The genetic and epigenetic landscapes of the epithelium in asthma

Asthma susceptibility genes in airway epithelium

Other studies detected GPRA (also known as Neuropeptide S Receptor 1; NPSR1, and GPR154) on chromosome 7p15-p14 [19–21]. Both GPRA, which belongs to the G protein-coupled receptor family, and its agonist, Neuropeptide S (NPS) are co-expressed in bronchial epithelium and specific activation of the GPRA-A isoform with NPS inhibits cell growth [19, 20]. Since the balance between epithelial cell proliferation and regeneration is dysregulated in asthmatics [4, 36], GPRA likely plays an important role in the pathogenesis of disease [19, 20]. Further studies identified the SPINK5 gene on chromosome 5q31-35 which encodes a multidomain serine protease inhibitor known as lympho-epithelial Kazal-type-related inhibitor (LEKTI). LEKTI has been shown to be a major physiological inhibitor of multiple serine proteinases, including the exogenous serine proteases trypsin, plasmin, subtilisin A, cathepsin G and neutrophil elastase [37]. SPINK5 is essential in the epidermal barrier function through regulating protease activity [38] and LEKTI plays a crucial role in skin homeostasis by selectively inhibiting human kallikrein-related peptidase genes including, KLK5, KLK7 and KLK14 [39]. LEKTI may therefore protect the epithelium against allergens or inflammatory related proteases. However, the exact function of SPINK5 in airway epithelium remains to be elucidated. Another asthma susceptibly gene is IRAKM, which is located on chromosome 12q13-24. IRAK-M regulates NF-kB and inflammation via suppressing Toll-like receptor/IL-1R pathways. When IRAK-M function is hampered, overproduction of inflammatory cytokines in the lung in response to infection/allergens may result in a Th2-mediated allergic response and/or Th1-dependent exacerbation of asthma symptoms [26].

Further technological advances led to GWAS [40], and associated single nucleotide polymorphisms (SNPs) [13], which detected a completely different set of genes; interleukin (IL) 1 receptor-like 1 (IL1RL1) and IL18 receptor 1 (IL18R1), IL33, HLA-DQ, SMAD3, thymic stromal lymphopoietin (TSLP), ORM1-like 3 (ORMDL3) and gasdermin B (GSDMB) as asthma susceptibility genes expressed in airway epithelium (Table 1) [13, 14, 41–43]. IL1RL1 and IL18R1 contain SNP rs3771166 on chromosome 2 [13, 44]. IL1RL1 (also known as T1, ST2, DER4, or FIT-1) belongs to the IL-1 superfamily and is the receptor for IL-33 [45]. IL33 with SNP rs1342326 located on chromosome 9 is also associated with atopic asthma [13, 46, 47]. IL-33 possesses potent transcriptional-repressive properties and is constitutively expressed in epithelial cells [48]. It has been shown that IL-33 activates NF-kB and mitogen-activated protein (MAP) kinases, and induces production of T-helper (Th) 2-associated cytokines, including IL-4, IL-5, and IL-13 [49]. In this context, IL-33 functions as a prototypical ‘alarmin’ and an endogenous ‘danger’ signal to alert the immune system after epithelial cell damage during trauma or infection [50] and plays an essential role in pro-inflammatory pathway in asthma [13]. IL18R1 encodes the receptor for IL-18 [45]. IL-18 modulates innate and adaptive immune responses by increasing interferon (IFN)-γ production by Th1 and natural killer (NK) cells or by activating IgE production and Th2 cell differentiation [45].

Another candidate gene detected by GWAS is HLA-DQ region of the major histocompatibility (MHC) gene located on chromosome 6, which contains SNP rs9273349 [51]. The airway epithelium expresses MHC class II; a heterodimer molecule that consists of an α- and a β-chain in one of three HLA loci: DR, DP and DQ [52], on their surface [53]. Immune response to allergens is also related to specific HLA-DR and DQ haplotypes [13], and is associated with asthma induced by house dust mite, aspirin, soybean, and occupational triggers [54]. However, the exact role of HLA-DQ in airway epithelium still remains unclear.

SMAD3, with SNP rs744910 located on chromosome 15, is another asthma susceptibility gene [13]. SMAD3 is an essential signal transducer in transforming growth factor (TGF)-β signalling, which is elevated in airway epithelial cells of some asthmatics [55]. TGF-β1 induces epithelial–mesenchymal transition (EMT) in airway epithelial cells via a SMAD3-dependent transcription factor snail1 (SNAI1) which transcriptionally supresses E-cadherin [36, 56]. Furthermore, the TGF-β/SMAD3 pathway play essential roles in the airway epithelial response to respiratory viral infection [57–59], including increasing replication of both respiratory syncytial virus [57, 58] and rhinovirus [59].

Among asthma susceptibility genes, ORMDL3 and GSDMB, with SNP rs2305480 at chromosome 17q21, are associated with childhood asthma [60]. ORMDL3 is a member of a gene family that encodes transmembrane proteins anchored in the endoplasmic reticulum of airway epithelial cells, predominantly [60, 61]. Allergens induce ORMDL3 expression in airway epithelium leading to increased expression of asthma-associated chemokines, metalloproteases and the unfolded protein response (UPR), which may implicate the potential link between ORMDL3 and asthma [61]. ORMDL regulates ORM protein expression in airway epithelial cells, which is induced in response to allergen challenge [62]. ORM proteins are important homeostatic regulators of sphingolipid metabolism [63], which is associated with the pathogenesis of asthma [64]. Sphingolipids are pivotal in maintenance of cell structure and signaling pathways in physiological and pathological processes; e.g. proliferation, apoptosis and migration [63, 65] and have been shown to contribute to BHR in experimental models of asthma [66]. Further meta-analysis has showed that SNP rs7216389 in the ORMDL3 may play essential and independent predisposing roles in ethnically diverse populations for both childhood and adult-onset asthma [41]. GSDMB is adjacent to ORMDL3 and is a member of gasdermin family that encodes gasdermin B protein which has roles in secretory pathways, epithelial cell differentiation, cell cycle control and apoptosis [67, 68]. Furthermore, there are several response elements for interferon regulatory factors present in the GSDMB promoter region and epithelial interferon-α induces GSDMB gene and protein in human nasal epithelial cells, in vitro [69]. GSDMB has been proposed to be the causative gene associated with asthma [70].

Among the candidate genes identified by GWAS, TSLP on chromosome 5 plays protective roles against the risk of asthma, atopic asthma and BHR across various ethnic groups [42, 43, 71–73]. The rs1837253 SNP may be directly involved in the regulation of TSLP secretion in primary nasal epithelial cells [42]. TSLP is an IL-7 like cytokine that induces myeloid dendritic cells to stimulate the differentiation of naive CD4⁺ T cells to Th2 cells. TSLP mRNA and protein are highly expressed in the asthmatic airway epithelium [72–74]. TSLP has been shown to induce bronchial epithelial cell proliferation and increases repair responses to injury through IL-13 production [74].

Collectively these genes are important in epithelial cell damage, innate and adaptive immunity, and airway inflammation, which are pivotal in the pathology of asthma. Furthermore, some of the products associated with these genes can determine the phenotype of asthma. For instance, the level of IL-33 is highly elevated and widely distributed in bronchial epithelial cells of moderate and severe asthmatics [48]. IL-18 may also contribute to asthma exacerbations in mild and moderate asthmatics through activation of immunologic responses [51]. Given the relationship to asthma endotypes, these genes may indicate pathways for therapeutic intervention. In fact, Phase II trials are currently proceeding using an anti-TSLP antibody; AMG 157 from Amgen Corp., to neutralise the TSLP cytokine for the treatment of allergic diseases as asthma [75].

Notably, only a few genes, such as IL33 and TSLP, are shared among all asthmatics [42, 76, 77] and may play roles as potential biomarkers. Furthermore, while the association of the 17q21 locus (ORMD/GSDMB) with asthma is the most consistent finding from different studies, there is limited evidence to validate certain SNPs [14]. Integrative genomics defined as identification of causal genes and variants, with improved statistical power, is a promising new approach. By using gene expression as a phenotype and examining how DNA polymorphisms contribute to both gene expression (expression quantitative trait loci; eQTLs) and disease phenotypes, true causal relationships can be discovered [78–80]. Although GWAS have identified loci that are strongly associated with asthma, the molecular mechanisms underlying these associations rely on other technology such as eQTLs [78].

One eQTL study showed that chromosome 17q21, which contained strong GWAS hits, also regulates expression levels of cyclin-dependent kinase 12 (CDK12), protein phosphatase 1 regulatory subunit 1B (PPP1R1B), titin-cap (TCAP) and StAR-related lipid transfer (START) domain containing 3 (STARD3) genes in the airway epithelium [78]. CDK12 is a member of the cyclin-dependent kinase (CDK) family, which are serine/threonine kinases regulating cell cycle progression [78, 81]. Airway epithelial cells from asthmatics overexpress the CDK inhibitor; p21^waf [82], which may explain the abnormal repair responses of the airway epithelium of asthmatics after wounding [82]. However, the role of TCAP, PPP1R1B, STARD3 in asthma are still unknown [78]. Furthermore, epithelial eQTL detected Cystatin SN (CST1) on chromosome 20p11.21, which contains SNP rs16856186 [78]. CST1 may neutralise cystatin C; a potent cathepsin B inhibitor, and increase cell proliferation [83]. CST1 is expressed differentially in airway cells of asthmatics with exercise-induced bronchoconstriction (EIB) compared to asthmatics without EIB [84]. eQTL also confirmed cadherin-related family member 3 (CDHR3) gene, as an epithelial susceptibly gene for severe exacerbations in childhood asthma [78, 85]. CDHR3 encodes a hemophilic cell adhesion molecule, which may be involved in maintaining cell integrity by forming cell-cell junctions. Furthermore, functional disruption of CDHR3 has been reported in human rhinovirus-induced asthma exacerbation [86]. Also, epithelial eQTL supported SPINK5 as an asthma susceptibility gene [86], as described earlier.

It is also essential to note that in a disease as complex as asthma, it is unlikely that one or a few functional gene variants will be responsible for all pathophysiological events. While GWAS have been useful and continue to identify novel genes for allergic diseases through increased sample sizes and phenotype refinement, further approaches to integrate analyses of rare variants, eQTL approaches, and epigenetic mechanisms will likely lead to greater insight into the genetic basis of the disease.

The advent of whole genome sequencing (WGS), which includes copy number variants (CNVs) and low-frequency variants, has been proposed to overcome the drawbacks of the earlier technologies [15, 76]. CNVs, which are genetic variants including the deletion or duplication of more than 50 bp of gene sequence [15], are one the most recent advances to detect asthma susceptibility genes. Recently, an association between a 6 kbp deletion in an intron of NEDD4L with increased risk of asthma was reported but only in Hutterites [15]. NEDD4L is expressed in bronchial epithelial cells, and NEDD4L knockout mice showed severe airway inflammation and mucus accumulation [15].

To adequately assess the entire genome, a large number of genetic polymorphisms (250,000 to 1 million) is required and the number of polymorphisms will vary between studies due to different levels of linkage disequilibrium [14]. Currently, WGS is neither affordable nor feasible on the large number of individuals to acquire sufficient power for detecting associations with asthma [15].

Effect of environmental exposure on asthma

Environmental factors play essential roles in asthma aetiology. The increase in the prevalence of asthma worldwide during recent decades, the substantial variations in populations with a similar racial and ethnic background but exposed to different environmental stimuli, and the significant increase in the frequency of occupational asthma are all pointing out toward the important role of environmental factors [87].

Environmental stimuli affecting asthma are categorised to outdoor and indoor factors. Outdoor stimuli that trigger or exacerbate asthma include microbial and viral pathogens, airborne particulates, ozone, diesel exhaust particles, pollens, outdoor moulds, environmental tobacco smoke, cold air, and humidity [87, 88]. Indoor environmental factors include allergens derived from dust mites, cockroaches, mice and pets which has been shown to induce airway inflammation; particles generated from indoor burning of tobacco, wood, and biomass; and biological agents such as indoor endotoxin, products from gram-positive bacteria, and 1,3-β-glucans from moulds [87, 88].

In particular relation to asthma susceptibility, the exposure to specific environmental factors can play key factor in the induction or suppression of asthma-related genes. The main areas of studies in regards to the impact of gene-environment interactions on asthma development and pathogenesis have been so far related to smoking, air pollution, and microbial exposures. Maternal smoking is one of the major risk factor for asthma in offspring. Maternal smoking substantially enhances the strength of the linkage signal on chromosome 5q31–34 to asthma in the children [89, 90]. Furthermore, polymorphic variation in candidate genes known to be involved in asthma, for example TNF-308 and glutathione-S-transferase M1 (GSTM1; involved in detoxification of oxidative stress and lung function growth in children), are predictors of BHR to passive smoking [89, 91]. The most well-known interaction between environmental factors and gene is between endotoxin with Toll like receptor (TLR)-4 with further impact on adaptive immune response, epithelial and smooth muscle cells through NF-kB. Polymorphism in TLR-4 is related to asthma and it is proposed that the other TLRs (e.g. TLR-9 and -3) present the similar polymorphic associations with other environmental stimuli, such as CpG methylation of TLR-9 and double-stranded RNA (dsRNA) for TLR-3 [89, 92]. These reports point out to the importance of early life environmental factors, such as passive smoking, pollutant exposure and viral infections, as a perverse factor on the developing asthma in childhood.

However little is known about the effect of environmental-gene interactions in airway epithelium of asthmatics. It has been shown that particulate matter with a diameter of <10 μm diameter (known as PM10) increases HAT activity and the level of acetylated histone 4 (H4) through oxidative stress. PM10 induced histone acetylation is associated with promoter region of the IL-8 resulting in increased IL-8 gene and protein release from alveolar epithelial (A549) cells [93]. Interestingly, butyrate; a fermentation product of intestinal bacteria, also showed to enhance histone acetylation by inhibition of HDAC enzymes leading to an increase in gene expression of inflammatory cytokines in intestinal epithelial cells [94]. Cigarette smoke-induced oxidative stress also reduces HDAC2 and increases cytokines expression in alveolar macrophages [95] but the effect on airway epithelium is yet to be determined.

Most importantly, many of the indoor and outdoor asthma triggers also have demonstrable reprogramming effects on the immature airway during early life, leading to altered asthma risk in later life. Asthma hence is not a homogeneous disease but a condition influenced by interactions between genetic and environmental factors through epigenetic mechanisms that influence gene expression.

Epigenetic regulatory factors in airway epithelium

Environmental challenges can affect gene expression through epigenetic mechanisms. Epigenetics is described as a heritable regulation of gene transcription that does not require alterations in gene sequence [16]. Epigenetic changes may form stable heritable changes in gene expression and in a tissue-specific fashion [96, 97]. Particularly, epigenetic regulation affects gene expression through three main mechanisms, including DNA modifications, histone modifications, and non-coding RNAs (Fig. 1 and Table 2).

DNA modification signatures
Gene	Status	Function	References
KRT5a	Hypo-methylation	Epithelial homeostasis	[97, 99]
STAT5Aa	Hyper-methylation	Immune system, Cell proliferation	[99]
CRIP1a	Hyper-methylation	Epithelial homeostasis, transcription	[99]
ARG2a	Hyper-methylation	Reduced FeNO	[100]
IL-6a	Hypo-methylation	Increased FeNO	[98]
iNOSa	Hypo-methylation	Increased FeNO	[98]
ADAM33	Hyper-methylation	BHR	[101]
Histone modification signatures
HDAC/HAT	Status	Function	References
H3K18	Acetylation	Increases the expression of ΔNp63, EGFR and STAT6 affecting epithelial homeostasis	[109]
HDAC2a	De-acetylation	Anti-inflammatory	[112]
miRNA signatures
miRNA	Status	Function	References
let-7fb	Overexpressed	unknown	[121]
miR-487bb
miR-181cb
miR-203b	Suppressed	Targeting p63 and c-Abl	[121–123]
miR-34/449 family	Suppressed	Targeting NOTCH1 mRNA and affecting cell homeostasis	[125]
miR-18a	Suppressed	activation/signalling of IL-6 and IL-8	[129]
miR-27a
miR-128
miR-155
miR-19ac	Overexpressed	Targeting TGF-β receptor 2 mRNA and affecting cell homeostasis	[130]

^aChildren

^bMild asthma

^cSevere asthma

miRNA expression in epithelium of asthmatics

There are a limited number of studies examining miRNA expression in epithelium of asthmatics or non-asthmatics (Table 2). miRNA microarray performed on primary bronchial epithelial cells cultured at air-liquid interface (ALI) showed higher expression of let-7f, miR-181c* and miR-487b but lower expression of miR-203 in mild asthmatics compared with healthy controls [121]. miR-203 has been shown to play a potent role in the (keratinocyte) self-renewal program during epidermal differentiation by targeting p63, facilitating cell cycle exit and promoting differentiation [122]. Given that the epithelium of asthmatics expresses higher levels of p63 [36], it is intriguing to speculate that there is a direct link between miR-203 and p63 (Fig. 2). Recently, miR-203 has also been shown to inhibit airway smooth cell proliferation through targeting the non-receptor tyrosine kinase c-Abl (Abelson tyrosine kinase, Abl, ABL1) [123]. In particular, Abl kinases regulate cell-cell adhesion in epithelial cells and fibroblasts through cadherin-mediated adhesion signals via regulating the activities of the Rac and Rho GTPases [124]. Since, c-Ab1 plays important roles in regulation of the actin cytoskeleton and hence different cellular functions such as proliferation, cell adhesion and migration as well as growth and development [123], the reduced level of miR-203 in epithelial cells of asthmatics may promote cell proliferation, increase goblet cell hyperplasia and/or decrease ciliated cells. Genome wide profiling of bronchial epithelial brushings also revealed four members of the miR-34/449 family (miR-34b-5p, miR-34c-5p, miR-449a, and miR-449b-5p) were significantly suppressed in asthma [125]. Interestingly, no clear relationship was observed between these differentially expressed miRNA and serum IgE level in asthmatics [125], which indicates the role of these miRNAs at cellular and molecular levels rather than inflammatory and allergic responses. Furthermore, inhaled corticosteroids showed only minor effects on miRNA expression, and failed to restore miRNA levels to healthy control levels [125]. The miR-34/449 family are closely associated with regulation of epithelial cell proliferation and differentiation. Specifically, miR-449 is essential in regulation of airway ciliated cells by targeting NOTCH1 [126]. Notch signaling triggers airway mucous metaplasia and inhibits alveolar development (ciliated cells) [127]. The low level of miR-449 in epithelial cells of asthmatics may therefore shift the fate of these cells toward more mucous production (Fig. 2). Furthermore, miR-34/449-deficient mice suffer from primary ciliary dyskinesia (PCD). These abnormalities have been shown to be mediated by Cp110, a centriolar protein suppressing cilia assembly, that is a target of miR-34/449 [128]. A panel of miRNAs including miR-18a, miR-27a, miR-128 and miR-155 were also down-regulated in the epithelium of asthmatics. These miRNAs are involved in activation/signaling of IL-6 and IL-8 [129].

Recently, miR-19a was reported to be the only miRNA that differentiates severe from mild asthma. miR-19a is up-regulated in severe asthmatic epithelial cells and is not restored by corticosteroids [130]. Elevated levels of miR-19a further stimulate cell proliferation of epithelial cells by targeting TGF-β receptor 2 mRNA. Overexpression of miR-19a results in reduction in phosphorylated SMAD3 whereas suppression of miR-19a facilitates SMAD3 phosphorylation through TGF-β receptor 2 signaling and restore epithelial cells proliferation [130]. These may suggest the potential of miR-19a as a therapeutic target in airway epithelium of asthmatics to re-establish epithelial regeneration.

There are limited numbers of investigations on miRNAs expression in asthmatic epithelium of children, possibly due to difficulty in obtaining epithelial samples. Higher level of miR-148b; a member of miR-152 family, was reported in airway epithelial cells of adults with an asthmatic mother and correlates with sHLA-G levels in the BAL fluid of these subjects [131]. The prenatal effects of maternal asthma on the regulation of foetal genes in airway cells may persist well into adulthood through miRNA regulatory mechanisms.