The
new engl a nd jour nal
o f
medicine
n engl j med 363;13 nejm.org september 23, 2010
1211
original article
A Large-Scale, Consortium-Based
Genomewide Association Study of Asthma
Miriam F. Moffatt, D.Phil., Ivo G. Gut, Ph.D., Florence Demenais, M.D.,
David P. Strachan, M.D., Emmanuelle Bouzigon, M.D., Ph.D., Simon Heath, Ph.D.,
Erika von Mutius, M.D., Martin Farrall, F.R.C.Path., Mark Lathrop, Ph.D.,
and William O.C.M. Cookson, M.D., D.Phil., for the GABRIEL Consortium*
From the National Heart and Lung In-
stitute, Imperial College (M.F.M.,
W.O.C.M.C.), the Division of Community
Health Sciences, St. George’s, University
of London (D.P.S.), and Royal Brompton
and Harefield NHS Foundation Trust
(W.O.C.M.C.) — all in London; Commis-
sariat à l’Energie Atomique, Institut de
Génomique, Centre National de Géno-
typage, Evry, France (I.G.G., S.H., M.L.);
INSERM, Unité 946, Fondation Jean-
Dausset–Centre d’Etude du Polymor-
phisme Humain (CEPH) (F.D., E.B.), Fon-
dation Jean Dausset–CEPH (F.D., E.B.,
M.L.), and Université Paris Diderot Paris 7,
Institut Universitaire d’Hématologie (F.D.,
E.B.) — all in Paris; University Children’s
Hospital, Asthma and Allergy Depart-
ment, Ludwig Maximilians University,
Munich, Germany (E.M.); and Wellcome
Trust Centre for Human Genetics, Univer-
sity of Oxford, Oxford, United Kingdom
(M.F.). Address reprint requests to Dr.
Cookson at Imperial College London, Roy-
al Brompton Campus, Guy Scadding Bldg.,
Dovehouse St., London SW3 6LY, United
Kingdom, or at w.cookson@imperial.ac.uk.
Drs. Moffatt, Gut, and Demenais con-
tributed equally to this article.
*Members of the GABRIEL (A Multidisci-
plinary Study to Identify the Genetic
and Environmental Causes of Asthma
in the European Community) Consor-
tium are listed in the Supplementary
Appendix, available at NEJM.org.
N Engl J Med 2010;363:1211-21.
Copyright © 2010 Massachusetts Medical Society.
Abstr act
Background
Susceptibility to asthma is influenced by genes and environment; implicated genes
may indicate pathways for therapeutic intervention. Genetic risk factors may be use-
ful in identifying subtypes of asthma and determining whether intermediate pheno-
types, such as elevation of the total serum IgE level, are causally linked to disease.
Methods
We carried out a genomewide association study by genotyping 10,365 persons with
physician-diagnosed asthma and 16,110 unaffected persons, all of whom were matched
for ancestry. We used random-effects pooled analysis to test for association in the
overall study population and in subgroups of subjects with childhood-onset asthma
(defined as asthma developing before 16 years of age), later-onset asthma, severe
asthma, and occupational asthma.
Results
We observed associations of genomewide significance between asthma and the fol-
lowing single-nucleotide polymorphisms: rs3771166 on chromosome 2, implicat-
ing IL1RL1/IL18R1 (P =3×10
−9
); rs9273349 on chromosome 6, implicating HLA-DQ
(P = 7×10
−14
); rs1342326 on chromosome 9, flanking IL33 (P = 9×10
−10
); rs744910 on
chromosome 15 in SMAD3 (P = 4×10
−9
); and rs2284033 on chromosome 22 in IL2RB
(P = 1.1×10
−8
). Association with the ORMDL3/GSDMB locus on chromosome 17q21 was
specific to childhood-onset disease (rs2305480, P = 6×10
−23
). Only HLA-DR showed a
significant genomewide association with the total serum IgE concentration, and loci
strongly associated with IgE levels were not associated with asthma.
Conclusions
Asthma is genetically heterogeneous. A few common alleles are associated with dis-
ease risk at all ages. Implicated genes suggest a role for communication of epithe-
lial damage to the adaptive immune system and activation of airway inflammation.
Variants at the ORMDL3/GSDMB locus are associated only with childhood-onset dis-
ease. Elevation of total serum IgE levels has a minor role in the development of asthma.
(Funded by the European Commission and others.)
The New England Journal of Medicine
Downloaded from www.nejm.org at UNIVERSITY LIBRARY UTRECHT on November 11, 2010. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour nal
o f
medicine
n engl j med 363;13 nejm.org september 23, 2010
1212
A
sthma is a syndrome of unknown
origins that is characterized by abnor-
mal and inflamed mucosa of the airways,
wheezing, and shortness of breath. In some pa-
tients, irreversible airway remodeling and intrac-
table airflow limitation may develop. The disease
has a high prevalence and a chronic relapsing
course. Although some effective therapies exist for
mild asthma, severe asthma remains difficult
to treat. The societal cost of the disease is sub-
stantial.
1,2
Asthma runs strongly in families, and its heri-
tability has been estimated as 60%.
3
Genetic stud-
ies offer a structured means of understanding the
causes of asthma as well as identifying targets
that can be used to treat the syndrome.
Childhood asthma is more common in boys
than in girls and may persist throughout life. It
is often associated with atopy, and as a conse-
quence, much of the research into asthma has fo-
cused on atopic mechanisms.
4
However, the link
between atopic sensitization and asthma symp-
toms in children is absent in many populations,
5
calling into question a causal role of IgE produc-
tion in the disease. Adult-onset asthma typically
develops in middle age and is more common in
women than in men. It is less obviously associated
with allergy and may be resistant to treatment.
Occupational asthma results from workplace ex-
posure to dusts or chemicals and is the most com-
mon occupational lung disease in the European
Community.
An adequately powered genomewide associ-
ation study is currently the method of choice for
identifying genes that influence complex disease.
6
The GABRIEL (A Multidisciplinary Study to Iden-
tify the Genetic and Environmental Causes of
Asthma in the European Community) Consortium
conducted a previous, comparatively small genome-
wide association study, which showed that mul-
tiple markers on chromosome 17q21 were strongly
and reproducibly associated with childhood-onset
asthma.
7,8
We report here the results of a genome-
wide association study conducted with a sample
10 times the size of that used in our first-gener-
ation study
.
In addition to reporting on analyses
of childhood-onset asthma, we report on analyses
of later-onset and occupational asthma. Genetic
studies can determine whether disease-related
phenotypes cause the disease or result from it.
9,10
Consequently, we wanted to determine whether
there was a concordance between single-nucle-
otide polymorphisms (SNPs) associated with asth-
ma and SNPs associated with total serum IgE
levels. We also wanted to ascertain the extent to
which the genetic variants that we observed to be
associated with asthma determine the individual
risk of asthma development, and we modeled the
effects these variants have on the burden of dis-
ease in the populations we studied.
Methods
Subjects
The study consisted of 10,365 case subjects and
16,110 controls recruited from 23 studies (
Table 1
;
and the Supplementary Appendix, available with
the full text of this article at NEJM.org). Data on
case subjects and population-matched controls
were obtained from clinics and from cohort and
cross-sectional population surveys in Europe (
Ta-
ble 1
in the Supplementary Appendix). We included
some subjects from family-based studies, case
subjects and controls from our first-generation
genomewide association study,
7
and case subjects
and controls of European descent from Canadian,
Australian, and U.S. surveys. Asthma was con-
sidered to be present if it had been diagnosed by
a physician. We defined childhood-onset asthma
as the presence of the disease in a person young-
er than 16 years of age and later-onset asthma as
disease that developed at 16 years of age or older.
Some surveys contributed samples to both age-
at-onset groups. Other subgroups consisted of
subjects with asthma that developed at an un-
known age, subjects with occupational asthma,
and subjects with severe asthma (for details, see
the Supplementary Appendix). All participants or
their parents provided written informed consent
for their participation in the study, in accordance
with the rules of local ethics committees. The
study was performed in accordance with the pro-
tocol.
Genotyping
We genotyped 582,892 SNPs for each subject, pro-
ducing approximately 15 billion genotypes. See
the Supplementary Appendix for a description of
the handling of DNA and for further details on
genotyping.
Statistical Analysis
The study was well powered to detect variants
with an allele frequency of 10% and an odds ratio
The New England Journal of Medicine
Downloaded from www.nejm.org at UNIVERSITY LIBRARY UTRECHT on November 11, 2010. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
A Genomewide Association Study of Asthma
n engl j med 363;13 nejm.org september 23, 2010
1213
greater than 1.20 or those with an allele frequency
of 45% and an odds ratio greater than 1.10. See the
Supplementary Appendix for a detailed description
of the statistical analyses.
After filtering the genotypes for quality, we
controlled for population admixture using prin-
cipal-components analysis for case subjects and
controls in individual populations, removing any
outliers. The principal-components analysis for
European studies showed good discrimination of
country of origin (Fig. 1 in the Supplementary Ap-
pendix), including subjects from multicenter stud-
ies such as the European Community Respiratory
Health Survey. The mean genomic control param-
eter after principal-components analysis was 1.02
(range, 1.01 to 1.05).
Results
Associations with Asthma
We examined the full sample of persons with asth-
ma (including subjects for whom age at onset was
Table 1. Subjects and Populations in Studies of Patients with Asthma Onset in Childhood or Later or with Unknown Age at Onset.*
Study or Study Group Subjects with Asthma Controls
Country Total
Childhood
Onset Later Onset Total
For Childhood
Onset For Later Onset
number (percent male)
Known age at onset
ALSPAC United Kingdom 607 607 (57) — 609 609 (51) —
B58C United Kingdom 445 213 (56) 232 (38) 420 200 (44) 220 (50)
BAMSE Sweden 239 239 (63) — 246 246 (48) —
BUSSELTON Australia 398 188 (51) 210 (33) 809 390 (45) 419 (41)
CAPPS and SAGE Canada 88 88 (66) — 182 182 (51) —
ECRHS Multiple countries 664 279 (56) 385 (34) 1546 620 (50) 926 (50)
EGEA France 676 482 (61) 194 (49) 1252 598 (48) 654 (48)
GABRIEL Advanced Surveys Germany 841 841 (63) — 851 851 (52) —
GAIN Multiple countries 1475 1475 (56) — 1668 1668 (50) —
Karelia Allergy Study Finland 73 33 (61) 40 (0) 75 36 (61) 39 (0)
KSMU Russia 291 112 (66) 179 (39) 277 116 (47) 161 (52)
MAGICS Germany 630 630 (67) — 572 572 (50) —
MAS Germany 171 171 (57) — — — —
MRCA-UKC United Kingdom 177 177 (62) — 399 399 (29) —
PIAMA Holland 172 172 (59) — 187 187 (46) —
SAPALDIA Switzerland 608 237 (62) 371 (39) 913 356 (52) 557 (50)
SLSJ Familial Collection Canada 586 373 (50) 213 (34) 613 390 (49) 223 (43)
TOMSK Russia 259 197 (65) 62 (29) 422 91 (56) 331 (52)
UFA Russia 330 269 (67) 61 (41) 348 209 (49) 139 (78)
Total 8730 6783 1947 11,389 7720 3669
Unknown age at onset
CAPPS and SAGE Canada 179 (46) 372 (48)
Occupational asthma Multiple countries 529 (83) 698 (85)
AUGOSA United Kingdom 637 (48) 2677 (67)
SEVERE United Kingdom 290 (39) 974 (49)
Total 1635 4721
* Data are shown for subjects who were successfully genotyped and whose genotypes passed all tests of quality control. For a list of the com-
plete names of each study or study group, see the Supplementary Appendix.
The New England Journal of Medicine
Downloaded from www.nejm.org at UNIVERSITY LIBRARY UTRECHT on November 11, 2010. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour nal
o f
medicine
n engl j med 363;13 nejm.org september 23, 2010
1214
unknown and those with occupational asthma or
severe asthma) and controls (Fig. 1 and
Table 2
,
and
Table 1
in the Supplementary Appendix). The
full set of results is available at www.cng.fr/
gabriel, and the results and data have been depos-
ited in the European Genome–Phenome Archive
(EGA) at the European Bioinformatics Institute
(accession number, EGAS00000000077). We also
tested for association among subjects with child-
hood-onset asthma (6783 case subjects and 7720
controls) and those with later-onset asthma (1947
case subjects and 3669 controls) (
Table 2
and
Fig. 2, and
Tables 1 and 2
in the Supplementary
Appendix). We then examined loci with genome-
wide significance in the full sample in the sub-
groups of subjects with occupational asthma and
those with severe asthma (these samples were too
small to justify full genomewide analyses) (
Table 1
in the Supplementary Appendix). We identified loci
with genomewide significance in the overall sam-
ple on chromosomes 2, 6, 9, 15, 17, and 22 (Fig. 1
and
Table 2
, and
Table 1
in the Supplementary
Appendix).
The strongest association on chromosome 2
was with a SNP within the IL18R1 gene (rs3771166)
(P = 3.4×10
−9
; odds ratio, 0.87). IL18R1 is in close
proximity to IL1RL1, which also contained SNPs
showing significant association with asthma
(
Table 1
in the Supplementary Appendix); these
SNPS were in high linkage disequilibrium with
rs3771166 (r
2
= 0.96 for the correlation between
rs3771166 and each of the three most significant
SNPs on the IL1RL1 gene). A locus centered on
rs9273349 in the HLA-DQ region of the major his-
tocompatibility complex gene (MHC) also showed
strong evidence of association in the overall sam-
ple (P = 7×10
−14
; odds ratio, 1.18). We observed
significant association between disease and
rs1342326 (on chromosome 9), flanking the IL33
gene (P = 9×10
−10
; odds ratio, 1.20) (
Table 2
). Other
SNPs showing significant association with dis-
ease were rs744910 on chromosome 15 (P = 4×10
−9
;
odds ratio, 0.89), within the SMAD3 gene, and
rs2284033, within IL2RB on chromosome 22
(P = 1×10
−8
; odds ratio, 0.89) (
Table 2
).
The odds ratios for these loci generally sug-
gested more pronounced effects in childhood-
onset asthma than in later-onset asthma. The
HLA-DQ region was observed to have a slightly
stronger association with later-onset disease
(P = 4×10
−8
; odds ratio, 1.26) than with childhood-
onset disease (P = 2×10
−5
; odds ratio, 1.14). How-
ever, none of these differences were significant
(
Table 2
and Fig. 2, and
Table 2
in the Supple-
mentary Appendix).
The span of the SNPs associated with disease at
the ORMDL3/GSDMB locus on chromosome 17q12
was approximately 380 kb (
Table 2
, and
Table 1
in the Supplementary Appendix). As in previous
studies,
7,8
we observed for 17q12 SNPs signifi-
cant evidence of heterogeneity between associ-
ations with childhood-onset and disease and
associations with significant later-onset dis-
ease (P<0.001) (Fig. 2). Significant associations
were confined to childhood-onset disease (
Ta-
ble 1
in the Supplementary Appendix), and a
separate analysis of subjects with childhood-
onset disease (and controls) yielded a maximum
association at rs2305480 within the GSDMB
gene (P = 6×10
−23
; odds ratio, 0.76) (
Table 2
).
Among subjects with childhood-onset disease,
forward stepwise regression established an in-
dependent association with rs3894194 within
GSDMA (P = 0.0015), a member of the same gene
family as GSDMB, in addition to rs2305480. We
P Value (−Log
10
)
15
5
10
0
54321 109876 1514131211
19
18
1716
20
21
22
X
Chromosome
HLA-DQ
IL33
IL13
IL18R1
SLCA22A5
IL2RB
SMAD3
RORA
ORMDL3/GSDMB
Figure 1. Manhattan Plot of Study Results.
The results of genomewide association testing are shown for all subjects.
The horizontal line indicates the stringent genomewide significance threshold
(P≤7.2×10
−8
).
11
Gene names are provided for loci at this level of significance
and are also provided for loci at the less stringent threshold of P≤5×10
−7
used by the Wellcome Trust Case-Control Consortium.
6
The New England Journal of Medicine
Downloaded from www.nejm.org at UNIVERSITY LIBRARY UTRECHT on November 11, 2010. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
A Genomewide Association Study of Asthma
n engl j med 363;13 nejm.org september 23, 2010
1215
Table 2. Loci with Genomewide Significance for Asthma and Tests of Heterogeneity across Studies.*
Chromosome Location Allele Childhood Onset Later Onset Total Sample†
Gene† Marker Position‡ Ref Alt Alt Fq
Odds Ratio
(95% CI)
P Value,
Random
Effects
Odds Ratio
(95% CI)
P Value,
Random
Effects
Odds Ratio
(95% CI)
P Value,
Random
Effects
P Value,
Hetero-
geneity
P Value,
Fixed
Effects
Loci with genomewide significance — P random ≤7.2×10
−8
2
IL18R1
rs3771166§ 102352654 G A 0.38 0.85 (0.81–0.90) 1.1×10
−8
0.94 (0.85–1.04) 1.9×10
−1
0.87 (0.83–0.91) 3.4×10
−9
1.8×10
−1
3.5×10
−12
6
HLA-DQ
rs9273349 32733847 T C 0.58 1.14 (1.08–1.22) 1.9×10
−5
1.26 (1.16–1.37) 3.9×10
−8
1.18 (1.13–1.24) 7.0×10
−14
5.0×10
−1
7.0×10
−14
9
IL33
rs1342326 6180076 A C 0.16 1.27 (1.17–1.38) 1.6×10
−8
1.12 (0.99–1.26) 6.8×10
−2
1.20 (1.13–1.28) 9.2×10
−10
2.2×10
−1
8.7×10
−12
15
SMAD3
rs744910 65233839 G A 0.51 0.89 (0.84–0.93) 8.1×10
−6
0.94 (0.87–1.0) 1.4×10
−1
0.89 (0.86–0.92) 3.9×10
−9
8.5×10
−1
3.9×10
−9
17
GSDMB
rs2305480 35315722 G A 0.45 0.76 (0.72–0.81) 6.4×10
−23
1.03 (0.94–1.13) 4.9×10
−1
0.85 (0.81–0.90) 9.6×10
−8
9.2×10
−4
—
17
GSDMA
rs3894194 35375519 G A 0.45 1.26 (1.19–1.33) 3.0×10
−17
1.02 (0.94–1.11) 6.0×10
−1
1.17 (1.11–1.23) 4.6×10
−9
2.0×10
−2
—
22
IL2RB
rs2284033 35863980 G A 0.44 0.92 (0.87–0.97) 1.6×10
−3
0.86 (0.80–0.94) 4.2×10
−4
0.89 (0.86–0.93) 1. 2×10
−8
9.2×10
−1
1.2×10
−8
Additional loci — P random ≤5×10
−7
5
SLC22A5
rs2073643 131751187 T C 0.55 0.89 (0.84–0.93) 7.6×10
−6
0.94 (0.87–1.0) 1.5×10
−1
0.90 (0.87–0.94) 2.2×10
−7
8.3×10
−1
2.2×10
−7
5
IL13
rs1295686 132023742 T C 0.80 0.85 (0.79–0.90) 3.3×10
−7
0.94 (0.85–1.04) 2.6×10
−1
0.87 (0.83–0.92) 1.4×10
−7
3.3×10
−1
1.4×10
−8
15
RORA
rs11071559 58857280 C T 0.14 0.88 (0.81–0.95) 1.0×10
−3
0.78 (0.69–0.88) 5.7×10
−5
0.85 (0.80–0.90) 1.1×10
−7
8.4×10
−1
1.1×10
−7
* Markers are identified according to their standard rs numbers. Only single-nucleotide polymorphisms (SNPs) with significant or independent effects (according to stepwise regression
analysis) on the risk of asthma are shown. The total sample includes all case subjects and controls. A complete list of SNPs with genomewide significance is available in Table 2 in the
Supplementary Appendix. P values are shown for tests of association under a random-effects model, tests of heterogeneity across studies with the use of Cochran’s Q test, and tests of
association under a fixed-effects model when there was no significant evidence of heterogeneity. For the calculation of these odds ratios, alternate alleles (Alt) were designated as risk
alleles. Alt Fq denotes alternative allele frequency, CI confidence interval, and Ref reference allele.
† The nominal gene was taken from the Illumina SNP database.
‡ Position is based on the number of base pairs from the start of the chromosome.
§ The rs3771166 SNP was imputed in the MRC Asthma and United Kingdom Control panel (MRCA-UKC) and in the Multicenter Asthma Genetics in Childhood Study (MAGICS) panel.
The New England Journal of Medicine
Downloaded from www.nejm.org at UNIVERSITY LIBRARY UTRECHT on November 11, 2010. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
