Genome-wide association studies of atopic dermatitis
Laboratory for Respiratory and Allergic Diseases, Center for Integrative Medical Sciences, The Institute of Physical and Chemical Research (RIKEN), Kanagawa, Japan
Correspondence: Mayumi Tamari, M.D., Ph.D., Laboratory for Respiratory and Allergic Diseases, Center for Integrative Medical Sciences, Institute of Physical and Chemical Research (RIKEN), 1-7-22 Suehiro, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan. Email: email@example.com
Atopic dermatitis is a common inflammatory disease caused by a combination of genetic and environmental factors. Genome-wide association study (GWAS) is a comprehensive and unbiased approach to identify the genetic components of human diseases and to discover the cellular pathways underlying them. GWAS and recent immunochip analysis of atopic dermatitis have identified a total of 19 associated loci with a genome-wide level of significance (P < 5 × 10−8). The candidate genes identified by GWAS suggest a role for epidermal barrier functions, innate-adaptive immunity, interleukin-1 family signaling, regulatory T cells, the vitamin D pathway and the nerve growth factor pathway in the pathogenesis of atopic dermatitis. Combinations of these genetic factors may influence a wide range of phenotypes of atopic dermatitis among individuals. Although a more complete collection of associated genes and pathways is needed, genetic components revealed by GWAS provide valuable insights into the pathophysiology of atopic dermatitis.
Atopic dermatitis is a complex disease caused by a combination of genetic and environmental factors.[1, 2] Recent advances in genomic medicine have improved our understanding of the human genome's contribution to health and disease. Genome-wide association study (GWAS) is a powerful method for identifying disease susceptibility genes for common human diseases and has begun to reveal underlying cellular pathways and point to new therapeutic approaches.[3, 4] This review focuses on GWAS and immunochip analysis of atopic dermatitis. Genes located within and around the susceptible loci and their possible roles in pathogenesis of atopic dermatitis are also discussed.
Genomic variation and GWAS
Human genetic variants are referred to as either common or rare. Common variants, also called polymorphisms, are defined as variants with a minor allele frequency of at least 1% in the population. Single nucleotide polymorphisms (SNP) are the most prevalent class of genetic variations among individuals. GWAS comprehensively surveys associations between common SNP in common complex disorders.
The international HapMap (haplotype map) project (http://hapmap.ncbi.nlm.nih.gov/) focuses on common SNP with minor allele frequencies of more than 1% and defines these patterns across the entire genome.[7-9] The haplotype structure of the human genome implies that a limited set of 1 000 000 SNP can capture approximately 90% of the genetic variation in the population. Pairwise linkage disequilibrium (LD) is calculated based on the genotyping data from the international HapMap project, and a set of markers in a haplotype are said to be in LD. Genotyping arrays have also been developed with reference to the LD, and can now assay up to 4.3 million variants. The data from the HapMap project and the development of dense genotyping chips have enabled us to conduct GWAS effectively on a large number of samples. However, if large numbers of SNP are examined simultaneously in a GWAS, the statistical significance threshold must be adjusted to reduce the chance of making a type 1 error (false positive). A genome-wide significance threshold level of 5 × 10−8 is generally used to adjust for 1–2 million independent tests. To date, GWAS have identified more than 1000 genetic loci associated with disease susceptibility and other disease-related phenotypes.
Genotype imputation is a statistical approach deducing the allelic states at polymorphic sites not actually genotyped from the surrounding SNP data, based on LD and likelihood estimates. To conduct rigorous statistical and comprehensive GWAS meta-analyses for common disorders, a number of GWAS consortiums have been founded. However, GWAS are often performed using different genotyping platforms. Imputation is useful for merging distinct GWAS and enables powerful combined analyses of GWAS for the discovery of novel associations.[12, 13]
Although recent GWAS have provided valuable findings about genetic factors in common human diseases, they explain little about heritability and the associated genetic variants have small effect sizes.[14, 15] Unmapped common and rare variants are considered a source of missing heritability.[14, 15] Actually, the chips currently used in GWAS contain relatively few rare SNP in the coding and promoter regions. The 1000 Genomes Project aims to find essentially all variants with frequencies of more than 1% across the genome and more than 0.1% in protein-coding regions and expands the catalog of variations for the next generation of GWAS. Using data from the 1000 Genomes Project and other available disease-specific re-sequencing data, the immunochip, a custom Illumina Infinium high-density array containing 196 524 variants, has been developed.[17, 18] The variants on the chip cover intervals with established GWAS.[17, 18] However, the immunochip does not cover the whole genome and is designed for use in white European populations. Thus, it would seem to be less informative for Asian populations.
Genetic Studies of Atopic Dermatitis
Systemic, well-powered, genome-wide surveys using GWAS and immunochip analysis have explored the relationship between SNP and susceptibility to atopic dermatitis.[19-23] It is well established that common loss-of-function variants of filaggrin (FLG) are a major predisposing factor for atopic dermatitis.[24, 25] Because FLG mutations involved in atopic dermatitis are featured in another review article in this special issue, this review focuses on GWAS and immunochip analysis for atopic dermatitis. To date, a total of 19 genome-wide significant (P <5 × 10−8) susceptibility loci have been identified through GWAS and immunochip analysis, and Table 1 summarizes the findings.
Table 1. A total of 19 genome-wide significant susceptibility loci identified by genome-wide association and immunochip studies for atopic dermatitis
The first GWAS of atopic dermatitis was reported in 2009. The GWAS analyzed 939 cases and 975 controls as well as 270 complete nuclear families with two affected siblings. A total of 342 303 markers were assessed, and a susceptibility region on chromosome 11q13 located 38 kb downstream of C11orf30 was identified. The peak association was observed 68 kb upstream of leucine rich repeat containing 32 (LRRC32). LRRC32 was previously reported to have specific expression in activated human naturally occurring regulatory T cells.[26, 27] The GWAS also confirmed the important role of FLG mutations in atopic dermatitis. In the study, another association reaching genome-wide significance was observed at rs6661961 in the epidermal differentiation complex (EDC) on chromosome 1q21. Rs6661961 is in high LD with the FLG mutations, and exclusion of FLG mutation carriers abolished the observed association.
In 2011, a GWAS reported two new susceptibility loci, 5q22.1 and 20q13.33, for atopic dermatitis in the Chinese Han population. The GWAS was conducted using 1012 cases and 1362 controls followed by a replication study in an additional 3624 cases and 12 197 controls. The results were validated using 1806 cases and 3256 controls from Germany. The 5q22.1 region contains transmembrane protein 232 (TMEM232) and solute carrier family 25, member 46 (SLC25A46), and the thymic stromal lymphopoietin (TSLP) gene is located adjacent (~300 kb downstream) to the associated region. TSLP promotes T-helper (Th)2 cell responses associated with immunity to some helminth parasites and the pathogenesis of a number of inflammatory diseases such as atopic dermatitis.[28, 29] A variety of stimuli, including viral infections, inflammatory cytokines, protease allergens, bacteria and bacterial products, induce TSLP from barrier epithelial cells.[28, 29] Another study has reported that genetic variants in TSLP are associated with atopic dermatitis, and the association is stronger in patients with atopic dermatitis who have a history of eczema herpeticum. The 20q13.33 region contains the tumor necrosis factor receptor superfamily, member 6b (TNFRSF6B) gene that encodes decoy receptor 3 (DcR3). Interestingly, DcR3 binds to LIGHT (TNFSF14), which has been shown to bind a herpesvirus entry mediator (HVEM), and LIGHT is a target for asthmatic airway remodeling.[31, 32] Herpes simplex viruses (HSV-1 and -2) attach to the cell surface via this herpesvirus entry mediator. It has been known that eczema herpeticum, a disseminated viral infection after HSV infection, is a serious complication of atopic dermatitis. Although further investigation is required, genetic variants in these loci may influence the immune response to HSV.
Large sample sizes are required to ensure sufficient statistical power, and collaborative studies on samples from a large number of cohorts have provided new insights about the genetic bases of common diseases. In 2011, a meta-analysis of GWAS for atopic dermatitis of 5606 cases and 20 565 controls and a validation study in an additional 5419 cases and 19 833 controls were conducted. The studies identified a total of three novel risk loci for atopic dermatitis, 11q13.1 (OVOL1), 19p13.2 (ACTL9) and 5q31.1 (KIF3A/IL4/IL13). The 11q13.1 locus contains OVO homolog-like 1 (OVOL1), which functions as a c-Myc repressor in superbasal cells and is required for proliferation exit of committed epidermal progenitor cells. OVOL1 knockout mice show keratinocyte hyperproliferation, hair shaft abnormalities, kidney cysts and defective spermatogenesis.[34, 35] The most significant association in the 19p13.2 region was observed between the a disintegrin and metalloproteinase with thrombospondin motifs 10 (ADAMTS10) and actin-like 9 (ACTL9) genes. ADAMTS proteins are secreted zinc-dependent metalloproteinases and play a role in connective tissue remodeling and extracellular matrix turnover.[36, 37] Actin proteins are important for the maintenance of epithelial morphology and cell migration.[38, 39] The 5q31.1 region contains a clustered family of Th2 cytokine genes, including IL13 and IL4. The adaptive immune response in atopic dermatitis is associated with increased expression of the Th2 cytokines, interleukin (IL)-4 and IL-13. Recent studies have shown a significantly greater number of IL4, IL5 and IL13 mRNA-expressing cells in acute skin lesions and an increased number of Th2 cells expressing IL4 and IL13 mRNA in unaffected skin of atopic dermatitis patients.[41, 42] This GWAS suggests that atopic dermatitis is caused by both epidermal barrier dysfunctions and immunological features.
GWAS for atopic dermatitis in the Japanese population
We conducted a GWAS in the Japanese population with 1472 cases and 7971 controls followed by a replication study in an additional 1856 cases and 7021 controls. We assessed a total of 606 164 SNP loci and finally found a total of eight novel susceptibility loci for atopic dermatitis.
Chromosome 2q12 contains the receptors of the IL-1 family cytokines IL1RL1, IL18R1 and IL18RAP. Many of the IL-1 family receptors are clustered in a small region (~0.5 Mb; Fig. 1), and IL1RL1 (also known as ST2) is a component of the IL-33 receptor.[43, 44] IL-1 family members are abundantly expressed in the skin, and the IL-33/IL-1RL1 axis triggers the release of several pro-inflammatory mediators and induces systemic Th2-type inflammation.[45, 46] IL-33 and its receptors, ST2 and IL-1RAcP, are increased in lesional skin of patients with atopic dermatitis, and topical tacrolimus treatment suppresses the increased expression of IL-33 and ST2 caused by irritants, allergens and staphylococcal enterotoxin B. Interestingly, expression levels of IL-1α and IL-1β in the stratum corneum are increased in patients with atopic dermatitis with FLG mutations, and these levels are inversely correlated with natural moisturizing factor. Associated variants in the loci may influence the pathophysiology of atopic dermatitis through dysfunction of the IL-1, IL-18 and IL-33 signaling pathways.
In our study, the most significant association was at rs176095 in the major histocompatibility complex (MHC) class III region, and conditional logistic regression analysis showed two independent association signals in the MHC class I and class III regions. The MHC is associated with a number of autoimmune diseases, and the MHC class I and III regions are the shared GWAS autoimmune loci.[49, 50] In recent GWAS, the MHC class I region was found to be associated with psoriasis and vitiligo, and the MHC class III region with vitiligo.[51-58] Furthermore, immunoglobulin (Ig)E antibodies against keratinocytes and endothelial cells are observed in serum specimens from subjects with severe atopic dermatitis. The genetic components in the MHC region may influence the development of autoimmune conditions in chronic inflammation in patients with atopic dermatitis.
Chromosomal region 11p15.4 contains nucleotide-binding domain, leucine-rich repeat and pyrin domain containing protein 10 (NLRP10), which belongs to the NALP family but lacks the leucine-rich repeat region. A recent study has shown that NALP10 has an essential role to initiate adaptive immunity by dendritic cells. Nlrp10−/− mice show a profound defect in helper T-cell-driven immune responses to diverse adjuvants and a dendritic cell intrinsic defect in emigration from inflamed tissues. This suggests that NLRP10 plays a critical role in diverse types of innate immune stimulation and the immune function of mature dendritic cells.
The 3p21.33 region is located adjacent to the chemokine (C-C motif) receptor 4 (CCR4) gene, which encodes a Th2-associated chemokine receptor for chemokine (C-C motif) ligand 22 (CCL22) and CCL17 (thymus and activation-regulated chemokine [TARC]). A keratinocyte-derived TSLP induces dendritic cells to produce TARC, and CCR4 mediates skin-specific recruitment of T cells during inflammation. A recent study has shown that the TARC in stratum corneum from patients with atopic dermatitis is correlated with diverse clinical phenotypes: the severity of local skin lesions, the severity scoring of atopic dermatitis (SCORAD) index, serum TARC level, serum IgE level, serum lactate dehydrogenase level, IL-4 producing T-cell ratio and blood eosinophil count. Furthermore, CCR4 is a skin-homing receptor of Th22 cells, and significant intensification of Th22 and Th2 cytokines is observed in acute and chronic lesions of atopic dermatitis. Associated SNP may also play a role in the Th22 immune response in atopic dermatitis.
The 3q13.2 locus contains CCDC80, which is involved in induction of C/EBPα and peroxisome proliferator-activated receptor (PPAR). C/EBPα and C/EBPβ are co-expressed in basal keratinocytes and upregulated while keratinocytes exit the basal layer and undergo terminal differentiation. PPARγ acts as a negative regulator in immune cells, and a PPARγ agonist markedly suppresses both expression of TSLP in the skin and maturation and migration of dendritic cells in a murine model of atopic dermatitis.
The 7p22 region contains caspase recruitment domain-containing protein 11 (CARD11), which encodes CARD-containing MAGUK protein 1 (CARMA1). CARMA1 plays an essential role in T-cell differentiation and has a critical role in the regulation of JunB and GATA3 transcription factors and subsequent production of Th2 cell-specific cytokines. Furthermore, mice homozygous for the Carma-1 mutation show gradual development of atopic dermatitis with hyper-IgE.
Chromosomal region 10q21.2 contains three genes and, interestingly, contains early growth response protein 2 (EGR2), a key molecule for anergy induction that activates the expression of genes involved in the negative regulation of T-cell proliferation and inflammation. A recent study has shown that CD4+CD25–LAG3+ regulatory T cells (Treg) characteristically express Egr-2. Interestingly, the presence of environmental microbiota influences the number of CD4+CD25–LAG3+ Treg.
The 20q13 region includes cytochrome p450, family 24, subfamily A, polypeptide 1 (CYP24A1), which initiates the degradation of 1,5-dihydroxyvitamin D3, the active form of vitamin D3. Vitamin D is a modulator of innate and adaptive immune system functions and plays a role in epithelial antimicrobial defense of the skin and intestine. An association between vitamin D deficiency and the severity of atopic dermatitis has been reported.
Immunochip analysis of atopic dermatitis
Rare large-effect mutations have been recognized as causes of many different common medical conditions. The immunochip was developed to conduct deep replication of major autoimmune and inflammatory diseases, and fine mapping of established loci identified by GWAS.[17, 18] A recent immunochip analysis for atopic dermatitis revealed four new susceptibility loci: 4q27 (IL-2/IL21), 11p13 (PRR5L), 16p13.13 (CLEC16A/DEXI) and 17q21.32 (ZNF652). A replication study using a Japanese population confirmed the associations at the 11p13, 16p13.13 and 17q21.32 loci. Because the top signal in European populations is not always replicated in Asian populations, further fine-mapping analysis of the locus is necessary in the Japanese population. The 4q27 region contains the IL2 and IL21 genes. IL-2 has pleiotropic immunoregulatory functions and controls the proliferation and survival of regulatory T cells. Tacrolimus (FK506) is a well-known medication for atopic dermatitis and blocks IL-2 gene activation in T cells. Interestingly, the TNF receptor associated factors (TRAF6) (11p13), recombination-activating gene 1 (RAG1) (11p13), RAG2 (11p13), suppressor of cytokine signaling 1 (SOCS1) (16p13.13) and nerve growth factor receptor (NGFR) (17q21.32) genes are located adjacent to the associated regions.
GWAS for total IgE levels, allergic sensitization and eosinophil numbers
Elevated total IgE levels are considered markers of helminth exposure and allergic inflammation,[80, 81] and there is a strong genetic contribution to the variability of the total IgE level.[82, 83] GWAS for total IgE levels have been conducted and identified several susceptibility loci at genome-wide significance levels.[84, 85] A GWAS of total serum IgE levels found two susceptibility loci, 1q23 (FCER1A) and 5q31 (RAD50) in 1530 individuals from the population-based study following replication studies in four independent population-based study samples (n = 9769). The study showed that a functional SNP in the region significantly influenced the cell surface expression of Fc fragment of IgE, high affinity I, receptor for; alpha polypeptide (FCER1A) on basophils and the regulatory mechanism of FCER1A expression via GATA2. Another GWAS in the Framingham Heart Study (FHS) reported a total of three loci containing 1q23 (FCER1A), 12q13 (STAT6) and 5q31 (IL13) with genome-wide significance. Furthermore, a meta-analysis combining the GWAS of the FHS and replication cohorts identified additional loci, 6p21.3 (HLA-G, HLA-DQA2 and HLA-A) and 1q21 (Duffy blood group, atypical chemokine receptor [DARC]), although the DARC association seems not to be independent of SNP in the nearby FCER1A gene. A GWAS of diverse populations including African American and Latino subjects validated findings from previous GWAS (1q23, 5q31, 6p21.3 and 12q13) and found a variant near HLA-DQB1 as a predictor of the total serum IgE levels in multiple race-ethnic groups, but not at the level of genome-wide significance (Pcombined = 2.45 × 10−7).
Allergen-specific IgE plays a crucial role in the pathogenesis of allergic diseases by binding allergens and initiating immunological responses. A meta-analysis of GWAS of allergic sensitization was conducted for 5789 cases and 10 056 controls, and the top SNP at each of 26 loci was validated using 6114 cases and 9920 controls. In the study, the sensitization status was assessed for common food and inhalant allergens evaluated by either the level of allergen-specific IgE in blood or a positive skin prick test. Finally, a total of 10 loci influencing allergic sensitization were identified: 11q13.5 (C11orf30), 12q13.3 (STAT6), 5q22.1 (SLC25A46), 6p21.32 (HLA-DQB1), 2q12.1 (IL1RL1/IL18R1), 4p14 (TLR1/TLR6/TLR10), 3q28 (LPP), 8q24.21 (MYC/PVT1), 4q27 (IL2/ADAD1) and 6p21.33 (HLA-B/MICA). 2q12.1, 4q27, 5q22.1, 6q21.3 and 11q13.5 are overlapping susceptibility regions between atopic dermatitis and allergic sensitization (Table 2).
Table 2. Genetic susceptibility loci overlapping atopic dermatitis and other allergy-related phenotypes identified by GWAS
Other allergy-related phenotypes
GWAS, genome-wide association study; Ig, immunoglobulin.
Eosinophilic infiltration is a characteristic feature of allergic diseases and parasitic infestation. Tissue eosinophilia is a typical characteristic of atopic dermatitis, especially seen in lesions with pronounced epidermal hyperplasia. The number of eosinophils in the skin correlates with disease severity, and eosinophils are involved in acute, spongiotic dermatitis. A recent GWAS has shown that the 2q12 (IL1RL1), 2q13 (IKZF2), 3q21 (GATA2), 5q31 (IL5) and 12q24 (SH2B3) loci are associated with the number of blood eosinophils. The study reported an association between an SNP at the IL1RL1 locus at 2q12 and asthma with a genome-wide significance level.
Loci overlapping atopic dermatitis and other allergy-related phenotypes
The allergic march is the natural history of atopic manifestations such as atopic dermatitis, bronchial asthma and allergic rhinitis. Recent GWAS have revealed a number of susceptibility loci for bronchial asthma, allergic rhinitis, the number of eosinophils, allergic sensitization and total IgE levels, and there are significant genome-wide susceptibility regions overlapping atopic dermatitis and other allergy-related phenotypes (Table 2).[92-96] These genetic components may play a role in the atopic march or atopic phenotypes. The 11q13.5 region, which contains LRRC32 (also called GARP), has been reported to be a susceptibility locus for atopic dermatitis, bronchial asthma, allergic rhinitis, the total serum IgE level and allergic sensitization.[19, 95, 96] Foxp3+ Treg play a crucial role in maintaining tolerance to allergens and inhibiting allergies.[97, 98] LRRC32 mRNA is highly expressed in activated Foxp3+ regulatory T cells, and LRRC32 is essential for the surface expression of latent transforming growth factor-β on the cells. Thus, the findings of GWAS also help to enhance our understanding of the allergic march.
Possible roles of the candidate genes for atopic dermatitis
Candidate genes for atopic dermatitis within and around the susceptibility loci suggest roles for epidermal barrier functions, innate and adaptive immunity, IL-1 signaling, the inflammatory response, regulatory T cells, the vitamin D pathway and NGF signaling in the pathogenesis of atopic dermatitis.[19-23] Pruritus is a major symptom of atopic dermatitis and the itch–scratch cycle can damage the epidermal keratinocytes.[1, 40] Recent immunological studies have shown that epithelial-derived cytokines activate innate immune cells such as basophils, mast cells and innate lymphoid cells (ILC).[100-103] Interestingly, the candidate genes for atopic dermatitis identified by the GWAS and immunochip are involved with those innate immune cells.[19-23] Group 2 ILC (comprising ILC2) contribute to promotion of type 2 responses as potent sources of Th2-associated cytokines.[104, 105]. ILC2-derived cytokines play physiological roles in wound healing and immunity to helminths, and are also involved in allergic diseases. Mast cells and basophils, which are also associated with type 2 immune responses, are activated through IL-18R, IL-33R and TSLPR, and these cells produce cytokines such as IL-2, IL-13 and TSLP.
Various human infections, such as West Nile virus, Japanese encephalitis, Tsutsugamushi disease and malaria parasites, are transmitted via insect bites. Basophils and mast cells release histamine, an important mediator of itching, and itching is the essential hallmark of atopic dermatitis. Itching and irritation from skin contact are likely to be a protective response to insect stings. NGF has been suggested to play a role in the pathogenesis of itching in atopic dermatitis. Increased serum levels of NGF and substance P have been shown to correlate with the severity of atopic dermatitis, and increased NGF and its receptors, trkA and NGFR (p75NTR), in atopic dermatitis have been reported in an immunohistochemical study. The 17q21.32 region is located adjacent to the NGFR gene. A dysregulated NGF pathway and dysfunctions of mast cells and basophils through the genetic variations may be involved in coordinating the itch–scratch cycle in atopic dermatitis.
Multiple clinical phenotypes of atopic dermatitis
It is known that there are multiple clinical phenotypes of atopic dermatitis. There are great differences in disease severity between individuals, and mild atopic dermatitis can be outgrown in later childhood. Furthermore, some patients have the atopic march, and some patients are prone to skin infections such as Staphylococcus aureus, eczema herpeticum and Malassezia. Approximately 80% of patients with atopic dermatitis have elevated serum IgE and/or immediate skin test reactivity to allergens. To date, a total of 19 susceptibility loci have been identified at a genome-wide level of significance (Table 1). Combinations of these genetic factors may influence diverse phenotypes of atopic dermatitis among individuals.
Recent GWAS that comprehensively assess genes related to multifactorial diseases in a non-biased manner across the whole genome have improved our understanding of human atopic dermatitis. The susceptibility loci contain a number of genes encoding markers of immune cells and inflammation that are suppressed by topical corticosteroids and calcineurin inhibitors. There may be a novel therapeutic target gene within or around the associated regions. Further studies to elucidate the functional links between associated variants and phenotypic traits are important to improve our understanding of atopic dermatitis. The findings of GWAS are also helpful to highlight the genes and establish an experimental animal model that mimics human atopic dermatitis. Cross-disciplinary studies combining genetics, immunology, bioinformatics, epidemiology and clinical allergology are necessary for translation of the research into clinical practice.