You are here

Article Text

Article menu
Clinical and epidemiological research
Extended report
The TRAF1-C5 region on chromosome 9q33 is associated with multiple autoimmune diseases
  1. Fina A S Kurreeman1,
  2. George N Goulielmos2,
  3. Behrooz Z Alizadeh3,
  4. Blanca Rueda4,
  5. Jeanine Houwing-Duistermaat5,
  6. Elena Sanchez4,
  7. Marianna Bevova3,
  8. Timothy R Radstake6,
  9. Madelon C Vonk6,
  10. Emmanouil Galanakis2,
  11. Norberto Ortego7,
  12. Willem Verduyn8,
  13. Maria I Zervou2,
  14. Bart O Roep8,
  15. Barbara Dema9,
  16. Laura Espino9,
  17. Elena Urcelay9,
  18. Dimitrios T Boumpas2,
  19. Leonard H van den Berg10,
  20. Cisca Wijmenga3,11,
  21. Bobby P C Koeleman3,
  22. Tom W J Huizinga1,
  23. Rene E M Toes1,
  24. Javier Martin4
  25. the AADEA Group and SLEGEN Consortium
  1. 1Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands
  2. 2Department of Medicine, University of Crete, Greece
  3. 3Department of Biomedical Genetics, Utrecht University Medical Center, Utrecht, The Netherlands
  4. 4Instituto de Biomedicina López-Neyra, CSIC, Granada, Spain
  5. 5Department of Medical Statistics, Leiden University Medical Center, Leiden, The Netherlands
  6. 6Department of Rheumatology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands
  7. 7Department of Internal Medicine, Hospital Clínico San Cecilio, Granada, Spain
  8. 8Department of Immunohaematology and Bloodbank, Leiden University Medical Center, Leiden, The Netherlands
  9. 9Department of Immunology, Hospital Clínico San Carlos, Madrid, Spain
  10. 10Department of Neurology, Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands
  11. 11Department of Human genetics, Groningen University Medical Center and University of Groningen, Groningen, The Netherlands
  1. Correspondence to Rene E M Toes, Department of Rheumatology, C1R, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands; r.em.toes{at}lumc.nl

Abstract

Objectives The TRAF1-C5 locus has recently been identified as a genetic risk factor for rheumatoid arthritis (RA). Since genetic risk factors tend to overlap with several autoimmune diseases, a study was undertaken to investigate whether this region is associated with type 1 diabetes (TID), celiac disease (CD), systemic sclerosis (SSc) and systemic lupus erythematosus (SLE).

Methods The most consistently associated SNP, rs10818488, was genotyped in a total of 735 patients with T1D, 1049 with CD, 367 with SSc, 746 with SLE and 3494 ethnically- and geographically-matched healthy individuals. The replication sample set consisted of 99 patients with T1D, 272 with SLE and 482 healthy individuals from Crete.

Results A significant association was detected between the rs10818488 A allele and T1D (OR 1.14, p=0.027) and SLE (OR 1.16, p=0.016), which was replicated in 99 patients with T1D, 272 with SLE and 482 controls from Crete (OR 1.64, p=0.002; OR 1.43, p=0.002, respectively). Joint analysis of all patients with T1D (N=961) and all patients with SLE (N=1018) compared with 3976 healthy individuals yielded an allelic common OR of 1.19 (p=0.002) and 1.22 (p=2.6×10−4), respectively. However, combining our dataset with the T1D sample set from the WTCCC resulted in a non-significant association (OR 1.06, p=0.087). In contrast, previously unpublished results from the SLEGEN study showed a significant association of the same allele (OR 1.19, p=0.0038) with an overall effect of 1.22 (p=1.02×10−6) in a total of 1577 patients with SLE and 4215 healthy individuals.

Conclusion A significant association was found for the TRAF1-C5 locus in SLE, implying that this region lies in a pathway relevant to multiple autoimmune diseases.

https://doi.org/10.1136/ard.2008.106567

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    The region encompassing tumour necrosis factor (TNF) receptor-associated factor 1 (TRAF1) and complement component 5 (C5) has recently been reported to be a genetic risk factor involved in rheumatoid arthritis (RA).1 The robustness of this association is demonstrated by its prevalent risk in Dutch, Swedish, Crete2 and American populations, and is further corroborated by a genome-wide association study3 and an extensive fine-mapping study which we have undertaken.4 Interestingly, one consistent RA association signal defined by rs10818488 or its perfect proxy (rs3761847 and rs7021049, linkage disequilibrium R2 >0.98) has been identified in this region by three independent studies and has been further confirmed by a large European family-based study.5 Moreover, an association of the opposite allele of rs3761847 in RA has been reported in the Japanese population but no association has been found in a Korean population.6 7 In addition, we and others have shown that this polymorphism is associated with juvenile idiopathic arthritis (JIA),8 9 indicating its possible role in other autoimmune diseases.

    From a functional perspective, TRAF1 is likely to be a negative regulator of TNF-receptor signalling10 and C5 is a central component of the complement pathway.11 Both molecules are potent immune mediators and, so far, the question remains whether this region is restricted to arthritis as such or whether it lies in a biological pathway common to other autoimmune diseases. In the present study we aimed to investigate this hypothesis further by considering the role of this locus in four diseases including type 1 diabetes (T1D), celiac disease (CD), systemic sclerosis (SSc) and systemic lupus erythematosus (SLE).

    Methods

    Sample sets

    DNA was obtained from cohorts of patients with T1D, CD, SSc and SLE from The Netherlands and Spain as well as patients with T1D and SLE from Crete. Control sets were matched to each specific population of origin.

    The T1D sample set consisted of 556 white Dutch and 306 Spanish patients, the CD sample set comprised 496 Spanish and 553 Dutch patients, the SSc sample set consisted of 138 Dutch and 229 Spanish patients and the SLE sample set included 161 Dutch and 585 Spanish patients. All patients were Caucasian. A common set of controls was used consisting of 1396 Dutch and 2098 Spanish geographically- and ethnically-matched healthy individuals.

    The replication sample set consisted of 272 patients with SLE, 99 with T1D and 482 geographically- and ethnically-matched controls originating from Crete.

    All patients were diagnosed using the appropriate classification criteria and are described extensively in the online supplement.

    Genome-wide association study data

    Genotype counts were obtained for rs10118357 (R2=1 with rs10818488) from the publicly available Wellcome Trust Case Control Consortium (WTCCC) study12 consisting of 1960 patients with T1D and 2930 ethnically-matched healthy individuals from the UK population. For the SLE sample set, genotype counts were obtained from the SLE Genetics consortium (SLEGEN) study13 for rs3761847 (r2=1 with rs10818488) from 720 patients with SLE and 2337 healthy individuals. These sample sets are described in further detail elsewhere.12 13

    Genotyping

    All samples were genotyped using the Taqman assay (Applied Biosystems, Foster City, California, USA) according to the manufacturer's instructions. Each plate consisted of at least eight positive and eight negative controls. At least 10% of the samples were genotyped in duplicate with no discrepancies observed. The Dutch control set consisted of 511 healthy individuals who had been previously genotyped.1 An additional 535 Dutch healthy unrelated individuals were genotyped using allele-specific kinetic PCR14 as well as 715 Dutch controls from Utrecht using the Taqman assay. To compare genotyping methods, at least 50 samples were genotyped on each platform (Taqman, allele-specific kinetic PCR and RFLP) and revealed a concordance rate of >99%.

    Statistical analysis

    In the controls, the frequencies were in Hardy–Weinberg equilibrium as determined by the observed versus expected genotype counts. Genotype counts were analysed using SPSS Version 12.0. Odds ratios and CIs were calculated using Statcalc. Combining odds ratios across sample sets was performed using the Cochran–Mantel–Haenszel test as implemented in EasyMA and the meta package in R.15 No evidence of heterogeneity of risk effect was observed using the Breslow and Day method (p>0.05) when combining OR. All power calculations were performed using Quanto version 1.2 (http://hydra.usc.edu/gxe). p Values <0.05 were considered significant.

    Results

    T1D, CD, SSc and SLE samples were genotyped from both the Spanish and Dutch populations for the strongest and most consistent association signal in the TRAF1-C5 region characterised by SNP rs10818488. Since the Dutch study was largely underpowered to detect modest effect sizes (OR ~1.2), we opted for a combined analysis which consisted of a total of 3494 controls and 735 patients with T1D, 1049 with CD, 367 with SSc and 746 with SLE. This combined dataset enhanced the power to ≥85% to detect effect sizes of 1.2 at p<0.05, with the exception of SSc which only achieved 64% power to detect an odds ratio of 1.2.

    The frequency of the rs10818488 A allele was significantly increased in patients with T1D (OR 1.14; 95% CI 1.02 to 1.28; p=0.027; table 1). Similarly, we found a significant difference in the prevalence of the A allele in patients with SLE, resulting in an OR of 1.16 (95% CI 1.03 to 1.31, p=0.016). Patients harbouring two copies of the A allele had a 1.3-fold (p=0.01) and a 1.4-fold (p=0.04) increased risk for T1D and SLE, respectively, compared with those who carried none. We observed no association with CD (allelic OR 1.07, 95% CI 0.97 to 1.18, p=0.18) and SSc (allelic OR 1.02, 95% CI 0.87 to 1.19, p=0.84). While the possibility remains that the absence of an association in SSc may be due to lack of power to detect an OR of ≤1.2, the direction of association in the individual studies was opposite (Spanish: allelic OR 1.16, 95% CI 0.94 to 1.43; Dutch: allelic OR 0.85, 95% CI 0.65 to 1.09), suggesting that the absence of association may be more likely.

    View this table:
    Table 1

    Association of rs10818488 in the TRAF1-C5 locus with autoimmune diseases

    We then proceeded to replicate our significant findings in T1D and SLE in the genetically homogeneous population of Crete. Since this largest island of Greece consists of 650 000 inhabitants who share the same genetic and cultural background as well as a common environment, it represents a ‘geographically isolated’ gene pool which may enhance the detection of risk alleles that may be diluted in larger continental populations.16 We observed an 11% increase and an 8% increase in the A allele in T1D and SLE patients, respectively, compared with controls (table 2). This resulted in a 1.6-fold increased risk for T1D (p=0.002) and a 1.4-fold increased risk for SLE (p=0.002) in the Crete population. Overall analysis of T1D and SLE in all three datasets (Spanish, Dutch and Greek) reveals a common OR of ~1.2 (p=0.002 and p=2.6×10−4, respectively) for both diseases.

    View this table:
    Table 2

    Meta-analysis of rs10818488 in the TRAF1-C5 locus with genome-wide association studies

    We also analysed our data in combination with rs10118357 (perfect proxy of rs10818488, R2=1) for T1D obtained from the WTCCC study12 as well as a meta-analysis of the effect of SNP rs3761847 (perfect proxy of rs10818488, R2=1) in the SLEGEN study13 (table 2). The WTCCC T1D dataset comprising 1960 patients and 2930 healthy individuals showed no association with an OR of 0.99 (95% CI 0.91 to 1.08, p=0.837). Combining all four datasets (2794 patients and 6906 controls) generated an overall OR of 1.06 (95% CI 0.99 to 1.13, p=0.087) indicating that the effect size is very modest and that our study is largely underpowered to detect a significant effect. Previously unpublished data from the SLEGEN study, in contrast, show a significant increase in the rs10818488 A allele frequency in 720 patients with SLE (43%) compared with 2337 healthy individuals (39%) (OR 1.19, 95% CI 1.06 to 1.35, p=0.0038), independently replicating the association of this locus with SLE. Combining all datasets consisting of 1577 patients with SLE and 4215 controls showed a significant association with OR 1.22 (95% CI 1.12 to 1.31, p=1.02×10−6).

    Discussion

    We report here for the first time reproducible association of the TRAF1-C5 region with SLE, complementing the already consistent finding of this variant with RA and JIA. Remarkably, the same allele that predisposes to RA and JIA also predisposes to SLE, lending support to the hypothesis that this region may contribute to a shared pathway involved in RA, JIA and SLE. While the possibility also remains that other additional alleles at this locus may be involved in these diseases, complementary studies undertaking further fine mapping as well as sequencing will yield further insight into the most likely causal alleles at this locus.

    We also observed a difference in allele frequencies in healthy populations of Dutch (44%), Spanish (36%) and Greek (29%) origins. To address this difference in population, each patient-control sample set was geographically and ethnically matched. Data from HapMap (www.hapmap.org) support these observations, with the G allele (minor allele in Caucasians) frequency of rs3761847 (perfect proxy of rs10818488) varying between 48% in Caucasians of European descent, 31% in Gujarati Indians in Texas, 42% in Japanese in Tokyo, 66% in Mexicans in Los Angeles, 59% in Yorubans in Nigeria and 74% in Luhya in Kenya. However, underlying population stratification is difficult to account for as panels of markers thoroughly characterising each population in our study have not yet been described. Association of the TRAF1-C5 locus was observed in more than one sample set, considerably reducing the chances of false-positive findings due to population stratification.

    The association of the TRAF1-C5 region was not reported in the recently published genome-wide association studies (GWAS) in SLE13 17 18 and the GWAS of T1D in the WTCCC study.12 However, while an association with RA was initially not detected in the WTCCC study, a follow-up report showed a significant association of the same allele in the UK population, although with a lower effect size.19 In line with this observation, a meta-analysis of GWAS in RA has recently shown that much larger sample sizes are required to detect the commonly observed low penetrance of genes in autoimmune diseases like RA.20 It is also not uncommon that modest effects are not detected in such large-scale studies as exemplified by the association of TNFSF4, a gene identified by family-based association studies,21 that did not surface in the recently performed GWAS. Likewise BANK1, a gene identified in the GWA scan by Kozyrev et al,18 was not identified by the other two GWA studies that employed 500K SNPs. However, while our study combined with the GWAS of the WTCCC does not provide sufficient evidence of the role of the TRAF1-C5 locus in T1D, data obtained from the SLEGEN study confirms our findings of this locus in SLE. Interestingly, this association is absent in a well-powered Japanese case-control study and a small study in the Columbian population.6 22 As with many genetic loci, it is highly likely that ethnic differences exist in the contribution of this locus to SLE.

    TRAF1 has been reported to be a negative regulator of the TNF receptor signalling cascade10 23 and high levels of TNF have been detected in both human and murine SLE. It is therefore likely that dysregulation of the function and/or expression of this molecule could be involved in the inflammatory processes in SLE.10 24 Controlled clinical trials involving the blockade of TNF are currently being conducted.25 Although the haplotype block encompassing rs10818488 does not encompass the C5 coding region,4 this central component of the complement system is also a likely functional candidate in SLE. The prominent role of complement activation in humans and murine models of SLE, as well as the beneficial effect of blocking C5 anaphylatoxin in murine lupus models, point to the likely role of this molecule in SLE.26 27

    In summary, we report here for the first time an association of the TRAF1-C5 locus with SLE which, in combination with previous findings of an increased risk of RA and JIA, indicates that this region is likely to be part of a shared mechanism underlying several autoimmune diseases.

    References

    1. ↵
      1. Kurreeman FA,
      2. Padyukov L,
      3. Marques RB,
      4. et al
      . A candidate gene approach identifies the TRAF1/C5 region as a risk factor for rheumatoid arthritis. PLoS Med 2007;4:e278.
      OpenUrlCrossRefPubMed
    2. ↵
      1. Zervou MI,
      2. Sidiropoulos P,
      3. Petraki E,
      4. et al
      . Association of a TRAF1 and a STAT4 gene polymorphism with increased risk for rheumatoid arthritis in a genetically homogeneous population. Hum Immunol 2008;69:567–71.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Plenge RM,
      2. Seielstad M,
      3. Padyukov L,
      4. et al
      . TRAF1-C5 as a risk locus for rheumatoid arthritis – a genomewide study. N Engl J Med 2007;357:1199–209.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Chang M,
      2. Rowland CM,
      3. Garcia VE,
      4. et al
      . A large-scale rheumatoid arthritis genetic study identifies association at chromosome 9q33.2. PLoS Genet 2008;4:e1000107.
    5. ↵
      1. Kurreeman FA,
      2. Rocha D,
      3. Houwing-Duistermaat J,
      4. et al
      . Replication of the tumor necrosis factor receptor-associated factor 1/complement component 5 region as a susceptibility locus for rheumatoid arthritis in a European family-based study. Arthritis Rheum 2008;58:2670–4.
      OpenUrlCrossRefPubMed
    6. ↵
      1. Nishimoto K,
      2. Kochi Y,
      3. Ikari K,
      4. et al
      . Association study of TRAF1-C5 polymorphisms with susceptibility to rheumatoid arthritis and systemic lupus erythematosus in Japanese. Ann Rheum Dis 2010;69:368–73.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Lee HS,
      2. Korman BD,
      3. Le JM,
      4. et al
      . Genetic risk factors for rheumatoid arthritis differ in Caucasian and Korean populations. Arthritis Rheum 2009;60:364–71.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Behrens EM,
      2. Finkel TH,
      3. Bradfield JP,
      4. et al
      . Association of the TRAF1-C5 locus on chromosome 9 with juvenile idiopathic arthritis. Arthritis Rheum 2008;58:2206–7.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Albers HM,
      2. Kurreeman FA,
      3. Houwing-Duistermaat JJ,
      4. et al
      . The TRAF1/C5 region is a risk factor for polyarthritis in juvenile idiopathic arthritis. Ann Rheum Dis 2008;67:1578–80.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Wajant H,
      2. Henkler F,
      3. Scheurich P
      . The TNF-receptor-associated factor family: scaffold molecules for cytokine receptors, kinases and their regulators. Cell Signal 2001;13:389–400.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Schmidt RE,
      2. Gessner JE
      . Fc receptors and their interaction with complement in autoimmunity. Immunol Lett 2005;100:56–67.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      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.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Harley JB,
      2. Alarcón-Riquelme ME,
      3. Criswell LA,
      4. et al
      . Genome-wide association scan in women with systemic lupus erythematosus identifies susceptibility variants in ITGAM, PXK, KIAA1542 and other loci. Nat Genet 2008;40:204–10.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Begovich AB,
      2. Carlton VE,
      3. Honigberg LA,
      4. et al
      . A missense single-nucleotide polymorphism in a gene encoding a protein tyrosine phosphatase (PTPN22) is associated with rheumatoid arthritis. Am J Hum Genet 2004;75:330–7.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Cucherat M,
      2. Boissel JP,
      3. Leizorovicz A,
      4. et al
      . EasyMA: a program for the meta-analysis of clinical trials. Comput Methods Programs Biomed 1997;53:187–90.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Arnaiz-Villena A,
      2. Iliakis P,
      3. González-Hevilla M,
      4. et al
      . The origin of Cretan populations as determined by characterization of HLA alleles. Tissue Antigens 1999;53:213–26.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Hom G,
      2. Graham RR,
      3. Modrek B,
      4. et al
      . Association of systemic lupus erythematosus with C8orf13-BLK and ITGAM-ITGAX. N Engl J Med 2008;358:900–9.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Kozyrev SV,
      2. Abelson AK,
      3. Wojcik J,
      4. et al
      . Functional variants in the B-cell gene BANK1 are associated with systemic lupus erythematosus. Nat Genet 2008;40:211–6.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Barton A,
      2. Thomson W,
      3. Ke X,
      4. et al
      . Re-evaluation of putative rheumatoid arthritis susceptibility genes in the post-genome wide association study era and hypothesis of a key pathway underlying susceptibility. Hum Mol Genet 2008;17:2274–9.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Raychaudhuri S,
      2. Remmers EF,
      3. Lee AT,
      4. et al
      . Common variants at CD40 and other loci confer risk of rheumatoid arthritis. Nat Genet 2008;40:1216–23.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Cunninghame Graham DS,
      2. Graham RR,
      3. Manku H,
      4. et al
      . Polymorphism at the TNF superfamily gene TNFSF4 confers susceptibility to systemic lupus erythematosus. Nat Genet 2008;40:83–9.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Palomino-Morales RJ,
      2. Rojas-Villarraga A,
      3. González CI,
      4. et al
      . STAT4 but not TRAF1/C5 variants influence the risk of developing rheumatoid arthritis and systemic lupus erythematosus in Colombians. Genes Immun 2008;9:379–82.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Tsitsikov EN,
      2. Laouini D,
      3. Dunn IF,
      4. et al
      . TRAF1 is a negative regulator of TNF signaling. enhanced TNF signaling in TRAF1-deficient mice. Immunity 2001;15:647–57.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Aringer M,
      2. Smolen JS
      . The role of tumor necrosis factor-alpha in systemic lupus erythematosus. Arthritis Res Ther 2008;10:202.
      OpenUrlCrossRefPubMed
    25. ↵
      1. Aringer M,
      2. Smolen JS
      . TNF inhibition in SLE: where do we stand? Lupus 2009;18:5–8.
      OpenUrlFREE Full Text
    26. ↵
      1. Manderson AP,
      2. Botto M,
      3. Walport MJ
      . The role of complement in the development of systemic lupus erythematosus. Annu Rev Immunol 2004;22:431–56.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Walport MJ
      . Complement and systemic lupus erythematosus. Arthritis Res 2002;4(Suppl 3):S279–93.
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Web Only Data ard.2008.106567

      Files in this Data Supplement:

    Footnotes

    • GNG and BZA contributed equally.

    • Participants The AADEA (Andalusian Association of Autoimmune Diseases) Group participants are as follows: Juan Jiménez-Alonso (Servicio Medicina Interna, Hospital Virgen de las Nieves, Granada); Julio Sanchez-Román (Servicio de Medicina Interna, Hospital Virgen del Rocio, Sevilla); Enrique De-Ramon and Mayte Camps (Servicio Medicina Interna, Hospital Carlos Haya, Málaga); M Angeles Aguirre (Servicio de Reumatología, Hospital Reina Sofía, Córdoba); Rosa García-Portales (Servicio Medicina Interna, Hospital Virgen de la Victoria, Málaga). The SLEGEN consortium members are as follows: John B Harley, Lindsey A Criswell, Timothy Vyse, Robert Kimberly, Chaim Jacob, Kathy Moser, Carl Langfeldt, Marta E Alarcón-Riquelme and Betty Tsao.

    • Funding This study was supported by grants from the Dutch Arthritis Foundation, grant SAF2006–00398 from the Spanish Ministerio de Educacion y Ciencia, grant CTS1180 from Junta de Andalucía, the European Community's FP6 funding project 018661 Autocure and the Center for Medical Systems Biology in The Netherlands. Support was also obtained from funding for a VIDI grant for REMT and VICI grant 918–66-620 from the Dutch organisation for scientific research (NWO) as well as the Celiac Disease Consortium (an innovative cluster approved by The Netherlands Genomics Initiative and partly funded by the Dutch government (grant BSIK03009)).

    • Competing interests None.

    • Ethics approval Informed consent was obtained from all subjects and the study was approved by the local ethics committee of each centre.

    • Provenance and peer review Not commissioned; externally peer reviewed.