Associations between the 17q21 region and allergic rhinitis in 5 birth cohorts

The high heritability and comorbidity between allergic rhinitis and asthma suggest common etiologies and genetic susceptibility loci. Among the most robust signals for asthma is the chromosomal locus 17q21.1Moffatt M.F. Kabesch M. Liang L. Dixon A.L. Strachan D. Heath S. et al.Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma.Nature. 2007; 448: 470-473Crossref PubMed Scopus (1219) Google Scholar Studies that examined whether this strong asthma locus is also associated with allergic rhinitis have yielded conflicting results. A recent large genome-wide association study (GWAS)2Ferreira M.A.R. Matheson M.C. Tang C.S. Granell R. Ang W. Hui J. et al.Genome-wide association analysis identifies 11 risk variants associated with the asthma with hay fever phenotype.J Allergy Clin Immunol. 2014; 133: 1564-1571Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar and a candidate gene study3Tomita K. Sakashita M. Hirota T. Tanaka S. Masuyama K. Yamada T. et al.Variants in the 17q21 asthma susceptibility locus are associated with allergic rhinitis in the Japanese population.Allergy. 2013; 68: 92-100Crossref PubMed Scopus (32) Google Scholar identified associations between genetic variants at the 17q21 locus and allergic rhinitis, in contrast to null findings in previous GWASs on allergic rhinitis4Andiappan A.K. Wang D.Y. Anantharaman R. Parate P.N. Suri B.K. Low H.Q. et al.Genome-wide association study for atopy and allergic rhinitis in a Singapore Chinese population.PLoS One. 2011; 6: e19719Crossref PubMed Scopus (58) Google Scholar, 5Ramasamy A. Curjuric I. Coin L.J. Kumar A. McArdle W.L. Imboden M. et al.A genome-wide meta-analysis of genetic variants associated with allergic rhinitis and grass sensitization and their interaction with birth order.J Allergy Clin Immunol. 2011; 128: 996-1005Abstract Full Text Full Text PDF PubMed Scopus (169) Google Scholar and self-reported allergy.6Hinds D.A. McMahon G. Kiefer A.K. Do C.B. Eriksson N. Evans D.M. et al.A genome-wide association meta-analysis of self-reported allergy identifies shared and allergy-specific susceptibility loci.Nat Genet. 2013; 45: 907-911Crossref PubMed Scopus (172) Google Scholar Large, well-defined studies are needed to clarify these inconsistencies. We pooled data from 5 birth cohorts not included in the aforementioned GWASs (Children, Allergy, Milieu, Stockholm, Epidemiological Survey [BAMSE], Canadian Asthma Primary Prevention Study [CAPPS], German Infant study on the influence of Nutritional Intervention plus environmental and genetic influences on allergy development [GINIplus]/Lifestyle-related factors, Immune System and the development of Allergies in East and West Germany plus the influence of traffic emissions and genetics study [LISAplus], Prevention and Incidence of Asthma and Mite Allergy [PIAMA], and Study of Asthma Genes and the Environment [SAGE]; Ntotal = 4624 children, 92.7% white) and examined whether 7 single nucleotide polymorphisms (SNPs) at the 17q21 locus are associated with allergic rhinitis from early childhood to adolescence, and the effects of comorbidity with asthma on the associations. Allergic rhinitis was defined as concomitantly reported rhinitis and positive sensitization to any aeroallergen (age range 4-16 years in pooled data). Seven SNPs, including top GWAS hits for asthma (rs7216389 and rs2305480),1Moffatt M.F. Kabesch M. Liang L. Dixon A.L. Strachan D. Heath S. et al.Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma.Nature. 2007; 448: 470-473Crossref PubMed Scopus (1219) Google Scholar were extracted from genotyped (BAMSE) or imputed genome-wide (other cohorts) data and coded as 0, 1, and 2 according to the number of effect alleles. Cohort-specific study designs, outcome definitions, and genotyping/imputation are provided in this article's Methods section and Table E1 in the Online Repository at www.jacionline.org. The effect of each SNP on allergic rhinitis over time was analyzed using generalized estimating equation models (logit link and exchangeable correlation structure) adjusted for age and cohort (pooled models only). Associations were calculated per cohort and in the combined pooled data. Odds ratios with corresponding 95% CIs are presented. Allergic rhinitis was reported at least once by 969 of the 4624 children (Table I). The proportion of cases generally increased with age (see Table E2 in this article's Online Repository at www.jacionline.org). Effect allele frequencies were very similar across cohorts for all SNPs (see Table E3 in this article's Online Repository at www.jacionline.org). The correlation between 5 of the SNPs (rs2305480, rs7216389, rs4065275, rs8076131, and rs12603332) was high (r2 > 0.7, D′ > 0.9) and more moderate between these 5 SNPs and the other 2 (r2 < 0.3, D′ > 0.9; rs17608925 and rs3744246).Table IPooled longitudinal associations between 7 SNPs at the 17q21 locus and health outcomes∗Models were adjusted for age and cohort.SNPGeneEffect alleleEAFAllergic rhinitis†Age-specific controls were defined as those without allergic rhinitis. P values of significant pooled associations ranged from .002 for rs2305480 to .017 for rs3744246. Bonferroni-corrected P value corresponds to .05/7 = .007. The associated SNPs are in linkage disequilibrium (r2 > 0.2, D′ > 0.9). (case ever = 969 of 4624)Asthma‡Age-specific controls were defined as those without asthma. (case ever = 571 of 4619)Allergic rhinitis with concomitant asthma§Age-specific controls were defined as those without allergic rhinitis or asthma. (case ever = 234 of 4019)Allergic rhinitis without concomitant asthma§Age-specific controls were defined as those without allergic rhinitis or asthma. (case ever = 816 of 4425)NOR (95% CI)NOR (95% CI)NOR (95% CI)NOR (95% CI)rs2305480GSDMBA0.4546050.85 (0.77-0.94)P < .05.46000.83 (0.72-0.96)P < .05.40020.72 (0.59-0.88)P < .05.44070.89 (0.80-0.99)P < .05.rs7216389GSDMBT0.4946041.15 (1.04-1.27)P < .05.45991.15 (1.01-1.32)P < .05.40011.35 (1.12-1.63)P < .05.44081.10 (0.99-1.22)rs4065275ORMDL3G0.5044211.14 (1.03-1.26)P < .05.44171.13 (0.98-1.29)38591.35 (1.11-1.64)P < .05.42661.09 (0.98-1.22)rs8076131ORMDL3G0.4544170.86 (0.77-0.95)P < .05.44130.85 (0.73-0.98)P < .05.38560.73 (0.60-0.89)P < .05.42620.89 (0.80-1.00)P < .05.rs12603332ORMDL3C0.5046181.15 (1.04-1.27)P < .05.44261.13 (0.98-1.29)40141.34 (1.11-1.61)P < .05.44201.10 (0.99-1.23)rs17608925ORMDL3C0.1144140.88 (0.74-1.05)44100.99 (0.79-1.25)38530.94 (0.67-1.32)42590.89 (0.74-1.07)rs3744246ORMDL3T0.2046070.86 (0.75-0.97)P < .05.46020.89 (0.75-1.06)40040.83 (0.65-1.06)44110.89 (0.77-1.02)Case ever, Whether a child ever had a positive report at any follow-up; EAF, effect allele frequency; GSDMB, gasdermin B; N, number of children included in the model; OR, odds ratio; ORMDL3, ORM1-like protein 3.∗ Models were adjusted for age and cohort.† Age-specific controls were defined as those without allergic rhinitis. P values of significant pooled associations ranged from .002 for rs2305480 to .017 for rs3744246. Bonferroni-corrected P value corresponds to .05/7 = .007. The associated SNPs are in linkage disequilibrium (r2 > 0.2, D′ > 0.9).‡ Age-specific controls were defined as those without asthma.§ Age-specific controls were defined as those without allergic rhinitis or asthma.|| P < .05. Open table in a new tab Case ever, Whether a child ever had a positive report at any follow-up; EAF, effect allele frequency; GSDMB, gasdermin B; N, number of children included in the model; OR, odds ratio; ORMDL3, ORM1-like protein 3. Allergic rhinitis and 6 of the 7 studied SNPs were significantly associated in the pooled data (Table I). Because the SNPs are in high linkage disequilibrium, the associations are not independent. Cohort-specific risk estimates were consistent across all cohorts except the smallest cohort CAPPS (see Table E4 in this article's Online Repository at www.jacionline.org). Associations stratified by sex were significant only among males, although the direction of effects was similar in males and females (data not shown). Sensitivity analyses were conducted to assess whether associations with allergic rhinitis were modified by the copresence of asthma. Associations between the 17q21 variants and allergic rhinitis with concomitant asthma appeared more pronounced (although not significantly) than when allergic rhinitis or asthma was modeled independently (Table I). Risk estimates for allergic rhinitis without concomitant asthma appeared smaller but were also significant/borderline significant for 5 SNPs in high linkage disequilibrium. Associations with allergic rhinitis were similar after adjustment for asthma during early life (3-8 years). However, stratified analyses indicated that associations between the SNPs and allergic rhinitis were significant only among those with a history of asthma (Fig 1). In our study, risk estimates for the 17q21 locus with concomitant allergic rhinitis and asthma appeared largest, which is intuitive given that significant associations were observed for asthma and allergic rhinitis modeled independently. These results have been observed by others2Ferreira M.A.R. Matheson M.C. Tang C.S. Granell R. Ang W. Hui J. et al.Genome-wide association analysis identifies 11 risk variants associated with the asthma with hay fever phenotype.J Allergy Clin Immunol. 2014; 133: 1564-1571Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar and may suggest that the 17q21 locus is involved in the development of multiple clinical manifestations or a more severe type of disease, possibly via a causal pathway involving asthma. An asthma-dependent mechanism is supported by the stratified analyses in which associations between the SNPs and allergic rhinitis were most pronounced for those with a history of asthma. Furthermore, the 17q21 locus appears to be a stronger risk factor for asthma than for allergic rhinitis.2Ferreira M.A.R. Matheson M.C. Tang C.S. Granell R. Ang W. Hui J. et al.Genome-wide association analysis identifies 11 risk variants associated with the asthma with hay fever phenotype.J Allergy Clin Immunol. 2014; 133: 1564-1571Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar Nevertheless, because weaker borderline significant longitudinal associations for allergic rhinitis without concomitant asthma were also observed, we are unable to exclude the possibility that any effect of 17q21 variation on allergic rhinitis may also be mediated through an asthma-independent pathway, as has been suggested by others.3Tomita K. Sakashita M. Hirota T. Tanaka S. Masuyama K. Yamada T. et al.Variants in the 17q21 asthma susceptibility locus are associated with allergic rhinitis in the Japanese population.Allergy. 2013; 68: 92-100Crossref PubMed Scopus (32) Google Scholar However, the null associations with allergic rhinitis for those without any history of asthma observed in this study do not support this hypothesis (Fig 1). Allele-specific differences in gene expression at 17q21 are being investigated.7Verlaan D.J. Berlivet S. Hunninghake G.M. Madore A.-M. Larivière M. Moussette S. et al.Allele-specific chromatin remodeling in the ZPBP2/GSDMB/ORMDL3 locus associated with the risk of asthma and autoimmune disease.Am J Hum Genet. 2009; 85: 377-393Abstract Full Text Full Text PDF PubMed Scopus (216) Google Scholar, 8Hao K. Bossé Y. Nickle D.C. Paré P.D. Postma D.S. Laviolette M. et al.Lung eQTLs to help reveal the molecular underpinnings of asthma.PLoS Genet. 2012; 8: e1003029Crossref PubMed Scopus (216) Google Scholar However, how these genes (and other genetic, epigenetic, and gene-environment influences) may affect asthma or allergic rhinitis development remains under study. A functional role for ORM1-like protein 3 in regulating eosinophil trafficking, recruitment, and degranulation via the regulation of integrins and CD48 was recently identified.9Ha S.G. Ge X.N. Bahaie N.S. Kang B.N. Rao A. Rao S.P. et al.ORMDL3 promotes eosinophil trafficking and activation via regulation of integrins and CD48.Nat Commun. 2013; 4: 2479Crossref PubMed Scopus (82) Google Scholar Because eosinophils are involved in immune and inflammatory responses to allergens, this may represent a possible mechanism by which 17q21 variation affects allergic asthma and allergic rhinitis. Our study results are based on a substantially larger sample size and longer follow-up than any previous candidate-gene study for allergic rhinitis studying this region. It is noteworthy that the cohort-specific analyses, although consistent in trend, did not independently reach statistical significance in almost all cases. This highlights the need to carefully pool or meta-analyze homogenous data to achieve sufficient statistical power. It nevertheless remains possible that our study may have had less power to detect associations that were found to be without or with borderline significance. The SNPs studied were limited to those available for the largest subset of the cohorts, which included top hits from previous asthma GWASs.1Moffatt M.F. Kabesch M. Liang L. Dixon A.L. Strachan D. Heath S. et al.Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma.Nature. 2007; 448: 470-473Crossref PubMed Scopus (1219) Google Scholar It is reassuring that we observed significant associations for 6 of the 7 SNPs (which were in high linkage disequilibrium) and that associations were consistent across cohorts despite slightly different outcome definitions (doctor diagnoses [CAPPS, GINI/LISA, and SAGE], rhinitis symptom reports [PIAMA], or both [BAMSE]). This consistency also argues against the interpretation that the results are due to chance. Furthermore, associations between allergic rhinitis and the 17q21 locus would have remained significant after a very conservative Bonferroni correction (Table I). In summary, genetic variants at the 17q21 locus were significantly associated with allergic rhinitis, primarily with concomitant asthma, in a pooled longitudinal analysis of 5 birth cohorts. These results support the hypothesis of a shared genetic susceptibility between asthma and allergic rhinitis. We thank all the children and parents for their cooperation, and all technical and administrative support staff and medical and fieldwork teams. We acknowledge the Mutation Analysis Core Facility at Karolinska Institutet for genotyping in BAMSE, Dr I. M. Nolte and Dr O. Savenije for their contribution to the PIAMA imputation, and Denise Daley and the AllerGen Genetics team for assistance with CAPPS and SAGE data management and transfer. The cohort-specific rhinitis and asthma definitions, aeroallergens tested, and frequencies of follow-up are provided in Table E1. Parent-completed questionnaires were used to collect data on rhinitis and asthma. Aeroallergen sensitization was assessed as any specific IgE level of 0.35 kU/L or more for BASME, GINI/LISA, and PIAMA. For CAPPS and SAGE, sensitization was assessed by using skin prick testing. A positive reaction was defined as having a wheal diameter of 3 mm or more. Although not all cohorts had information on all aeroallergens, several aeroallergens were tested in each cohort (Table E1). Each cohort received ethical approval from its local authorized institutional review board. Statistical analyses were conducted in R, version 2.13.1 (www.r-project.org). The 7 SNPs examined in this study were in Hardy-Weinberg equilibrium (P > .01). For BAMSE, these SNPs were genotyped in a subset of the cohort (N = 2033 with DNA extracted from blood samples available) by Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (SEQUENOM, Inc, San Diego, Calif). For the CAPPS, GINI/LISA, PIAMA, and SAGE cohorts, information on these 7 SNPs was extracted from imputed genome-wide data. All imputed allele dosage data were recoded to hard coding (0, 1, and 2) before further analysis. As a sensitivity analysis, the longitudinal model results for the GINI/LISA cohort were compared using hard and allele dosage coding for the SNPs and the results were very consistent. For GINI/LISA, DNA samples from 1027 children and 69 children were analyzed using the Affymetrix Human SNP Arrays 5.0 and 6.0, respectively. Genotypes were called using the BRLMM-P (Affymetrix 5.0) or BIRDSEED V2 algorithm (Affymetrix 6.0). The genotype data were subjected to quality control filters on the variant and individual levels. Variants were excluded on the basis of the following criteria: a call rate below 95%, a minor allele frequency below 1%, or a Hardy-Weinberg equilibrium P value below 1 × 10−5 (∼18% excluded). Only those individuals who had a call rate above 95%, a heterozygosity value within ±4 SDs of the mean, and who passed a sex check and the similarity quality control step based on multidimensional scaling plots were retained for subsequent imputation. The genotype data were prephased using SHAPEIT v2E1Delaneau O. J. A for of 2012; Scopus Google Scholar, O. J. for and genetic 2013; PubMed Scopus Google Scholar and imputed using B.N. J. A and genotype for the of genome-wide association Genet. 2009; PubMed Scopus Google Scholar against from the variant individuals from all 2012; limited to variants with more than 1 minor allele All 7 SNPs used for the analysis were imputed with quality > For PIAMA, DNA was extracted from blood or DNA of children was genotyped on the and DNA of children was genotyped with the both at the Facility of the DNA of children was genotyped at the as of the M. I. Strachan D. Heath S. et al.A association study of J PubMed Scopus Google Scholar SNPs were by to and of for each or SNPs or SNPs that large differences in allele frequencies were quality a of individuals remained and was per using against the data of the of the SNPs of high quality were 1 data using and used for further analysis. For CAPPS and SAGE, data on only SNPs (rs2305480, rs7216389, and were In samples from parents and children were genotyped using the SNP Genotypes were subjected to the following quality control filters for variants and individual Variants were excluded on the basis of the following criteria: a call rate below 95%, a minor allele frequency below 1%, of Hardy-Weinberg equilibrium P value below 1 × and SNPs with more than 2 samples were retained had a call rate of more than and a heterozygosity value within SDs of the mean, and samples had to and scaling was used to check for and to = 2) and was using Y. J. using and genotype data to and PubMed Scopus Google Scholar and SNPs needed to be imputed with an to be used in the of of of children with health outcome information for at least 1 time and genotype at which both rhinitis and aeroallergen sensitization data were and allergic rhinitis be Allergy, Milieu, Stockholm, Epidemiological M. G. The of a longitudinal birth cohort Allergy Immunol. PubMed Scopus Google birth cohort with or and after to or or a medical of allergic rhinitis previous least of in the or at least 1 of with or of in the Asthma Primary Prevention M. J. A. W. A. A study on the of a in the of asthma in PubMed Scopus Google study with asthma of allergic rhinitis assessed at of asthma assessed at from only the from these 2 combined German cohorts were included in the analysis, to as GINI/LISA the because genome-wide genetic data were available only for that study and were Infant study on the influence of Nutritional Intervention plus environmental and genetic influences on allergy A. S. A. I. A. et in to results from a birth cohort 2007; Full Text Full Text PDF PubMed Scopus Google birth for of allergic rhinitis or hay fever during the of asthma during the from only the from these 2 combined German cohorts were included in the analysis, to as GINI/LISA the because genome-wide genetic data were available only for that study and were factors, Immune System and the development of Allergies in East and West Germany plus the influence of traffic emissions and genetics J. G. J. I. et and on and in blood of J. PubMed Scopus Google birth of allergic rhinitis or hay fever during the of asthma during the and Incidence of Asthma and Mite J. J. J. Postma D. et and of asthma and allergy birth cohort and Allergy Immunol. PubMed Scopus Google birth for during the without or of asthma during the only at and only at and of Asthma Genes and the A.L. A. P.D. M. A study to the of asthma in children 2009; PubMed Scopus Google cohort with asthma of allergic rhinitis assessed at of asthma assessed at of children with health outcome information for at least 1 time and genotype at which both rhinitis and aeroallergen sensitization data were and allergic rhinitis be from only the from these 2 combined German cohorts were included in the analysis, to as GINI/LISA the because genome-wide genetic data were available only for that study and were only at and Open table in a new tab Table and cohort-specific of cases and controls controls were defined as those without allergic rhinitis. rhinitis with concomitant controls were defined as those without allergic rhinitis or asthma. rhinitis without concomitant controls were defined as those without allergic rhinitis or asthma. Age-specific controls were defined as those without allergic Age-specific controls were defined as those without allergic rhinitis or asthma. Open table in a new tab Table information for the 7 SNPs at the 17q21 from genotyped from imputed from imputed from imputed from imputed SNPs are in linkage disequilibrium (r2 > 0.2, D′ > Effect allele frequency; GSDMB, gasdermin B; N, number of children with health outcome information for at least 1 time and genotype ORMDL3, ORM1-like protein 3.∗ Genotypes from genotyped Genotypes from imputed data. Open table in a new tab Table associations between 7 SNPs at the 17q21 locus and allergic were adjusted for age except for CAPPS and SAGE for which only 1 time was The SNPs are in linkage disequilibrium (r2 > 0.2, D′ > to 7 to to (95% CI)NOR (95% CI)NOR (95% CI)NOR (95% CI)NOR (95% < < of children included in the model; OR, odds Models were adjusted for age except for CAPPS and SAGE for which only 1 time was The SNPs are in linkage disequilibrium (r2 > 0.2, D′ > P < .05. Open table in a new tab The SNPs are in linkage disequilibrium (r2 > 0.2, D′ > 0.9). EAF, Effect allele frequency; GSDMB, gasdermin B; N, number of children with health outcome information for at least 1 time and genotype ORMDL3, ORM1-like protein 3. N, of children included in the model; OR, odds

Перевод пока недоступен