Coeliac disease (CD), or gluten sensitive enteropathy, is a common chronic inflammatory disease with chronic enteropathy. The disease is influenced by environmental factors, such as gluten and breast milk, as well as genetic factors, such as human leucocyte antigen (HLA) class II genes and non-HLA genes (1).
Genetically CD is characterised by a strong HLA component. More than 90% of patients with CD carry the HLA-DQA1*0501 and HLA-DQB1*02 alleles (coding for DQ2) compared with 20% to 30% of controls (2). This HLA class II molecule is either encoded in cis in individuals who have the DR3–DQ2 haplotype, or in trans in individuals who are DR5-DQ7 and DR7-DQ2 heterozygous. It is shown that there is an increased risk for CD among individuals who are DR3–DQ2 homozygous and DR3-DQ2/DR7-DQ2 heterozygous (3).
The possibility of other genes in the HLA complex involved in CD has been studied, but no convincing evidence has been found (4,5). Even if HLA class II genes seem necessary for disease development, they contribute only 40% of the genetic risk, indicating the importance of genes outside the HLA complex.
One of the regions outside the HLA region that has shown repeated association to CD is the CD28-CTLA4-ICOS gene region located on chromosome 2q33. Because CTLA4 plays an important role in maintaining immunological tolerance to self-antigens, this is an interesting candidate gene. Some studies have implicated a functional role of CTLA4. The G allele of the CTLA4 +49 A/G dimorphism has been shown to be associated with reduced control of T cell proliferation (6), and in a large study on type 1 diabetes mellitus (T1DM) and Graves disease different haplotypes were associated with different splice variants (7).
We have previously reported genetic association with CTLA4 +49 A/G, with preferential transmission of the A allele (8). This was also found in a French population (9). In a Finnish study, linkage was found for the marker D2S116, located adjacent to the CTLA4 gene and later association to the ICOS gene (10). In a study on Italian and Tunisian populations, no linkage to the 2q33 region or association with the CTLA4 +49 A/G polymorphism was found (11). However, a later Italian study showed association to the A allele, especially in DQ2-negative patients (12). In a haplotype analysis our group found an association with the MH30*G:−1147*T:+49*A:CT60*G:CT61*A haplotype (13).
Several genome-wide scan studies on CD have shown significant linkage on chromosome 5 (14–17). The 5q31-33 region was shown to be associated with other immunological diseases such as Crohn disease (18), T1DM (19), and asthma (20). Our group found the strongest evidence for linkage, except for the HLA class II region, on chromosome 5q31-33 (21,22). The 5q31-33 region comprises more than 200 coding genes, many of which are of immunological importance. Some have been analysed in CD, for example, IL-4, IL-5, IL-12, IL-13, and IL-14, without significant association (23,24).
The phenotype that is the clinical presentation of CD is widely variable, ranging from dramatic symptoms with diarrhoea, weight loss, and malnutrition to clinically silent or asymptomatic forms. In Sweden, the clinical presentation has been changing in the last few years, from predominantly young children with classical symptoms during the 1980s to adolescents and adults with more diffuse symptoms during the 1990s (25). The same changes were previously also reported in Finland (26). These rapid changes point to the complexity of the interaction among environment, genotype, and phenotype.
Our knowledge of the interaction between genotype and phenotype in CD is limited. Several studies have failed to establish associations between HLA genotypes and phenotypes (27,28). In 2 studies, however, a gene-dose effect was found on some phenotypes (29,30). So far, no studies have been published about interactions between CTLA4 polymorphisms or 5q31-33 loci and the clinical presentation in CD. However, 1 study showed no association between IL-12B and IRF1 genes on 5q31-33 and enteropathy grading according to the Marsh criteria (31).
The aim of this study was to investigate a possible interaction between the phenotypes and the genotypes, HLA class II risk alleles, the CTLA4 +49 A/G polymorphism, the haplotype MH30*G:−1147*T:+49*A:CT60*G:CT61*A, and the 5q31-33 locus in CD. A condition for this analysis is a significant heritability that is the proportion of the variance of a phenotype that is due to genetic factors.
PATIENTS AND METHODS
One hundred and seven families with at least 2 affected children (affected sib pair) were collected from Sweden and southern Norway to perform a genome-wide scan for CD genes (21). The strategy for recruiting families, collecting diagnostic data, and making clinical presentation has been described in detail previously (32).
To avoid inclusion of falsely diagnosed cases, strict inclusion and exclusion criteria were applied. Only families in which all siblings with CD were diagnosed according to the revised European Society for Paediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) criteria (33) were included. Children younger than 2 years at the time of diagnosis were included only if they fulfilled the original ESPGHAN criteria of 3 biopsies on different diets (34). One hundred and seven families with 224 siblings with CD fulfilled the diagnostic criteria and were included in the study. The median age at diagnosis was 2 years (range 1–58 years), whereas the mean age was 6.9 years. The distribution showed a female domination of 2.1:1 as expected (152 females, 72 males) (35).
The ethics committee of the Medical Faculty of Göteborg University approved the study. Informed consent was obtained from participants or their parents.
The clinical presentation of the patients was classified in 3 symptom grades: grade 1 patients with classical CD symptoms such as diarrhoea, vomiting, abdominal distension, malabsorption, and growth failure; grade 2 patients with milder presentation of symptoms typical of CD; and grade 3 patients with clinically silent CD, defined as asymptomatic individuals with positive antibodies and gluten enteropathy. Only 3 patients presented with atypical symptoms and they were included in grade 2 (Table 1). To make the grouping consistent the classification was done by the same person (A.H.G.). In addition to the clinical classification the following variables were recorded: probands that are the first individual diagnosed in a family, age at diagnosis, and sex.
HLA-DRB1, DQA1, and DQB1 genotyping was done by polymerase chain reaction–sequence-specific oligonucleotide (PCR-SSOP), as previously described (36). The serological typing implies DR3-DQ2 (DRB1*03-DQA1*05-DQB1*02), DR7-DQ2 (DRB1*07-DQA1*0201-DQB1*02), DR5-DQ7 (DRB1*11/12-DQA1*05-DQB1*0301), and DR4-DQ8 (DRB1*04-DQA1*03-DQB1*0302). The patients were divided in 3 risk groups according to their HLA class II status: high when a patient had either of the 2 genotypes, DR3-DQ2/DR3-DQ2 or DR3-DQ2/DR7-DQ2, intermittent when DR3-DQ2/X (X is any other genotype) or DR5-DQ7/DR7-DQ2, and low when the patient had any other HLA genotype (36).
At the CTLA4 locus all CD families were previously typed for the single nucleotide polymorphism CTLA4 A/G at position +49 (8). In addition, several single nucleotide polymorphisms surrounding CTLA4 were genotyped including MH30, −1147, CT60, and CT61 as described previously (17).
For 5q31-33, identity by descent (IBD) status of all families was determined at locus D5S436, where the linkage peak was the highest (Zlr = 3.96) (22).
The heritability of the symptom grade was established using the software SOLAR (37). To technically achieve this, we coded all healthy relatives as having unknown symptom grade.
We analysed the possible association between phenotypes and genotypes at the CTLA4 and HLA loci. Only 1 sibling from each family was sampled because phenotypes can be familiar for genetic as well as for non-genetic reasons. To get a balanced group size the siblings with the more rare symptom grade 3 were primarily chosen and then equally many siblings with symptom grades 1 and 2 were sampled. This sampling was repeated 10,000 times and for each sample the χ2 statistic was calculated. The P value corresponding to the median χ2 is reported.
To analyse possible associations between IBD sharing at CTLA4 and 5q31-33 on one hand and the phenotype on the other, we used the nonparametric linkage (NPL) statistics. NPL analysis is based on excess IBD sharing among affected relatives (38). Families were divided into 2 groups according to symptom gradation, 1 containing symptom concordant sibpairs and 1 with symptom discordant sibpairs (Table 2). For each group, NPL was calculated using the Allegro 1.2 software (deCODE Genetics, Reykjavik, Iceland) (39), thus obtaining 2 statistics, NPLconc and NPLdisc. A high NPLconc and low NPLdisc would indicate that phenotype is linked to the chromosome region. Permutation analysis with 10,000 iterations was further used to assess significance of difference between NPLconc and NPLdisc.
Correlations between age at diagnosis and genotypes were analysed using Spearman ρ statistics. Associations between sex versus phenotypes and sex versus genotypes were analysed using χ2 statistics.
To investigate the hypothesis that probands are more often present with symptom grade 1 than their affected siblings, symptom discordant sibpairs were analysed applying a binomial test. The nonproband sibling was chosen arbitrarily in families with more than 2 siblings with CD.
The heritability of the phenotype was statistically significant and was estimated at 0.45 (P = 0.02). Most patients, 91%, had symptoms grade 1 or 2 (Table 1).
Association analysis (Table 3) showed a statistically significant dependence between CTLA4 +49 A/G genotypes and the different symptom grades (P = 0.014), with more AA genotypes than expected in the clinically silent group.
No significant association between symptom grade and the haplotype MH30*G:−1147*T:+49*A:CT60*G:CT61*A (P = 0.29) or HLA alleles (P = 0.58) was found.
The stratified linkage analysis showed a difference in allele sharing at CTLA4 +49A/G among symptom concordant sibship families (NPLconc = 2.0) compared with the symptom discordant sibship families (NPLdisc = −0.7) (P = 0.04). The 5q31-33 locus, however, had no significant difference (P = 0.13) in allele sharing between the 2 groups.
No significant correlation was found between age at diagnosis in the probands and the genotypes or between sex and the genotypes. Neither was any significant association between sex and symptom grade found (P = 0.64).
Proband–phenotype analysis showed that in 34 out of 37 sibpairs discordant for symptom grade 1 and the other grade 2 or grade 3, the proband sibling had more severe symptoms than its nonproband sibling (P = <10−7). The probands were in these cases the younger sibling and the older siblings with milder symptoms were diagnosed later.
CD is a complex disease with a significant genetic component and is also under the influence of environmental factors. Our aim was to investigate the association between several genotypes and the phenotypes. The prerequisite for the study, a significant heritability of the symptoms that is the variance of the phenotype caused by genetic factors was verified. Our group has previously shown association and/or linkage to 4 genetic regions: HLA-DQA1 and −HLA-DQB1 on chromosome 6, 5q31-33 region (21), the A allele in CTLA4 +49 A/G gene polymorphism on chromosome 2q33 (8), and the haplotype MH30*G:-1147*T:+49*A:CT60*G:CT61*A on chromosome 2q33 (13). MYO9B on 19p13 was identified as a susceptibility gene for CD in the Dutch population (40). However, this finding was not verified in our Scandinavian material (41) or in other populations (42–45). For that reason we refrained from analysing MYO9B in this study.
We found a significant difference between different CTLA4 +49 A/G genotypes and symptom grades when we used both association and linkage analysis methods, indicating that this gene may influence the phenotype of the disease. Surprisingly, we found that the AA genotype was correlated to the asymptomatic phenotype. In multiple sclerosis, genotype-phenotype studies have shown that AA and AG genotypes are associated with multiple sclerosis progression (46).
No other associations were found between the investigated genotypes and the clinical presentation in this study. Furthermore, no significant correlation was found between the different genotypes on one hand and sex or age at diagnosis on the other. Also, no significant correlation was found between sex and symptom grades.
Our knowledge of the interaction between genotype and phenotype in CD is limited. However, a few studies focusing on HLA have been published as well as 1 study on IL-12B and IRF1 genes and enteropathy grading. For the HLA class II genes our results are in accordance with some of the previous studies (27,28). However, Zubillaga et al (29) found a moderate overrepresentation of classical presentation, female sex, earlier age at diagnosis, and a shorter delay between onset of symptoms and diagnosis in patients homozygous for DQB1*02. Furthermore, Karinen et al (47) recently presented a correlation between gene dose of HLA DQB1*0201 and severity of the villous atrophy, lower age of diagnosis, and lower hemoglobin values at the time of diagnosis.
How can these diverging results be understood? One explanation may be that the association is rather weak, which would make it difficult to reproduce. A second reason may be different inclusion criteria. To avoid falsely diagnosed cases, which could jeopardize the results of the study, we used strict and highly demanding diagnostic criteria for inclusion. A third reason may be the problem with the classifications of different phenotypes. A plausible classification from a clinical basis may be genetically irrelevant. In this study, different modes of clinical presentation were graded as phenotypes in accordance with previous studies by others (29) and us.
Another methodological problem is the retrospective approach. The ideal way to look at, for example, age in CD is the age at onset. Usually, however, it is more or less impossible to decide due to the generally insidious start of the disease. Therefore, we preferred to use age at diagnosis as a proxy for age at onset.
The result for symptom discordant sibpairs supported the hypothesis that the younger probands have more marked symptoms than their older nonprobands.
An interesting observation was that 8 individuals had another autoimmune disease: 5 with T1DM, 2 with psoriasis, and 1 with hypothyroidism. None of them were homozygous for the CTLA4 +49 A allele or shared 2 haplotypes IBD with their cosibling in the 5q31-33 region. This is in accordance with the findings that the CTLA4 +49 G allele is overrepresented in T1DM (48). However, because the number of patients in this study is small no firm conclusions should be made from this observation.
In conclusion, we found that the CTLA4 +49 AA genotype is associated with the clinically silent presentations of CD.
Thanks to all the families that participated in this study and all the paediatrics that identified the families. Special thanks to Britt-Marie Käck, paediatric nurse, who made it possible to collect blood samples and clinical information from all-round Sweden. Åsa Hellqvist for statistical analyses. We are also grateful to Benedicte A. Lie and Ludvig M. Sollid for scientific contribution to this work.
1. Green PH, Cellier C. Coliac disease. New Engl J Med 2007; 357:1731–1743.
2. Sollid LM, Markussen G, Ek J, et al
. Evidence for a primary association of coliac disease to a particular HLA-DQ alpha/beta heterodimer. J Exp Med 1989; 169:345–350.
3. Sollid LM, McAdam SN, Molberg O, et al
. Genes and environment in coliac disease. Acta Odontol Scand 2001; 59:183–186.
4. Louka AS, Sollid LM. HLA in coeliac disease
: unravelling the complex genetics of a complex disorder. Tissue Antigens 2003; 61:105–117.
5. Bolognesi E, Karell K, Percopo S, et al
. Additional factor in some HLA DR3/DQ2 haplotypes confers a fourfold increased genetic risk of coliac disease. Tissue Antigens 2003; 61:308–316.
6. Kouki T, Sawai Y, Gardine CA, et al
. CTLA-4 gene polymorphism at position 49 in exon 1 reduces the inhibitory function of CTLA-4 and contributes to the pathogenesis of Graves' disease. J Immunol 2000; 165:6606–6611.
7. Ueda H, Howson JM, Esposito L, et al
. Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 2003; 423:506–511.
8. Naluai AT, Nilsson S, Samuelsson L, et al
. The CTLA4/CD28 gene region on chromosome 2q33 confers susceptibility to coliac disease in a way possibly distinct from that of type 1 diabetes and other chronic inflammatory disorders. Tissue Antigens 2000; 56:350–355.
9. Djilali-Saiah I, Schmitz J, Harfouch-Hammoud E, et al
. CTLA-4 gene polymorphism is associated with predisposition to coeliac disease
. Gut 1998; 43:187–189.
10. Haimila K, Smedberg T, Mustalahti K, et al
. Genetic association of coeliac disease
susceptibility to polymorphisms in the ICOS gene on chromosome 2q33. Genes Immun 2004; 5:85–92.
11. Clot F, Fulchignoni-Lataud MC, Renoux C, et al
. Linkage and association study of the CTLA-4 region in coeliac disease
for Italian and Tunisian populations. Tissue Antigens 1999; 54:527–530.
12. Mora B, Bonamico M, Indovina P, et al
. CTLA-4 +49 A/G dimorphism in Italian patients with coliac disease. Hum Immunol 2003; 64:297–301.
13. Amundsen SS, Naluai AT, Ascher H, et al
. Genetic analysis of the CD28/CTLA4/ICOS (COLIAC3) region in coeliac disease
. Tissue Antigens 2004; 64:593–599.
14. Greco L, Corazza G, Babron MC, et al
. Genome search in coliac disease. Am J Hum Genet 1998; 62:669–675.
15. Greco L, Babron MC, Corazza GR, et al
. Existence of a genetic risk factor on chromosome 5q in Italian coeliac disease
families. Ann Hum Genet 2001; 65:35–41.
16. Holopainen P, Mustalahti K, Uimari P, et al
. Candidate gene regions and genetic heterogeneity in gluten sensitivity. Gut 2001; 48:696–701.
17. Zhong F, McCombs CC, Olson JM, et al
. An autosomal screen for genes that predispose to coliac disease in the western counties of Ireland. Nat Genet 1996; 14:329–333.
18. Rioux JD, Daly MJ, Silverberg MS, et al
. Genetic variation in the 5q31 cytokine gene cluster confers susceptibility to Crohn disease. Nat Genet 2001; 29:223–228.
19. Morahan G, Huang D, Ymer SI, et al
. Linkage disequilibrium of a type 1 diabetes susceptibility locus with a regulatory IL12B allele. Nat Genet 2001; 27:218–221.
20. Yokouchi Y, Nukaga Y, Shibasaki M, et al
. Significant evidence for linkage of mite-sensitive childhood asthma to chromosome 5q31-q33 near the interleukin 12 B locus by a genome-wide search in Japanese families. Genomics 2000; 66:152–160.
21. Naluai AT, Nilsson S, Gudjonsdottir AH, et al
. Genome-wide linkage analysis of Scandinavian affected sib-pairs supports presence of susceptibility loci for coliac disease on chromosomes 5 and 11. Eur J Hum Genet 2001; 9:938–944.
22. Adamovic S, Amundsen SS, Lie BA, et al
. Fine mapping study in Scandinavian families suggests association between coeliac disease
and haplotypes in chromosome region 5q32. Tissue Antigens 2008; 71:27–34.
23. Louka AS, Torinsson Naluai A, D'Alfonso S, et al
. The IL12B gene does not confer susceptibility to coeliac disease
. Tissue Antigens 2002; 59:70–72.
24. Ryan AW, Thornton JM, Brophy K, et al
. Chromosome 5q candidate genes in coeliac disease
: genetic variation at IL4, IL5, IL9, IL13, IL17B and NR3C1. Tissue Antigens 2005; 65:150–155.
25. Ivarsson A. The Swedish epidemic of coeliac disease
explored using an epidemiological approach–some lessons to be learnt. Best Pract Res Clin Gastroenterol 2005; 19:425–440.
26. Maki M, Holm K. Incidence and prevalence of coeliac disease
in Tampere. Coeliac disease
is not disappearing. Acta Paediatr Scand 1990; 79:980–982.
27. Greco L, Percopo S, Clot F, et al
. Lack of correlation between genotype
in coliac disease. J Pediatr Gastroenterol Nutr 1998; 26:286–290.
28. Mustalahti K, Holopainen P, Karell K, et al
. Genetic dissection between silent and clinically diagnosed symptomatic forms of coeliac disease
in multiplex families. Dig Liver Dis 2002; 34:842–845.
29. Zubillaga P, Vidales MC, Zubillaga I, et al
. HLA-DQA1 and HLA-DQB1 genetic markers and clinical presentation in coliac disease. J Pediatr Gastroenterol Nutr 2002; 34:548–554.
30. Nistico L, Fagnani C, Coto I, et al
. Concordance, disease progression, and heritability
of coeliac disease
in Italian twins. Gut 2006; 55:803–808.
31. Seegers D, Borm ME, van Belzen MJ, et al
. IL12B and IRF1 gene polymorphisms and susceptibility to coliac disease. Eur J Immunogenet 2003; 30:421–425.
32. Gudjonsdottir AH, Nilsson S, Ek J, et al
. The risk of coliac disease in 107 families with at least two affected siblings. J Pediatr Gastroenterol Nutr 2004; 38:338–342.
33. Walker-Smith J. Revised criteria for diagnosis of coeliac disease
. Report of Working Group of European Society of Paediatric Gastroenterology and Nutrition. Arch Dis Child 1990; 65:909–911.
34. Meeuwisse G. Diagnostic criteria in coeliac disease
. Acta Paediatric Scand 1970; 59:461–463.
35. Ivarsson A, Persson LA, Nystrom L, et al
. The Swedish coeliac disease
epidemic with a prevailing twofold higher risk in girls compared to boys may reflect gender specific risk factors. Eur J Epidemiol 2003; 18:677–684.
36. Louka AS, Nilsson S, Olsson M, et al
. HLA in coeliac disease
families: a novel test of risk modification by the ‘other’ haplotype when at least one DQA1*05-DQB1*02 haplotype is carried. Tissue Antigens 2002; 60:147–154.
37. Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 1998; 62:1198–1211.
38. Kruglyak L, Daly MJ, Reeve-Daly MP, et al
. Parametric and nonparametric linkage analysis: a unified multipoint approach. Am J Hum Genet 1996; 58:1347–1363.
39. Gudbjartsson DF, Jonasson K, Frigge ML, et al
. Allegro, a new computer program for multipoint linkage analysis. Nat Genet 2000; 25:12–13.
40. Monsuur AJ, de Bakker PI, Alizadeh BZ, et al
. Myosin IXB variant increases the risk of coliac disease and points toward a primary intestinal barrier defect. Nat Genet 2005; 37:1341–1344.
41. Amundsen SS, Monsuur AJ, Wapenaar MC, et al
. Association analysis of MYO9B gene polymorphisms with coliac disease in a Swedish/Norwegian cohort. Hum Immunol 2006; 67:341–345.
42. Hunt KA, Monsuur AJ, McArdle WL, et al
. Lack of association of MYO9B genetic variants with coeliac disease
in a British cohort. Gut 2006; 55:969–972.
43. Nunez C, Marquez A, Varade J, et al
. No evidence of association of the MYO9B polymorphisms with coliac disease in the Spanish population. Tissue Antigens 2006; 68:489–492.
44. Cirillo G, Di Domenico MR, Corsi I, et al
. Do MYO9B genetic variants predispose to coeliac disease
? An association study in a cohort of South Italian children. Dig Liver Dis 2007; 39:228–231.
45. Giordano M, Marano C, Mellai M, et al
. A family-based study does not confirm the association of MYO9B with coliac disease in the Italian population. Genes Immun 2006; 7:606–608.
46. Bilinska M, Frydecka I, Noga L, et al
. Progression of multiple sclerosis is associated with exon 1 CTLA-4 gene polymorphism. Acta Neurol Scand 2004; 110:67–71.
47. Karinen H, Karkkainen P, Pihlajamaki J, et al
. Gene dose effect of the DQB1*0201 allele contributes to severity of coeliac disease
. Scand J Gastroenterol 2006; 41:191–199.
48. Zalloua PA, Abchee A, Shbaklo H, et al
. Patients with early onset of type 1 diabetes have significantly higher GG genotype
at position 49 of the CTLA4 gene. Hum Immunol 2004; 65:719–724.