and Immunity (2008) 9, 259–263 & 2008 Nature Publishing Group All rights reserved 1466-4879/08 $30.00 www.nature.com/gene

ORIGINAL ARTICLE IL2RA and IL7RA genes confer susceptibility for in two independent European populations

F Weber1,10, B Fontaine2,3,4,10, I Cournu-Rebeix2,3,4, A Kroner5, M Knop1, S Lutz1,FMu¨ ller-Sarnowski1, M Uhr1, T Bettecken1, M Kohli1, S Ripke1, M Ising1, P Rieckmann5,11, D Brassat6, G Semana7, M-C Babron8,9, S Mrejen2,3,4, C Gout2,3,4, O Lyon-Caen2,3,4, J Yaouanq7, G Edan7, M Clanet6, F Holsboer1, F Clerget-Darpoux8,9,12 and B Mu¨ ller-Myhsok1,12 1Max Planck Institute of Psychiatry, Munich, Germany; 2Fe´de´ration des Maladies du Syste`me Nerveux, Assistance Publique-Hoˆpitaux de Paris, G.H. Pitie´-Salpeˆtrie`re, Paris, France; 3UMR546, INSERM, Paris, France; 4UMR S546, University Pierre et Marie Curie-Paris 6, Paris, France; 5Department of Neurology, University of Wu¨rzburg, Wu¨rzburg, Germany; 6Fe´de´ration de Neurologie, University of Toulouse, Toulouse, France; 7Department of Neurology, CHU Pontchaillou, Rennes, France; 8U535, INSERM, Villejuif, France and 9UMRS535, Univ Paris Sud, Villejuif, France

Multiple sclerosis (MS) is the most common chronic inflammatory neurologic disorder diagnosed in young adults and, due to its chronic course, is responsible for a substantial economic burden. MS is considered to be a multifactorial disease in which both genetic and environmental factors intervene. The well-established human leukocyte antigen (HLA) association does not completely explain the genetic impact on disease susceptibility. However, identification and validation of non-HLA-genes conferring susceptibility to MS has proven to be difficult probably because of the small individual contribution of each of these genes. Recently, associations with two single nucleotide polymorphisms (SNPs) in the IL2RA (rs12722489, rs2104286) and one SNP in the IL7RA gene (rs6897932) have been reported by several groups. These three SNPs were genotyped in a French and a German population of MS patients using the hME assay by the matrix-assisted laser desorption/ionization time of flight technology (Sequenom, San Diego, CA, USA). We show that these SNPs do contribute to the risk of MS in these two unrelated European MS patient populations with odds ratios varying from 1.1 to 1.5. The discovery and validation of new genetic risk factors in independent populations may help toward the understanding of MS pathogenesis by providing valuable information on biological pathways to be investigated. Genes and Immunity (2008) 9, 259–263; doi:10.1038/gene.2008.14; published online 20 March 2008

Keywords: multiple sclerosis; genetics; -2 receptor a-subunit; interleukin-7 receptor a-subunit; risk factors; genetic predisposition to disease

Introduction descendants which strengthens the hypothesis of genetic factors contributing to MS.2 Multiple sclerosis (MS)—the most common cause of Epidemiological studies performed in twins and chronic neurologic disability beginning in early to siblings clearly indicate the impact of genetic factors.3,4 middle adult life—is considered to be a multifactorial In addition, the geographical distribution of MS and disease involving both genetic and environmental studies of migrants also suggest a considerable contribu- factors. Evidence of genetic contribution to MS suscept- tion of environmental factors that may not be evenly ibility arose from familial aggregation, twin studies and distributed over the world. high incidence in populations of northern European Current evidence indicates that there is not only one origin, when compared with African and Asian groups.1 gene conferring susceptibility, but rather a number of The prevalence pattern may be explained, at least in part, genes—each contributing only a small amount.5 A large by migrations of northern European groups and their number of studies performed so far confirmed HLA- DRB1 on 6p21 to represent a major risk 6 Correspondence: Professor Dr med F Weber, Max Planck Institute factor for MS in all populations tested. However, it does of Psychiatry, Kraepelinstr. 2-10, Munich D-80804, Germany. not fully explain the genetic basis of the disease.7,8 Other E-mail: [email protected] loci outside the human leukocyte antigen (HLA) region 10These are the equal first authors. were suspected by the results of several candidate genes 11 Current address: MS Clinic, Division of Neurology, UBC Hospital, or genome-wide approaches, but could not be validated Vancouver, Canada. 12These are the equal last authors. in other samples—until recently. Variations in the Received 10 December 2007; revised 5 February 2008; accepted 7 interleukin-2 receptor-a gene (IL2RA) and the interleu- February 2008; published online 20 March 2008 kin-7 receptor-a chain (IL7RA) were demonstrated to IL2RA and IL7RA as risk factors for MS F Weber et al 260 confer a risk for MS in candidate gene approaches and a positive trend in the French trios (Table 2). Linkage whole genome screen.9–11 disequilibrium between these two SNPs is strong (D0 ¼ 1 The association of MS with IL7RA was originally in the German cases, German controls, and French case reported in a candidate gene approach by Zhang et al.12 2 , with D0 estimated at 0.989 in the French years ago, and is now replicated on large additional control chromosomes), but far from perfect (r2 ¼ 0.500, samples.9–11 Interestingly, linkage to chromosome 5p13, 0.552, 0.555 and 0.417, for the above-mentioned samples, precisely where IL7RA is located, was also reported in a respectively). The SNP rs6897932 located within the Canadian MS cohort with a maximum logarithm (base IL7RA gene exhibits a significant association in the 10) of odds score of 4.2413 and supported later by French trios and a positive trend in the German sample. others.14 With respect to IL2RA, genome-wide linkage Combining the results of both samples a significant P- analysis has shown potential linkage on chromosome value is obtained (Table 2). 10p15, which contains the IL2RA gene.15 Subsequently, a candidate approach on this gene first suggested an association between MS and IL2RA.16 This result was strongly supported by the whole genome approach on Discussion extremely large samples.11 This study reports the results obtained in two The identification of non-HLA genes conferring suscept- independent European MS samples (French and Ger- ibility for MS is difficult. Linkage analysis in nuclear man) by genotyping the three single nucleotide poly- families failed to identify additional loci with genome- morphisms (SNPs) showing the strongest association wide significance,8,17 but highlighted the IL2R and IL7R with MS in the study of the International Multiple regions13–15 and the involvement of these two genes were Sclerosis Genetics Consortium.11 Two SNPs are within suggested by candidate gene approaches.12,16 Recent the IL2RA gene (rs12722489, rs2104286) and one in the work9–11 on extremely large samples rendered IL2RA IL7RA gene (rs6897932). and IL7RA highly attractive and motivated further investigations. The risks associated to the SNPs of IL2RA and IL7RA however are modest at best (odds ratios Results between 1.1 and 1.5), so that only replications in a large number of different samples will make their involvement The three SNPs (rs12722489, rs2104286 and rs6897932) in MS susceptibility convincing. were typed in the German sample, consisting of 206 MS We confirm here the positive association with the same patients and 605 age and sex-matched controls, and in risk alleles within the IL2RA and the IL7RA genes in two the French sample, consisting of 540 MS trios. Basic independent European samples. The odds ratios are demographics of both samples are given in Table 1. A consistent with those previously published.9–11 The two statistical comparison showed significant difference SNPs typed in the IL2RA gene are in strong, but not between the samples for age at time of analysis, disease perfect, linkage disequilibrium, which could explain the course and Expanded Disability Status Scale (EDSS). observed difference in P-values and odds ratios ob- However, these differences do not alter the overall tained. Therefore, our findings convincingly corroborate genetic findings discussed below. the recently published results. Risk allele frequency (RAF) for the three SNPs, given In the meantime, another independent replication on a in Table 2, were estimated from the German control Canadian sample18 published similar results. In addition, sample and from the French untransmitted allele sample, association of IL2RA with MS susceptibility was also respectively. No significant difference was found be- recently concluded by Matesanz et al.16 with other SNPs tween our two populations, and the RAFs are very located within the IL2RA gene. Further studies on this similar to that of the original report.11 gene are needed to make a link between those different We first estimated the expected power of our two signals.19 In addition, it will be very important to take samples, under two scenarios (see Materials and meth- advantage of both the linkage and association informa- ods section) represented by the allele frequencies (0.75 tion for a better understanding and a refined modeliza- for the two first SNPs and 0.85 for the third one) and the tion of the effects of both the IL2RA and IL7RA genes.20,21 corresponding odds ratio (1.2 and 1.25 respectively) Together with the well-known HLA association these given in the original report. In both scenarios, the results fit very well the assumption that MS is an expected power for the French trio sample and the autoimmune disease caused by lymphocytes and mono- German sample was 56 and 38%, respectively. This cytes infiltrating the central nervous system.22 Both clearly demonstrates the interest in the joint analysis of cytokines are of central importance in promoting the samples. growth and differentiation of T and B cells. IL-2 is an Since our aim is to replicate the signals found in other essential T-cell growth factor required for division of populations, all the P-values reported here are based on a antigen-activated T cells. Activation of T cells induces the one-sided test. In the trios, 213 transmissions were synthesis of IL2RA and the formation of the high-affinity informative for rs12722489, 355 were informative for IL-2 receptor, allowing the cell to respond to very low rs2104286 and 389 were informative for rs6897932, concentrations of IL-2. IL-7 is important for memory T respectively. The SNP rs2104286 within the IL2RA gene cells and the development of g/D T cells, which are demonstrated a significant association in the French MS present in inflammatory lesions of MS patients.23,24 The trios and the German sample—whether analysed sepa- a-chain of its receptor expresses differently according to rately or combined in both samples (Table 2). The SNP the IL7RA genotypes.25 Interestingly, the likely causal rs12722489, also located within the IL2RA gene, shows SNP in the IL7RA gene has been shown to be responsible significant association in the German sample and a for increased skipping, which results in the

Genes and Immunity IL2RA and IL7RA as risk factors for MS F Weber et al 261 Table 1 Demographic characteristics of the patients

French German P-value

Patients nos. 540 206

Sex 0.08 (95 % lower bound) Female–male ratio 2.29:1 1.68:1

Age (year) At time of analysis o 1 Â10À6

Mean 41.65 37.8 -value Odds ratio Range 19–67 19–59 P

At onset of disease (year) ND Mean 26.45 ND Range 9–47 ND

Disease course no. (%) 4.5 Â 10À10

Relapsing remitting 340 (63) 111 (54) -values and 95% confidence intervals are one-sided, Secondary progressive 86 (16) 42 (20) P Primary progressive 61 (11) 2 (1) Unknown 53 (10) 51 (25) (95 % lower bound)

Expanded disability status scale 5.8 Â 10À9 Score no. (%)

o3 217 (40) 85 (41) -value Odds ratio 3too6 154 (29) 54 (26) P 6 54 (10) 8 (4) 6.5 18 (3) 8 (4) X7 49 (9) 2 (1) Unknown 48 (9) 49 (24) RAF

Abbreviations: ND, not determined; nos., numbers. P-values are given for a w2 contingency table test for sex distribution, disease course and EDSS and for a t-test for age at time of analysis. chain; RAF, risk allele frequency. production of more soluble IL-7 receptor than those of (95 % lower bound) the membrane-bound isoform.10 a Only marginal effect of these genes has been tested. They may interact with other genes. Even if, as under- 26 lined in the elegant editorial of Peltonen, these two -value Odds ratio P genes do not provide clues for unsuspected pathways, focusing on genes which are biologically interacting with IL2RA and IL7RA will be the next step. As stated by 27 Marrosu, ‘the implication of in suscept- RAF ibility to MS might lead to new avenues of investigation’. Taking into account the complexity of MS including the participation of other immune cells, for example monocytes, other cytokines, adhesion molecules, growth factors and neurodegenerative processes, there is a gene; IL7RA, interleukin-7 receptor- high probability that additional pathways are involved a in pathogenesis. Unravelling the MS mystery is just beginning.

Materials and methods Patients and controls Demographic features of patients and controls are given in Table 1. All patients were diagnosed as MS according to the criteria of Poser et al.28 or McDonald et al.29 In total, 206 German MS patients were recruited at the Max Planck Institute of Psychiatry, Munich and at the Association of IL2RA and IL7RA with MS in two independent European populations Department of Neurology, University of Wu¨ rzburg. In total, 605 healthy controls matched for ethnicity, age and -values, odds ratios and 95 % confidence intervals were calculated separately for each population and after combining both samples. Both Chromosome Gene Marker Risk allele French German Combined Table 2 sex were recruited at the Max Planck Institute of 10p1510p155p13Abbreviations: IL2RA IL2RA, interleukin-2 receptor- P IL2RAthat rs12722489 is IL7RA the 95% lower rs2104286 bound for the odds C ratio rs6897932 is given. T C 0.874 0.755 0.0974 0.742 0.00018 0.0385 1.194 (0.95) 1.47 (1.22) 1.19 (1.01) 0.852 0.752 0.041 0.746 0.022 0.17 1.35 (1.02) 1.32 (1.05) 1.14 (0.91) 0.065 0.000215 0.035 1.17 1.33 (0.99) (1.17) 1.15 (1.01)

Genes and Immunity IL2RA and IL7RA as risk factors for MS F Weber et al 262 Psychiatry, Munich. Controls were selected randomly frequency of 0.75 for the associated allele and a relative from a Munich-based community sample and screened risk of 1.2, second with an allele frequency of 0.85 and a for the absence of anxiety and affective disorders using relative risk of 1.25. the Composite International Diagnostic-Screener.30 His- tory of somatic and neurological disorders of the subjects and their first-degree relatives was recorded using a self- report questionnaire. Only individuals negative in the Acknowledgements screening questions for psychiatric disorders and with a negative personal and family history for MS were We thank the members of REFGENSEP, a national included. All participants underwent a brief physical network for the study of MS genetics in France for their examination, including blood analysis and drug mon- active contribution and for referring patients. REFGEN- itoring. In total, 540 MS trios (case and their parents) SEP is supported by grants from INSERM, AFM, ARSEP, were provided by the French network dans le cadre du Groupe Malakoff and Biogen-Idec, and received help re´seau franc¸ais d’e´tude ge´ne´tique de la SEP (REFGEN- from Genethon and CIC Pitie´-Salpeˆtrie`re. The DNA bank SEP). Cases were reviewed by a board-certified neuro- from Wu¨ rzburg was supported by a grant from the logist and diagnosed with criteria referred to above. Only Gemeinnu¨ tzige Hertie Stiftung (GHS). We gratefully cases with a defined diagnosis were included. Patients acknowledge expert technical assistance of S Damast, S originated from all over France even if most of the Sauer and M Koedel. samplings were performed in three centres (Paris, Rennes and Toulouse). Parents were asked about their medical history before the blood puncture was per- formed. The study was approved by the ethics commit- References tee of the Ludwig Maximilian University (Munich, 1 Rosati G. The prevalence of multiple sclerosis in the world: an Germany), the Julius Maximilians University (Wu¨ rzburg, update. Neurol Sci 2001; 22: 117–139. Germany) and the Groupe Hospitalier Pitie´-Salpeˆtrie`re 2 Sotgiu S, Pugliatti M, Sotgiu A, Sanna A, Rosati G. Does the (Paris, France) respectively, and performed according to ‘hygiene hypothesis’ provide an explanation for the high European Community, German and French Bioethics prevalence of multiple sclerosis in Sardinia? Laws. Written informed consent was obtained from all 2003; 36: 257–260. subjects according to a protocol validated by ethics 3 Dyment DA, Willer CJ, Scott B, Armstrong H, Ligers A, Hillert committees. J et al. Genetic susceptibility to MS: a second stage analysis in Canadian MS families. Neurogenetics 2001; 3: 145–151. Sample preparation and statistical analysis 4 Sadovnick AD, Ebers GC, Dyment DA, Risch NJ. Evidence for genetic basis of multiple sclerosis. The Canadian Collaborative DNA was extracted from peripheral blood leucocytes Study Group. Lancet 1996; 347: 1728–1730. and concentration adjusted using standard protocols. 5 Lindsey JW. Familial recurrence rates and genetic models of SNP genotyping was performed with the hME assay by multiple sclerosis. Am J Med Genet A 2005; 135: 53–58. the matrix-assisted laser desorption/ionization time of 6 Oksenberg JR, Barcellos LF. Multiple sclerosis genetics: flight technology (Sequenom, San Diego, CA, USA) as leaving no stone unturned. Genes Immun 2005; 6: 375–387. described.31 All three SNPs were genotyped together 7 Haines JL, Terwedow HA, Burgess K, Pericak-Vance MA, with SNPs of a completely different project in high Rimmler JB, Martin ER et al. Linkage of the MHC to familial multiplexed iPlexes. Both IL2RA SNPs (rs12722489, multiple sclerosis suggests genetic heterogeneity. The Multiple rs2104286) were part of an iPlex containing 28 SNPs. Sclerosis Genetics Group. Hum Mol Genet 1998; 7: 1229–1234. 8 Hillert J. Multiple sclerosis: postlinkage genetics. Clin Neurol The IL7RA SNP rs6897932 was in an iPlex comprising 10 Neurosurg 2006; 108: 220–222. SNPs. All subjects’ DNAs were typed on eight different 9 Lundmark F, Duvefelt K, Iacobaeus E, Kockum I, Wallstrom E, 384-well plates. The mean call rates (standard) over all Khademi M et al. Variation in receptor alpha eight plates per SNP were 99.8% (0.3%) for rs12722489, chain (IL7R) influences risk of multiple sclerosis. Nat Genet 99.3% (0.9%) for rs2104286 and 98.8% (1.8%) for 2007; 39: 1108–1113. rs6897932. All three SNPs were in Hardy–Weinberg 10 Gregory SG, Schmidt S, Seth P, Oksenberg JR, Hart J, Prokop A equilibrium (no P-value below 0.05) on each single plate. et al. Interleukin 7 receptor alpha chain (IL7R) shows allelic PCR and extension primer sequences may be obtained and functional association with multiple sclerosis. Nat Genet upon request. 2007; 39: 1083–1091. Demographic data were tested for differences using 11 The International Multiple Sclerosis Genetics Consortium. 2 Risk alleles for multiple sclerosis identified by a genomewide w -analysis of contingency tables. study. N Engl J Med 2007; 357: 851–862. Allele frequencies were estimated from the German 12 Zhang Z, Duvefelt K, Svensson F, Masterman T, Jonasdottir G, control sample and from the French untransmitted allele Salter H et al. Two genes encoding immune-regulatory sample. molecules (LAG3 and IL7R) confer susceptibility to multiple Statistical analysis was performed using the UN- sclerosis. Genes Immun 2005; 6: 145–152. PHASED Software v 3.0.10. which allows joint analysis 13 Ebers GC, Kukay K, Bulman DE, Sadovnick AD, Rice G, of data from trios and unrelated individuals.32 As in the Anderson C et al. A full genome search in multiple sclerosis. original report,11 we used the multiplicative (log addi- Nat Genet 1996; 13: 472–476. tive) model for risk. Since our aim is to replicate the 14 Oturai A, Larsen F, Ryder LP, Madsen HO, Hillert J, Fredrikson S et al. Linkage and association analysis of signals found in other populations, all the P-values susceptibility regions on chromosomes 5 and 6 in 106 reported are based on a one-sided test. Power for a one- Scandinavian sibling pair families with multiple sclerosis. sided test with a ¼ 0.05 was calculated with Genetic Ann Neurol 1999; 46: 612–616. Power Calculator (http://pngu.mgh.harvard.edu/~pur- 15 Akesson E, Coraddu F, Marrosu MG, Massacesi L, Hensiek A, cell/gpc/) under two scenarios, assuming first an allele Harbo HF et al. Refining the linkage analysis on chromosome

Genes and Immunity IL2RA and IL7RA as risk factors for MS F Weber et al 263 10 in 449 sib-pairs with multiple sclerosis. J Neuroimmunol 25 McKay FC, Swain LI, Schibeci SD, Rubio JP, Kilpatrick TJ, 2003; 143: 31–38. Heard RN et al. Haplotypes of the interleukin 7 receptor alpha 16 Matesanz F, Caro-Maldonado A, Fedetz M, Fernandez O, gene are correlated with altered expression in whole blood Milne RL, Guerrero M et al. IL2RA/CD25 polymorphisms cells in multiple sclerosis. Genes Immun 2008; 9: 1–6. contribute to multiple sclerosis susceptibility. J Neurol 2007; 26 Peltonen L. Old suspects found guilty—the first genome 254: 682–684. profile of multiple sclerosis. N Engl J Med 2007; 357: 927–929. 17 Sawcer S. A new era in the genetic analysis of multiple 27 Marrosu MG. Susceptibility to multiple sclerosis: the role of sclerosis. Curr Opin Neurol 2006; 19: 237–241. interleukin genes. Lancet Neurol 2007; 6: 846–847. 18 Ramagopalan SV, Anderson C, Sadovnick AD, Ebers GC. 28 Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Genomewide study of multiple sclerosis. N Engl J Med 2007; Ebers GC et al. New diagnostic criteria for multiple sclerosis: 357: 2199–2200. guidelines for research protocols. Ann Neurol 1983; 13: 19 Matesanz F, Fernandez O, Alcina A. Genomewide study of 227–231. multiple sclerosis. N Engl J Med 2007; 357: 2200–2201. 29 McDonald WI, Compston A, Edan G, Goodkin D, Hartung 20 Bourgain C, Genin E, Cox N, Clerget-Darpoux F. Are genome- HP, Lublin FD et al. Recommended diagnostic criteria for wide association studies all that we need to dissect the genetic multiple sclerosis: guidelines from the international panel on component of complex human diseases? Eur J Hum Genet the diagnosis of multiple sclerosis. Ann Neurol 2001; 50: 2007; 15: 260–263. 121–127. 21 Clerget-Darpoux F, Elston RC. Are linkage analysis and the 30 Robins LN, Wing J, Wittchen HU, Helzer JE, Babor TF, Burke J collection of family data dead? Prospects for family studies in et al. The Composite International Diagnostic Interview. An the age of genome-wide association. Hum Hered 2007; 64: 91–96. epidemiologic Instrument suitable for use in conjunction with 22 Hafler DA, Slavik JM, Anderson DE, O’Connor KC, De Jager different diagnostic systems and in different cultures. Arch P, Baecher-Allan C. Multiple sclerosis. Immunol Rev 2005; 204: Gen Psychiatry 1988; 45: 1069–1077. 208–231. 31 Binder EB, Salyakina D, Lichtner P, Wochnik GM, Ising M, 23 Seddon B, Tomlinson P, Zamoyska R. Interleukin 7 and Putz B et al. Polymorphisms in FKBP5 are associated with receptor signals regulate homeostasis of CD4 memory cells. increased recurrence of depressive episodes and rapid Nat Immunol 2003; 4: 680–686. response to antidepressant treatment. Nat Genet 2004; 36: 24 He YW, Malek TR. Interleukin-7 receptor alpha is essential for 1319–1325. the development of gamma delta + T cells, but not natural 32 Dudbridge F. Pedigree disequilibrium tests for multilocus killer cells. J Exp Med 1996; 184: 289–293. haplotypes. Genet Epidemiol 2003; 25: 115–121.

Genes and Immunity