease in Caucasians [25]. Although Japanese do not have the 1858T variant, the haplotype distribution differed significantly between T1D patients and controls, as evidenced by the permutation procedure used in this study. One rare haplotype, termed haplotype 7, was found to be significantly more frequent in T1D, whereas another rare haplotype (haplotype 5) was absent in T1D. These two haplotypes were not described in previous reports on Caucasian populations. Haplotype 7 contains the rs1310182 risk allele, whereas haplotype 5 does not. By contrast, haplotypes 3 and 4 containing the risk allele of rs1310182 did not appear to confer susceptibility to T1D. In addition, haplotypes 1, 2, and 6 do not contain the rs1310182 risk allele, and they did not appear to be protective. These results support the hypothesis that the rs1310182 variant itself is not responsible for susceptibility to or protection from T1D, and that the frequency differences in the above two haplotypes may partially explain the significant differences in the frequency of the rs1310182 allele. As both haplotypes described as being associated with T1D are rare, the association may not be clinically relevant. Our results suggest an association of the C allele of rs1310182 with T1D in our population. A haplotype analysis, including the five SNPs showed over-representation and under-representation, respectively, of two rare haplotypes in Japanese T1D patients. When combined with the results of previous studies, our data indicate the associations between haplotypes of the PTPN22 gene and AID to be complicated. Identification of precise risk or protective alleles in the region of the PTPN22 gene linked to the associated haplotypes requires further investigation. Recently, a loss-of-function mutation in the PTPN22 gene that does not exist in Asian populations was shown to confer protection against human SLE [28]. Finally, it has also been proposed that significant LD may extend for several hundred kilobases surrounding PTPN22, and thus potentially encompass additional genes, including RSBN1 and DCLRE1B [23].
Table 2 Genotype and allele frequencies of five SNPs in T1D patients and controls db SNP ID
Allele
T1D
(%)
Control
(%)
p Value
rs12760457
C T
339 29
92.1 7.9
335 23
93.6 6.4
NS
T C
208 160
56.5 43.5
218 140
60.9 39.1
T C
272 96
73.9 26.1
291 67
81.3 18.7
G A
306 62
83.1 16.8
315 43
88.0 12.0
A C
287 81
78.0 22.0
295 63
82.4 17.6
rs2797415
rs1310182
rs2476599
rs3789604
SNP, single nucleotide polymorphism; T1D, type 1 diabetes; NS, not significant.
NS
0.017
NS
NS
Genotype
T1D
(%)
Control
(%)
C/C C/T T/T T/T T/C C/C T/T T/C C/C G/G G/A A/A A/A A/C C/C
155 29 0 65 78 41 103 66 15 129 48 7 115 57 12
84.2 15.8 0.0 35.3 42.4 22.3 56.0 35.9 8.2 70.1 26.1 3.8 62.5 31.0 6.5
157 21 1 71 76 32 120 51 8 138 39 2 124 47 8
87.7 11.7 0.6 40.0 42.5 17.9 67.0 28.5 4.5 77.1 21.8 1.1 69.3 26.3 4.5
p Value
NS
NS
NS
NS
NS
798
M. Taniyama et al. / Human Immunology 71 (2010) 795–798
Table 3 Frequencies of five SNP haplotypes Haplotype no.
SNP 18-25-27-33-37
DM
Control
1 2 3 4 5 6 7
C-T-T-G-A C-C-T-G-C C-C-C-A-A T-C-C-G-A C-C-T-G-A C-T-T-G-C C-C-C-G-A
0.536 0.195 0.139 0.071 0.000 0.021 0.039
0.592 0.170 0.113 0.061 0.050 0.012 0.003
Permutation p value
⬍0.0001 0.0013
SNP numbers are those designated by Carlton et al. [11]. 18: rs12760457, 25: rs2797415, 27: rs1310182, 33: rs2476599, 37: rs3789604.
Therefore, the possibility that a gene other than PTPN22 is responsible for the association with T1D cannot be ruled out. Acknowledgments This work was supported in part by a Showa University Grantin-aid for Innovative Collaborative Research Projects (to Y.B.), the Showa University Medical Foundation (to Y.B.), a grant from the Showa University School of Medicine Alumni Association (to Y.B.), and a grant from the Yamaguchi Endocrine Research Association (to Y.B.). References [1] Notokins L, Lernmark Å. Autoimmune type 1 diabetes: Resolved and unresolved. J Clin Invest 2001;108:1247–52. [2] Pugliese A, Eisenbarth GS. Type 1 diabetes mellitus of man: genetic susceptibility and resistance. Adv Exp Med Biol 2004;552:170 –203. [3] She JX. Susceptibility to type 1 diabetes: HLA-DQ and DR revisited. Immunol Today 1996;17:323–9. [4] Wellcome Trust. Case Control Consortium Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 2007;447:661–78. [5] Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K, Plagnol V, et al. Robust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 2007;39:857– 64. [6] T1D Base. Available at: http://www.T1DBase.org. Last accessed 2010. [7] Howson JMM, Walker NM, Smyth DJ, Todd JA, and the Type I Diabetes Genetics Consortium. Analysis of 19 genes for association with type I diabetes in the Type I Diabetes Genetics Consortium families. Genes Immun 2009;10:S74 – 84. [8] Podolsky R, Prasad Linga-Reddy MV, She J-X, and the Type I Diabetes Genetics Consortium. Analyses of multiple single-nucleotide polymorphisms in the SUMO4/IDDM5 region in affected sib-pair families with type I diabetes. Genes Immun 2009;10:S16 –20. [9] Baranzini SE. The genetics of autoimmune diseases: A networked perspective. Curr Opin Immunol 2009;21:596 – 605.
[10] Bottini N, Musumeci L, Alonso A, Rahmouni S, Nika K, Rostamkhani M, et al. A functional variant of lymphoid tyrosine phosphatase is associated with type I diabetes. Nat Genet 2004;36:337– 8. [11] Bottini N, Vang T, Cucca F, Mustelin T. Role of PTPN22 in type 1 diabetes and other autoimmune diseases. Semin Immunol 2006;18:207–13. [12] Ban Y, Tozaki T, Taniyama M, Tomita M, Ban Y. The codon 620 single nucleotide polymorphism of the protein tyrosine phosphatase-22 gene does not contribute to autoimmune thyroid disease susceptibility in the Japanese. Thyroidology 2005;15:1115– 8. [13] Ikari K, Momohara S, Inoue E, Tomatsu T, Hara M, Yamanaka H, Kamatani N. Haplotype analysis revealed no association between the PTPN22 gene and RA in a Japanese population. Rheumatology 2006;45:1345– 8. [14] Kawasaki E, Awata T, Ikegami H, Kobayashi T, Maruyama T, Nakanishi K, et al. Systematic search for single nucleotide polymorphisms in a lymphoid tyrosine phosphatase gene (PTPN22): Association between a promoter polymorphism and type 1 diabetes in Asian populations. Am J Med Genet 2006;140:586 –93. [15] Cinek O, Hradsky O, Ahmedov G, Slavcev A, Kolouskova S, Kulich M, Sumnik Z. No independent role of the -1123 G⬎C and⫹2740 A⬎G variants in the association of PTPN22 with type 1 diabetes and juvenile idiopathic arthritis in two Caucasian populations. Diabetes Res Clin Pract 2007;76:297–303. [16] Viken MK, Olsson M, FlÇm ST, Førre O, Kvien TK, Thorsby E, Lie BA. The PTPN22 promoter polymorphism -1123G⬎C association cannot be distinguished from the 1858C⬎T association in a Norwegian rheumatoid arthritis material. Tissue Antigens 2007;70:190 –7. [17] Carlton VE, Hu X, Chokkalingam AP, Schrodi SJ, Brandon R, Alexander HC, et al. PTPN22 genetic variation: Evidence for multiple variants associated with rheumatoid arthritis. Am J Hum Genet 2005;77:567– 81. [18] Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997. [19] SNPAlyze. Available at: http://www.dynacom.co.jp/e/products/package/ snpalyze/. Last accessed 2010. [20] Baschal EE, Eisenbarth GS. Extreme genetic risk for type 1A diabetes in the post-genome era. J Autoimmun 2008;31:1– 6. [21] Ikegami H, Kawabata Y, Noso S, Fujisawa T, Ogihara T. Genetics of type 1 diabetes in Asian and Caucasian populations. Diabetes Res Clin Pract 2007; 77(Suppl 1):S116 –21. [22] Vang T, Congia M, Macis MD, Musumeci L, OrrÛ V, Zavattari P, et al. Autoimmuneassociated lymphoid tyrosine phosphatase is a gain-of-function variant. Nat Genet 2005;37:1317–19. [23] Onengut-Gumuscu S, Buckner JH, Concannon P. A haplotype-based analysis of the PTPN22 locus in type 1 diabetes. Diabetes 2006;55:2883–9. [24] Zoledziewska M, Perra C, OrrÚ V, Moi L, Frongia P, Congia M, et al. Further evidence of a primary, causal association of the PTPN22 620W variant with type 1 diabetes. Diabetes 2008;57:229 –34. [25] Heward JM, Brand OJ, Barrett JC, Carr-Smith JD, Franklyn JA, Gough SC. Association of PTPN22 haplotypes with Graves’ disease. J Clin Endocrinol Metab 2007;92:685–90. [26] Ichimura M, Kaku H, Fukutani T, Koga H, Mukai T, Miyake I, et al. Associations of protein tyrosine phosphatase nonreceptor 22 (PTPN22) gene polymorphisms with susceptibility to Graves’ disease in a Japanese population. Thyroidology:18:625–30. [27] Hinks A, Eyre S, Barton A, Thomson W, Worthington J. Investigation of genetic variation across the protein tyrosine phosphatase gene in patients with rheumatoid arthritis in the UK. Ann Rheum Dis 2007;66:683– 6. [28] OrrÛ V, Tsai SJ, Rueda B, Fiorillo E, Stanford SM, Dasgupta J, et al. A loss-offunction variant of PTPN22 is associated with reduced risk of systemic lupus erythematosus. Hum Mol Genet 2009;18:569 –79.