The Autoimmune Disease-Associated KIF5A, CD226 and SH2B3 Gene Variants Confer Susceptibility for Multiple Sclerosis

ORIGINAL ARTICLE The autoimmune disease-associated KIF5A, CD226 and SH2B3 gene variants confer susceptibility for multiple sclerosis

A Alcina1,16, K Vandenbroeck2,3,16, D Otaegui4,16, A Saiz5,16, JR Gonzalez6, O Fernandez7, ML Cavanillas8,MCCeńit8, R Arroyo8, I Alloza2, M Garcıá-Barcina9, A Antigu¨ edad10, L Leyva11, G Izquierdo12, M Lucas13, M Fedetz1, MJ Pinto-Medel11, J Olascoaga14, Y Blanco5, M Comabella15, X Montalban15, E Urcelay8,16 and F Matesanz1,16 1Instituto de Parasitologıá y Biomedicina ‘Lo´pez Neyra’, Consejo Superior de Investigaciones Cientı´ficas (CSIC), Granada, Spain; 2Neurogenomiks Group, Universidad del Paı´s Vasco (UPV/EHU), Leioa, Spain; 3IKERBASQUE, Basque Foundation for Science, Bilbao, Spain; 4A´ rea de Neurociencias, Instituto de Investigacioń Sanitaria Biodonostia, San Sebastiań, Spain; 5Neurology Service, Hospital Clinic and Institut d’Investigacio´ Biome`dica Pi i Sunyer (IDIBAPS), Barcelona, Spain; 6Centre de Recerca en Epidemiologia Ambiental (CREAL), CIBERESP, Institut Municipal d’Investigacio´ Me`dica (IMIM), Barcelona, Spain; 7Servicio de Neurologıá and Instituto de Neurociencias Clıńicas del Hospital Regional Universitario Carlos Haya, Ma´laga, Spain; 8Immunology and Neurology Department, Hospital Clıńico S. Carlos, Madrid, Spain; 9Servicio de Gene´tica, Hospital de Basurto, Bilbao, Spain; 10Servicio de Neurologıá, Hospital de Basurto, Bilbao, Spain; 11Laboratorio de Investigacioń, Instituto de Neurociencias Clıńicas del Hospital Regional Universitario Carlos Haya, Ma´laga, Spain; 12Unidad de Esclerosis Mu´ltiple, Hospital Virgen Macarena, Sevilla, Spain; 13Servicio de Biologıá Molecular, Hospital Virgen Macarena, Sevilla, Spain; 14Servicio de Neurologıá, Unidad de esclerosis mu´ltiple, Hospital Donostia, San Sebastiań, Spain and 15Centre d’Esclerosi Mu´ltiple de Catalunya, CEM-Cat, Unitat de Neuroimmunologia Clıńica, Hospital Universitari Vall d’Hebron, Barcelona, Spain

Genome-wide association studies (GWAS) have revealed that different diseases share susceptibility variants. Twelve single- nucleotide polymorphisms (SNPs) previously associated with different immune-mediated diseases in GWAS were genotyped in a Caucasian Spanish population of 2864 multiple sclerosis (MS) patients and 2930 controls. Three SNPs were found to be associated with MS: rs1678542 in KIF5A (P ¼ 0.001, odds ratio (OR) ¼ 1.13, 95% confidence interval (CI) ¼ 1.05–1.23); rs3184504 in SH2B3 (P ¼ 0.00001, OR ¼ 1.19, 95% CI ¼ 1.10–1.27) and rs763361 in CD226 (P ¼ 0.00007, OR ¼ 1.16, 95%CI ¼ 1.08–1.25). These variants have previously been associated with rheumatoid arthritis and type 1 diabetes. The SH2B3 polymorphism has additionally been associated with systemic lupus erythematosus. Our results, in addition to validating some of these loci as risk factors for MS, are consistent with shared genetic mechanisms underlying different immune-mediated diseases. These data may help to shape the contribution of each pathway to different disorders. Genes and Immunity (2010) 11, 439–445; doi:10.1038/gene.2010.30; published online 27 May 2010

Keywords: multiple sclerosis (MS); rheumatoid arthritis (RA); type 1 diabetes (T1D); KIF5A; CD226; SH2B3

Introduction common central nervous system disease in young adults of developing countries. Although the etiology Multiple sclerosis (MS, MIM 126200) is a chronic remains elusive, disease susceptibility is thought to be autoimmune disease of the central nervous system conferred by a combination of genetic and environmental with a complex pathogenesis in which demyelination factors.4,2 and neurodegeneration are the main contributors to 1,2 The disease shows familial clustering, and about one- disability. Prevalence and incidence rates in Spain fifth of affected individuals have a family history of MS. are around 77/100 000 habitants and 5.3/100 000 habi- 3 Siblings to MS patients exhibit a 20–40 times higher tants per year, respectively. MS is estimated to affect relative risk of MS than the general population, and twin over 2 million individuals worldwide, and is the most studies have revealed that a large proportion of family clustering can be attributed to shared genes.5 Correspondence: Dr F Matesanz, Instituto de Parasitologıáy Genetic studies in autoimmune diseases have for long Biomedicina ‘Lo´pez Neyra’, Consejo Superior de Investigaciones supported a model in which the major risk factor resides Cientı´ficas (CSIC), Parque Tecnolo´gico de Ciencias de la Salud, in the HLA region.2 However, recent application of Avda. del Conocimiento s/n. 18100, Granada, Spain. genome-wide association studies (GWAS) to large sam- E-mail: [email protected] 16These authors contributed equally to this work. ple sets have provided a convincing support for addi- Received 10 March 2010; revised and accepted 22 April 2010; tional risk loci outside the HLA region, all with low-risk published online 27 May 2010 contribution to the disease.6–8 KIF5A, CD226 and SH2B3 variants associate with MS A Alcina et al 440 It is relatively common that patients affected by one Table 1 Demographic characteristics of the patients autoimmune disease suffer from another autoimmune disease, and that members of the same family suffer from Case Control different autoimmune diseases. For example, co-occurrence of type 1 diabetes (T1D), autoimmune thyroiditis Gender no. (%) (Graves’ disease and Hashimoto’s disease), rheumatoid Female 1907 (66.6) 2004 (68.4) 9 Male 957 (33.4) 926 (31.6) arthritis (RA), celiac disease and MS has been observed. Female to male ratio 2 2.1 In families with systemic lupus erythematosus (SLE), other autoimmune diseases, such as MS and RA, occur Age (years). more frequently than in the general population.10 Such Age at onset observations suggest the presence of shared genes and Mean 29.9±10.35 involvement of common disease pathways, and are Range 6–68 Age at analysis supported by a number of reports on genes that are Mean 36±12.23 38±10 associated with more than one autoimmune disease, Range 13–85 10–90 such as PTPN22 in RA and SLE,11 IRF5 in MS,12 RA13 and SLE,14 and more recently IL2RA/CD25 in T1D15 and MS,16 Disease course no. (%) though with a different risk allele. RR 1682 (75.4) As GWAS constitute powerful tools to detect common SP 401 (18) PP 120 (5.3) genetic susceptibility variants in complex diseases, these PR 16 (0.7) screens have further revealed that different diseases CIS 17 (0.8) share associated variants.17 In fact, the variants asso- No details 628 ciated in one disease are now routinely genotyped in related diseases, thus providing the rationale for the Abbreviations: CIS, clinically isolated syndrome; PP, primary candidate gene approach that we have applied to MS in progressive; PR, progressive relapsing; RR, relapse remitting; SP, this study. As such, 12 candidate genes that have been secondary progressive. identified recently in the GWAS published over the past The samples were assembled by REEM (Spanish network for the 2 years and validated in different immune-mediated study of multiple sclerosis (MS); www.reem.es) and consisted of pathologies, such as RA, SLE, ankylosing spondylitis, 5794 Caucasian individuals collected from four Spanish regions. Of psoriasis, T1D and MS, were selected and analyzed for this number, 2864 were case samples with clinically defined MS, their association in our cohort of MS patients and controls. according to Poser’s criteria,32 and 2930 were blood donor controls. By location, 460 cases and 462 controls were from Hospital Clinic and Vall d’Hebron of Barcelona; 535 cases and 536 controls from Hospital Clıńico of Madrid; 777 cases and 706 controls from the Results and discussion Basque region, that is, Hospital de Basurto of Bilbao and Hospital of In this work, we have shown for the first time an Donostia of San Sebastiań; 1092 cases and 1226 controls from four association of the rs1678542/KIF5A variant with MS in a hospitals and a blood bank of Andalucıá, specifically, Hospital ´ Caucasian Spanish cohort comprising 2864 MS cases and Virgen Macarena of Sevilla, Hospital Carlos Haya of Malaga, Hospital Clıńico and Virgen de las Nieves of Granada and the Blood 2930 controls (Table 1), and confirmed the recently found Bank of Granada. The study was approved by the Ethics associations of the rs763361/CD226 and rs3184504/ Committees of each of the hospitals participating in the study and SH2B3 variants. For this study, we selected 12 single- written informed consent was obtained from all participants. nucleotide polymorphisms (SNPs); Table 2) that had been shown to be associated with RA (rs2812378/ CCL2118 and rs1678542/KIF5A18), SLE (rs10516487/ (0.9–1.13)).22 Subsequent candidate studies provided BANK119 and rs13277113/FAM167A-BLK (C80RF13- convincing evidence for association with MS17,25 (sum- BLK)20), ankylosing spondylitis (rs17561/IL1A21 and marized in Table 3). rs30187/ERAP122), T1D (rs3184504/SH2B3,23 rs1893217/ As illustrated in Figure 1, a meta-analysis of our data PTPN223 and rs763361/CD22623), psoriasis (rs2066808/ was performed in combination with the results obtained STAT224) and the last two from a GWAS based on 12 374 for these two variants in the study of the International non-synonymous SNPs carried out by the Wellcome Multiple Sclerosis Genetics Consortium,17 which in- Trust Case–Control Consortium in MS (rs6542517/ cludes five populations (Belgium, Norway, Sweden, UK STEAP122 and rs7162473/HERC122). There were no and USA). Two other studies have also reported statistically significant differences in the genotyping association of rs763361/CD226 with MS,25,26 however, failure rate between the cases and controls for any of because of some sample overlapping they were not these markers. None of the markers deviated from included in our analysis. The overall effect estimates for Hardy–Weinberg equilibrium (HWE). the rs763361/CD226 and rs3184504/SH2B3 variants were As shown in Table 2, four SNPs were associated with statistically significant with combined ORs (95% con- MS. The strongest association corresponded to SNPs fidence interval (CI)) of 1.14 (1.09–1.18) and 1.13 (1.08– rs763361 in the CD226 gene (P ¼ 0.000069) and rs3184504 1.17), respectively. Heterogeneity variance was estimated in the SH2B3 gene (P ¼ 0.000011). These two variants not to be significant for any of the polymorphisms were first identified in a T1D23 GWAS, but were, (P ¼ 0.889 and P ¼ 0.526, respectively). however, not detected as risk factors for MS in an In addition to T1D, the rs763361/CD226 variant was association study of 14 500 SNPs carried out by the also found to be associated with other autoimmune Wellcome Trust Case–Control Consortium in 1000 MS diseases such as RA25 and Graves’ disease.25 Tag-SNP patients and 1500 controls (rs763361, P ¼ 0.09, odds ratio studies in this gene for T1D and MS suggest that (OR) ¼ 1.1 (0.98–1.24); rs3184504, P ¼ 0.84, OR ¼ 1.012 rs763361/CD226, which is a non-synonymous SNP,

Genes and Immunity Table 2 Summary of association results for multiple sclerosis case–control study

CH Gene SNP A4a Risk DS Ref. Power n Control genotypes (%) n MS genotypes (%) P* PBH * OR (95% CI) allele AA Aa aa AA Aa aa

2q14 IL1A rs17561G4TGAS20 0.99 2919 1530 (52.4) 1169 (40.0) 221 (7.6) 2850 1467 (51.6) 1146 (40.2) 237 (8.3) 0.31 0.52 0.97 (0.90–1.05) 4q23 BANK1 rs10516487 C4T C SLE 18 0.99 2902 1564 (53.9) 1126 (38.8) 212 (7.3) 2831 1474 (52.1) 1134 (40.1) 223 (7.9) 0.16 0.38 1.15 (0.94–1.11) 5q15 ERAP1 rs30187 C4TTAS21 0.99 2860 992 (34.7) 1369 (47.9) 499 (17.4) 2776 978 (35.2) 1321 (47.6) 577 (17.2) 0.65 0.78 1.01 (0.94–1.09) 7q21 STEAP1 rs6542517 A4TT MS21 0.99 2921 2795 (95.7) 121 (4.1) 5 (0.2) 2860 2719 (95.8) 117 (4.1) 2 (0.1) 0.77 0.84 0.93 (0.73–1.19) 8p23 FAM167A-BLK rs13277113 G4A A SLE 19 0.99 2893 1828 (63.2) 935 (32.3) 130 (4.5) 2844 1792 (63.0) 934 (32.0) 118 (4.1) 0.88 0.88 1.00 (0.91–1.10) 9p13 CCL21 rs2812378 T4CC RA17 0.81 2891 1426 (49.3) 1198 (41.4) 267 (9.2) 2809 1428 (50.8) 1171 (41.7) 210 (7.5) 0.047 0.14 0.92 (0.85–1.00) 12q13 KIF5A rs1678542 G4CG RA17 0.87 2863 1302 (45.5) 1243 (43.4) 318 (11.1) 2757 1359 (49.3) 1148 (41.6) 250 (9.1) 1.1 Â 10À3 4 Â 10À3 1.13 (1.05–1.23) 12q13.3 STAT2 rs2066808 T4CT PS 23 0.96 2906 2598 (89.4) 297 (10.2) 11 (0.4) 2823 2548 (90.3) 269 (9.5) 6 (0.2) 0.24 0.48 0.91 (0.77–1.07) 12q24 SH2B3 rs3184504 C4T T T1D 22 0.99 2851 930 (32.6) 1341(47.0) 580 (20.3) 2762 755 (27.3) 1366 (49.0) 641 (23.2) 1.1 Â 10À5 1.3 Â 10À4 1.19 (1.10–1.27) 15q22 HERC1 rs7162473 C4GG MS21 0.99 2912 2367 (81.3) 517 (17.8) 28 (1.0) 2853 2297 (80.5) 532 (18.6) 24 (0.8) 0.53 0.78 1.03 (0.91–1.16) 18p11.3 PTPN2 rs1893217 G4T T T1D 22 0.99 2915 2114 (72.5) 756 (25.9) 45 (1.5) 2843 2058 (72.4) 724 (25.5) 61 (2.1) 0.61 0.78 0.99 (0.89–1.10) 18q13 CD226 rs763361 G4A A T1D 22 0.99 2838 955 (33.0) 1377 (47.5) 565 (19.5) 2843 824(29.0) 1371 (48.3) 643 (22.7) 6.9 Â 10À5 4 Â 10À4 1.16 (1.08–1.25) I5,CD226 KIF5A, Alcina A Abbreviations: A4a, major4minor allele; AS, ankylosing spondylitis; BANK1, B cell scaffold protein with ankyrin repeats; BLK, B lymphoid tyrosine kinase; CCL21, chemokine (C-C motif) ligand 21; CD226, cluster of differentiation 226; CH, chromosome; CI, confidence interval; DS, disease firstly described to be associated with the corresponding SNP; ERAP1, endoplasmic reticulum aminopeptidase 1; FAM167A, family with sequence similarity 167, member A; HERC1, homologous to the E6-AP (UBE3A) carboxyl terminus; IL1A, interleukin-1a; Kif5A, kinesin tal et family member 5; MS, multiple sclerosis; n, number of samples; OR, odds ratio for the risk allele originally reported; PBH ; P corrected with Benjamini–Hochberg; P, uncorrected P-value; PS, psoriasis; PTPN2, protein tyrosine phosphatase non-receptor type 2; RA, rheumatoid arthritis; ref., reference; SLE, systemic lupus erythematosus; STAT2, signal transducer and activator of transcription 2; SH2B3, SH2B adaptor protein 3; SNP, single-nucleotide polymorphism; STEAP1, six transmembrane epithelial antigen of the prostate 1; T1D, type 1 diabetes. and

*Bold labels indicate significant associations. SH2B3 Departure from Hardy–Weinberg equilibrium (HWE) was tested using the Fisher’s exact test. For individual SNP association analyses, genotype frequencies were assessed assuming an additive model (e.g., multiplicative allelic effect). Logistic regression models, adjusted for region, were used to estimate crude ORs and 95%CI. P-values were computed using a likelihood ratio 33 test. Analysis was performed using the SNPassoc R package (Catalan Institute of Oncology; http://bioinfo.iconcologia.net/index.php?module=SNPassoc). To avoid false-positive results due MS with associate variants to multiple testing, and considering that the SNPs analyzed are not in complete linkage disequilibrium, we applied the Benjamini–Hochberg (BH) method.34 Power was calculated for the effect detected in previous reports with different diseases at an a level of 0.05, assuming a multiplicative allelic effects model, and a sample number of 2864 cases and 2930 controls. Genotyping was performed using the MassARRAY SNP genotyping system (Sequenom 3595 John Hopkins Court, San Diego, CA, USA) according to the manufacturer’s instructions at the Spanish National Genotyping Center’s (CeGen, Santiago de Compostela). Two CEPH trios were genotyped as part of the quality control (QC) (CEPH, Na10830, Na10831, Na12147, Na10860, Na10861 and Na11984). The genotypes of these samples confirmed absence of Mendelian inconsistencies and corresponded with the ones deposited in HapMap. On the other hand, a 10% of random samples were subjected to re-genotyping. Final data were concordant with an average accuracy of 499.9%. ee n Immunity and Genes 441 KIF5A, CD226 and SH2B3 variants associate with MS A Alcina et al 442 Table 3 Summary of data from different well-powered studies performed in different population and diseases

Gene SNP Risk allele Disease Populations cases/controlsref. P-value, OR

IL1A rs17561 G AS 2675/259221 1.9 Â 10À5, 1.21 T1D 8010/973331 NS MS Present work NS BANK1 rs10516487 C SLE 2003/196819 4.74 Â 10À11, 1.42 SLE 1892/265235 2.84 Â 10À5, 1.21 SLE 1579/172636 0.006, 1.2 RA 1080/136837 0.002, 1.11 T1D 8010/973331 NS MS Present work NS ERAP1 (ARTS1) rs30187 T AS 1000/150022 2.1 Â10À7, 1.31 T1D 8010/973331 0.004, 1.07 MS Present work NS STEAP1 rs6542517 T MS 1000/150022 3.2 Â 10À4, 2.13 MS Present work NS FAM167A-BLK rs13277113 A SLE 2104/419720 10À10, 1.39 SLE 1579/172636 5.1 Â10À7, 1.34 RA 1635/190638 NS RAa 5022/738639 5.69 Â 10À9, 1.19 SC 1050/69440 3.6 Â 10À4, 1.32 APSa 133/46841 1.8 Â 10À6, 2.06 T1D 8010/973331 NS MS Present work NS CCL21 rs2812378 C RA 7322/1826718 2.8 Â 10À7, 1.12 RA 1128/102242 NS RA 3962/353143 0.04, 1.08 T1D 8010/973331 NS MS Present work NS KIF5A rs1678542 G RA 7322/1826718 8.8 Â 10À8, 1.12 T1D 8010/973331 3 Â 10À4, 1.08 RA 1368/168344 NS MS Present work 0.001, 1.13 STAT2 rs2066808 T PS 5048/504124 1 Â10À9,1.34 MS Present work NS SH2B3 rs3184504 T T1D 6100/680023 1.73 Â 10À21,1.33 MS 5737/10296/2369 Trio17 4.4 Â 10À6, 1.08 SLE 1963/432928 4.0 Â 10À4, 1.08 RA 1368/168344 1.15 Â 10À5, 131 MS Present work 10À5, 1.19 HERC1 rs7162473 G MS 1000/150022 5.3 Â 10À4, 1.3 MS Present work NS PTPN2 rs1893217 T T1D 6100/680023 1.49 Â 10À14, 1.3 GD 2200/360023 0.025, 1.13 MS 5737/10296/2369 Trio17 0.003, 1.1 MS Present work NS CD226 rs763361 A T1D 6100/680023 2.82 Â 10À8, 1.16 GD 2200/360025 0.018, 1.1 RA 3595/321425 0.017, 1.09 MS 5737/10296/2369 Trio17 5.4 Â 10À8, 1.13 MS Present work 7 Â 10À5, 1.16

Abbreviations: AS, ankylosing spondylitis; BANK1, B cell scaffold protein with ankyrin repeats; BLK, B lymphoid tyrosine kinase; CCL21, chemokine (C-C motif) ligand 21; CD226, cluster of differentiation 226 ; ERAP1, endoplasmic reticulum aminopeptidase 1; FAM167A family with sequence similarity 167, member A; GD, Graves’ disease; HERC1, homologous to the E6-AP (UBE3A) carboxyl terminus; IL1A, interleukin-1a; Kif5A, kinesin family member 5; MS, multiple sclerosis; NS, not significant; PS, psoriasis; PTPN2, protein tyrosine phosphatase non-receptor type 2; RA, rheumatoid arthritis; SC, systemic sclerosis; SLE, systemic lupus erythematosus; STAT2, signal transducer and activator of transcription 2; SH2B3, SH2B adaptor protein 3; STEAP1, six transmembrane epithelial antigen of the prostate 1; T1D, type 1 diabetes. aProxy rs2736340.

Gly307Ser), could be the causal variant.23,25 This SNP could alter the signaling cascade by affecting the two might affect the expression and/or biological activity of known phosphorylation sites at positions 322 and 329.23 CD226 by disruption of an exon splicing silencer The identification of rs3184504/SH2B3 as a risk factor sequence. The minor allele A maintains the sequence for T1D followed that of rs17696736, originally uncov- motif of an exon splicing silencer in the final 30exon, ered in a T1D GWAS.27 The latter SNP was also found to whereas the G allele disrupts it. The allele A expression be associated with SLE in a very large cohort (P ¼ 0.014, product is therefore likely to result in a putative CD226 OR ¼ 1.08).28 Both polymorphisms are located within a isoform, acting either as a nonfunctional protein or as a large (41.2 Mb) linkage disequilibrium block that con- protein with a novel function. This amino-acid substitution tains several genes of possible functional relevance for

Genes and Immunity KIF5A, CD226 and SH2B3 variants associate with MS A Alcina et al 443 rs703842/methyltransferase-like protein 1 30 UTR with 2 IMSGC, 2009 a linkage disequilibrium r of 0.184. The latter polymorphism is the most strongly associated SNP with MS Belgium in the locus in the GWAS carried out by the Australia and Norway New Zealand Multiple Sclerosis Genetics Consortium.7 Sweden A candidate gene study by the T1D Genetics Consortium UK also provides evidence of association of this region with US T1D.31 As suggested in the Australia and New Zealand Pooled Multiple Sclerosis Genetics Consortium work,7 assuming Present study a common causal autoimmune susceptibility gene underlying these associations, the strongest candidate gene CYP27B1 COMBINED would be ; however, a fine mapping of the region will be required to assign the actual causal gene in all of these diseases. The rs2812378 positioned 113 nucleotides downstream 1.00 1.05 1.10 1.15 1.20 1.26 1.32 1.38 from the CCL21 gene were also associated with MS rs763361/CD226 (T), ODDS RATIO (P ¼ 0.047, OR ¼ 0.92, 95% CI ¼ 0.85–1.00). However, it did not withstand correction for multiple testing, and the risk allele obtained in the present study was contrary to 18 IMSGC, 2009 the one reported in the original RA study. Belgium We observed a lack of association of IL1A, BANK1, Norway ERAP1 (ARTS1), FAM167A-BLK, CCL21, PTPN2, STEAP1, Sweden HERC1 and STAT2 variants with MS. The STEAP and UK HERC1 variants have been associated with MS in a recent US GWAS carried out by the Wellcome Trust Case–Control Pooled Consortium based on 12 374 non-synonymous SNPs,22 Present study but were not validated in our sample set. This discrepancy probably reflects a spurious association in the original COMBINED GWAS. In Table 3, we have summarized data from different well-powered studies carried out in different population and diseases. It is interesting to note that KIF5A, SH2B3 and CD226 have been found to be associated with most 0.870.91 0.95 1.00 1.05 1.10 1.15 1.20 1.26 1.32 of the studied diseases that suggest their involvement in rs3184504/SH2B3 (T), ODDS RATIO the basic mechanisms of immunopathologies. Figure 1 Summary effect estimate for SNPs rs763361/CD226 The confirmation of the common and specific asso- (upper panel) and rs3184504/SH2B3 (lower panel) using informa- ciated genes in the different immune-mediated diseases tion from the IMSGC17 study. Combined analysis of our study will be crucial in future attempts at shaping the results on 2864 Spanish MS cases and 2930 Spanish controls with contribution of each pathway to the different disorders those obtained from 398 cases and 461 controls from Belgium, 644 and the identification of novel therapies. cases and 1023 controls from Norway, 1014 cases and 1173 controls from Sweden, 2370 cases and 7078 controls from UK and 1311 cases and 561 controls from US. Conflict of interest The authors declare no conflict of interest. autoimmune diseases. A follow-up analysis in T1D of four SNPs in this region pointed the rs3184504/SH2B3 as the most probable causal variant, as this non-synon- Acknowledgements ymous SNP is sufficient to model the association of the entire region.23 We thank the patients with multiple sclerosis and The SH2B3 gene product (Lnk) is an adaptor protein persons who acted as controls for making this study expressed primarily in the lymphocytes and hematopoietic possible. Financial support for the study was provided precursor cells. The possible role of the SH2B3 variant by Fondos Europeos de Desarrollo Regional (FEDER), ´ associated in these studies to different autoimmune Ministerio de Ciencia e Innovacion (SAF2006-02023 and ´ diseases is at the moment undetermined, mainly because SAF2009-11491), Junta de Andalucıa (P07-CVI-02551), ´ of the lack of knowledge on the functional role of the Fondo de Investigacion Sanitaria (RETICS-REEM RD07/ causal variant. However, the overexpression as well as 0060, PI081636, PI070353, PS0902105), Ikerbasque, Bas- ´ the deficiency of this gene has consequences that could que Foundation for Science, Bilbao and Fundacion be of importance in the development of autoimmune ILUNDAIN. Genotyping services were provided by the diseases. SH4B3 overproduction in lymphoid precursors Spanish ‘Centro Nacional de Genotipado’ CEGEN-USC resulted in a reduction in the number of both B and Tcells, (www.cegen.org). whereas the lack of the SH4B3 gene showed enhanced B-cell production.29 References The association of rs1678542/KIF5A variant with MS, uncovered for the first time in this work, was first 1 Hauser SL, Oksenberg JR. The neurobiology of multiple associated with RA in European populations.18,30 The sclerosis: genes, inflammation, and neurodegeneration. Neu- rs1678542/KIF5A is located 194 kb apart from the ron 2006; 52: 61–76.

Genes and Immunity KIF5A, CD226 and SH2B3 variants associate with MS A Alcina et al 444 2 Ramagopalan SV, Knight JC, Ebers GC. Multiple sclerosis and 21 Sims AM, Timms AE, Bruges-Armas J, Burgos-Vargas R, Chou the major histocompatibility complex. Curr Opin Neurol 2009; CT, Doan T et al. Prospective meta-analysis of interleukin 1 22: 219–225. gene complex polymorphisms confirms associations with 3 Oliver B, Mayorga C, Fernańdez V, Leyva L, Leoń A, Luque G ankylosing spondylitis. Ann Rheum Dis 2008; 67: 1305–1309. et al. Multiple sclerosis in Gypsies from southern Spain: 22 Burton PR, Clayton DG, Cardon LR, Craddock N, Deloukas P, prevalence, mitochondrial DNA haplogroups and HLA class Duncanson A, et al., Wellcome Trust Case Control Con- II association. Tissue Antigens 2008; 71: 426–433. sortium; Australo-Anglo-American Spondylitis Consortium 4 McElroy JP, Oksenberg JR. Multiple sclerosis genetics. Curr (TASC). Association scan of 14 500 nonsynonymous SNPs in Top Microbiol Immunol 2008; 318: 45–72. four diseases identifies autoimmunity variants. Nat Genet 5 Willer CJ, Dyment DA, Risch NJ, Sadovnick AD, Ebers GC, 2007; 39: 1329–1337. Canadian Collaborative Study Group. Twin concordance and 23 Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K, sibling recurrence rates in multiple sclerosis. Proc Natl Acad Sci Plagnol V, et al., Wellcome Trust CaseControl Consortium. USA 2003; 100: 12877–12882. Robust associations of four new chromosome regions from 6 Baranzini SE, Wang J, Gibson RA, Galwey N, Naegelin Y, genome-wide analyses of type 1 diabetes. Nat Genet 2007; 39: Barkhof F et al. Genome-wide association analysis of suscept- 857–864. ibility and clinical phenotype in multiple sclerosis. Hum Mol 24 Nair RP, Duffin KC, Helms C, Ding J, Stuart PE, Goldgar D Genet 2009; 18: 767–778. et al. Genome-wide scan reveals association of psoriasis 7 Australia New Zealand Multiple Sclerosis Genetics Consor- with IL-23 and NF-kappaB pathways. Nat Genet 2009; 41: tium (ANZgene). Genome-wide association study identifies 199–204. new multiple sclerosis susceptibility loci on chromosomes 12 25 Hafler JP, Maier LM, Cooper JD, Plagnol V, Hinks A, and 20. Nat Genet 2009; 41: 824–828. Simmonds MJ et al. CD226 Gly307Ser association with multi- 8 International Multiple Sclerosis Genetics Consortium. Risk ple autoimmune diseases. Genes Immun 2009; 10: 5–10. alleles for multiple sclerosis identified by a genomewide 26 De Jager PL, Jia X, Wang J, de Bakker PI, Ottoboni L, Aggarwal study. N Engl J Med 2007; 357: 851–862. NT et al. Meta-analysis of genome scans and replication 9 Somers EC, Thomas SL, Smeeth L, Hall AJ. Autoimmune identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis diseases co-occurring within individuals and within families: susceptibility loci. Nat Genet 2009; 41: 776–782. a systematic review. Epidemiology 2006; 17: 202–217. 27 Wellcome Trust CaseControl Consortium. Genome-wide 10 Corporaal S, Bijl M, Kallenberg CG. Familial occurrence of association study of 14 000 cases of seven common diseases autoimmune diseases and autoantibodies in a Caucasian and 3000 shared controls. Nature 2007; 447: 661–678. population of patients with systemic lupus erythematosus. 28 Gateva V, Sandling JK, Hom G, Taylor KE, Chung SA, Sun X Clin Rheumatol 2002; 21: 108–113. et al. A large-scale replication study identifies TNIP1, PRDM1, 11 Lee YH, Rho YH, Choi SJ, Ji JD, Song GG, Nath SK et al. The JAZF1, UHRF1BP1 and IL10 as risk loci for systemic lupus PTPN22 C1858T functional polymorphism and autoimmune erythematosus. Nat Genet 2009; 41: 1228–1233. diseases-a meta-analysis. Rheumatology 2007; 46: 49–56. 29 Takaki S, Tezuka Y, Sauer K, Kubo C, Kwon SM, Armstead E 12 Kristjansdottir G, Sandling JK, Bonetti A, Roos IM, Milani L, et al. Impaired lymphopoiesis and altered B cell subpopula- Wang C et al. Interferon regulatory factor 5 (IRF5) gene tions in mice overexpressing Lnk adaptor protein. J Immunol variants are associated with multiple sclerosis in three distinct 2003; 170: 703–710. populations. J Med Genet 2008; 45: 362–369. 30 Barton A, Thomson W, Ke X, Eyre S, Hinks A, Bowes J et al. 13 Sigurdsson S, Padyukov L, Kurreeman FA, Liljedahl U, Rheumatoid arthritis susceptibility loci at chromosomes Wiman AC, Alfredsson L et al. Association of a haplotype in 10p15, 12q13 and 22q13. Nat Genet 2008; 40: 1156–1159. the promoter region of the interferon regulatory factor 5 31 Fung EY, Smyth DJ, Howson JM, Cooper JD, Walker NM, gene with rheumatoid arthritis. Arthritis Rheum 2007; 56: Stevens H et al. Analysis of 17 autoimmune disease-associated 2202–2210. variants in type 1 diabetes identifies 6q23/TNFAIP3 as a 14 Sigurdsson S, Nordmark G, Go¨ring HH, Lindroos K, Wiman susceptibility locus. Genes Immun 2009; 10: 188–191. AC, Sturfelt G et al. Polymorphisms in the tyrosine kinase 2 32 Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, and interferon regulatory factor 5 genes are associated Ebers GC et al. New diagnostic criteria for multiple sclerosis: with systemic lupus erythematosus. Am J Hum Genet 2005; guidelines for research protocols. Ann Neurol 1983; 13: 76: 528–537. 227–231. 15 Lowe CE, Cooper JD, Brusko T, Walker NM, Smyth DJ, 33 Gonza´lez JR, Armengol L, Sole´ X, Guino´ E, Mercader JM, Bailey R et al. Large-scale genetic fine mapping and Estivill X et al. SNPassoc: an R package to perform whole genotype-phenotype associations implicate polymorphism genome association studies. Bioinformatics 2007; 23: 644–645. in the IL2RA region in type 1 diabetes. Nat Genet 2007; 39: 34 Benjamini Y, Hochberg Y. Controlling the false discovery rate: 1074–1082. a practical and powerful approach to multiple testing. J Roy 16 Alcina A, Fedetz M, Ndagire D, Fernańdez O, Leyva L, Stat Soc Ser B 1995; 57: 289–300. Guerrero M et al. IL2RA/CD25 gene polymorphisms: uneven 35 Guo L, Deshmukh H, Lu R, Vidal GS, Kelly JA, Kaufman KM association with multiple sclerosis (MS) and type 1 diabetes et al. Replication of the BANK1 genetic association with (T1D). PLoS One 2009; 4: e4137. systemic lupus erythematosus in a European-derived popula- 17 International Multiple Sclerosis Genetics Consortium tion. Genes Immun 2009; 10: 531–538. (IMSGC). The expanding genetic overlap between multiple 36 Suarez-Gestal M, Calaza M, Endreffy E, Pullmann R, Ordi-Ros sclerosis and type I diabetes. Genes Immun 2009; 10: 11–14. J, Domenico Sebastiani G et al. Identified systemic lupus 18 Raychaudhuri S, Remmers EF, Lee AT, Hackett R, Guiducci C, erythematosus genetic associations: a case-control study. Burtt NP et al. Common variants at CD40 and other loci confer Arthritis Res Ther 2009; 11: R69. risk of rheumatoid arthritis. Nat Genet 2008; 40: 1216–1223. 37 Orozco G, Abelson AK, Gonza´lez-Gay MA, Balsa A, Pascual- 19 Kozyrev SV, Abelson AK, Wojcik J, Zaghlool A, Linga Reddy Salcedo D, GarcıáAet al. Study of functional variants of the MV, Sanchez E et al. Functional variants in the B-cell gene BANK1 gene in rheumatoid arthritis. Arthritis Rheum 2009; 60: BANK1 are associated with systemic lupus erythematosus. 372–379. Nat Genet 2008; 40: 211–216. 38 Suarez-Gestal M, Calaza M, Dieguez-Gonzalez R, Perez- 20 Hom G, Graham RR, Modrek B, Taylor KE, Ortmann W, Pampin E, Pablos JL, Navarro F et al. Rheumatoid arthritis Garnier S. Association of systemic lupus erythematosus with does not share most of the newly identified systemic C8orf13-BLK and ITGAM-ITGAX. N Engl J Med 2008; 358: lupus erythematosus genetic factors. Arthritis Rheum 2009; 900–909. 60: 2558–2564.

Genes and Immunity KIF5A, CD226 and SH2B3 variants associate with MS A Alcina et al 445 39 Gregersen PK, Amos CI, Lee AT, Lu Y, Remmers EF, 42 Lee HS, Korman BD, Le JM, Kastner DL, Remmers EF, Kastner DL et al. REL, encoding a member of the NF-kappaB Gregersen PK et al. Genetic risk factors for rheumatoid family of transcription factors, is a newly defined risk locus arthritis differ in Caucasian and Korean populations. Arthritis for rheumatoid arthritis. Nat Genet 2009; 41: 820–823. Rheum 2009; 60: 364–371. 40 Gourh P, Agarwal SK, Martin E, Divecha D, Rueda B, Bunting 43 Orozco G, Eyre S, Hinks A, Ke X, Wilson AG, Bax DE et al. H et al. Association of the C8orf13-BLK region with systemic Association of CD40 with rheumatoid arthritis confirmed sclerosis in North-American and European populations. in a large UK case-control study. Ann Rheum Dis 2010; 69: J Autoimmun 2010; 34: 155–162. 813–816. 41 Yin H, Borghi MO, Delgado-Vega AM, Tincani A, Meroni PL, 44 Coenen MJ, Trynka G, Heskamp S, Franke B, van Diemen CC, Alarco´n-Riquelme ME. Association of STAT4 and BLK, but not Smolonska J et al. Common and different genetic background BANK1 or IRF5, with primary antiphospholipid syndrome. for rheumatoid arthritis and coeliac disease. Hum Mol Genet Arthritis Rheum 2009; 60: 2468–2471. 2009; 18: 4195–4203.

Genes and Immunity