Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

No Evidence of Association or Interaction between the IL4RA, IL4, and IL13 Genes in Type 1 Diabetes

2005; Elsevier BV; Volume: 76; Issue: 3 Linguagem: Inglês

10.1086/428387

ISSN

1537-6605

Autores

Lisa M. Maier, Juliet Chapman, Joanna M. M. Howson, David Clayton, Rebecca Pask, David P. Strachan, Wendy L. McArdle, Rebecca C.J. Twells, John A. Todd,

Tópico(s)

IL-33, ST2, and ILC Pathways

Resumo

Attempts to identify susceptibility loci that, on their own, have marginal main effects by use of gene-gene interaction tests have increased in popularity. The results obtained from analyses of epistasis are, however, difficult to interpret. Gene-gene interaction, albeit only marginally significant, has recently been reported for the interleukin-4 and interleukin-13 genes (IL4 and IL13) with the interleukin-4 receptor A gene (IL4RA), contributing to the susceptibility of type 1 diabetes (T1D). We aimed to replicate these findings by genotyping both large family and case-control data sets and by using previously published data. Gene-gene interaction tests were performed using linear regression models in cases only. We did not find any single-locus associations with T1D and did not obtain evidence of gene-gene interaction. Additional support from independent samples will be even more important in the study of gene-gene interactions and other subgroup analyses. Attempts to identify susceptibility loci that, on their own, have marginal main effects by use of gene-gene interaction tests have increased in popularity. The results obtained from analyses of epistasis are, however, difficult to interpret. Gene-gene interaction, albeit only marginally significant, has recently been reported for the interleukin-4 and interleukin-13 genes (IL4 and IL13) with the interleukin-4 receptor A gene (IL4RA), contributing to the susceptibility of type 1 diabetes (T1D). We aimed to replicate these findings by genotyping both large family and case-control data sets and by using previously published data. Gene-gene interaction tests were performed using linear regression models in cases only. We did not find any single-locus associations with T1D and did not obtain evidence of gene-gene interaction. Additional support from independent samples will be even more important in the study of gene-gene interactions and other subgroup analyses. The genetic analysis of common, multifactorial diseases, such as type 1 diabetes (T1D [MIM 222100]), that are highly clustered in families is proving to be a challenging task (Altmuller et al. Altmüller et al., 2001Altmüller J Palmer LJ Fischer G Scherb H Wjst M Genomewide scans of complex human diseases: true linkage is hard to find.Am J Hum Genet. 2001; 69: 936-950Abstract Full Text Full Text PDF PubMed Scopus (422) Google Scholar; Hirschhorn et al. Hirschhorn et al., 2002Hirschhorn JN Lohmueller K Byrne E Hirschhorn K A comprehensive review of genetic association studies.Genet Med. 2002; 4: 45-61Crossref PubMed Scopus (1363) Google Scholar; Ioannidis et al. Ioannidis et al., 2003Ioannidis JP Trikalinos TA Ntzani EE Contopoulos-Ioannidis DG Genetic associations in large versus small studies: an empirical assessment.Lancet. 2003; 361: 567-571Abstract Full Text Full Text PDF PubMed Scopus (496) Google Scholar; Wang et al. Wang et al., 2005Wang WYS Barratt BJ Clayton DG Todd JA Genome-wide association studies: theoretical issues and practical concerns.Nat Rev Genet. 2005; 6: 109-118Crossref PubMed Scopus (883) Google Scholar). There are many reasons for this, including statistically underpowered small sample sizes; lack of coverage of the genome due to technical, and thus cost, limitations in genotyping; and statistical issues, such as subgroup analyses and very low prior probability of obtaining a true result (Dahlman et al. Dahlman et al., 2002Dahlman I Eaves IA Kosoy R Morrison VA Heward J Gough SC Allahabadia A Franklyn JA Tuomilehto J Tuomilehto-Wolf E Cucca F Guja C Ionescu-Tirgoviste C Stevens H Carr P Nutland S McKinney P Shield JP Wang W Cordell HJ Walker N Todd JA Concannon P Parameters for reliable results in genetic association studies in common disease.Nat Genet. 2002; 30: 149-150Crossref PubMed Scopus (198) Google Scholar; Thomas and Clayton Thomas and Clayton, 2004Thomas DC Clayton DG Betting odds and genetic associations.J Natl Cancer Inst. 2004; 96: 421-423Crossref PubMed Scopus (139) Google Scholar; Wang et al. Wang et al., 2005Wang WYS Barratt BJ Clayton DG Todd JA Genome-wide association studies: theoretical issues and practical concerns.Nat Rev Genet. 2005; 6: 109-118Crossref PubMed Scopus (883) Google Scholar). One possibility, which is gaining popularity among some authors (Culverhouse et al. Culverhouse et al., 2002Culverhouse R Suarez BK Lin J Reich T A perspective on epistasis: limits of models displaying no main effect.Am J Hum Genet. 2002; 70: 461-471Abstract Full Text Full Text PDF PubMed Scopus (295) Google Scholar; Moore Moore, 2003Moore JH The ubiquitous nature of epistasis in determining susceptibility to common human diseases.Hum Hered. 2003; 56: 73-82Crossref PubMed Scopus (574) Google Scholar; Hoh and Ott Hoh and Ott, 2004Hoh J Ott J Genetic dissection of diseases: design and methods.Curr Opin Genet Dev. 2004; 14: 229-232Crossref PubMed Scopus (39) Google Scholar), is that susceptibility gene effects may be only identified in analyses of statistical interaction, since the individual marginal effects may be close to null. However, these extreme models of epistasis are difficult to explain biologically. Furthermore, the interpretation of statistical interaction in terms of "epistatic" mechanisms is problematic (Cordell Cordell, 2002Cordell HJ Epistasis: what it means, what it doesn't mean, and statistical methods to detect it in humans.Hum Mol Genet. 2002; 11: 2463-2468Crossref PubMed Scopus (747) Google Scholar). Whether testing for interactions or subgroup analyses increases the power to detect disease susceptibility genes or not, such approaches may exacerbate the problem of false positives due to the even lower prior probability of detecting a true positive (Thomas and Clayton Thomas and Clayton, 2004Thomas DC Clayton DG Betting odds and genetic associations.J Natl Cancer Inst. 2004; 96: 421-423Crossref PubMed Scopus (139) Google Scholar; Wang et al. Wang et al., 2005Wang WYS Barratt BJ Clayton DG Todd JA Genome-wide association studies: theoretical issues and practical concerns.Nat Rev Genet. 2005; 6: 109-118Crossref PubMed Scopus (883) Google Scholar). Subsequently, for reliable gene-gene interaction results, very small P-value thresholds, less than P<10−6, and larger sample sizes, in addition to replication in independent samples, may be required. Recently, in a sample of 90 cases of T1D and 94 Filipino population-based controls, Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) reported evidence of an interaction between 10 SNPs in the interleukin-4 receptor A gene (IL4RA [MIM 147781]) on chromosome 16p11-p12 (SNPs 5′ −3223C→T [rs2057768], 5′ −1914C→T [rs2107356], I50V [rs1805010], N142N [rs2234895], E375A [rs1805011], L389L [rs2234898], C406R [rs1805012], S478P [rs1805015], Q551R [rs1801275], and S761P [rs1805014]) and 5 SNPs in the adjacent interleukin-4 and interleukin-13 genes (IL4 [MIM 147780] and IL13 [MIM 147683]) on chromosome 5q31 (SNPs 5′ −524T→C [rs2243250] in IL4 and 5′ −1512A→C [rs1881457], 5′ −1112C→T [rs1800925], +1923C→T/intron 3 [rs1295686], and R110Q [rs20541] in IL13). Of these 15 SNPs, 4 showed some evidence of primary disease association—namely, E375A (P=.02), L389L (P=.001), and C406R (P=.05) in IL4RA and the 5′ −1512A→C SNP in IL13 (P=.05). Furthermore, at IL4RA, a 7-locus haplotype (P=.005) and a 10-locus haplotype (P=.001) showed some evidence of association. Additional effects of five-locus haplotypes at IL4 and IL13 (smallest P value = .004) were observed, as well as gene-gene interaction of the IL4 and IL13 loci with IL4RA (P<.045) (Bugawan et al. Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). Note that, to correct for multiple testing, Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) permutated genotypes at chromosomes 5 and 16 within patients and controls, keeping genotype frequencies constant. The P values were <.05, but, nevertheless, if the reported effect were true, then a much larger study should confirm the result. Given the linkage and association results from studies of T1D and other diseases that are inherited in a similarly complex way, it is likely that non–human leukocyte antigen T1D-susceptibility loci will have effect sizes with odds ratios (ORs) <2.0 (Wang et al. Wang et al., 2005Wang WYS Barratt BJ Clayton DG Todd JA Genome-wide association studies: theoretical issues and practical concerns.Nat Rev Genet. 2005; 6: 109-118Crossref PubMed Scopus (883) Google Scholar). However, since the sample size used by Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) is very small (90 cases and 94 controls) for purposes of detecting a true effect, the effect size would need to be in considerable excess of OR 2.0 for the minor allele. Since the effects reported here for minor alleles are much smaller than this, the posterior probability that Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) have detected a true effect is very small. In this study, we attempted replication of the main findings reported by Bugawan et al., using a larger sample of up to 748 multiplex families and up to 1,616 cases and 1,829 controls. First, we tested the primary, single-locus association of the 15 SNPs from IL4, IL13, and IL4RA in a white British case-control sample. The case-control DNA set consisted of 1,616 white individuals with T1D who were recruited from across Britain for the Juvenile Diabetes Research Foundation/Wellcome Trust–funded U.K. GRID study (see U.K. GRID Study Web site), and 1,829 population-based controls from the 1958 British Birth Cohort (see National Child Development Study Web site). The mean age at onset of the patients, who were all <16 years old at diagnosis, is 7.5 years (SD 4 years). We obtained no evidence of association with T1D (table 1). The SNPs included IL13 R110Q, which showed some evidence of association in previous studies (Bugawan et al. Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). Genotyping of IL4RA L389L was not performed, since this variant is in very strong linkage disequilibrium with its neighboring SNPs E375A, C406R, and Q551R (Bugawan et al. Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). We employed the Invader (Third Wave Technologies) and TaqMan (Applied Biosystems) genotyping technologies. Primer and probe sequences are shown in Table A1, Table A2 (online only). The IL4RA results were consistent with our previous study of 3,475 families with T1D (Maier et al. Maier et al., 2003Maier LM Twells RC Howson JM Lam AC Clayton DG Smyth DJ Savage D Carson D Patterson CC Smink LJ Walker NM Burren OS Nutland S Rance H Tuomilehto-Wolf E Tuomilehto J Guja C Ionescu-Tirgoviste C Undlien DE Ronningen KS Cucca F Todd JA Testing the possible negative association of type 1 diabetes and atopic disease by analysis of the interleukin 4 receptor gene.Genes Immun. 2003; 4: 469-475Crossref PubMed Scopus (10) Google Scholar), in that no P values <.05 were obtained. There was also no evidence for the previously suggested IL13 R110Q and E375A genotype associations: the OR was 0.90 (95% CI 0.77–1.06) for the IL13 R110Q AG heterozygote and 1.05 (95% CI 0.88–1.26) for the E375A AC heterozygote. Furthermore, no evidence was obtained for the previously reported associations of the IL4RA −3223C→T SNP (P>.05) (Maier et al. Maier et al., 2003Maier LM Twells RC Howson JM Lam AC Clayton DG Smyth DJ Savage D Carson D Patterson CC Smink LJ Walker NM Burren OS Nutland S Rance H Tuomilehto-Wolf E Tuomilehto J Guja C Ionescu-Tirgoviste C Undlien DE Ronningen KS Cucca F Todd JA Testing the possible negative association of type 1 diabetes and atopic disease by analysis of the interleukin 4 receptor gene.Genes Immun. 2003; 4: 469-475Crossref PubMed Scopus (10) Google Scholar) or for that of a six-locus IL4RA haplotype (P>.05) (Mirel et al. Mirel et al., 2002Mirel DB Valdes AM Lazzeroni LC Reynolds RL Erlich HA Noble JA Association of IL4R haplotypes with type 1 diabetes.Diabetes. 2002; 51: 3336-3341Crossref PubMed Scopus (55) Google Scholar; Bugawan et al. Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar; Maier et al. Maier et al., 2003Maier LM Twells RC Howson JM Lam AC Clayton DG Smyth DJ Savage D Carson D Patterson CC Smink LJ Walker NM Burren OS Nutland S Rance H Tuomilehto-Wolf E Tuomilehto J Guja C Ionescu-Tirgoviste C Undlien DE Ronningen KS Cucca F Todd JA Testing the possible negative association of type 1 diabetes and atopic disease by analysis of the interleukin 4 receptor gene.Genes Immun. 2003; 4: 469-475Crossref PubMed Scopus (10) Google Scholar). We have analyzed two haplotypes, one consisting of IL4RA I50V (rs1805010), E375A (rs1805011), C406R (rs1805012), S411L (rs1805013), S478P (rs1805015), and Q551R (rs1801275) and the other consisting of IL13 −1512A→C (rs1881457), IL13 −1112C→T (rs1800925), IL13 +1923C→T (rs1295686), IL13 R110Q (rs20541), and IL4 −524C→T (rs2243250). However, no results with P<.05 were obtained, and all 95% CIs overlapped 1. Results from haplotype analyses for IL4RA and IL4/IL13 are given in Table A3, Table A4 (online only).Table 1Association Analysis of IL4, IL13, and IL4RA SNPs in a T1D Case-Control SampleMAF forGene and SNPdbSNPNucleotide ChangeNo. of Controls with GenotypeNo. of Cases with GenotypeControlsCasesPIL4: 5′ −524rs2243250T→C1,6221,557.13.14.47IL13: 5′ −1512rs1881457A→C1,6551,578.19.17.87 5′ −1112rs1800925C→T1,6601,583.18.17.83 +1923rs1295686C→T1,6241,559.18.18.43 R110Qrs20541G→A1,6531,559.17.18.41IL4RA: 5′ −3223rs2057768C→T1,5781,592.30.29.46 I50Vrs1805010A→G1,5821,583.46.45.48 E375Ars1805011A→C1,6221,565.12.12.79 C406Rrs1805012T→C1,6741,575.11.11.49 S411Lrs1805013C→T1,6531,587.05.05.16 S478Prs1805015T→C1,6391,584.18.17.35 Q551Rrs1801275A→G1,6731,590.22.22.73 S761Prs1805014T→C1,6441,590.01.01.87Note.—To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Note that, for every SNP, genotyping of DNA samples from 1,829 controls and 1,616 patients with T1D was attempted. The numbers of cases and controls, MAFs of cases and controls, and P values were obtained by association analyses with T1D. Open table in a new tab Note.— To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Note that, for every SNP, genotyping of DNA samples from 1,829 controls and 1,616 patients with T1D was attempted. The numbers of cases and controls, MAFs of cases and controls, and P values were obtained by association analyses with T1D. We note that the minor-allele frequencies (MAFs) of all SNPs except for IL4RA I50V and S478P are lower in our U.K. and U.S. populations than in the 184 Filipino subjects studied elsewhere (Bugawan et al. Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). However, the IL4RA S761P SNP, which was monomorphic in Filipinos, was polymorphic in our population (MAF 0.01). Nevertheless, our larger data set and, hence, increased statistical power should compensate for the loss of power due to the lower MAFs. Second, we tested for evidence of an interaction (i.e., deviation from a multiplicative model of epistasis) between eight IL4RA SNPs and the one IL4 and four IL13 SNPs in a case-only analysis (table 2) (Piegorsch et al. Piegorsch et al., 1994Piegorsch WW Weinberg CR Taylor JA Non-hierarchical logistic models and case-only designs for assessing susceptibility in population-based case-control studies.Stat Med. 1994; 13: 153-162Crossref PubMed Scopus (372) Google Scholar; Umbach and Weinberg Umbach and Weinberg, 1997Umbach DM Weinberg CR Designing and analysing case-control studies to exploit independence of genotype and exposure.Stat Med. 1997; 16: 1731-1743Crossref PubMed Scopus (152) Google Scholar). We assume markers are in linkage equilibrium in the general population (i.e., the controls). A convenient and powerful approach is to test the correlation coefficient between the genotypes at the two loci, scored 0, 1, and 2 (Piegorsch et al. Piegorsch et al., 1994Piegorsch WW Weinberg CR Taylor JA Non-hierarchical logistic models and case-only designs for assessing susceptibility in population-based case-control studies.Stat Med. 1994; 13: 153-162Crossref PubMed Scopus (372) Google Scholar; Umbach and Weinberg Umbach and Weinberg, 1997Umbach DM Weinberg CR Designing and analysing case-control studies to exploit independence of genotype and exposure.Stat Med. 1997; 16: 1731-1743Crossref PubMed Scopus (152) Google Scholar). We genotyped a DNA collection consisting of 1,616 cases and 1,829 controls from Great Britain, for the previously associated candidate SNPs. The candidate SNPs for IL4 and IL13 were IL4 −524C→T (rs2243250), IL13 −1512A→C (rs1881457), IL13 −1112C→T (rs1800925), IL13 +1923C→T (rs1295686), and IL13 R110Q (rs20541). We did not observe any evidence of interaction between single SNPs in our data set (P>.05) (table 2).Table 2P Values Obtained by Regression Analyses of Interactions of IL13 and IL4 SNPs with Eight IL4RA SNPs for Susceptibility to T1DIL4RA SNP (dbSNP)Gene and SNP (dbSNP)5′ 3223 (rs2057768)I50V (rs1805010)E375A (rs1805011)C406R (rs1805012)S411L (rs1805013)S478P (rs1805015)Q551R (rs1801275)S761P (rs1805014)IL4: 5′ −524 (rs2243250).92.56.98.96.45.93.61.55IL13: 5′ −1512 (rs1881457).63.87.95.54.80.70.34.43 5′ −1112 (rs1800925).48.68.99.52.98.81.52.43 +1923 (rs1295686).41.37.98.47.75.39.61.79 R110Q (rs20541).38.43.26.53.64.35.62.45Note.—Results are shown for a sample of 1,616 white patients. To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Open table in a new tab Note.— Results are shown for a sample of 1,616 white patients. To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Third, we considered the possibility that there is a true association and that the etiological variant has not been identified yet. To ensure that we captured the information in the IL4 and IL13 regions, we genotyped tagSNPs for both genes, using 748 white families with T1D from the United Kingdom and the United States, to perform further interaction analyses with IL4RA. The families with T1D each included two parents and at least one affected child. The 748 families with T1D comprised 472 multiplex families from the U.K. Warren 1 repository (Bain et al. Bain et al., 1990Bain SC Todd JA Barnett AH The British Diabetic Association—Warren repository.Autoimmunity. 1990; 7: 83-85Crossref PubMed Scopus (74) Google Scholar) and 276 multiplex families from the Human Biological Data Interchange ascertained in the United States (Lernmark Lernmark, 1991Lernmark A Human cell lines from families available for diabetes research.Diabetologia. 1991; 34: 61Crossref PubMed Scopus (8) Google Scholar), with inclusion criteria described elsewhere (Vella et al. Vella et al., 2004Vella A Howson JMM Barratt BJ Twells RCJ Rance HE Nutland S Tuomilehto-Wolf E Tuomilehto J Undlien DE Ronningen KS Guja C Ionescu-Tîrgoviste C Savage DA Todd JA Lack of association of the Ala45Thr polymorphism and other common variants of the NeuroD gene with type 1 diabetes.Diabetes. 2004; 53: 1158-1161Crossref PubMed Scopus (22) Google Scholar). The sequencing data and genotypes of a sequencing panel required for the tagging approach were obtained from the University of Washington–Fred Hutchinson Cancer Research Center (UW-FHCRC) Cancer Variation Discovery Resource (SeattleSNPs), published at the UWFHCRC Web site. From this resource, we downloaded data on common polymorphisms of exons and introns from 23 white individuals (from Centre d'Etude du Polymorphism Humain [CEPH]) and used these to select tagSNPs. For IL4 and IL13, 12 and 6 tagSNPs, respectively, with MAFs >0.05, were selected as described elsewhere (Chapman et al. Chapman et al., 2003Chapman JM Cooper JD Todd JA Clayton DG Detecting disease associations due to linkage disequilibrium using haplotype tags: a class of tests and the determinants of statistical power.Hum Hered. 2003; 56: 18-31Crossref PubMed Scopus (347) Google Scholar; Clayton et al. Clayton et al., 2004Clayton D Chapman J Cooper J The use of unphased multilocus genotype data in indirect association studies.Genet Epidemiol. 2004; 27: 415-428Crossref PubMed Scopus (165) Google Scholar) and are listed in Table A5, Table A6 (online only). With the chosen subset of tagSNPs, the remaining SNPs were required to be predicted with a minimum locus R2 of 0.80. In total, 18 tagSNPs were selected and genotyped in the same 748 families with T1D investigated for IL4RA association by Maier et al. (Maier et al., 2003Maier LM Twells RC Howson JM Lam AC Clayton DG Smyth DJ Savage D Carson D Patterson CC Smink LJ Walker NM Burren OS Nutland S Rance H Tuomilehto-Wolf E Tuomilehto J Guja C Ionescu-Tirgoviste C Undlien DE Ronningen KS Cucca F Todd JA Testing the possible negative association of type 1 diabetes and atopic disease by analysis of the interleukin 4 receptor gene.Genes Immun. 2003; 4: 469-475Crossref PubMed Scopus (10) Google Scholar). Note that, in the selection of tagSNPs for IL4, the IL4 −524 variant was selected as a tagSNP. Similarly, for IL13, the IL13 −1512A→C (rs1881457), IL13 −1112C→T (rs1800925), and IL13 +1923C→T (rs1295686) variants were selected as tagSNPs. The multilocus test (Chapman et al. Chapman et al., 2003Chapman JM Cooper JD Todd JA Clayton DG Detecting disease associations due to linkage disequilibrium using haplotype tags: a class of tests and the determinants of statistical power.Hum Hered. 2003; 56: 18-31Crossref PubMed Scopus (347) Google Scholar; Clayton et al. Clayton et al., 2004Clayton D Chapman J Cooper J The use of unphased multilocus genotype data in indirect association studies.Genet Epidemiol. 2004; 27: 415-428Crossref PubMed Scopus (165) Google Scholar) P values for IL4 and IL13 were .58 and .74, respectively, indicating that neither locus was associated with T1D in this U.K. and U.S. family collection, which is consistent with the data obtained in our U.K. case-control collection. We tested for interaction, a deviation from the model of multiplicative effect, between single SNPs in one region or gene and a set of tagSNPs in a second region, by regressing the genotype at the single SNP on all tag genotypes in cases. For each IL4RA locus, we used this regression technique to test for an interaction between specific IL4RA loci and the IL4 region, as well as the IL13 region. As shown in table 3, neither IL4 nor IL13 tagSNPs showed P values <.05 with eight IL4RA SNPs for an interaction effect. For completeness, we also performed the same test employed by Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) in both a case-control and case-only design for interaction between IL4RA SNPs and IL13 tagSNPs and between IL4RA SNPs and IL4 tagSNPs, but obtained no evidence of interaction (P>.05).Table 3P Values for Interaction Tests Obtained from Regression Analyses of IL13 and IL4 tagSNPs with Eight IL4RA SNPsPValue forIL4RA SNP (dbSNP)IL13 tagSNPsIL4 tagSNPs5′ −3223 (rs2057768).72.97I50V (rs1805010).14.37E375A (rs1805011).21.62C406R (rs1805012).55.20S411L (rs1805013).39.43S478P (rs1805015).71.97Q551R (rs1801275).65.51S761P (rs1805014).55.24Note.—Results are shown for a U.K. and U.S. family collection of up to 748 families with T1D. Open table in a new tab Note.— Results are shown for a U.K. and U.S. family collection of up to 748 families with T1D. We conclude that there is no evidence of interaction between the IL4RA and the IL13 or IL4 loci in susceptibility to T1D, with regard to the investigated variants in our populations of European descent. The previously published result obtained in 184 individuals from the Philippines may be specific to that sample population, but a study of such small size has minimal power to detect a true interaction effect. In comparing the study of Bugawan et al. (Bugawan et al., 2003Bugawan TL Mirel DB Valdes AM Panelo A Pozzilli P Erlich HA Association and interaction of the IL4R, IL4, and IL13 loci with type 1 diabetes among Filipinos.Am J Hum Genet. 2003; 72: 1505-1514Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar) with our study, it is evident that the present study has substantially greater statistical power. This is because we not only used a larger sample size but also employed regression analyses that use only cases, and this approach has been shown elsewhere to be more powerful when the assumption of independence is true (Piegorsch et al. Piegorsch et al., 1994Piegorsch WW Weinberg CR Taylor JA Non-hierarchical logistic models and case-only designs for assessing susceptibility in population-based case-control studies.Stat Med. 1994; 13: 153-162Crossref PubMed Scopus (372) Google Scholar; Umbach and Weinberg Umbach and Weinberg, 1997Umbach DM Weinberg CR Designing and analysing case-control studies to exploit independence of genotype and exposure.Stat Med. 1997; 16: 1731-1743Crossref PubMed Scopus (152) Google Scholar). Failure to replicate genetic association studies is well documented in the literature and, even though several reasons have been reported—including allelic heterogeneity, true variation in disease association between populations, modifying genetic and/or environmental factors, and misclassification of outcome—the most important factor is probably insufficient and/or inadequate sample sizes in the context of a very-low prior probability of detecting a true effect (Clayton and McKeigue Clayton and McKeigue, 2001Clayton D McKeigue PM Epidemiological methods for studying genes and environmental factors in complex diseases.Lancet. 2001; 358: 1356-1360Abstract Full Text Full Text PDF PubMed Scopus (394) Google Scholar; Dahlman et al. Dahlman et al., 2002Dahlman I Eaves IA Kosoy R Morrison VA Heward J Gough SC Allahabadia A Franklyn JA Tuomilehto J Tuomilehto-Wolf E Cucca F Guja C Ionescu-Tirgoviste C Stevens H Carr P Nutland S McKinney P Shield JP Wang W Cordell HJ Walker N Todd JA Concannon P Parameters for reliable results in genetic association studies in common disease.Nat Genet. 2002; 30: 149-150Crossref PubMed Scopus (198) Google Scholar; Hirschhorn et al. Hirschhorn et al., 2002Hirschhorn JN Lohmueller K Byrne E Hirschhorn K A comprehensive review of genetic association studies.Genet Med. 2002; 4: 45-61Crossref PubMed Scopus (1363) Google Scholar; Colhoun et al. Colhoun et al., 2003Colhoun HM McKeigue PM Davey Smith G Problems of reporting genetic associations with complex outcomes.Lancet. 2003; 361: 865-872Abstract Full Text Full Text PDF PubMed Scopus (974) Google Scholar; Ioannidis et al. Ioannidis et al., 2003Ioannidis JP Trikalinos TA Ntzani EE Contopoulos-Ioannidis DG Genetic associations in large versus small studies: an empirical assessment.Lancet. 2003; 361: 567-571Abstract Full Text Full Text PDF PubMed Scopus (496) Google Scholar; Lohmueller et al. Lohmueller et al., 2003Lohmueller KE Pearce CL Pike M Lander ES Hirschhorn JN Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease.Nat Genet. 2003; 33: 177-182Crossref PubMed Scopus (1583) Google Scholar; Thomas and Clayton Thomas and Clayton, 2004Thomas DC Clayton DG Betting odds and genetic associations.J Natl Cancer Inst. 2004; 96: 421-423Crossref PubMed Scopus (139) Google Scholar; Wang et al. Wang et al., 2005Wang WYS Barratt BJ Clayton DG Todd JA Genome-wide association studies: theoretical issues and practical concerns.Nat Rev Genet. 2005; 6: 109-118Crossref PubMed Scopus (883) Google Scholar). Such problems will be much more extreme for reported gene-gene interactions; power to detect interaction is low and the problems of low prior probability and subgroup analyses are more extreme. These results highlight the necessity of attempting replication in independent samples before drawing conclusions about a potential disease-susceptibility locus or gene-gene interaction. We acknowledge the use of DNA from the 1958 British Birth Cohort collection, funded by Medical Research Council grant G0000934 and Wellcome Trust grant 068545/Z/02. We gratefully acknowledge the participation of all patients, controls, and family members and the support of the Juvenile Diabetes Research Foundation and the Wellcome Trust. L.M.M. was a Wellcome Trust Prize student. Note.— NA = sequence not available. Note.— Data are for six SNPs: IL4RA I50V (rs1805010), E375A (rs1805011), C406R (rs1805012), S411L (rs1805013), S478P (rs1805015), and Q551R (rs1801275). SNP S761P (rs1805014) was excluded from analysis because of low MAF. To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Note.— Data are for IL13 −1512A→C (rs1881457), IL13 −1112C→T (rs1800925), IL13 +1923C→T (rs1295686), IL13 R110Q (rs20541), and IL4 −524C→T (rs2243250). To correct for regional variation in allele frequencies, the analyses were stratified according to 12 broad geographical regions within Great Britain (D.G.C., unpublished data). Note.— tagSNPs and minimal R2 values are given for SNPs with MAFs >0.05. Note that SNPs with MAFs 0.05. Note that SNPs with MAFs <0.05 are not shown. See UW-FHCRC Web site for SeattleSNPs.

Referência(s)