Abstract
Background
Genome-wide association studies (GWASs) have identified various genes associated with asthma, yet, causal genes or single nucleotide polymorphisms (SNPs) remain elusive. We sought to dissect functional genes/SNPs for asthma by combining expression quantitative trait loci (eQTLs) and GWASs.
Methods
Cis-eQTL analyses of 34 asthma genes were performed in cells from human bronchial epithelial biopsy (BEC, n =107) and from bronchial alveolar lavage (BAL, n = 94).
Results
For TSLP-WDR36 region, rs3806932 (G allele protective against eosinophilic esophagitis) and rs2416257 (A allele associated with lower eosinophil counts and protective against asthma) were correlated with decreased expression of TSLP in BAL (P = 7.9x10−11 and 5.4x10−4, respectively) and BEC, but not WDR36. Surprisingly, rs1837253 (consistently associated with asthma) showed no correlation with TSLP expression levels. For ORMDL3-GSDMB region, rs8067378 (G allele protective against asthma) was correlated with decreased expression of GSDMB in BEC and BAL (P = 1.3x10−4 and 0.04) but not ORMDL3. rs992969 in the promoter region of IL33 (A allele associated with higher eosinophil counts and risk for asthma) was correlated with increased expression of IL33 in BEC (P = 1.3x10−6) but not in BAL.
Conclusions
Our study illustrates cell-type-specific regulation of the expression of asthma-related genes documenting SNPs in TSLP, GSDMB, IL33, HLA-DQB1, C11orf30, DEXI, CDHR3, and ZBTB10 affect asthma risk through cis-regulation of its gene expression. Whenever possible, disease-relevant tissues should be used for transcription analysis. SNPs in TSLP may affect asthma risk through up-regulating TSLP mRNA expression or protein secretion. Further functional studies are warranted.
Keywords: asthma susceptibility genes, bronchial alveolar lavage, bronchial epithelial cells, eQTL, GWAS
Introduction
34 genes associated with asthma susceptibility (MIM 610906) have been identified by GWAS (1). Ten of these genes in six major regions have been consistently replicated: ORM1-like 3 and gasdermin B (ORMDL3-GSDMB) region (2–4), interleukin 33 (IL33) (3–5), interleukin 1 receptor-like 1 and interleukin 18 receptor 1 (IL1RL1-IL18R1) region (3–5), RAD50 homolog and interleukin 13 (RAD50-IL13) region (3, 6), thymic stromal lymphopoietin and WD repeat domain 36 region (TSLP-WDR36) region (3–5, 7), and major histocompatibility complex class II DR/DQ region (HLA-DR/DQ) (3, 6, 7). The linkage disequilibrium (LD) structure of these regions makes the task of identifying the actual disease causing genes/SNPs quite difficult.
Most of the trait-associated SNPs identified by GWASs are not coding-change variants (1). Instead, SNPs associated with complex traits are more likely to be associated with expression quantitative trait locus (eQTL) (8). eQTL analysis is an efficient approach to identify functional SNPs regulating expression of disease-associated genes. For example, the first GWAS asthma gene, ORMDL3, was identified through the combined approaches of GWAS and eQTL in lymphoblastoid cell lines (LCLs) (2, 9, 10). The purpose of this study was to determine which SNPs, known to confer asthma susceptibility, do so through alteration in mRNA expression levels.
Most eQTLs related to asthma have been performed in LCLs (2, 11), rarely in lung tissues (12). There may be extensive overlap in cis-eQTL signals across multiple tissues particularly with regard to housekeeping genes (13, 14). However, a study indicated that the mean levels of genetic correlation for gene expression in LCLs and whole blood were near zero (14). In addition, the expression of non-housekeeping genes were more likely to be tissue-specific than housekeeping genes (14). Thus, we hypothesize that eQTLs in target tissues which physiologically change during disease progression should display differential regulation. Selection of asthma-relevant tissues, such as bronchial epithelial cells (BEC) and bronchial alveolar lavage (BAL), may be required to discover functional variants underlying asthma risk. Therefore, we performed eQTL in asthma-relevant tissues (BEC and BAL) for the first time.
Methods
Study subjects
Subjects with mild to severe asthma and healthy controls recruited at four NHLBI funded Severe Asthma Research Program (SARP) centers were carefully characterized, including baseline spirometry with a medication withhold before testing (15). All studies were approved by the appropriate Institutional Review Boards at the participating sites including appropriate informed consent.
Primary human bronchial epithelial cells and bronchoalveolar lavage cells were obtained as part of the SARP projects from a range of asthmatics and healthy controls by bronchoscopy with endobronchial epithelial brushings and lavage. Sample preparation and array procedures were performed as previously described by us (16–18). Briefly, total RNA was extracted from BEC and BAL suspended in Qiazol solution using the QIACube system (QIAGEN Inc, Valencia, CA, USA). RNA quality was determined using the Agilent Bioanalyzer 2100 (Agilent Technologies Inc, Santa Clara, CA, USA). Complementary RNA (cRNA) labeled with the Cy5 fluorescent dye was hybridized to 4X44K v2 Whole Human Genome Microarrays. The microarrays were scanned using the Agilent Microarray Scanner and the data was extracted using the Agilent Feature Extraction software v9.5.
Genomic DNA was isolated from whole blood using DNA purification kits (QIAGEN Inc, Valencia, CA, USA). SNP genotyping was performed using the Illumina HumanHap1M BeadChip or the Illumina HumanOmniExpress700k BeadChip. Genotyping for all studies was performed using BeadStudio or GenomeStudio (Illumina, Inc., San Diego, CA, USA) (19, 20).
Statistical Analyses
The following 34 candidate genes in 23 regions were selected for analysis based on SNPs achieving genome-wide significance (P < 5x10−8) or P-value < 10−6 with replication which were reported by NIH GWAS database (http://www.genome.gov/gwastudies/) (1) and the published literatures: ORMDL3-GSDMB, IL33, IL1RL1-IL18R1, RAD50-IL13, TSLP-WDR36, HLA-DQB1, HLA-DPA1, TNIP1, PDE4D, DENND1B, SMAD3, IL2RB, RORA, PYHIN1, NOTCH4-AGER-C6orf10-PBX2, USP38-GAB1, GATA3, IKZF4-CDK2, IL6R, C11orf30-LRRC32, CDHR3, ZBTB10, and CLEC16A-DEXI.
Quality control processes of genotypes were described previously (19, 20). In brief, subjects were removed if they 1) had genotyping call rates < 95%, 2) were discrepant or ambiguous for genetic sex, 3) failed the check for family relatedness (PI_HAT > 0.25), or 4) were detected as an outlier (> 6 standard deviations for the first or second principal component generated from whole-genome genotyping data). After subjects meeting these criteria were excluded, SNPs were removed if 1) call rates < 95%), 2) inconsistent with Hardy-Weinberg Equilibrium (HWE) (P < 10−5), or 3) minor allele frequency (MAF) < 0.05.
Gene expression data was normalized using a cyclic loess algorithm authored in the R programming environment using the Bioconductor suite of tools as described (18, 21). The data discussed in this manuscript have been deposited in NCBI’s Gene Expression Omnibus database (GEO), and are accessible through GEO series accession number GSE67940 (http://www.ncbi.nlm.nih.gov/geo/) (18). The expression values were cyclic LOESS normalized and log2 transformed. An inverse normalization transformation was applied to the residuals of expression data (adjusted for age, gender, asthma status, and the first and second components from the multidimensional scaling analysis) to remove outliers and normalize the data. A linear additive model was chosen to test association between expression levels and GWAS-identified candidate SNPs and also cis-SNPs (within 100kb upstream or downstream of candidate genes or regions) of 34 asthma genes using PLINK software (http://pngu.mgh.harvard.edu/purcell/plink/) (22). BEC was used as primary dataset because it is composed primarily of epithelial cells (> 90%). BAL is a mix of cell types, and thus was used as secondary dataset. Multiple tests adjustment was done by 10,000 permutations within candidate genes/regions using label-swapping and max(T) procedures incorporated in PLINK (22). P-values < 0.05 after permutations were considered significant for cis-SNPs. eQTLs in LCLs of 34 asthma genes were extracted from public available databases of the GABRIEL study (11, 23) or GENEVAR study (24) and compared with our eQTLs in BEC and BAL.
Results
SARP is a cohort enriched for subjects with severe asthma, but also well balanced with mild to moderate asthma and healthy controls (Table 1). After quality control was performed as described, data from 107 and 94 subjects with GWAS and expression data from BEC and BAL, respectively, was analyzed. The population was European Americans (> 60%), African Americans (> 25%), and Hispanics (< 15%) with over 75% asthmatics and greater than 40% severe asthmatics. BAL consists of macrophages, lymphocytes, neutrophils, and eosinophils (Table 2). The percentages of macrophages and neutrophils were lower and higher, respectively, in severe asthma than controls or mild/moderate asthma. The percentage of eosinophils was higher in asthma or severe asthma than controls (Table 2).
Table 1.
Demographics of subjects with both GWAS data and expression data in BEC or BAL cells
| Bronchial epithelial cells | Bronchial alveolar lavage | |
|---|---|---|
| N | 107 | 94 |
| Age (y) | 37.1 ± 12.6 | 33.6 ± 12.3 |
| Sex (% female) | 68 | 61 |
| Race (non-Hispanic white/African American/others) | 63/29/15 | 59/23/12 |
| Asthma (Yes/No) | 88/19 | 66/28 |
| ATS classification (Mild/Moderate/Severe) | 31/18/39 | 28/10/28 |
| FEV1 (%) | 76.8 ± 22.4 | 84.4 ± 21.6 |
| FVC (%) | 86.7 ± 18.0 | 92.0 ± 17.2 |
| FEV1/FVC | 0.72 ± 0.12 | 0.75 ± 0.12 |
| Log total IgE (geometric mean) | 1.93 ± 0.75 (85.5) | 1.71 ± 0.93 (51.1) |
Table 2.
The composition of cell types in BAL
| Controls* (n=28) | Mild/moderate* (n=37) | Severe* (n=28) | Cases* (n=65) | P value
|
|||
|---|---|---|---|---|---|---|---|
| Controls vs. Cases | Controls vs. Severe | Mild/moderate vs. Severe | |||||
| Macrophages (%) | 86.5 (8.69) | 88.2 (7.69) | 80.7 (12.1) | 84.9 (10.4) | 0.48 | 0.043 | 0.0034 |
| Lymphocytes (%) | 9.36 (6.10) | 8.56 (6.82) | 10.7 (8.95) | 9.48 (7.81) | 0.94 | 0.51 | 0.28 |
| Neutrophils (%) | 3.61 (4.19) | 2.00 (1.62) | 6.42 (6.61) | 3.91 (4.98) | 0.79 | 0.063 | 0.0002 |
| Eosinophils (%) | 0.40 (0.66) | 1.07 (2.06) | 2.17 (3.55) | 1.55 (2.83) | 0.038 | 0.012 | 0.12 |
Mean (standard deviation) of percentage of different cell types in BAL was reported.
To study asthma gene expression in disease-relevant tissues, BEC and BAL were obtained by bronchoscopy with endobronchial epithelial brushings and lavage. The SNPs and expression probes of 34 candidate genes reported by GWASs of asthma are listed in Table 3. One gene may be represented by one or more probes on the microarray. In this study, the probe with the most significant eQTL was used to represent the specific gene (Table 3) (2–7, 19, 25–30). The expression levels of 34 candidate genes were moderately correlated with asthma status (data will be presented in a separate manuscript). In brief, only the expression levels of IL18R1 were consistently higher in asthma than controls in BEC and BAL (p < 0.05).
Table 3.
List of SNPs and probes of 34 genes identified through GWAS of asthma
| Gene | Chr | GWAS (reported, ref.*) | Population (reported) | SNP (reported) | SNP (tested) | Probe (tested) |
|---|---|---|---|---|---|---|
| ORMDL3 | 17 | 2 | European | rs7216389 | rs7216389 | A_23_P38190 |
| IL1RL1 | 2 | 5 | European | rs1420101 | rs1420101 | A_23_P51126 |
| WDR36 | 5 | 5 | European | rs2416257 | rs2416257 | A_23_P110598 |
| PDE4D | 5 | 25 | European American | rs1588265 | rs1588265 | A_23_P124456 |
| DENND1B | 1 | 26 | European American, African American | rs2786098 | rs2786098 | A_23_P201605 |
| RAD50 | 5 | 6 | European American | rs2244012 | rs2244012 | A_24_P226198 |
| IL13 | 5 | 6 | European American | rs20541 | rs20541 | A_23_P251031 |
| HLA-DQB1 | 6 | 6 | European American | rs1063355 | rs1063355 | A_23_P8108 |
| GSDMB | 17 | 3 | European | rs2305480 | rs2305480 | A_23_P66454 |
| IL33 | 9 | 3 | European | rs1342326 | rs992969 (r2=0.58, CEU) | A_23_P31945 |
| IL18R1 | 2 | 3 | European | rs3771166 | rs3771166 | A_23_P39735 |
| SMAD3 | 15 | 3 | European | rs744910 | rs744910 | A_23_P48936 |
| IL2RB | 22 | 3 | European | rs2284033 | rs2284033 | A_24_P203000 |
| RORA | 15 | 3 | European | rs11071559 | rs9920560 (r2=0.94, CEU) | A_23_P26124 |
| HLA-DPA1 | 6 | 27 | Japanese | rs987870 | rs987870 | A_23_P30913 |
| TSLP | 5 | 4 | European American, African American, Hispanic | rs1837253 | rs1837253 | A_23_P121987 |
| PYHIN1 | 2 | 4 | African American | rs1101999 | rs856090 (r2=0.64, YRI) | A_23_P365834 |
| NOTCH4 | 6 | 7 | Japanese | rs404860 | rs404860 | A_23_P365614 |
| AGER | 6 | 7 | Japanese | rs204993 | rs204993 | A_23_P93360 |
| C6orf10 | 6 | 7 | Japanese | rs3129943 | rs3129943 | A_23_P136734 |
| PBX2 | 6 | 7 | Japanese | rs204993 | rs204993 | A_23_P214658 |
| IKZF4 | 12 | 7 | Japanese | rs1701704 | rs1701704 | A_23_P358904 |
| CDK2 | 12 | 7 | Japanese | rs2069408 | rs2069408 | A_23_P98898 |
| USP38 | 4 | 7 | Japanese | rs7686660 | rs7686660 | A_23_P44734 |
| GAB1 | 4 | 7 | Japanese | rs3805236 | rs3805236 | A_23_P18505 |
| GATA3 | 10 | 7 | Japanese | rs10508372 | rs10508372 | A_23_P75056 |
| IL6R | 1 | 28 | European | rs4129267 | rs4129267 | A_24_P379413 |
| LRRC32 | 11 | 28 | European | rs7130588 | rs7130588 | A_24_P389916 |
| C11orf30 | 11 | 28 | European | rs7130588 | rs7130588 | A_23_P380834 |
| TNIP1 | 5 | 19 | European American | rs10036748 | rs10036748 | A_23_P19036 |
| CDHR3 | 7 | 29 | European | rs6967330 | rs17152490 (r2=0.58, CEU) | A_32_P204239 |
| ZBTB10 | 8 | 30 | European | rs7009110 | rs1543857 (r2=0.97, CEU) | A_24_P64071 |
| CLEC16A | 16 | 30 | European | rs62026376 | rs12919083 (r2=0.34, CEU) | A_32_P194246 |
| DEXI | 16 | 30 | European | rs62026376 | rs12919083 (r2=0.34, CEU) | A_24_P144377 |
The first GWAS studies reported the novel genes/SNPs were cited and GWAS studies were in chronological order.
eQTL analyses of six most consistently replicated regions in asthma
10 genes in six regions were consistently replicated by GWASs at least twice, and thus, these genes were prioritized in this study (Table 4). rs8067378 in ORMDL3-GSDMB region was significantly correlated with the expression levels of GSDMB in BEC (P = 1.3x10−4) and was replicated in BAL (P = 0.04). No SNPs were correlated with the expression levels of ORMDL3 in BEC or BAL. Multiple SNPs (including rs3806932, rs3806933, and rs2289276) in TSLP-WDR36 region were significantly correlated with the expression levels of TSLP but not WDR36 in BEC and BAL. Our results indicated that GSDMB and TSLP were more likely to be functional genes in ORMDL3-GSDMB and TSLP-WDR36 regions in BEC and BAL, respectively. SNPs in the promoter region of IL33 (including rs992969 and rs3939286) were significantly correlated with the expression levels of IL33 only in BEC. rs1063355, located at 3′ UTR of HLA-DQB1, was significantly correlated with the expression levels of HLA-DQB1 in BEC. rs12999517, located in the intron of IL1RL1, was significantly correlated with the expression levels of IL1RL1 in BAL. No SNPs were correlated with the expression levels of IL18R1, IL13, or RAD50.
Table 4.
eQTL of six most consistently replicated regions for asthma susceptibility
| SNP | Gene | Chr | Location | Distance to Gene | Minor Allele | MAF | GSDMB | ORMDL3 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||
| BEC | BAL | BEC | BAL | |||||||||||
|
| ||||||||||||||
| Beta* | P-value (adjusted) | Beta* | P-value (adjusted) | Beta* | P-value | Beta* | P-value | |||||||
| rs8067378 | GSDMB | 17 | 3′ | −9500 | G | 0.50 | −0.52 | 1.3x10−4 (1.9x10−3) | −0.35 | 0.04 | −0.05 | 0.74 | 0.11 | 0.52 |
| rs2305480 | GSDMB | 17 | coding | 96_42 | T | 0.35 | −0.45 | 4.4x10−4 (4.7x10−3) | −0.23 | 0.12 | 0.15 | 0.27 | −0.05 | 0.75 |
| rs7216389 | GSDMB | 17 | intron | −1199 | C | 0.39 | −0.43 | 1.1x10−3 (0.01) | −0.24 | 0.10 | 0.07 | 0.62 | 0.00 | 0.99 |
|
| ||||||||||||||
| TSLP | WDR36 | |||||||||||||
|
| ||||||||||||||
| rs1837253 | TSLP | 5 | 5′ | −5518 | T | 0.26 | −0.24 | 0.15 | −0.03 | 0.86 | −0.07 | 0.69 | 0.05 | 0.80 |
| rs3806932 | TSLP | 5 | 5′ | −1715 | G | 0.50 | −0.39 | 2.3x10−3 (0.05) | −0.85 | 7.9x10−11 (1.0x10−4) | −0.07 | 0.61 | −0.04 | 0.76 |
| rs3806933 | TSLP | 5 | 5′ | −648 | T | 0.36 | −0.43 | 7.8x10−4 (0.02) | −0.82 | 1.5x10−9 (1.0x10−4) | −0.06 | 0.66 | −0.01 | 0.94 |
| rs2289276 | TSLP | 5 | 5UTR | 117_81 | T | 0.24 | −0.51 | 5.2x10−4 (0.01) | −0.66 | 2.7x10−5 (8.0x10−4) | −0.21 | 0.15 | −0.09 | 0.57 |
| rs2416257 | WDR36 | 5 | intron | −798 | A | 0.12 | −0.17 | 0.44 | −0.81 | 5.5x10−4 (0.02) | 0.29 | 0.20 | 0.20 | 0.41 |
|
| ||||||||||||||
| IL13 | RAD50 | |||||||||||||
|
| ||||||||||||||
| rs2244012 | RAD50 | 5 | intron | −6166 | C | 0.38 | −0.21 | 0.11 | 0.19 | 0.18 | 0.10 | 0.47 | 0.05 | 0.71 |
| rs20541 | IL13 | 5 | coding | 97_10 | T | 0.25 | 0.05 | 0.76 | 0.32 | 0.07 | 0.17 | 0.27 | 0.12 | 0.48 |
|
| ||||||||||||||
| IL1RL1 | IL18R1 | |||||||||||||
|
| ||||||||||||||
| rs1420101 | IL1RL1 | 2 | coding | 68_101 | A | 0.38 | −0.16 | 0.30 | −0.22 | 0.21 | 0.02 | 0.88 | −0.10 | 0.57 |
| rs12999517 | IL1RL1 | 2 | intron | −236 | C | 0.14 | −0.02 | 0.90 | −0.68 | 4.4x10−4 (9.5x10−3) | 0.33 | 0.08 | −0.12 | 0.56 |
| rs3771166 | IL18R1 | 2 | intron | −1694 | T | 0.41 | 0.15 | 0.27 | −0.14 | 0.31 | −0.16 | 0.23 | −0.06 | 0.68 |
|
| ||||||||||||||
| IL33 | ||||||||||||||
|
|
||||||||||||||
| rs992969 | IL33 | 9 | 5′ | −31981 | A | 0.27 | 0.69 | 1.3x10−6 (1.0x10−4) | 0.11 | 0.53 | ||||
| rs3939286 | IL33 | 9 | 5′ | −31579 | A | 0.33 | 0.50 | 2.0x10−4 (5.9x10−3) | 0.21 | 0.18 | ||||
|
|
||||||||||||||
| HLA-DQB1 | ||||||||||||||
|
|
||||||||||||||
| rs1063355 | HLA-DQB1 | 6 | 3UTR | 50_298 | A | 0.46 | 0.63 | 8.0x10−7 (1.0x10−4) | 0.05 | 0.71 | ||||
|
|
||||||||||||||
Beta is the correlation coefficient of SNP vs. gene expression value on the basis of minor allele from linear additive model.
eQTL analyses of 24 other asthma genes
24 genes in 17 regions were identified by GWAS just once (Table 3 and Table 5). Six SNPs (rs2513525, etc.) in C11orf30-LRRC32 region were significantly correlated with the expression levels of C11orf30 in BAL but not in BEC (Table 5). rs2513525 was in moderate LD (r2=0.35) with rs7130588 which was identified through GWAS of asthma (28). No SNPs were correlated with the expression levels of LRRC32 in BEC or BAL. Our results indicated that C11orf30 was more likely to be functional gene in C11orf30-LRRC32 region. Similarly, rs12919083, in the intron of CLEC16A, was correlated with the expression levels of DEXI in BAL, but not CLEC16A (Table 5). rs12919083 was in moderate LD (r2=0.34) with rs62026376 which was identified through GWAS of asthma with hay fever (30). Our results indicated that DEXI was more likely to be the functional gene in the CLEC16A-DEXI region, as previously reported (30). eQTL SNPs were identified for CDHR3 (rs17152490) and ZBTB10 (rs1543857) in BEC, and these eQTL SNPs were in LD with the GWAS SNPs (rs6967330 in CDHR3 (29) and rs7009110 in ZBTB10 (30), respectively). eQTL SNPs were also identified for TNIP1 (rs871269), NOTCH4 (rs2071279), and SMAD3 (rs4776890); however, these eQTL SNPs were not in LD with the GWAS SNPs (rs10036748, rs404860, and rs744910 for TNIP1 (19), NOTCH4 (7), and SMAD3 (3), respectively). No eQTL SNPs were identified for the other 17 genes (HLA-DPA1, PDE4D, DENND1B, IL2RB, RORA, PYHIN1, AGER-C6orf10-PBX2, USP38-GAB1, GATA3, IKZF4-CDK2, IL6R, LRRC32, and CLEC16A).
Table 5.
eQTL of 7 of 24 other asthma genes (with significant eQTL)
| SNP | Gene | Location | Distance to Gene | Chr | Position | Major Allele | Minor Allele | MAF | TNIP1 | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||
| BEC | BAL | |||||||||||
|
| ||||||||||||
| Beta* | P-value (adjusted) | Beta* | P-value (adjusted) | |||||||||
| rs10036748 | TNIP1 | intron | −2295 | 5 | 150458146 | C | T | 0.43 | 0.02 | 0.87 | 0.04 | 0.74 |
| rs871269 | TNIP1 | intron | −568 | 5 | 150432388 | C | T | 0.38 | 0.10 | 0.49 | −0.47 | 1.5x10−3 (0.03) |
|
| ||||||||||||
| NOTCH4 | ||||||||||||
|
| ||||||||||||
| rs404860 | NOTCH4 | intron | −377 | 6 | 32184345 | T | C | 0.2103 | 0.21 | 0.22 | 0.11 | 0.5518 |
| rs2071279 | NOTCH4 | intron | −25 | 6 | 32164874 | G | T | 0.28 | −0.13 | 0.37 | −0.54 | 3.3x10−4 (0.05) |
|
| ||||||||||||
| C11orf30 | ||||||||||||
|
| ||||||||||||
| rs4300410 | C11orf30 | intron | −150 | 11 | 76163151 | C | T | 0.43 | −0.08 | 0.52 | −0.48 | 2.6x10−4 (0.01) |
| rs10793169 | C11orf30 | intron | −346 | 11 | 76164012 | G | A | 0.43 | −0.08 | 0.52 | −0.48 | 2.6x10−4 (0.01) |
| rs4245443 | C11orf30 | intron | −40 | 11 | 76183924 | G | A | 0.42 | −0.09 | 0.51 | −0.46 | 6.1x10−4 (0.02) |
| rs2508740 | C11orf30 | intron | −4 | 11 | 76227182 | A | G | 0.39 | −0.04 | 0.75 | −0.45 | 8.4x10−4 (0.03) |
| rs2513513 | C11orf30 | 3UTR | 342_1056 | 11 | 76261533 | G | A | 0.42 | −0.08 | 0.52 | −0.44 | 1.0x10−3 (0.04) |
| rs2513525 | C11orf30 | 3′ | −4119 | 11 | 76266708 | C | A | 0.42 | −0.06 | 0.62 | −0.43 | 9.1x10−4 (0.03) |
| rs7130588 | C11orf30 | 3′ | −8094 | 11 | 76270683 | A | G | 0.32 | −0.23 | 0.11 | −0.26 | 0.09 |
|
| ||||||||||||
| SMAD3 | ||||||||||||
|
| ||||||||||||
| rs4776890 | SMAD3 | intron | −34347 | 15 | 67393045 | T | G | 0.36 | 0.48 | 1.0x10−4 (5.3x10−3) | −0.06 | 0.68 |
| rs744910 | SMAD3 | intron | −10448 | 15 | 67446785 | G | A | 0.40 | 0.09 | 0.57 | 0.11 | 0.48 |
|
| ||||||||||||
| CDHR3 | ||||||||||||
|
| ||||||||||||
| rs17152490 | CDHR3 | 3′ | −3361 | 7 | 105677579 | G | A | 0.14 | 0.52 | 0.01 | −0.08 | 0.69 |
|
| ||||||||||||
| ZBTB10 | ||||||||||||
|
| ||||||||||||
| rs1543857 | ZBTB10 | 5′ | −103224 | 8 | 81295224 | A | G | 0.48 | −0.32 | 0.02 | −0.12 | 0.41 |
|
| ||||||||||||
| DEXI | ||||||||||||
|
| ||||||||||||
| rs12919083 | CLEC16A | intron | −25542 | 16 | 11188930 | A | C | 0.37 | 0.18 | 0.28 | 0.63 | 1.0x10−4 (3.9x10−3) |
Beta is the correlation coefficient of SNP vs. gene expression value on the basis of minor allele from linear additive model.
Cell-type-specific eQTL
eQTLs of 34 asthma genes in BEC, BAL, and LCLs (11, 23, 24) are summarized in Table 6. 15 GWAS SNPs for 17 genes (TSLP-WDR36, SMAD3, PDE4D, DENND1B, IL2RB, RORA, PYHIN1, NOTCH4, USP38-GAB1, GATA3, IL6R, TNIP1, C11orf30-LRRC32, and CLEC16A) were not eQTL SNPs in BEC, BAL, or LCLs. 15 GWAS SNPs for 17 genes (ORMDL3-GSDMB, IL33, IL1RL1-IL18R1, RAD50-IL13, HLA-DQB1, HLA-DPA1, AGER-C6orf10-PBX2, IKZF4-CDK2, CDHR3, ZBTB10, and DEXI) were eQTL SNPs, but not consistently observed in BEC, BAL, and LCLs. For example, rs992969 was correlated with the expression levels of IL33 in BEC, but not in BAL and LCLs. Our study shows that eQTLs of asthma-related genes in BEC, BAL, and LCLs often do not overlap.
Table 6.
eQTL comparison of BEC, BAL, and LCLs
| SNP | Chr | Location | Expressed Gene | Bronchial epithelial cells (BEC) | Bronchial alveolar lavage (BAL) | Lymphoblastoid cell lines (LCLs) Gabriel/GENEVAR (11, 23, 24) | Consistency |
|---|---|---|---|---|---|---|---|
| rs2305480 | 17 | coding | GSDMB | Yes | No | Yes/No | No |
| rs7216389 | 17 | intron | ORMDL3 | No | No | Yes/Yes | No |
| rs1837253 | 5 | 5′ | TSLP/WDR36 | No | No | No/No | Yes |
| rs2244012 | 5 | intron | RAD50/IL13 | No | No | Yes/No | No |
| rs3771166 | 2 | intron | IL18R1 | No | No | Yes/No | No |
| rs12999517 | 2 | intron | IL1RL1 | No | Yes | NA/No | No |
| rs992969 | 9 | 5′ | IL33 | Yes | No | No (rs2381416: r2=1 with rs992969)/No | No |
| rs1063355 | 6 | 3UTR | HLA-DQB1 | Yes | No | No (rs9273349: r2=0.46 with rs1063355)/NA | No |
| rs10036748 | 5 | intron | TNIP1 | No | No | NA/No | Yes |
| rs7130588 | 11 | 3′ | C11orf30/LRRC32 | No | No | No/No | Yes |
| rs404860 | 6 | intron | NOTCH4 | No | No | No/No | Yes |
| rs744910 | 15 | intron | SMAD3 | No | No | No/No | Yes |
| rs1588265 | 5 | 5′ | PDE4D | No | No | No/No | Yes |
| rs9920560 | 15 | intron | RORA | No | No | No (rs11071559: r2=0.94 with rs9920560)/No | Yes |
| rs2284033 | 22 | intron | IL2RB | No | No | No/No | Yes |
| rs856090 | 1 | intron | PYHIN1 | No | No | No (rs1101999: r2=0.64 with rs856090)/No | Yes |
| rs4129267 | 1 | intron | IL6R | No | No | No/No | Yes |
| rs2786098 | 1 | intron | DENND1B | No | No | No/No | Yes |
| rs1701704 | 12 | 5′ | IKZF4 | No | No | Yes/Yes | No |
| rs2069408 | 12 | intron | CDK2 | No | No | Yes/No | No |
| rs7686660 | 4 | 5′ | USP38 | No | No | No/No | Yes |
| rs3805236 | 4 | intron | GAB1 | No | No | No/No | Yes |
| rs10508372 | 10 | 5′ | GATA3 | No | No | No/No | Yes |
| rs204993 | 6 | intron | PBX2/AGER | No | No | Yes/NA | No |
| rs3129943 | 6 | intron | C6orf10 | No | No | Yes/NA | No |
| rs987870 | 6 | 5′ | HLA-DPA1/DPB1 | No | No | Yes/NA | No |
| rs17152490 | 7 | 3′ | CDHR3 | Yes | No | No/No | No |
| rs1543857 | 8 | 5′ | ZBTB10 | Yes | No | No/No | No |
| rs12919083 | 16 | intron | CLEC16A-DEXI | No | Yes | Yes/No | No |
Discussion
In this study, we performed eQTL analysis in asthma-relevant tissues (BEC and BAL) for the first time, and compared our eQTL results with published eQTL in LCLs. eQTLs of asthma-related genes in BEC, BAL, and LCLs often differed (Table 6), indicating that the expression of asthma-associated genes are cell-type-specific. Tissue-specific transcriptional regulation is not uncommon. Previous studies have shown tissue-specific or cell-type specific eQTLs between whole blood and LCLs (14), among blood, liver, subcutaneous tissue, visceral adipose tissue, and skeletal muscle (31), and among primary fibroblasts, T cells and LCLs (32). In addition, SNPs associated with complex diseases more often affect gene expression in a tissue-dependent manner (31).
The ORMDL3-GSDMB region (including rs7216398 and rs2305480 (2, 3)) is the most reproduced loci associated with asthma susceptibility. Identification of functional genes/SNPs in ORMDL3-GSDMB region is difficult due to very strong and long LD in this region and co-expression of ORMDL3 and GSDMB in LCLs. An allele-specific chromatin remodeling study indicated that rs12936231 and rs8067378 (in complete LD) might be functional SNPs (9). In our study, rs8067378 was identified as the most significant eQTL SNP for GSDMB in BEC (P = 1.3x10−4) and BAL (P = 0.04). However there was no correlation with ORMDL3 in either BEC or BAL (Table 4). The G allele of rs8067378 (protective against asthma (2–4)) was correlated with decreased expression of GSDMB, implying that A allele is the risk allele for asthma by up-regulating the expression of GSDMB. A recent study indicated that ORMDL3 or GSDMB might affect childhood asthma through interaction with human rhinovirus wheezing illnesses (10). BEC provides the first line of defense against viral/bacterial infection. SNPs in this region were correlated with the expression of GSDMB in BEC but not ORMDL3, indicating that GSDMB is more likely to be the functional gene in this process.
IL33 may sense the damage of epithelial cells, induce the expression of Th2-type cytokines, and lead to eosinophil infiltration into the airway. SNPs in IL33 have been reported to be associated with eosinophil counts in blood (rs3939286) and asthma (rs992969) (3, 5). All asthma-associated SNPs in IL33 region are located in the 5′ or first intron, making them good candidates for eQTL SNPs; however, no published data support this speculation (33). Our study showed for the first time that SNPs in the promoter region of IL33 (including rs3939286 and rs992969) were correlated with IL33 expression in BEC but not in BAL, confirming its ‘alarmin’ role in BEC (Table 4). We further tested the correlation between rs3939286 and blood eosinophil percentage using 107 SARP subjects with BEC data. The A allele of rs3939286 was associated with increased eosinophil percentage in blood with p value of 0.0077 (GG: 2.88%; AG: 3.92%; AA: 4.94%). The A allele of rs3939286 may be associated with higher eosinophil counts and affect asthma risk by up-regulating IL33 expression. Published eQTL analyses showed that rs996929 was not correlated with IL33 expression in LCLs (Table 5) (23, 24), emphasizing the importance to perform eQTL in disease-relevant tissues. rs1420101 and rs3771166 in IL1RL1-IL18 region are associated with eosinophil counts in blood and asthma (3, 5). In this study, a different SNP, rs12999517 was correlated with IL1RL1 (the receptor of IL33) expression in BAL (Table 4), which was not in LD with asthma or eosinophil associated SNPs.
SNPs in TSLP-WDR36 region are associated with eosinophil counts in blood (rs2416257) (5), eosinophilic esophagitis (rs3806932) (34), and asthma (rs1837253) (4, 7). We found multiple SNPs correlated with the expression levels of TSLP but not WDR36 in both BEC and BAL (Table 4). Decreased expression of TSLP in BAL was correlated with the G allele of rs3806932 (protective against eosinophilic esophagitis) and the A allele of rs2416257 (associated with lower eosinophil counts and protective against asthma) at P-values of 7.9x10−11 and 5.4x10−4, respectively. rs1837253, which is most consistently associated with asthma (4, 5, 7), and correlated with the secretion levels of TSLP protein from human nasal epithelial cells (35), was not correlated with TSLP mRNA expression levels, nor was it in LD with any other SNPs in this region. Our findings and previous evidence suggest that eosinophilic or atopic asthma may be associated with up-regulation of TSLP mRNA expression while asthma may also be affected via mechanisms other than TSLP mRNA expression, such as the regulation of TSLP protein secretion (35).
eQTL is a good approach for detecting SNPs regulating gene expression; however, other mechanisms that do not require changes in gene expression, such as protein structure changes and post-translational modification, will also be important for the manifestation and progression of disease. Thus, the aim of this study is to identify functional SNPs through cis-eQTL, not to exclude the potential candidate asthma genes simply based on the negative findings of eQTLs. The purpose of this study was to determine which SNPs associated with asthma susceptibility were also correlated with changes in mRNA expression in tissues which are important in asthma physiology. Since a relatively small sample size was available, exploration of trans-eQTLs was not performed and provided the rationale for analyzing only cis-eQTLs for known asthma genes. The expression data from different ethnic groups with or without asthma were included in this study to increase the sample size and power (Table 1). We have adjusted gene expression levels with asthma status and the first and second principal components generated from GWAS data to reduce the influence of ethnicity and disease status. The GWAS-driven focus of this work refines and identifies the subset of genes/SNPs which should be analyzed by reporter gene and quantitative PCR in the future.
BEC, the first line of defense against viral/bacterial infection, recognizes allergens and initiates airway inflammation. For example, we observed eQTL SNPs of IL33 (alarmin to induce Th2 pathway) and HLA-DQB1 (allergen recognition) in BEC but not in BAL. BAL consists of macrophages, lymphocytes, neutrophils, and eosinophils, and reflects current inflammatory process in the airway (Table 2). Since the composition of BAL is mixed cell types and may vary based on disease status, the interpretation of findings from BAL is not as straight forward. We identified eQTL SNPs of TSLP in both BEC and BAL and observed a much stronger effect size in BAL, suggesting that TSLP is involved in the early induction of Th2 pathway in BEC and highly induced in inflammatory cells in BAL. Unfortunately, we can not find additional expression datasets from BEC and BAL with available genotyping information to replicate our findings. Future replication studies are necessary to confirm our findings. Furthermore, other tissues such as bronchial smooth muscle are also asthma-relevant and may have different expression patterns as what we have observed in BEC and BAL.
Some asthma genes show cell-type-specific regulation of expression, and thus it is essential to study gene expression in disease-relevant tissues whenever possible. SNPs in IL33, GSDMB, TSLP, HLA-DQB1, C11orf30, DEXI, CDHR3, and ZBTB10 identified through GWASs affect asthma risk through cis-regulation of gene expression. For TSLP, two distinct mechanisms may affect asthma risk: one involving regulation of TSLP mRNA expression while the other may regulate TSLP protein secretion. The identification of functional genes/SNPs may help to gain understanding of molecular mechanisms underlying disease and target new therapeutic approaches, and thus, facilitate ‘personalized medicine’ for complex diseases. Such findings stimulate further functional studies of asthma genes in post-GWAS era.
Acknowledgments
Funding: Genetic studies for SARP were funded by NIH HL87665 and Go Grant RC2HL101487.
Footnotes
Author contribution: Drs Li, Ampleford, Meyers and Bleecker and Ms Li designed the genetic study and participated in the interpretation of results. Dr. Hawkins was responsible for isolating DNA, establishing the DNA database and genotyping all samples. Dr. Milosevic isolated RNA and with Dr. Kaminski was responsible for performing and analyzing gene expression microarrays on BEC and BAL. Drs Li, Ampleford, Kaminski, Wenzel, Meyers and Bleecker and Ms Li were responsible for the analytical plan including the data base, quality control and data analysis. Drs Moore, Hastie, Busse, Erzurum, Kaminski, Wenzel, Meyers, and Bleecker were responsible for the design of the SARP clinical study, recruitment and characterization of subjects with asthma and interpretation of SARP results. All authors contributed to the writing of the manuscript.
Conflict of interest: W.W. Busse has received consultancy fees from Novartis, GlaxoSmithKline, Genentech, Pfizer and Roche; has received consultancy fees from Circassia for the Data Monitoring Board; has received consultancy fees from Boston Scientific for the Data Monitoring Board and consultancy fees from ICON for the Study Oversight Committee; has received research support from the NIH/NIAID and NIH/NHLBI; and receives royalties from Elsevier.
References
- 1.Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS, et al. Potential etiologic and functional implications of genome-wide association loci for human diseases and traits. Proc Natl Acad Sci. 2009;106:9362–9367. doi: 10.1073/pnas.0903103106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Moffatt MF, Kabesch M, Liang L, Dixon AL, Strachan D, Heath S, et al. Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma. Nature. 2007;448:470–473. doi: 10.1038/nature06014. [DOI] [PubMed] [Google Scholar]
- 3.Moffatt MF, Gut IG, Demenais F, Strachan DP, Bouzigon E, Heath S, et al. A large-scale, consortium-based genomewide association study of asthma. N Engl J Med. 2010;363:1211–1221. doi: 10.1056/NEJMoa0906312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Torgerson DG, Ampleford EJ, Chiu GY, Gauderman WJ, Gignoux CR, Graves PE, et al. Meta-analysis of genome-wide association studies of asthma in ethnically diverse North American populations. Nat Genet. 2011;43:887–892. doi: 10.1038/ng.888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gudbjartsson DF, Bjornsdottir US, Halapi E, Helgadottir A, Sulem P, Jonsdottir GM, et al. Sequence variants affecting eosinophil numbers associate with asthma and myocardial infarction. Nat Genet. 2009;41:342–347. doi: 10.1038/ng.323. [DOI] [PubMed] [Google Scholar]
- 6.Li X, Howard TD, Zheng SL, Haselkorn T, Peters SP, Meyers DA, et al. Genome-wide association study of asthma identifies RAD50-IL13 and HLA-DR/DQ regions. J Allergy Clin Immunol. 2010;125:328–335. doi: 10.1016/j.jaci.2009.11.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hirota T, Takahashi A, Kubo M, Tsunoda T, Tomita K, Doi S, et al. Genome-wide association study identifies three new susceptibility loci for adult asthma in the Japanese population. Nat Genet. 2011;43:893–896. doi: 10.1038/ng.887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nicolae DL, Gamazon E, Zhang W, Duan S, Dolan ME, Cox NJ. Trait-associated SNPs are more likely to be eQTLs: annotation to enhance discovery from GWAS. PLoS Genet. 2011;6:e1000888. doi: 10.1371/journal.pgen.1000888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Verlaan DJ, Berlivet S, Hunninghake GM, Madore AM, 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;85:377–393. doi: 10.1016/j.ajhg.2009.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Calişkan M, Bochkov YA, Kreiner-Møller E, Bønnelykke K, Stein MM, Du G, et al. Rhinovirus wheezing illness and genetic risk of childhood-onset asthma. N Engl J Med. 2013;368:1398–1407. doi: 10.1056/NEJMoa1211592. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dixon AL, Liang L, Moffatt MF, Chen W, Heath S, Wong KC, et al. A genome-wide association study of global gene expression. Nat Genet. 2007;39:1202–1207. doi: 10.1038/ng2109. [DOI] [PubMed] [Google Scholar]
- 12.Hao K, Bosse Y, Nickle DC, Paré PD, Postma DS, Laviolette M, et al. Lung eQTLs to help reveal the molecular underpinnings of asthma. PLoS Genet. 2012;8:e1003029. doi: 10.1371/journal.pgen.1003029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ding J, Gudjonsson JE, Liang L, Stuart PE, Li Y, Chen W, et al. Gene expression in skin and lympholastoid cells: refined statistical method reveals extensive overlap in cis-eQTL signals. Am J Hum Genet. 2010;87:779–789. doi: 10.1016/j.ajhg.2010.10.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Powell JE, Henders AK, McRae AF, Wright MJ, Martin NG, Dermitzakis ET, et al. Genetic control of gene expression in whole blood and lymphoblastoid cell lines is largely independent. Genome Res. 2012;22:456–466. doi: 10.1101/gr.126540.111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Moore WC, Meyers DA, Wenzel SE, Teague WG, Li H, Li X, et al. Identification of asthma phenotypes using cluster analysis in the severe asthma research program. Am J Respir Crit Care Med. 2010;181:315–323. doi: 10.1164/rccm.200906-0896OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cabrera S, Selman M, Lonzano-Bolaños A, Konishi K, Richards TJ, Kaminski N, et al. Gene expression profiles reveal molecular mechanisms involved in the progression and resolution of bleomycin-induced lung fibrosis. Am J Physiol Lung Cell Mol Physiol. 2013;304:L593–601. doi: 10.1152/ajplung.00320.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Herazo-Maya JD, Noth I, Duncan SR, Kim S, Ma SF, Tseng GC, et al. Peripheral blood mononuclear cell gene expression profiles predict poor outcome in idiopathic pulmonary fibrosis. Sci Transl Med. 2013;5:205ra136. doi: 10.1126/scitranslmed.3005964. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Voraphani N, Gladwin MT, Contreras AU, Kaminski N, Tedrow JR, Milosevic J, et al. An airway epithelial iNOS-DUOX2-thyroid peroxidase metabolome drives Th1/Th2 nitrative stress in human severe asthma. Mucosal Immunol. 2014;7:1175–1185. doi: 10.1038/mi.2014.6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Li X, Ampleford EJ, Howard TD, Moore WC, Torgerson DG, Li H, et al. Genome-wide association studies of asthma indicate opposite immunopathogenesis direction from autoimmune diseases. J Allergy Clin Immunol. 2012;130:861–868. doi: 10.1016/j.jaci.2012.04.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Li X, Hawkins GA, Ampleford EJ, Moore WC, Li H, Hastie AT, et al. Genome-wide association study identifies TH1 pathway genes associated with lung function in asthmatic patients. J Allergy Clin Immunol. 2013;132:313–320. doi: 10.1016/j.jaci.2013.01.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wu W, Dave N, Tseng GC, Richards T, Xing EP, Kaminski N. Comparison of normalization methods for CodeLink Bioarray data. BMC Bioinformatics. 2005;6:309. doi: 10.1186/1471-2105-6-309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. Plink: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81:559–575. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Li L, Kabesch M, Bouzigon E, Demenais F, Farrall M, Moffatt MF, et al. Using eQTL weights to improve power for genome-wide association studies: a genetic study of childhood asthma. Front Genet. 2013;4:103. doi: 10.3389/fgene.2013.00103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Stranger BE, Montgomery SB, Dimas AS, Parts L, Stegle O, Ingle CE, et al. Patterns of cis regulatory variation in diverse human populations. PLoS Genet. 2012;8:e1002639. doi: 10.1371/journal.pgen.1002639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Himes BE, Hunninghake GM, Baurley JW, Rafaels NM, Sleiman P, Strachan DP, et al. Genome-wide association analysis identifies PDE4D as an asthma-susceptibility gene. Am J Hum Genet. 2009;84:581–593. doi: 10.1016/j.ajhg.2009.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sleiman PM, Flory J, Imielinski M, Bradfield JP, Annaiah K, Willis-Owen SA, et al. Variants of DENND1B associated with asthma in children. N Engl J Med. 2010;362:36–44. doi: 10.1056/NEJMoa0901867. [DOI] [PubMed] [Google Scholar]
- 27.Noguchi E, Sakamoto H, Hirota T, Ochiai K, Imoto Y, Sakashita M, et al. Genome-wide association study identifies HLA-DP as a susceptibility gene for pediatric asthma in Asian populations. PLoS Genet. 2011;7:e1002170. doi: 10.1371/journal.pgen.1002170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ferreira MA, Matheson MC, Duffy DL, Marks GB, Hui J, Le Souëf P, et al. Identification of IL6R and chromosome 11q13. 5 as risk loci for asthma. Lancet. 2011;378:1006–1014. doi: 10.1016/S0140-6736(11)60874-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Bønnelykke K, Sleiman P, Nielsen K, Kreiner-Møller E, Mercader JM, Belgrave D, et al. A genome-wide association study identifies CDHR3 as a susceptibility locus for early childhood asthma with severe exacerbations. Nat Genet. 2014;46:51–55. doi: 10.1038/ng.2830. [DOI] [PubMed] [Google Scholar]
- 30.Ferreira MA, Matheson MC, Tang CS, 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;133:1564–1571. doi: 10.1016/j.jaci.2013.10.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Fu J, Wolfs MG, Deelen P, Westra HJ, Fehrmann RS, Te Meerman GJ, et al. Unraveling the regulatory mechanisms underlying tissue-dependent genetic variation of gene expression. PLoS Genet. 2012;8:e1002431. doi: 10.1371/journal.pgen.1002431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Dimas AS, Deutsch S, Stranger BE, Montgomery SB, Borel C, Attar-Cohen H, et al. Common regulatory variation impacts gene expression in a cell type-dependent manner. Science. 2009;325:1246–1250. doi: 10.1126/science.1174148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Grotenboer NS, Ketelaar ME, Koppelman GH, Nawijn MC. Decoding asthma: translating genetic variation in IL33 and IL1RL1 into disease pathophysiology. J Allergy Clin Immunol. 2013;131:856–865. doi: 10.1016/j.jaci.2012.11.028. [DOI] [PubMed] [Google Scholar]
- 34.Rothenberg ME, Spergel JM, Sherrill JD, Annaiah K, Martin LJ, Cianferoni A, et al. Common variants at 5q22 associated with pediatric eosinophilic esophagitis. Nat Genet. 2010;42:289–291. doi: 10.1038/ng.547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hui CC, Yu A, Heroux D, Akhabir L, Sandford AJ, Neighbour H, et al. Thymic stromal lymphopoietin (TSLP) secretion from human nasal epithelium is a function of TSLP genotype. Mucosal Immunol. 2014 doi: 10.1038/mi.2014.126. [DOI] [PubMed] [Google Scholar]