香京julia种子在线播放

A variety of recent studies generally confirm that the heritability of childhood-onset asthma is high, and that this tends to be higher than that of later-onset asthma (Table 1). The classical twin studies of children, where the family members being utilized are exactly the same age and birth cohort, tend to give the highest estimates. It should be noted that studies do not find an increased prevalence of asthma in twins, particularly MZ twins, especially after adjusting for covariates such as birth weight (20). Thomsen et al. (21) were able to break down that twin sample by age at onset of asthma, and found the heritability for onset after age 20 to be 60%, compared to 80% in younger twins. And data reported by Paaso et al. (5) can be reanalyzed to estimate the genetic correlations between parental allergy and adult-onset and childhood-onset offspring asthma (first two entries of Table 2). These give a consistent pattern, where the allergy early-asthma correlation is higher, but there is still a significant genetic correlation between atopy and later-onset asthma.

The heritability estimates based on GWAS (Table 1) tend to be lower, given they only represent the contributions of common measured SNPs (“array,” “chip,” or “SNP” heritability). Even so, the estimates from LD regression seem to be lower again.

The Trans-national Asthma Genetic Consortium (30) compared SNP allele frequencies in 24,000 asthma cases and 119,000 controls to detect 22 asthma associated loci. The locus of greatest effect was that on chromosome 17q21 in the region of GSDMB and ORMDL3 genes. It has been known since 2008 (31) that this locus is associated with childhood onset asthma. Despite the association with early onset disease, these SNPs are not associated with atopy, and the asthma-increasing alleles probably reduce risk of other immunological conditions such as inflammatory bowel disease and Type 1 diabetes. If we take one of the functional 17q21 SNPs rs12936231 (actually within ZPBP2), in the UK BioBank the C allele increases “all” asthma OR = 1.1 (P = 1 × 10⁻⁵⁹) and self-reported “emphysema/chronic bronchitis” OR = 1.07 (P = 1 × 10⁻⁸, but not doctor-diagnosed emphysema, P = 0.5), as well as neutrophil count (P = 7 × 10⁻¹⁴²), but less strongly with eosinophil count (P = 3 × 10⁻⁵) (32). The effect sizes of these alleles are not large, so general inferences about childhood vs. adult asthma must be based on aggregated effects of multiple loci.

Demenais et al. (30) also confirmed the association of rs9272346 and rs9273349 in the region of HLA-DQ with asthma. These SNPs were reported more strongly correlated with adult-onset than childhood-onset disease (9, 33), but still exhibiting detectable association in early-onset cases. Like the 17q21 locus, are unrelated (or only weakly associated) to atopy. In the UKBB (32), the rs9272346^*A allele increases incidence of asthma, coeliac disease, Type 1 diabetes, hypothyroidism, rheumatoid arthritis, and diminishes risk of multiple sclerosis and ulcerative colitis. It has no association with self-reported COPD. Pividori et al. (34), see below, classify it as shared by childhood and adult onset asthma in UKBB.

Ferreira et al. (35) carried out a GWAS meta-analysis (13 studies, total N = 360838) to detect 99 loci (136 peak SNPs) for atopic disease, defined as any of asthma, hayfever or atopic dermatitis. Of these, 49 were novel. Several loci were disease-specific, such as Filaggrin (FLG) for atopic dermatitis, or a SNP near IL18R1, where the effect allele was increased in asthma and hayfever, but not atopic dermatitis. Several variants were associated earlier onset of asthma (7445 asthma-only cases)—the strongest being rs921650 in GSDMB (17q21 region).

Pividori et al. (34) and Ferreira et al. (36) have recently reported on childhood and adult onset asthma within the UK Biobank (UKBB) sample. Pividori et al. were able to detect 61 independent asthma loci comparing 9433 childhood-onset cases, 21564 adult-onset cases, and 318237 controls. Of these, “23 were childhood onset specific, one was adult onset specific, and 37 were shared.” Ferreira and coworkers supplemented the UKBB data with data from 23andMe (further 32,000 childhood-onset cases and 215,000 controls) and reported 123 childhood-onset (age <20 years) asthma loci, of which 98 were reproducible in the second dataset. Five of these loci did not affect risk of adult-onset asthma. For adult-onset asthma, they found 34 replicable loci, of which 3 were significantly weaker as predictors of childhood disease. They estimated the genetic correlation between adult and childhood-onset asthma as 0.67. They further tested for loci specifically for age at diagnosis within their childhood-onset sample (defined as <20 years old) and concluded that such effects were relatively small (5% of variance, P = 0.02; but N was only 14,000, so power was low). There were 25 novel childhood-onset asthma loci detected, in or near biologically important genes such as NOD2, IL4R, IL2RA, and IRF4. Because the UKBB dataset is generally available, we can contrast these findings to those published by Zhu et al. (26) who reported a total of 38 loci for “doctor diagnosed” asthma detectable using an earlier available version of the dataset containing 7908 cases and 76768 controls, and Johansson et al. (24) who reported 52 loci defined via 41934 cases and 239773 controls (Figure 2). The “pure” adult-onset SNP reported by Pividori et al. was rs12617922 on chromosome 2 near TEX41, which decreased risk of tobacco smoking (P = 5 × 10⁻²⁴), as commented on by Ferreira et al. (36). Figure 3 shows that such a strong negative association is not seen for any of the other asthma loci from Zhu et al. (26). A SNP in the same region near TEX41, rs10193706, was previously highlighted as associated with tobacco use in the subset of UKBB used for the BiLEVE study of COPD (37). In European populations, it is in moderate linkage disequilibrium with rs12617922 (r² = 0.37), but not associated with self-reported asthma in the full UKBB sample.

An even more intriguing observation is that the peak SNP for this region in terms of “any” asthma and smoking (and suggestively with caffeine intake as well) is rs10427255, previously flagged as a locus for the photic sneeze reflex (P = 1 × 10⁻¹¹) by Eriksson et al. (38) in the 23andMe dataset (see Table 3). These SNPs (rs12617922, rs10427255) are not associated with serum eosinophil level or atopy. Of the 32 SNPs known to be associated with photic sneeze reflex, only these near TEX41 predict asthma, so there is no general relationship between these two phenotypes.

Hobbs et al. (29) examined COPD in 25000 cases and 58000 controls. There were 22 associated loci (13 novel), none overlapping with asthma loci from the GABRIEL study or the NCBI GWAS catalog. Examining the UK BioBank via the GeneAtlas web site (32), one does find that five of their top COPD SNPs are associated with self-reported asthma at P < 1 × 10⁻⁵. The strongest association of these five is for rs2070600 (a protein coding change Gly82Ser in AGER on chromosome 6, P = 4 × 10⁻²⁹), where the T (Serine) allele (European frequency 5–10%) increases asthma (OR = 1.15). The Hobbs et al. (29) result for COPD is OR = 1.21 (P = 6 × 10⁻¹⁰) for the C (Glycine) allele, so (if replicable) this would lead to a negative genetic correlation between asthma and COPD due to this locus. The Ser82 allele reduces sRAGE levels and increases FEV1 (39).

Hayden et al. (40) describe the COPDGene study of 10200 current and former adult smokers. The GWAS analysis found known childhood asthma risk SNPs were associated with asthma, and childhood asthma (7% of the sample) increased the risk of later COPD 3.4-fold, and also reduced lung function.

The components of the polygenic risk score (PRS) approach have already been presented (e.g., Figure 3), in that the PRS for a trait is merely the weighted sum of the regression coefficients for a number of significantly associated SNPs. Richardson et al. (41) have made available a web site presenting results of such cross-trait analyses (162 PRS, 551 outcomes) using the UK Biobank dataset (see Table 4). For example, while the BMI PRS predicts asthma, the asthma PRS does not predict BMI, a finding we can interpret as implying BMI is causative of asthma risk.

A table of results for COPD PRS (spirometric phenotypes including FEV1/FVC were used to define COPD) has been presented by Shrine et al. (42) in yet another preprint. This score (294 SNPs) is a significantly predictor of self-reported asthma (P = 10⁻⁴¹). Examining the individual SNPs, we can see that 20 are significant (P < 1 × 10⁻⁵) independent predictors of UKBB self-reported asthma, and three are associated with hayfever/rhinitis (in IL1RL1, SUOX and LRP1).

The most widely used approach for this is the LD score regression method, and results from this are listed in Table 2. Speed and Balding (17) suggest that magnitude of estimates of genetic correlations tend to relatively similar across different methods, which is less true of heritability estimates, as seen above. We see that “any” asthma is more strongly correlated with atopic diseases such as allergic rhinitis (r_g 0.6–0.7) than it is with COPD (r_g 0.4).

Chen et al. (43) carried out a bidirectional MR analysis of the direction of causation between obesity and childhood asthma in Taiwan. While the obesity genetic risk score was a significant predictor of asthma, the asthma genetic risk score did not predict obesity. Granell et al. (44) similarly estimated the causal relative risk for the effect of BMI on asthma at 1.55 per kg/m², slightly more strongly for non-atopic childhood asthma. In 162124 adults from European studies (45), the causal OR for BMI on asthma was 1.07 (−9 ml FEV1, 16 ml FVC). There was no causal effect on allergic sensitization or serum IgE level—despite a positive phenotypic correlation between IgE and overweight in their study population. After all the above, however, Contreras et al. (46) found in a traditional longitudinal study that asthma diagnosis by age 4 was associated with obesity at age 8, which would seem consistent with the idea that asthma causes obesity, or at the very least that genes causative of asthma also cause obesity via a different pathway (acting at a different age) than that by which they cause asthma. Does this give an insight into adult-onset asthma? Multiple longitudinal studies in adults have found that overweight precedes and predisposes to asthma. Burgess et al. (47) showed BMI at age 7 years predicted development of asthma with an onset after age 21 years. So, in this case a causal pathway from overweight to asthma would be supported by both genetic and non-genetic studies.

Finally, the relationship between vitamin D deficiency and childhood and adult asthma is still unclear, with both cross-sectional epidemiological and animal model evidence making this plausible but results from supplementation studies unsupportive (48). MR studies (49, 50) have showed no relationship between vitamin D level SNPs and asthma or atopy.

This is the most well-studied environmental exposure with respect to genetic interactions in asthma in humans. Family studies have shown regions of linkage to asthma phenotypes to differ by environmental tobacco smoke (ETS) exposure in childhood (51–54). A study in 144 US families suggested that genes in chromosomal regions 1p, 5q, and 17p might interact with ETS to confer asthma risk in the exposed groups (51). The evidence for linkage for asthma and bronchial hyperresponsiveness (BHR) for chromosome 5q in the passive smoke-exposed groups during childhood was subsequently replicated in 200 Dutch families (52). A French study conducted in 295 families reported four regions, 1q43-q44, 4q34, 5p15, and 17p11, potentially involved in genetic susceptibility to BHR interacting with ETS in early life (53); however, none of these regions had been reported by previous genomes scans on gene-ETS interaction, except for the region in the 1q43-q44. Moreover, an interaction has been described between 17q21 variants and early-life exposure to ETS in early-onset asthma (54). The interaction between ORMDL3 variants and early-life exposure to ETS in childhood-onset asthma was later replicated (94, 95).

Most candidate gene studies have focused on examining variants in genes coding for antioxidant defenses and xenobiotic-metabolizing enzymes and childhood-onset asthma susceptibility to tobacco smoking, especially the glutathione-S transferase (GST) family of antioxidant enzymes. Childhood-onset asthma risk has been shown to differ by GSTM1 genotype in relation to both maternal smoking in pregnancy (55, 56) and childhood ETS exposure (57), with effects largely restricted to children with GSTM1 null genotype. Results from other studies have suggested an interaction between GSTP1 rs1695 A (Ile105) is a risk allele for childhood wheeze illness in relation to maternal smoking in early life, with an effect most clearly seen in children who are exposed to maternal smoking (59, 60). Most of the studies looking at interactions with GSTM1 and GSTP1 genetic variants have shown a positive finding, but not always in the same direction for GSTP1 (58).

Given the importance of inflammation in asthma, genes related to inflammation responses and innate immunity have also been examined. Genetic variation in tumor necrosis factor (TNF-α) may contribute to childhood asthma and that associations may be modified by parental smoking (60, 61). Ramadas et al. (62) showed interaction of the interleukin-1 receptor antagonist (IL1RN) gene polymorphism rs2234678 and maternal smoking during pregnancy increased the risk for childhood asthma. Salam et al. (63) found that children with the transforming growth factor beta-1 (TGFB1)−509TT genotype are at increased risk of asthma when they are exposed to maternal smoking in utero; however, no interaction was found for parental tobacco smoke exposure in childhood (96). Household ETS and interleukin-13 gene (IL-13) variants may have interactive effects on childhood asthma phenotypes (64). Interactions between the human intercellular adhesion molecule 1 (ICAM1) polymorphisms and ETS have also been associated with risk of childhood-onset asthma (65). In adults, Miyake et al. revealed that the combination of ever smoking with interleukin 3 (IL-3) genetic variants was significantly positively associated with adult asthma in Japanese women (66).

Several studies have focused on genes related to epithelial function. Reijmerink et al. (67) first reported gene–environment interaction of ADAM33 genotypes, the first identified asthma gene by positional cloning, and in utero tobacco smoke exposure with respect to childhood-onset asthma risk; however, no interaction was detected with postnatal ETS exposure (97). Moreover, Wang et al. (68) reported joint effects of ETS exposure and E-cadherin CDH1 genotypes associated with the development of childhood asthma.

To date, two GWIS on tobacco smoke exposure and asthma are available. Scholtens et al. (69) conducted the first GWIS specifically aiming to identify genetic polymorphisms that interact with two well-known environmental risk factors for childhood-onset asthma: in utero and childhood tobacco smoke exposure. The authors found that genes reported previously to interact with tobacco smoke exposure with respect to asthma development (i.e., GSTP1, TNF and ADAM33) were not among the most significant hits and showed suggestive interactions between rs8094633 near erythrocyte membrane protein band 4.1 like 3 (EPB41L3) and exposure to in utero tobacco smoke, and between rs1575472 in parkin coregulated gene (PACRG) and childhood ETS. Interestingly, these two SNPS had not been identified previously in GWAS on childhood asthma. Subsequently, Vonk et al. conducted the first hypothesis-free genome-wide study to identify SNPs that interact with active tobacco smoking with respect to adult-onset asthma using data of the GABRIEL consortium, showing suggestive evidence for an interaction with two intergenic markers (rs9969775 on chromosome 9 and rs5011804 on chromosome 12) with potential regulatory functions linked to gene expression regulation in lung tissue (70).

Overall, existing literature has shown that prenatal and postnatal exposure to maternal/paternal smoking interacts with genetic variants to increase the risk of childhood-onset asthma; interestingly these genes are related to regions of linkage to asthma phenotypes and coding for antioxidant defenses and xenobiotic-metabolizing enzymes, inflammation responses and innate immunity and epithelial function. However, GWIS approaches have failed to replicate those findings revealing additional loci.

Interactions between outdoor air pollution and genetic variants have been focused on candidate genes related to antioxidative stress and detoxification systems, inflammation, and innate immunity as pathogenic pathways for asthma. Interactive effects of variants in genes belonging to the GST family (GSTP1 Ile105Val, GSTM1 deletion) with levels of different outdoor air pollutants (e.g., NO₂, ozone, and particulate matter PM₁₀ and PM_2.5) have been described for childhood asthma (71–75) and traffic-related air pollution (76). Bowatte et al. (76) first reported significant effect modification of GSTT1 polymorphisms for the association of traffic-related air pollution exposure and childhood-onset asthma, in contrast to previous studies that found no evidence of interaction (73, 74). In addition, Bowatte et al. also first demonstrated effect modification by GSTT1 genetic variation, but not by GSTM1 or GSTP1 polymorphisms, on the association between traffic-related exposure and asthma in adults (77), in contrast to the study by Castro-Giner et al. that only detected an interaction with common polymorphisms in the NQO1 gene (78).

Genetic variation in few genes related to inflammation and innate immunity have also been investigated as effect modifiers between outdoor air pollution and asthma. In the study of Li et al., the common TNF-α−308 GG genotype was found to interact with annual average levels of ozone in childhood onset asthma (79); although results were not replicated (75). Polymorphisms in toll-like receptor genes (TLR2 and TLR4) have also been identified as potential effect modifiers of the association between outdoor PM2.5 levels and childhood asthma (80).

Recently, a GWIS of air pollution exposure and childhood asthma showed supportive evidence for interaction with outdoor NO₂ levels for the novel loci B4GALT5 and the previously lung disease associated loci ADCY2 and DLG2 (81).

In summary, results on interactions between exposure to outdoor air pollution and genetic variants on genes related to antioxidative stress, detoxification system, inflammation, and innate immunity on childhood and adult onset asthma are inconsistent across studies.

Several studies have investigated interaction between candidate genes and diverse indoor exposures. Gene-environment interactions between the IL-4 promoter and an indoor mold may play an important role in childhood asthma (82). Increased BHR in adults was associated with gas cooking (a major indoor source of NO2), but only among subjects with the GSTM1 null genotype (83). IL-13 variants were found to interact with household carpet use on the risk of asthma in Taiwanese children (84).

The number of gene interactions with a farming environment remains very limited, focusing primarily genes related to innate immunity. Bieli et al. reported an interaction between early farm milk consumption and a polymorphism in the CD14 gene (CD14/-1721) on the risk of childhood asthma (85). Effect modification by farm exposure in childhood on the association between polymorphism in TLR genes and asthma have been shown. Being exposed to a farm environment in childhood was protective against childhood-onset asthma for those with TLR2/-16934 T-allele (86), TLR6-rs1039559 T-allele and TLR6-rs5743810 C-allele (87). The most comprehensive GWIS in 1708 children from 4 rural regions of Central Europe for childhood asthma in relation to farm-related exposures did not reveal any significant interaction with common SNPs (88); however, strong interactions were found for rarer variants in 15 genes, particularly of the glutamate receptor, metabotropic 1 gene (GRM1).

Respiratory viral infections in the first few years of life increase risk of childhood asthma; however, not all children develop the disease, suggesting an interaction with the host genetic factors (98). Smit et al. (89) found that 17q21 genetic variants enhance the association between early respiratory infections and childhood-onset asthma. The presence of two common polymorphisms in the extracellular domain of the TLR4 has been found to modify the effect of endotoxins on asthma in adults (90). Moreover, multiple studies have characterized a similar interaction between asthma, CD14 variants and environmental endotoxin exposure, a marker of microbial exposure capable of inducing severe airway inflammation. Zambelli-Weiner et al. found CD14/260 polymorphism to interact with dust endotoxin on the risk of childhood-onset asthma (91). Similarly, in adults, Smit et al. showed occupational endotoxin exposure and wheeze in agricultural workers to be significantly modified by genetic variants in CD14 and MD2 (92). A recent systematic-review has highlighted the apparent modification of the effect of early life endotoxin exposure on risk of asthma in childhood, but not in adults, by the CD14/260 polymorphism (93).

Literature on inhaled exposures such as indoor air pollution and farming-related is scarce and mostly focused on genetic interactions with genes related to innate immunity and childhood onset asthma. Host innate immunity genetic variation might play a key role in childhood and adult onset asthma susceptibility in relation to microbial exposures including endotoxins.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.