Copy

Analysis of Candidate Genes on Chromosomes 5q and 19p inCeliac Disease

�Anna Latiano, yBarbara Mora, zMargherita Bonamico, yFrancesca Megiorni,yMaria Cristina Mazzilli, zSalvatore Cucchiara, �Orazio Palmieri, �Maria Rosa Valvano, and

�Vito Annese

Journal of Pediatric Gastroenterology and Nutrition45:180–186 # 2007 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition andNorth American Society for Pediatric Gastroenterology, Hepatology, and Nutrition

right © 2007 by

�U.O. Gastroenterologia e Lab. di Ricerca I.R.C.C.S. ‘‘Casa Sollievo della Sofferenza’’, San Giovanni Rotondo,{

respectively.

Symptoms may inclubloating, cramps, flafatigue.

Received January 19, 2Address correspondenc

Endoscopia Digestiva, OIRCCS, Viale Cappuccin(e-mail: v.annese@operap

Study supported by the

Dipartimento di Medicina Sperimentale, and {Dipartimento di Pediatria, Universita

‘‘La Sapienza’’ Roma, Italy

ABSTRACT

Background and Aim: Celiac disease (CD) is a multifactorialdisease with involvement of both environmental and geneticsusceptibility factors. The HLA-DQ loci account for <40% ofCD heritability, but linkage studies have delineated other loci atthe 5q31–33 (CELIAC2), and 19p13 regions (CELIAC4),similarly as in inflammatory bowel diseases. However, datain association studies are contradictory. To evaluate whethersingle nucleotide polymorphisms (SNPs) tagging the MYO9Bsusceptibility haplotype and the IBD5 locus (5q31–33) areinvolved in CD predisposition, we performed case-controland family-based analyses. Additionally, any possiblecorrelation with the HLA-DQ status was investigated.Finally, our data were pooled with the results of otherstudies by a meta-analysis.Patients and Methods: In all, 337 unrelated patients with CD,424 parents (212 sets), and 452 healthy individuals weregenotyped for the IGR2198a_1, rs12521868, rs1050152, andrs2631367 SNPs (IBD5 locus) and the rs962917, rs2305764,

Lippincott Williams & Wilkins.Un

de diarrhea, steatorrhea, abdominaltulence, weight loss, weakness, and

007; accepted March 20, 2007.e and reprint requests to Dr Vito Annese, U.O.spedale ‘‘Casa Sollievo della Sofferenza’’i, 1–71013 San Giovanni Rotondo, Italyadrepio.it).Italian Minister of Health (RC0702GA30).

180

Results: In comparison with healthy control individuals, theallele, genotype, and haplotype frequencies of all investigatedSNPswerenotdifferent in theCDpatients,norwasanycorrelationobserved with the HLA-DQ status or clinical presentation. Thetransmission disequilibrium test did not show a transmissiondistortion. Five other studies were available for meta-analysison MYO9B variants; by pooling of data, no significant associationwas demonstrated by the random effect model. A significantheterogeneity (P< 0.002) among the studies was present,mainly explained by a single study in the Dutch population.Conclusions: Our results and those of the meta-analysis(>2000 CD patients and 4000 control individuals) questionthe role of MYO9B at the CELIAC4 locus as a disease-causinggene. Moreover, none of the investigated SNPs explain thelinkage at the CELIAC2 locus. JPGN 45:180–186, 2007. KeyWords: Celiac disease—Organic cation transporter—MYO9B—Single nucleotide polymorphism—Inflammatorybowel disease—Human leukocyte antigen. # 2007 by

and rs1545620 SNPs of the MYO9B gene by the restriction

European Society for Pediatric Gastroen enzyme method and the TaqMan system ABI PRISM 7700,

terology, Hepatology,and Nutrition and North American Society for Pediatric

Gastroenterology, Hepatology, and Nutrition

INTRODUCTION

Celiac disease (CD) (MIM 212750) is an immune-mediated disorder characterized by intestinal inflam-mation induced by dietary wheat, rye, and barley.

Twin-based and family-based studies have shown astrong genetic component of CD development, withinherited risk attributable to HLA and non-HLA factors(1). To identify chromosomal regions likely to containdisease-predisposing genes (2), 11 genome-widesearches and numerous region-specific linkage studieshave been performed. A well-known genetic factor, theCELIAC1, lies in the human leukocyte antigen (HLA)class II region on chromosome 6p21 (3). Approximately80% of patients with CD share the HLA-DQA1�05,HLA-DQB1�02 combination, coding for the DQ2high-risk molecule, and approximately half of the

authorized reproduction of this article is prohibited.

remaining patients have the HLA-DQ8. Nevertheless,the HLA-DQ loci have been estimated to account for<40% of CD heritability (2,3).

Cop

MES

Other regions of interest have been also investigatedwithout definite conclusions (4–7). The linkage at the5q31–33 region (CELIAC2), originally found in anItalian genome scan (4), has been replicated inother populations (8,9). This chromosomal region,which contains a cytokine gene cluster, coincideswith a linkage region for inflammatory bowel disease(IBD5 locus) (10,11). More specifically, the TChaplotype (1507C>T and �207G>C) (rs1050152 andrs2631367, respectively) in the organic cation transpor-ter (OCTN) gene cluster within the IBD5 locus has beenreported in association with Crohn disease, anotherinflammatory disease affecting the small intestine(12). By contrast, no significant association with CDof the IBD5 risk haplotype has been found in 2 previousstudies (13,14).

The CELIAC4 locus on chromosome 19p13.1 hasalso been suggested as a non-HLA CD susceptibilitylocus (15,16). Interestingly, this region has also beenstrongly linked to Crohn disease in 2 independentgenome scans (10,17). A recent study in a Dutchpopulation (18) identified a genetic noncoding variant(rs2305764; P¼ 2.1� 10�6) in the 30 region of themyosin IXB gene (MYO9B), located on chromosome19p13.1 as a risk factor for CD: heterozygotesand homozygotes for the A allele had 1.66-fold and2.27-fold, respectively, increased risk for CD. However,this study was not confirmed in a subsequent case-control study (19) and several family-based associationstudies (20–22).

Given these conflicting findings, we performed acase-control and family-based association study inpediatric-onset Italian patients with CD by genotyping3 single nucleotide polymorphisms (SNPs) tagging theMYO9B susceptibility haplotype, and 4 SNPs at the IBD5locus, to evaluate whether any of these variations wasinvolved in disease predisposition. Furthermore, weinvestigated any possible correlation with the HLA-DQstatus. Finally, the results of the present investigationwere pooled with data available from other studies by ameta-analysis.

PATIENTS AND METHODS

A cohort of 337 unrelated patients with CD, recruited at thePediatric Department of the University of Rome ‘‘LaSapienza,’’ were included in the study. In addition, 424 parentswere also available for analysis (212 pairs). Healthy controlindividuals (n¼ 452) were recruited from asymptomaticindividuals (blood donors and staff members), all of whomwere of Italian origin. In all of the patients the diagnosis of CDwas made at the pediatric age, in accordance with the revisedcriteria of the European Society for Pediatric Gastroenterology,

CANDIDATE GENES AND CHROMOSO

yright © 2007 by Lippincott Williams & Wilkins.U

Hepatology, and Nutrition (23). All of the participants wereinformed about the objectives of the study. Written consent wasobtained from patients or parents.

Genotyping

Genomic DNA was extracted from whole-blood samples atthe Department of Experimental Medicine of the University ofRome by the salt-chloroform method. All of the participants werepreviously typed for HLA-DRB1, HLA-DQA1, and HLA-DQB1genes by the polymerase chain reaction sequence-specificprimers method with commercial kits (Dynal, UK).

The genotyping for the IBD5 locus and the MYO9B gene wasperformed at the Molecular Laboratory of the GastroenterologyUnit of the ‘‘Casa Sollievo della Sofferenza’’ Hospital in SanGiovanni Rotondo, Italy. Four variants of the IBD5 locus(IGR2198a_1, rs12521868, rs1050152, and rs2631367) weregenotyped by the restriction enzyme method and the TaqMansystem ABI PRISM 7700 (Applied Biosystems), respectively,as previously described (24,25). These variants have beenwidely used in studies investigating their involvement in IBDpredisposition (26). More specifically, they lie within a 250-kbgenomic region with a high linkage disequilibrium, whichhampered all of the attempts to better identify the disease-associated genetic variants (Fig. 1). Given the potential sharingof genetic variants between CD and IBD, we specifically aimedto investigate the same variants in patients with IBD.

Three MYO9B SNPs (rs962917, rs2305764, and rs1545620)were detected by the ABI 7700 TaqMan process. Polymerasechain reactions (15mL) were performed in 1X TaqMan UniversalPolymerase Chain Reaction Master Mix, 1X Genotyping AssayMix, and 50-ng genomic DNA. After denaturation at 508C for2 minutes and at 948C for 10 minutes, reaction was amplified for40 cycles: 15 seconds at 948C and 60 seconds at 608C.

Regarding the MYO9B variants, because more than 1000individuals were included in our study, and the known highlinkage disequilibrium (single haplotype 19-kB block) betweenthe 5 SNPs described by Monsuur et al (18), we selected theSNP rs2305764 that gave the smallest P value and strongestassociation in the study by Monsuur et al (18), and the SNPrs962917 evaluated in 3 of 5 previously published studies(18–22). In addition, the SNP rs1545620 inducing a codingchange (Ala1011Ser) within the third IQ domain of thegene (27) was investigated because it has also recently beenassociated with a predisposition to IBD (28). All 3 SNPs are intight linkage disequilibrium (Fig. 1).

Statistical Analysis

Comparison of allele and genotype frequencies wasperformed by the x2 and Fisher exact tests when appropriate.Tests for Hardy-Weinberg equilibrium, linkage disequilibrium,transmission disequilibrium, and haplotype frequencies wereperformed by the Haploview Software version 3.2 (http://www.broad.mit.edu/personal/jvbarret/haploview). Power calcu-lation was performed by use of the PS software (http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize).

Assuming similar results for the rs2305764 SNP as reportedin the Dutch population (18), with an allele frequency in controlsamples of 0.379, we estimated that our study had 98% power atP< 0.05 to detect association in Italian CD patients. Moreover,by assuming the lower 95% confidence intervals (CIs) of

5Q AND 19P IN CELIAC DISEASE 181

nauthorized reproduction of this article is prohibited.

genotypic odds ratios (ORs) reported in the Dutch population,we estimated an 81% power at P< 0.05 to detect association inour study.

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

Copy

of the

182 LATIANO ET AL.

Meta-analysis

We reviewed studies published before December 2006 toidentify articles related to the MYO9B gene and IBD5 locusregarding a putative association with CD. We used the followingmedical subject headings terms and/or text words: celiacdisease, MYO9B, IBD5, SLC22A5, SLC22A4. No restrictionswere placed on race, ethnicity, or geographic area. Each studywas included in the meta-analysis if it reported original data(independent among studies) and provided enough data toenable the calculation of ORs. Allele frequencies from eachrespective study were determined by the allele-countingmethod. A x2 test was used to determine whether the observedfrequencies of genotypes in control individuals conformed toHardy-Weinberg expectations. We examined the contrast ofthe effect of the variant allele against that of the common allele.

FIG. 1. Physical position and D0 values of linkage disequilibrium

right © 2007 by Lippincott Williams & Wilkins.Un

The point estimates of the risk, the OR, and its 95% CI wereestimated for each study. For each outcome (allele and genotypefrequencies), the between-study heterogeneity across all

FIG. 2. Linkage disequilibrium between investigated SNPs in con-trol individuals. Data given as D0 (within the boxes) and r2 values.

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

eligible comparisons was performed by use of the x2-basedQ statistics; this heterogeneity test assessed the null hypothesisthat all studies were evaluating the same effect. A significant Qstatistics (P< 0.10) indicated heterogeneity across studies. Datawere combined by use of both fixed and random effect models;random effects are more appropriate when heterogeneity ispresent. Analysis was conducted by the ComprehensiveMeta Analysis version 1.0.23 (1999 Biostat software; www.meta-analysis.com).

RESULTS

The clinical characteristics of the CD patients aresummarized in Table 1. The majority of patients(61.6%) had typical gastrointestinal symptoms; however,in one third of the patients, the symptoms were atypicaland 25 children were completely asymptomatic.

There were no significant deviations of genotypic datafrom the Hardy-Weinberg equilibrium at any variant, inboth control individuals and CD patients. A high linkagedisequilibrium for both markers of the IBD5 locus and

investigated SNPs on IBD5 locus and MYO9B gene.

authorized reproduction of this article is prohibited.

MYO9B gene was confirmed (Fig. 2).For all tested SNPs allele and genotype, frequencies

did not differ significantly between CD patients and

TABLE 1. Clinical characteristics of CD patients

Typical�

Atypicaly Silentz

Patients, % 61.6 31.0 7.4Age at diagnosis

(y, mean�SD)4.3� 4.2 6.7� 4.4 8.2� 4.6

Sex ratio (F/M) 1.5 1.9 2.2

�Gastrointestinal manifestations.

y Nongastrointestinal manifestations.z Asymptomatic.

Cop

ntroleg/DQ

CANDIDATE GENES AND CHROMOSOMES 5Q AND 19P IN CELIAC DISEASE 183

control individuals, as depicted in Table 2. Moreover, nocorrelation with clinical phenotype was apparent.

Table 3 summarizes the results of the haplotype-basedcase-control association analyses of 2 haplotype blocks

FIG. 3. Allele frequencies (%) of rs2305764 variant in healthy cohaplotype (at least 1 positive), and CD patients-carriers of DQ8-n

yright © 2007 by Lippincott Williams & Wilkins.U

(IBD5 and MYO9B). The haplotype frequencies of the4 SNPs of IBD5 locus (block 1) and the 3 combinedMYO9B variants (block 2) were similarly distributed in

TABLE 2. Genotype and allelic frequencies of the IBD5 locus andcontrol individu

Genotype CD (n¼ 337) n (%) HC n (%) P Alle

IBD5rs1050152 n¼ 354

CC 124 (37) 115 (32) 0.28 CCT 151 (45) 180 (51) TTT 62 (18) 59 (17)

rs2631367 n¼ 361GG 88 (26) 89 (25) 0.17 GCG 158 (47) 193 (53) CCC 91 (27) 79 (22)

IGR2198a_1 n¼ 422CC 125 (37) 150 (35) 0.36 CCG 148 (44) 205 (49) GGG 64 (19) 67 (16)

rs12521868 n¼ 424CC 127 (38) 155 (37) 0.22 CCA 148 (44) 208 (49) AAA 62 (18) 61 (14)

MYO9Brs962917 n¼ 452

GG 127 (38) 176 (39) 0.78 GAG 165 (49) 223 (49) AAA 45 (13) 53 (12)

rs1545620 n¼ 445TT 98 (29) 148 (33) 0.33 TGT 178 (53) 230 (52) GGG 61 (18) 67 (15)

rs2305764 n¼ 438GG 115 (34) 162 (37) 0.59 GAG 172 (51) 220 (50) AAA 50 (15) 56 (13)

patients with CD and control individuals. Moreover, thetransmission disequilibrium test analysis revealed noevidence of transmission distortion of all tested allelesin our CD samples (Table 4).

individuals (HC), CD patients, CD patients-carriers of DQ8/DQ22-neg haplotype. All P>0.05.

nauthorized reproduction of this article is prohibited.

All the 337 patients were HLA class II typed, and90.5% carried DQ high-risk molecules: 78.3% DQ2(DQA1�05, DQB1�02 in cis or trans position), 11.0%

MYO9B variants in patients with CD compared with healthyals (HC)

le CD n (%) HC n (%) P Odds ratio (95% CI)

399 (59) 410 (58) 1.00275 (41) 298 (42) 0.63 0.95 (0.77–1.17)

334 (50) 371 (51) 1.00340 (50) 351 (49) 0.49 1.08 (0.87–1.33)

398 (59) 505 (60) 1.00276 (41) 339 (40) 0.76 1.03 (0.84–1.27)

402 (60) 518 (61) 1.00272 (40) 330 (39) 0.57 1.06 (0.86–1.31)

419 (62) 575 (64)255 (38) 329 (36) 0.56 1.06 (0.87–1.31)

374 (55) 526 (59)300 (45) 364 (41) 0.15 1.16 (0.95–1.42)

402 (60) 544 (62)272 (40) 332 (38) 0.33 1.11 (0.90–1.36)

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

Copy

TABLE 3. Case-control association analysis of IBD5 andMYO9B haplotypes in CD probands vs control individuals

Haplotype frequencies (%) CD HC P

IBD5 (block 1)CGCC 49.6 48.4 0.66TCGA 37.9 37.2 0.77CCCC 7.7 8.0 0.83TCGC 1.6 1.9 0.73

MYO9B (block 2)GTG 52.6 56.6 0.14AGA 37.5 34.6 0.26GGG 6.7 5.0 0.17GTA 2.8 2.5 0.72

Block 1 is derived from SNPs rs1050152, rs2631367, IRG2198a_1,

TABLE 4. TDT analysis of the IBD5 and MYO9B markers

SNPs Allele T/U (%) P

IBD5IGR2198 G 110/98 (54) 0.41IGR2096 A 116/98 (54) 0.22rs1050152 T 115/99 (53) 0.27rs2631367 C 119/102 (54) 0.25

MYO9Brs962917 A 107/95 (53) 0.40rs1545620 G 108/94 (53) 0.32rs2305764 A 116/97 (64) 0.19

184 LATIANO ET AL.

DQ8 (DQA1�03-DQB1�0302) and 1.2% DQ2 and DQ8.Of the remaining 32 CD patients, 22 were DR7 positive,carrying only the DQB1�02 allele. We attempted todiscover any possible correlation between the HLA-DQ status and the MYO9B and IBD5 genotyping, butno significant difference was found. In Figure 3 thecorrelation between the HLA status and 1 of the MYO9BSNP (as an example) is depicted.

Finally, for the meta-analysis, 5 association studies forMYO9B (18–22) and 2 for the IBD5 (13,14) locus wereidentified in CD. We extracted frequency data of thers2305764 and rs962917 variants of the MYO9B gene andpooled them with the results of the present study. For the2 investigated SNPs of the MYO9B gene (Fig. 4), a

and rs12521868 of the IBD5 locus. Block 2 is obtained from SNPsrs962917, rs2305764, and rs1545620 of the MYO9B gene. Haplotypesare given on decreasing frequency.

right © 2007 by Lippincott Williams & Wilkins.Un

significant association with CD was found at the fixedmodel (OR¼ 1.1); however, because of the significantheterogeneity between studies (P< 0.002), when we

FIG. 4. Pooled data of the available studies in the literature and presentthe MYO9B gene in patients with CD. Data expressed as OR, 95% CI, an

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

evaluated the data with a random model, no significantassociation was appreciated. The meta-analysis of data ofthe single nucleotide polymorphisms of the IBD5 locuscould not be performed because the variants used in ourstudy differed from those in the 2 previously publishedstudies (13,14).

DISCUSSION

Besides the well-known association with HLA-DQgenes, numerous linkage studies point to genetic suscepti-bility loci for CD being located on chromosome 5q31–q33(CELIAC 2) and 19p13.1 (CELIAC 4). Interestingly, theycoincide with regions where linkage for IBD has beenestablished (29). CD and IBD are typical multifactorialdiseases with both environmental and genetic suscepti-bility factors involved in their pathogenesis. It is concei-

TDT indicates transmission disequilibrium test; T, transmittedalleles; U, untransmitted alleles.

authorized reproduction of this article is prohibited.

vable that these diseases, all with a clear involvement ofthe mucosal immune system, may share commonpredisposing genetic factors. Of note, in some families

study on allele frequency of the rs2305764 and rs962917 SNPs ofd P values. Values obtained with both fixed and random methods.

Cop

MES

(30), clusters of patients with IBD and CD have alsobeen described.

A highly conserved 250-kb haplotype, spanning thecytokine-rich gene cluster on 5q31 (IBD5), has beenreported to confer susceptibility to Crohn disease (31).More recently, within this locus the TC haplotype(1507C>T and �207G>C) in the OCTN gene clusterhas been suggested as a causative disease gene in Crohndisease (12), although the data are conflicting (32). In CDseveral plausible candidate genes in the same area havebeen tested (13,14,33), but no evidence of association hasbeen definitively provided.

Recently, Monsuur et al (18) reported a significantassociation of MYO9B gene variants within the CELIAC4locus and CD. The MYO9B is a good candidate genebecause it encodes for the myosin molecule, which playsa role in remodeling epithelial enterocytes and intestinalpermeability (18). In particular, the A allele of rs2305764SNP increases the risk for CD with an allele-dose effect(heterozygous OR 1.66; homozygous OR 2.27). Theoriginal study, performed in a Dutch cohort of CD patients,identified a significant linkage for CD to chromosome19p13.1 (34) by a genome-wide screen and also confirmedthe association of the D19S899 microsatellite marker(P¼ 1.3� 10�3), located at the intron 1 of the MYO9B)gene, in an independent case-control cohort. Interestingly,the same variants have been recently reported in thepredisposition to ulcerative colitis, another clinical formof IBD.

In the present study 4 genetic variations on IBD5 locus,2 of which tagged the OCTN gene cluster associated withCrohn disease (12), have been tested. In agreement withprevious findings (13,14), no association with CD orinteraction with HLA-DQ status was demonstrated.

Regarding the MYO9B gene, we tested 2 SNPs(rs962917 and rs2305764) that were shown to be stronglyassociated with CD in the Dutch cohort (18). Moreover, wealso investigated the nonsynonymous rs1545620 variant inexon 20, inasmuch as it induces a coding change (Ala1011-Ser) in the third calmodulin binding IQ domain of MYO9Band therefore might affect the function or expression of theprotein. A strong positive linkage disequilibrium betweenthe 3 SNPs was observed (Fig. 1). Although our study,carried out in a large cohort of pediatric CD patients, wasadequately powered, no evidence to support the originalresult of the Dutch study, either in the case-control study orin the family-based analysis, was found, in keepingwith the findings of other recent studies (19–22). Morespecifically, the minor allele of the rs2305764 SNP wasfound in our CD cohort at a similar frequency (40%) as inprevious studies (range 39%—46%) (18–22). Accord-ingly, the frequency in our control population (38%)was similar to that (40%–43%) in other studies

CANDIDATE GENES AND CHROMOSO

yright © 2007 by Lippincott Williams & Wilkins.U

(19–22), with the exception of the Dutch cohort (36%)(18). We also failed to detect any correlation between theMYO9B variants and the HLA-DQ status.

Lack of replication in the successive studies couldindicate a false-positive finding secondary to type I errorin the study by Monsuur et al (18). Although in that studya second independent case-control cohort and Bonferronicorrection to account for multiple testing was used, onlymultiple independent replication studies may excludethis possibility. Indeed, in the replication studies, includ-ing our own, the allele frequencies of the investigatedvariants were more similar across the patients than in thecontrol individuals. Accordingly, in this study we usedthe family-based strategy of analysis to avoid the dangerof an imperfect match with control individuals becausethis issue is particularly crucial in complex disease,especially with genes of moderate effect. Moreover, toaccount for the possibility that our study represented afalse negative because of inadequate power size, wecombined all of the available studies in a meta-analysis.A marginal significant association of MYO9B variantswith CD was detected (OR 1.1, CI 1.018–1.191);however, this finding was driven only by the Dutch study.By taking into account the significant heterogeneitybetween studies and evaluating the random model, weno longer detected an association, although more than2000 patients with CD were compared with 4000 healthycontrol individuals.

In conclusion, in our study cohort we cannot confirman association of the tagging SNPs of the IBD5 locus(included OCTN1 and 2 variants) with CD. Moreover,our data and the results of our meta-analysis question therole of MYO9B as a causative gene for CD.

Acknowledgments: The authors thank Ms Tiziana Latianoand Mr Corritore Giuspepe for genotyping assistance. They aregrateful to the patients and their parents who participated inthe study.

REFERENCES

1. Louka AS, Sollid LM. HLA in coeliac disease: unravelling thecomplex genetics of a complex disorder. Tissue Antigens 2003;61:105–17.

2. van Heel DA, Hunt K, Greco L, et al. Genetics in coeliac disease.Best Pract Res Clin Gastroenterol 2005;19:323–39.

3. Petronzelli F, Bonamico M, Ferrante P, et al. Genetic contribution ofthe HLA region to the familial clustering of coeliac disease. AnnHum Genet 1997;61:307–17.

4. Greco L, Babron MC, Corazza GR, et al. Existence of a genetic riskfactor on chromosome 5q in Italian coeliac disease families. AnnHum Genet 2001;65:35–41.

5. Naluai AT, Nilsson S, Samuelsson L, et al. The CTLA4/CD28 generegion on chromosome 2q33 confers susceptibility to celiac diseasein a way possibly distinct from that of type 1 diabetes and otherchronic inflammatory disorders. Tissue Antigens 2000;56:350–5.

6. Holopainen P, Mustalahti K, Uimari P, et al. Candidate gene regionsand genetic heterogeneity in gluten sensitivity. Gut 2001;48:696–701.

5Q AND 19P IN CELIAC DISEASE 185

nauthorized reproduction of this article is prohibited.

7. King AL, Fraser JS, Moodie SJ, et al. Coeliac disease: follow-uplinkage study provides further support for existence of a suscept-ibility locus on chromosome 11p11. Ann Hum Genet 2001;65:377–86.

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

Copy

ANO

8. Liu J, Juo SH, Holopainen P, et al. Genome-wide linkage analysis ofceliac disease in Finnish families. Am J Hum Genet 2002;70: 51–9.

9. Naluai AT, Nilsson S, Gudjonsdottir AH, et al. Genome-widelinkage analysis of Scandinavian affected sib-pairs supportspresence of susceptibility loci for celiac disease on chromosomes5 and 11. Eur J Hum Genet 2001;9:938–44.

10. Rioux JD, Silverberg MS, Daly MJ, et al. Genomewide search inCanadian families with inflammatory bowel disease reveals twonovel susceptibility loci. Am J Hum Genet 2000;66:1863–70.

11. vanHeelDA,Satsangi J,Carey AH,etal. Inflammatory boweldisease:progress toward a gene. Can J Gastroenterol 2000;14:207–18.

12. Peltekova VD, Wintle RF, Rubin LA, et al. Functional variants ofOCTN cation transporter genes are associated with Crohn’s disease.Nat Genet 2004;36:471–5.

13. Ryan AW, Thornton JM, Brophy K, et al. Haplotype variation at theIBD5/SLC22A4 locus (5q31) in coeliac disease in the Irish popula-tion. Tissue Antigens 2004;64:195–8.

14. Curley CR, Monsuur AJ, Wapenaar MC, et al. A functional candidatescreen for coeliac disease genes. Eur J Hum Genet 2006;14:1215–22.

15. Babron MC, Nilsson S, Adamovic S, et al. European GeneticsCluster on Coeliac Disease. Meta and pooled analysis of Europeancoeliac disease data. Eur J Hum Genet 2003;11:828–34.

16. Van Belzen MJ, Meijer JW, Sandkuijl LA, et al. A major non-HLAlocus in celiac disease maps to chromosome 19. Gastroenterology2003;125:1032–41.

17. van Heel DA, Dechairo BM, Dawson G, et al. The IBD6 Crohn’sdisease locus demonstrates complex interactions with CARD15 andIBD5 disease-associated variants. Hum Mol Genet 2003;12:2569–75.

18. Monsuur AJ, de Bakker PI, Alizadeh BZ, et al. Myosin IXB variantincreases the risk of celiac disease and points toward a primaryintestinal barrier defect. Nat Genet 2005;37:1341–4.

19. Hunt KA, Monsuur AJ, McArdle WL, et al. Lack of association ofMYO9B genetic variants with coeliac disease in a British cohort.

186 LATI

right © 2007 by Lippincott Williams & Wilkins.Un

20. Amundsen SS, Monsuur AJ, Wapenaar MC, et al. Associationanalysis of MYO9B gene polymorphisms with celiac disease ina Swedish/Norwegian cohort. Hum Immunol 2006;67:341–5.

J Pediatr Gastroenterol Nutr, Vol. 45, No. 2, August 2007

21. Giordano M, Marano C, Mellai M, et al. A family-based study doesnot confirm the association of MYO9B with celiac disease in theItalian population. Genes Immun 2006;7:606–8.

22. Nunez C, Marquez A, Varade J, et al. No evidence of association ofthe MYO9B polymorphisms with celiac disease in the Spanishpopulation. Tissue Antigens 2006;68:489–92.

23. Walker-Smith JA, Guandalini S, Schmitz J, et al. Revised criteriafor diagnosis of coeliac disease. Arch Dis Child 1990;65:909–11.

24. Latiano A, Palmieri O, Valvano RM, et al. Contribution of IBD5 locusto clinical features of IBD patients. Am J Gastroenterol 2006;101:318–25.

25. Palmieri O, Latiano A, Valvano R, et al. Variants of OCTN1-2cation transporter genes are associated with both Crohn’s diseaseand ulcerative colitis. Aliment Pharmacol Ther 2006;23:497–506.

26. Reinhard C, Rioux JD. Role of the IBD5 susceptibility locus in theinflammatory bowel diseases. Inflamm Bowel Dis 2006;12:227–38.

27. Altshuler D, Brooks LD, Chakravarti A, et al. A haplotype map ofhuman genome. Nature 2005;437:1299–320.

28. van Bodegraven AA, Curley CR, Hunt KA, et al. Genetic variation inmyosin IXB is associated with ulcerative colitis. Gastroenterology2006;131:1768–74.

29. Podolski DK. Inflammatory bowel disease. N Engl J Med 2002;347:417–29.

30. Yang A, Chen Y, Scherl E, et al. Inflammatory bowel disease inpatients with celiac disease. Inflamm Bowel Dis 2005;11:528–32.

31. Rioux JD, Daly MJ, Silverberg MS, et al. Genetic variation in the5q31 cytokine gene cluster confers susceptibility to Crohn disease.Nat Genet 2001;29:223–8.

32. Reinhard C, Rioux JD. Role of the IBD5 susceptibility locus in theinflammatory bowel disease. Inflamm Bowel Dis 2006;12:227–38.

33. Ryan AW, Thornton JM, Brophy K, et al. Chromosome 5q candi-date genes in coeliac disease: genetic variation at IL4, IL5, IL9,

ET AL.

IL13, IL17B and NR3C1. Tissue Antigens 2005;65:150–5.

Gut 2006;55:969–72.

authorized reproduction of this article is prohibited.

34. Van Belzen MJ, Meijer JW, Sandkuijl LA, et al. A major non-HLAlocus in celiac disease maps to chromosome 19. Gastroenterology2003;125:1032–41.