Genetic association of coeliac disease susceptibility to ... - Nature

Genes and Immunity (2004) 5, 85–92 & 2004 Nature Publishing Group All rights reserved 1466-4879/04 $25.00 www.nature.com/gene

FULL PAPER

Genetic association of coeliac disease susceptibility to polymorphisms in the ICOS gene on chromosome 2q33 K Haimila1, T Smedberg1, K Mustalahti2, M Ma¨ki2, J Partanen1 and P Holopainen1,3 Department of Tissue Typing and Research Laboratory, Division of Stem Cell and Transplantation Service, Finnish Red Cross Blood Service, Helsinki, Finland; 2Pediatric Research Centre, Medical School, Tampere University Hospital, Tampere, Finland

1

An interesting candidate gene region for coeliac disease (CD), a common multifactorial disease, is a segment on 2q33–37 harbouring the genes for the CD28, cytotoxic T-lymphocyte-associated antigen-4 (CTLA4), inducible costimulator (ICOS), and programmed death-1 (PD-1), all receptors that regulate lymphocyte activation. Several studies have suggested a role for this locus in immune-mediated diseases. To study further our previous finding of genetic linkage of this region to CD, we studied 25 polymorphic markers to identify the putative disease-associated polymorphism. Transmission/disequilibrium test in 106 Finnish families with CD indicated that only four polymorphisms, all located in the ICOS gene, showed evidence for genetic association. Strong linkage disequilibrium (LD), based on the analysis of 424 haplotypes, encompassed not only the associated ICOS markers but also many polymorphisms in the CTLA4 gene. Our results demonstrate that due to LD, it appears not easy to identify the genuine susceptibility factor in this region without larger multipopulation studies. Furthermore, the results did not support the evidence that polymorphisms in CTLA4 were the major susceptibility locus for CD. Genes and Immunity (2004) 5, 85–92. doi:10.1038/sj.gene.6364040 Published online 8 January 2004 Keywords: ICOS; CTLA4; CD28; PD-1; coeliac disease; genetic association

Introduction There is growing evidence that gene region on 2q33, harbouring adjacently located CD28, cytotoxic T-lymphocyte-associated antigen-4 (CTLA4) and inducible costimulator (ICOS) genes, contains a susceptibility locus or loci for a number of autoimmune and immunemediated diseases,1 including, for example, type I diabetes, autoimmune thyroiditis, multiple sclerosis and coeliac disease. It is of note that all these diseases also show strong genetic association to human leucocyte antigen (HLA) class II alleles. Hence, it is possible that these two loci, both involved in the crucial initiation steps of the cellular immune response, act together conferring susceptibility. The CD28, CTLA4 and ICOS molecules play a pivotal role as secondary signal molecules in the T-lymphocyte activation.2 Most studies thus far have focused to one particular single-nucleotide polymorphism (SNP) in the CTLA4 gene, usually called CTLA4 þ 49 A/G, as this SNP was the one first published and found to be associated with type I diabetes.3 However, there is no actual evidence that the CTLA4 þ 49 SNP as such were the genuine susceptibility factor. Owing to strong linkage disequilibrium (LD) in the region,4 the identification of the Correspondence: Dr J Partanen, Department of Tissue Typing, FRC Blood Service, Kivihaantie 7, Helsinki 00310, Finland. E-mail: [email protected] 3 Current address: Swiss Institute of Allergy and Asthma Research, Obere Strasse 22, Davos Platz CH-7270, Switzerland Received 27 July 2003; revised 13 October 2003; accepted 13 October 2003

primary susceptibility polymorphism may not be an easy task. Also, similar to the HLA loci, there can be multiple alleles or genes playing a role. Several studies have demonstrated that the expression level and the function of CTLA4 showed association to genetic variation in the region,5–9 although again due to the LD, it is not clear which polymorphism actually caused the functional variation. Important and interesting results have been provided by using in vitro mutagenesis that the CTLA4 þ 49 G (Ala) allele, which introduces a hydrophobic amino acid in a highly conserved position, changes glycosylation in one of the two possible sites. Ala homozygotes may express one-third less CTLA4 on the surface of their T cells than Thr (CTLA4 þ 49A) homozygotes.10 In addition to the known polymorphisms in the CTLA4 gene, the other two genes in the region are also polymorphic. The CD28 is known to have one microsatellite11 and ICOS two microsatellites and eight SNPs.12,13 However, their roles in disease associations have not been studied systematically, except for two recent studies.12,14 Coeliac disease (CD) is a common multifactorial disorder characterised by autoimmune-like small intestinal injury triggered by dietary gliadin and related proteins.15 Patients can also be clinically silent or present various extraintestinal symptoms. Genetic susceptibility is well established. So far, the only susceptibility locus confirmed in various studies is located in the major histocompatibility complex on chromosome 6p21.3. Practically all patients have the HLA class II alleles DQA1*05, DQB1*02 or DQB1*0302.16 The strong HLA association does not, however, explain the whole

ICOS in coeliac disease K Haimila et al

86

Figure 1

Gene and marker map of the region 2q33.

Table 1 Novel SNP markers and their frequencies Gene

Location

CD28 ICOS ICOS ICOS

594 693 +1720 +2033

Allele frequencies G G C G

0.69 0.78 0.98 0.83

A A T A

Genotype frequencies 0.31 0.22 0.02 0.17

G/G G/G C/C G/G

0.48 0.61 0.97 0.69

G/A G/A C/T G/A

0.42 0.35 0.03 0.29

N A/A A/A T/T A/A

0.10 0.04 0.00 0.02

134 134 130 118

The location is counted from A of the transcription initiator codon. The allele frequencies are calculated among unaffected founder individuals (N) in the present families.

susceptibility to the disease. The HLA locus has been estimated to account for B30% of the whole familial risk of CD leaving a definitive role for other genetic, as well as environmental factors in disease susceptibility.17–20 Eight whole genome screenings21–28 for the additional risk loci have been reported so far with inconsistent results between populations. In addition to systematic screening, many candidate gene approaches have been carried out. We reported genetic linkage between seven markers in 2q33 and CD in 100 Finnish families with affected sib-pairs.29 The same region has subsequently been confirmed to show linkage to CD in a Scandinavian study also,30 in which an association to CTLA4 allele þ 49*A was also observed, supporting the findings of a French case–control study.31 The association of the allele A was also detected in another Italian case–control and family study.32 Also, families from the UK showed a significant association, not with CTLA4 but microsatellite D2S221433 (Figure 1). In contrast, no evidence for this region was found in a study of Italian and Tunisian families.34 Despite the genetic linkage found in our previous study,29 the markers studied did not show allelic association in Finnish families. As an attempt to narrow down the interval carrying a putative disease predisposing gene and to study the LD in the region more systematically, we now report the genetic analysis of 24 polymorphic markers in a 3.4 Mb segment of 2q33 region and one SNP in the gene of programmed death-1 (PD-1) in 2q37 region in 106 CD families from Finland. PD-1 (called also PDCD1) is an activation-induced inhibitory receptor expressed on T cells, B cells and monocytes,35 and it has suggested to be an important factor in the prevention of autoimmune diseases.36

Results Initially, association between 23 genetically polymorphic markers located in the CD28–CTLA4–ICOS and PD-1 gene region and CD was studied in 106 families. The Genes and Immunity

markers and their locations are shown in Figure 1. The four novel SNP markers, which were found in the present study by sequencing, are described in Table 1. In the novel microsatellite marker CD28(CTTT)n, 14 alleles were observed, giving a heterozygosity of 80%. The shortest allele has an apparent amplicon length of 821 bp, the remaining ones ranged from 841 to 889 bp. Allele frequencies of all markers are given in Table 2. Transmission/disequilibrium test (TDT) suggested several nominally significant findings (Table 3). Near the CD28 and CTLA4 genes only one marker, D2S2214, showed statistically significant TDT: allele 7 (286 bp) showed a negative association (P ¼ 0.029). On the other hand, it was of particular interest that four SNPSs ICOS IVS þ 173T/C, ICOSc.602A/C, ICOSc.1624C/T and ICOSc.2373G/C in the ICOS gene showed statistically significant TDT results (Po0.05 in all cases). These test statistics, however, did not remain statistically significant when their significance was evaluated using permutation. The most significant finding was for allele T of ICOS IVS þ 173T/C: 35 transmitted versus 15 nontransmitted alleles (P ¼ 0.005). We also studied two CTLA4-linked SNPs, MH30 and CT60, which were reported to associate with Grave’s disease, autoimmune hypothyroidism and type I diabetes in a recent study,14 and genotyped the PD-1.3 marker in the PD-1 gene, which recently was reported to be associated with susceptibility to systemic lupus erythematosus.37 These three markers showed no evidence for association with CD by the TDT analysis (Table 3). To analyse the possible association with the ICOS markers further, we studied the transmission of ICOS haplotypes in the families (Table 4). Haplotypes including all the eight ICOS markers were constructed. Haplotype T-C-A-T-C-T-G-G, that is, the haplotype including all the four ICOS alleles showing the nominal association in the original TDT test (first, second, sixth and eighth marker in the haplotype), showed a T:NT ratio of 17:2 (P ¼ 0.0006 by TDT, relative risk 1.14, 95% CI 0.54–2.0). This suggests that the individual associated


Table 2

87

Allele frequencies for the parents (Freq P) and for the offspring (Freq O) in the present families

Marker

Allele

Freq P

Freq O

Marker

Allele

Freq P

Freq O

D2S116

1 2 3 4 5 6 7 8 9 10

0.01 0.09 0.12 0.07 0.34 0.34 0.27 0.04 0.03 0.01

0.01 0.10 0.14 0.02 0.06 0.36 0.23 0.06 0.02 o0.01

CTLA4(AT)n

13 14 15 16 17 18 1 2 1 2

0.02 0.01 0.01 0.01 0.01 o0.01 0.69 0.31 0.02 0.02

0.01 0.01 0.01 0 0.01 0 0.67 0.33 0.02 0.01

1 2 3 4 5 6 7 8 9

0.07 0.10 0.61 0.03 0.14 0.01 0.04 o0.01 o0.01

0.07 0.11 0.68 0.03 0.07 0.01 0.02 0 0.01

3 4 5 6 7 8 9 1 2

0.23 0.10 0.46 0.03 0.12 0.01 0.01 0.79 0.21

0.25 0.09 0.45 0.04 0.12 0.01 0.01 0.79 0.21

1 2 3 4 5 6 7 8 9 10 11 12 13 14

0.01 0.01 0.02 0.04 0.16 0.18 0.13 0.12 0.09 0.05 0.08 0.06 0.04 0.01

0.01 0.01 0.02 0.04 0.13 0.19 0.12 0.16 0.06 0.04 0.12 0.06 0.03 0.01

ICOS IVS+173T/C

1 2 1 2 1 2 1 2 1 2 1 2 1 2

0.86 0.14 0.83 0.17 0.95 0.05 0.91 0.09 0.82 0.18 0.91 0.09 0.99 0.01

0.79 0.21 0.78 0.22 0.91 0.09 0.88 0.12 0.79 0.21 0.88 0.12 0.99 0.01

CD28594A/G

1 2

0.33 0.67

0.30 0.70

ICOSc.1862A/G

1 2

0.91 0.09

0.87 0.13

CD28(CAA)n

1 2 3 4 5 6 7

0.01 0.84 0.01 0.02 0.04 0.08 o0.01

0.02 0.83 0.01 0.01 0.05 0.07 0.01

ICOSc.2033G/A

1 2 1 2 1 2 3

0.83 0.17 0.90 0.10 0.05 0.19 0.40

0.82 0.18 0.93 0.07 0.06 0.17 0.41

MH30C/G

1 2

0.71 0.29

0.65 0.35

4 5

0.20 0.09

0.18 0.08

CTLA4318C/T

1 2

0.89 0.11

0.90 0.10

6 7

0.05 0.02

0.06 0.03

CTLA4+49A/G

1 2

0.48 0.52

0.90 0.10

8 1

o0.01 0.01

0.01 0.01

1 2 3 4 5 6 7 8 9 10 11 12

0.33 0.36 0.04 0.05 o0.01 0.02 o0.01 0.01 0.02 0.05 0.03 0.03

0.33 0.34 0.05 0.06 0.01 0.02 0.01 0.01 0.01 0.04 0.04 0.04

2 3 4 5 6 7 8 9 10 1 2

0.15 0.63 0.01 0.02 0.12 0.02 0.01 0.02 0.01 0.94 0.06

0.17 0.61 0.01 0.01 0.16 0.02 0 0.01 0 0.96 0.04

D2S2214

CD28(CTTT)n

CTLA4(AT)n

CT60G/A D2S72

ICOS693G/A

ICOSc.602 A/C ICOSc.930G/A ICOSc.1459A/T ICOSc.1564T/C ICOSc.1624C/T ICOSc.1720C/T

ICOSc.2373G/C D2S1271

D2S2189

PD-1.3G/A

Genes and Immunity


88

Table 3 Results of TDT Marker

Allele

Trans

Nontrans

P-value

RR

D2S116 D2S2214 CD28(CTTT)n CD28594A/G CD28(CAA)n MH30G/C CTLA4318C/T CTLA4+49A/G CTLA4(AT)n CT60G/A D2S72 ICOS693G/A ICOS IVS+173T/C ICOSc.602 A/C ICOSc.930G/A ICOSc.1459A/T ICOSc.1564T/C ICOSc.1624C/T ICOSc.1720C/T ICOSc.1862A/G ICOSc.2033G/A ICOSc.2373G/C D2S1271 D2S2189 PD-1.3G/A

1–10 7 1–14 1–2 1–7 1–2 1–2 1–2 1–18 1–2 1–9 1–2 1 (T) 2 (C) 1–2 1–2 1–2 2 (T) 1–2 1–2 1–2 1 (G) 1–8 1–10 1–2

— 4 — — — — — — — — — — 35 37 — — — 22 — — — 24 — — —

— 13 — — — — — — — — — — 15 21 — — — 10 — — — 11 — — —

— 0.029 — — — — — — — — — — 0.005 0.036 — — — 0.034 — — — 0.028 — — —

— 1.59 — — — — — — — — — — 1.28 1.29 — — — 1.88 — — — 1.25 — — —

95% CI

1.19

2.92

1.32 1.25

4.37 1.79

1.84

3.1

1.61

4.54

In all, 25 markers in 106 families were studied. Actual numbers of transmitted (Trans) and nontransmitted (Nontrans) are given only for alleles giving statistically nominally significant TDT. Relative risk (RR) with 95% confidence intervals (CIs) are calculated for the associated alleles.

Table 4 Observed haplotypes of 8 ICOS polymorphisms Haplotype (markers within ICOS gene)a T C T C T T 16 other

A A C C A C different

G A G A A T G A G A G A haplotypes

T T C C C C

C C T C C C

A A G A A A

G G G C G C

Number

Frequency

Trans

Frequency

Nontrans

Frequency

223 20 19 11 9 9 30

0.7 0.06 0.06 0.03 0.03 0.03 0.09

134 7 17 4 6 6 16

0.71 0.04 0.09 0.02 0.03 0.03 0.08

89 13 2 7 3 3 14

0.68 0.10 0.02 0.05 0.02 0.02 0.11

a

ICOS IVS1+173T/C, c.602A/C, c.930G/A, c.1459T/C, c.1564C/T, c.1624C/T, c.1862A/G, c.2373G/C. Conserved segments in bold. Only complete haplotypes with no missing genotypes were collected from the haplotypes transmitted and nontransmitted to the affected offspring.

alleles actually formed a single haplotype, which showed transmission preference to affected children. Although the overall LD across the whole 2q33 region was surprisingly high both in the transmitted (Figure 2a) and nontransmitted (Figure 2b) haplotypes, in general, this particular haplotype appeared not to extend as a single conserved segment outside the ICOS gene. Hence, the susceptibility polymorphism should locate very close to, or within, the ICOS gene. To test this possibility, we sequenced the exons of the ICOS gene in one individual with this haplotype to probe, whether it contained any novel variation not detected in our original screening.13 Indeed, we found two novel SNPs in the noncoding region of the last exon: ICOS þ 1720C/T and þ 2033 G/A (Table 1). All the 106 families were genotyped for these two novel polymorphisms and TDT was performed, but no evidence for association was found (Table 3). Genes and Immunity

LD analysis of all haplotypes in the families indicated that there appear two segments with high gametic association (Figures 2a and b). First, markers located in the CD28 and CTLA4 genes or telomeric-flanking regions (D2S72 and ICOS-693G/A) were strongly associated with each other. Second, the markers within the ICOS gene showed strong LD.

Discussion Although the chromosome 2q33 region has been of considerable interest as a candidate susceptibility locus for many autoimmune-related diseases, only a few systematic studies encompassing all the region, including, for example, three related genes, CD28, CTLA4 and ICOS, have been carried out so far. Most studies have focused on the CTLA4 gene only—an obvious candidate


89

Figure 2

LD across the 2q33 region in transmitted (a) and nontransmitted haplotypes (b).

in the region, but the role of other genes cannot be automatically ruled out. In type I diabetes, the susceptibility gene on 2q33 was mapped by Marron et al11 into a 100 kb fragment harbouring CTLA4 gene. Ihara et al12 screened polymorphisms in all the three genes and concluded that CTLA4 polymorphisms were the best candidate for type I diabetes. Furthermore, Ueda et al14 could recently demonstrate that the susceptibility to type I diabetes, Grave’s disease and autoimmune hypothyroidism maps to polymorphisms in a noncoding 6.1 kb 30 -end region of CTLA4 and most interestingly, seems to correlate with alternatively spliced forms of CTLA4 transcript. In CD, we could not find any definitive evidence for CTLA4 or CD28 genes among nearly 800 European families with CD in a recent collaborative study; although linkage in the region was confirmed, the associated polymorphisms were rather inconsistent and population specific.38 In addition, we could not find evidence in the present study for the role of the two most interesting markers, CT60 or MH30,14 although we must note that the ability to detect minor susceptibility loci with low relative risks was low for the present family material. Taken these together, the chromosome 2q33-linked susceptibility locus in CD may differ from that in type I diabetes. In the present paper, preliminary evidence for the ICOS gene, rather than CTLA4, was provided in the Finnish families with CD. To our knowledge, the present study was the first one in CD studying systematically markers across the entire gene cluster and extended also to the PD-1 gene. ICOS can be regarded as a good

candidate in an HLA class II-associated disease, since it acts as a secondary stimulatory molecule in the T-cell activation.39–41 Its expression is induced on activated T-lymphocytes and its ligands are expressed on Blymphocytes and some nonimmune tissues. Of potential interest is that ICOS is critical to CD40-mediated antibody class switching,42 as CD is characterised by diagnostic anti-tissue transglutaminase autoantibodies of IgA class.43 It has been speculated that these autoantibodies might have a direct role in the pathogenesis of the disease,44 although the topic certainly is controversial. Whether any of the polymorphisms described in the present paper have effect on the function or expression of ICOS requires further studies. Also, based on some recent findings,14,45 a more detailed study of polymorphisms in the introns and other noncoding regions of ICOS is definitely warranted. LD is remarkable on 2q33 region, especially between intragenic polymorphisms in each of these three genes. In the present study, LD divides into two sections, strong LD appeared between polymorphisms in the genes of CD28 and CTLA4 and again within markers of ICOS, thus ICOS builds up into its own block of LD. Three separate blocks of LD, each including one of the genes CD28, CTLA4 and ICOS, in the region was detected in two recent papers.12,14 The higher level of LD in the present study may result from the founder effect typical to the Finnish population. Strong LD complicates the identification of the primary polymorphism, so more detailed haplotype analyses are required. Genes and Immunity


90

Materials and methods Patients A total of 106 Finnish CD families were studied. Of these, 83 of the families and five markers (D2S116, D2S2214, CTLA4 þ 49A/G, CTLA4(AT)n and D2S2189) were the same as studied in our original linkage report,29 and the same 83 families were included in the EU collaborative study35 in which seven markers (D2S116, D2S2214, CD28(CAA)n, CTLA4-318C/T, CTLA4 þ 49A/ G, CTLA4(AT)n and D2S72) overlapped with the present study. As only an association study was carried out, parent–child trios were under study. All families included one randomly selected affected child and both parents. The families were initially recruited through the Finnish Coeliac Society by advertising in the patients’ newsletter. The recruitment and clinical findings have been published elsewhere.46 Briefly, the earlier diagnoses were re-evaluated by scrutinising the medical records. Antibody screenings were performed to reveal the asymptomatic cases. The diagnosis was based on initial small bowel biopsy (or skin biopsy from dermatitis herpetiformis patients). The study protocol was accepted by the Ethics Review Board of the Tampere University Hospital. Genetic markers and novel polymorphisms One previously described and two novel CD28 gene polymorphisms were genotyped. Trinucleotide repeat CD28(CAA)n in the 30 -untranslated region was genotyped as originally described.11 The novel polymorphisms were identified in the present study by sequencing the 50 -flanking sequence of 770 bp upstream from the initiator codon. According to the public reference sequence in clone AF225899, a CT-rich stretch of about 315 nt in length is located at position 725 to 440 upstream from the CD28 transcription initiation site. The sequencing of the variable region revealed that the microsatellite is not a clear CTTT repeat but there are also some CCTT and CT repeats between CTTT repeats. The length of the sequence containing varying forms of repeats is about 315 bp. The difference between longest and the shortest allele is 69 bp. Both novel polymorphisms at CD28 promoter, the SNP 594*A/G and the tetranucleotide repeat CD28(CTTT)n, were genotyped from a PCR fragment with an average size of 850 bp amplified with forward primer 50 CCAGAGGGCTAC CTTATGTCCTACAC and reverse primer 50 CACTACAT GATGGGCTTATGGGAA. For CD28(CTTT)n genotyping, a fluorescently labelled forward primer was used, and the size of the alleles were determined by electrophoresis using ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). The dimorphism 594*A/G was typed by RFLP method, the amplified PCR product was digested with HinfI restriction enzyme, resulting in B530 and 317 bp fragments for allele G and an intact fragment of B850 bp for allele A. The SNP MH30 was genotyped with ABI PRISM SNaPshot Multiplex Kit (Applied Biosystems) using T-tailed SNaPshot primer 50 (T)10AATAAAGGTGAAAATTTTCT. The primers for PCR amplification were 50 GGCAGAAGGA GAAAATCAAAG and 50 TCACCTGAGGTTGGGAGTTC. Three known polymorphisms within the CTLA4 gene were genotyped. Two dimorphisms, þ 49*A/G in the exon 146 and –318*C/T in the promoter region,47 were

Genes and Immunity

typed by PCR/RFLP method, using primers 50 AACCC AGGTAGGAGAAACAC and 50 GCTCTACTTCCTGAA GACCT and BbvI digestion for CTLA4 þ 49*A/G, and primers 50 TTACGAGAAAGGAAGCCGTG and 50 AAA TGAATTGGACTGGATGGT and MseI digestion for CTLA4 þ 318*C/T. A polymorphic AT repeat, CTLA4(AT)n, in the 30 -untranslated region of CTLA4 gene48 was genotyped with fluorescent-labelled forward primer 50 GTGATGCTAAAGGTTGTATTGC and reverse primer 50 AAAACATACGTGGCTCTATGCAC using ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). The SNP CT60 (CTLA4 þ 6230G/ A) was genotyped by PCR/RFLP method, using primers 50 GGGTGCATGAAGGTTTCTGA and 50 AGCCACAGG GAAAAGCTCTA, and enzyme TaiI. Eight ICOS polymorphisms were genotyped as originally described.13 One novel polymorphism, SNP 693*G/A, was identified by sequencing the 50 -flanking sequence of 1150 bp upstream from the transcription initiation site. 693*G/A dimorphism was genotyped by PCR/RFLP method using a forward primer 50 ATCT TGGAAGCGCGTTCAGACTT and a reverse primer 50 ATTTCTATCTTATGCTAGGTGCTCCA. Digestion of the PCR product with the BseGI restriction enzyme produced 30, 185, 294 and 297 bp fragments for allele G, and 30, 185 and 591 bp fragments for allele A. Two other novel polymorphisms were identified by sequencing the exons of the ICOS gene in one individual carrying haplotype with the associated ICOS SNPs. ICOS þ 1720C/T was genotyped by PCR/RFLP method using primers 50 TTACCAAGACTTTAGATGCTTTCTT and 50 GAATCTTTCTAGCCAAATCATATTC following digestion with the restriction enzyme NcoI producing 520 and 304 bp fragments for allele C and a 824 bp fragment for allele T. þ 2033 G/A was genotyped by PCR amplification with allele-specific primers. The primers were 50 CCAGCTAAGAATAATCATATTCCAC for allele G and 50 CCAGCTAAGAATAATCATATTCCAT for allele A. The forward primer for both was 50 CATTATCTA TATGTTTTCATGGTGCTATT and an internal positive amplification control was included in each PCR reaction. To study the length of the linked region, five other microsatellite markers spanning B3400 kb region (Figure 1) were genotyped. Genetic maps of the Genome Database, National Center for Biotechnology Information and Genethon were used. Two microsatellites (D2S116 and D2S2214) are located centromeric to CD28, one (D2S72) is situated between genes CTLA4 and ICOS and two microsatellites (D2S1271 and D2S2189) are located telomeric to ICOS. The microsatellites were genotyped using standard PCR methods with fluorescentlabelled primers and ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). The SNP PD-1.3 was genotyped by PCR/RFLP method.37 Data analysis The TDT test was calculated by Genehunter 2.1 to investigate genetic linkage in the presence of allelic association. The TDT compares transmitted versus nontransmitted alleles from a heterozygous parent to affected offspring.49 Single and two, three and four locus TDT with nominal P-values were calculated. To take multiple testing into account, statistical significance of TDT statistics were also estimated by permutation, using


10 000 replicates. Test of LD was performed by ARLEQUIN program package50 for all pairs of markers among 212 independent founder haplotypes constructed by Genehunter 2.1. The test is an extension of Fisher exact probability test on contingency tables and it utilises a Markov chain to explore the tables. Power in the TDT was estimated using the Genetic Power Calculator (http://statgen.iop.kcl.ac.uk/cgi-bin/ powercalc/dtdt.cgi). We assumed that the disease prevalence was 0.01 and D0 was 0.99. Assuming (i) that the marker frequency was 0.4 and (ii) that the relative risk for disease allele heterozygote was 2 and for homozygote 4, the power was estimated to be 77, 80 and 20% when the frequency of the disease allele was 0.5, 0.3 and 0.1, respectively.

Acknowledgements We thank Drs Marta Alarco´n-Riquelme and Ludmila Prokunina of Uppsala University, Uppsala, Sweden, for kind advice in the PD-1 typing. This study has been partially funded by the Commission of the European Communities, specific RTD programme ‘Quality of Life and Management of Living Resources’, QLRT-199900037, ‘Evaluation of the prevalence of the coeliac disease and its genetic components in the European population’. It does not necessarily reflect its views and in no way anticipates the Commission’s future policy in this area. The study is also supported by the Sigrid Juselius Foundation, the Medical Research Fund of Finnish Red Cross Blood Transfusion Service, Maud Kuistila Memorial Foundation, the Medical Research Fund of Tampere University Hospital, Pa¨ivikki and Sakari Sohlberg Foundation and The Academy of Finland Research Council for Health, funding decision number 73489.

References 1 Kristiansen OP, Larsen ZM, Pociot F. CTLA-4 in autoimmune diseases—a general susceptibility gene to autoimmunity? Genes Immun 2000; 1: 170–184. 2 Carreno BM, Collins M. The B7 family of ligands and its receptors: new pathways for costimulation and inhibition of immune responses. Annu Rev Immunol 2002; 20: 29–53. 3 Nistico´ L, Buzzetti R, Pritchard LE et al. The CTLA-4 gene region of chromosome 2q33 is linked to, and associated with, type 1 diabetes. Hum Mol Genet 1996; 5: 1075–1080. 4 Holopainen PM, Partanen JA. Technical note: linkage disequilibrium and disease-associated ctla4 gene polymorphisms. J Immunol 2001; 167: 2457–2458. 5 Kouki T, Sawai Y, Gardine C, Fisfalen M-E, Alegre M-L, DeGroot LJ. 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. 6 Huang D, Giscombe R, Zhou Y, Pirskanen R, Lefvert AK. Dinucleotide repeat expansion in the CTLA-4 gene leads to T cell hyper-reactivity via the CD28 pathway in myasthenia gravis. J Neuroimmunol 2000; 105: 69–77. 7 Ligers A, Teleshova N, Masterman T, Huang WX, Hillert J. CTLA-4 gene expression is influenced by promoter and exon 1 polymorphisms. Genes Immun 2001; 2: 145–152. 8 Maurer M, Loserth S, Kolb-Maurer A et al. A polymorphism in the human cytotoxic T-lymphocyte antigen 4 (CTLA4) gene (exon 1 +49) alters T-cell activation. Immunogenetics 2002; 54: 1–8.

91 9 Wang XB, Zhao X, Giscombe R, Lefvert AK. A CTLA-4 gene polymorphism at position 318 in the promoter region affects the expression of protein. Genes Immun 2002; 3: 233–234. 10 Anjos S, Nguyen A, Ounissi-Benkalha H, Tessier MC, Polychronakos C. A common autoimmunity predisposing signal peptide variant of the cytotoxic T-lymphocyte antigen 4 results in inefficient glycosylation of the susceptibility allele. J Biol Chem 2002; 277: 46478–46486. 11 Marron MP, Zeidler A, Raffel LJ et al. Genetic and physical mapping of a type 1 diabetes susceptibility gene (IDDM12) to a 100-kb phagemid artificial chromosome clone containing D2S72–CTLA4–D2S105 on chromosome 2q33. Diabetes 2000; 49: 492–499. 12 Ihara K, Ahmed S, Nakao F et al. Association studies of CTLA4, CD28, and ICOS gene polymorphisms with type 1 diabetes in the Japanese population. Immunogenetics 2001; 53: 447–454. 13 Haimila KE, Partanen JA, Holopainen PM. Genetic polymorphism of the human ICOS gene. Immunogenetics 2002; 53: 1028–1032. 14 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. 15 Maki M, Collin P. Coeliac disease. Lancet 1997; 349: 1755–1759. 16 Karell K, Louka AS, Moodie SJ et al. Hla types in celiac disease patients not carrying the DQA1*05-DQB1*02 (DQ2) heterodimer: results from the european genetics cluster on celiac disease. Hum Immunol 2003; 64: 469–477. 17 Bevan S, Popat S, Braegger CP et al. Contribution of the MHC region to the familial risk of coeliac disease. J Med Genet 1999; 36: 687–690. 18 Petronzelli F, Bonamico M, Ferrante P et al. Genetic contribution of the HLA region to the familial clustering of coeliac disease. Ann Hum Genet 1997; 61: 307–317. 19 Risch N. Assessing the role of HLA-linked and unlinked determinants of disease. Am J Hum Genet 1987; 40: 1–14. 20 Rotter JI, Landaw EM. Measuring the genetic contribution of a single locus to a multilocus disease. Clin Genet 1984; 26: 529–542. 21 Zhong F, McCombs CC, Olson JM et al. An autosomal screen for genes that predispose to celiac disease in the western counties of Ireland. Nat Genet 1996; 14: 329–333. 22 Greco L, Corazza G, Babron MC et al. Genome search in celiac disease. Am J Hum Genet 1998; 62: 669–675. 23 King AL, Yiannakou JY, Brett PM et al. A genome-wide familybased linkage study of coeliac disease. Ann Hum Genet 2000; 64: 479–490. 24 Naluai AT, Nilsson S, Gudjonsdottir AH et al. Genome-wide linkage analysis of Scandinavian affected sib-pairs supports presence of susceptibility loci for celiac disease on chromosomes 5 and 11. Eur J Hum Genet 2001; 9: 938–944. 25 Liu J, Juo SH, Holopainen P et al. Genomewide linkage analysis of celiac disease in Finnish families. Am J Hum Genet 2002; 70: 51–59. 26 Woolley N, Holopainen P, Ollikainen V et al. A new locus for coeliac disease mapped to chromosome 15 in a population isolate. Hum Genet 2002; 111: 40–45. 27 Popat S, Bevan S, Braegger CP et al. Genome screening of coeliac disease. J Med Genet 2002; 39: 328–331. 28 Neuhausen SL, Feolo M, Camp NJ, Farnham J, Book L, Zone JJ. Genome-wide linkage analysis for celiac disease in North American families. Am J Med Genet 2002; 111: 1–9. 29 Holopainen P, Arvas M, Sistonen P et al. CD28/CTLA4 gene region on chromosome 2q33 confers genetic susceptibility to celiac disease. A linkage and family-based association study. Tissue Antigens 1999; 53: 470–475. 30 Naluai AT, Nilsson S, Samuelsson L et al. The CTLA4/CD28 gene region on chromosome 2q33 confers susceptibility to celiac disease in a way possibly distinct from that of type 1 diabetes and other chronic inflammatory disorders. Tissue Antigens 2000; 56: 350–355. Genes and Immunity


92 31 Djilali-Saiah I, Schmitz J, Harfouch-Hammoud E, Mougenot JF, Bach J-F, Caillat-Zucman S. CTLA-4 gene polymorphism is associated with predisposition to coeliac disease. Gut 1998; 43: 187–189. 32 Mora B, Bonamico M, Indovina P et al. CTLA-4 +49 A/G dimorphism in Italian patients with celiac disease. Hum Immunol 2003; 64: 297–301. 33 King AL, Moodie SJ, Fraser JS et al. CTLA-4/CD28 gene region is associated with genetic susceptibility to coeliac disease in UK families. J Med Genet 2002; 39: 51–54. 34 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. 35 Ishida Y, Agata Y, Shibahara K, Honjo T. Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily, upon programmed cell death. EMBO J 1992; 11: 3887–3895. 36 Nishimura H, Honjo T. PD-1: an inhibitory immunoreceptor involved in peripheral tolerance. Trends Immunol 2001; 22: 265–268. 37 Prokunina L, Castillejo-Lo´pez C, Berg F et al. A regulatory polymorphism in PDCD1 is associated with susceptibility to systemic lupus erythematosus in humans. Nat Genet 2002; 32: 666–669. 38 Holopainen P, Torinsson Naluai A, Moodie S et al. Candidate gene region 2q33 in European families with coeliac disease. Tissue Antigens 2004; 63: 1–11. 39 Dong C, Juedes AE, Temann UA et al. ICOS co-stimulatory receptor is essential for T-cell activation and function. Nature 2001; 409: 97–101. 40 Tafuri A, Shahinian A, Bladt F et al. ICOS is essential for effective T-helper-cell responses. Nature 2001; 409: 105–109.

Genes and Immunity

41 Yoshinaga SK, Whoriskey JS, Khare SD et al. T-cell costimulation through B7RP-1 and ICOS. Nature 1999; 402: 827–832. 42 McAdam AJ, Greenwald RJ, Levin MA et al. ICOS is critical for CD40-mediated antibody class switching. Nature 2001; 409: 102–105. 43 Sulkanen S, Halttunen T, Laurila K et al. Tissue transglutaminase autoantibody enzyme-linked immunosorbent assay in detecting celiac disease. Gastroenterology 1998; 115: 1322–1328. 44 Halttunen T, Maki M. Serum immunoglobulin A from patients with celiac disease inhibits human T84 intestinal crypt epithelial cell differentiation. Gastroenterology 1999; 116: 566–572. 45 Haaning Andersen AD, Lange M, Lillevang ST. Allelic variation of the inducible costimulator (ICOS) gene: detection of polymorphisms, analysis of the promoter region, and extended haplotype estimation. Tissue Antigens 2003; 61: 276–285. 46 Mustalahti K, Sulkanen S, Holopainen P et al. Coeliac disease among healthy members of multiple case coeliac disease families. Scand J Gastroenterol 2002; 37: 161–165. 47 Deichmann K, Heinzmann A, Bruggenolte E, Forster J, Kuehr J. An Mse I RFLP in the human CTLA4 promotor. Biochem Biophys Res Commun 1996; 225: 817–818. 48 Polymeropoulos MH, Xiao H, Rath DS, Merril CR. Dinucleotide repeat polymorphism at the human CTLA4 gene. Nucleic Acids Res 1991; 19: 4018. 49 Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene region and insulindependent diabetes mellitus (IDDM). Am J Hum Genet 1993; 52: 506–516. 50 Schneider S, Kueffer J-M, Roessli D. Excoffier LArlequin ver 11: A software for population genetic data analysis. (1.1). Genetics and Biometry Laboratory, University of Geneva: Switzerland, 1997.