Portal de Periódicos da CAPES

Elevated Expression and Genetic Association Links the SOCS3 Gene to Atopic Dermatitis

2006; Elsevier BV; Volume: 78; Issue: 6 Linguagem: Inglês

10.1086/504272

1537-6605

Elisabeth Ekelund, Annika Sääf, Maria Tengvall-Linder, Erik Melén, Jenny Link, Juliet N. Barker, Nick J. Reynolds, S.J. Meggitt, Juha Kere, Carl‐Fredrik Wahlgren, Göran Pershagen, Magnus Wickman, Magnus Nordenskjöld, Ingrid Kockum, Maria Bradley,

Urticaria and Related Conditions

In a systematic analysis of global gene-expression patterns, we found that SOCS3 messenger RNA was significantly more highly expressed in skin from patients with atopic dermatitis than in skin from healthy controls, and immunohistochemical analysis confirmed a similar elevation of SOCS3 protein. Furthermore, we found a genetic association between atopic dermatitis and a haplotype in the SOCS3 gene in two independent groups of patients (P<.02 and P<.03). These results strongly suggest that SOCS3, located in a chromosomal region previously linked to the disease (17q25), is a susceptibility gene for atopic dermatitis. In a systematic analysis of global gene-expression patterns, we found that SOCS3 messenger RNA was significantly more highly expressed in skin from patients with atopic dermatitis than in skin from healthy controls, and immunohistochemical analysis confirmed a similar elevation of SOCS3 protein. Furthermore, we found a genetic association between atopic dermatitis and a haplotype in the SOCS3 gene in two independent groups of patients (P<.02 and P<.03). These results strongly suggest that SOCS3, located in a chromosomal region previously linked to the disease (17q25), is a susceptibility gene for atopic dermatitis. Atopic dermatitis (AD [MIM 603165]), also known as "eczema" according to the new World Allergy Organization nomenclature,1Johansson SG Bieber T Dahl R Friedmann PS Lanier BQ Lockey RF Motala C Ortega Martell JA Platts-Mills TA Ring J Thien F Van Cauwenberge P Williams HC Revised nomenclature for allergy for global use: report of the Nomenclature Review Committee of the World Allergy Organization, October 2003.J Allergy Clin Immunol. 2004; 113: 832-836Abstract Full Text Full Text PDF PubMed Scopus (1715) Google Scholar commonly begins in infancy or early childhood and is characterized by itchy and inflamed skin. AD affects 10%–20% of children in Western societies and shows a strong familial aggregation.2Cookson WO Ubhi B Lawrence R Abecasis GR Walley AJ Cox HE Coleman R Leaves NI Trembath RC Moffatt MF Harper JI Genetic linkage of childhood atopic dermatitis to psoriasis susceptibility loci.Nat Genet. 2001; 27: 372-373Crossref PubMed Scopus (326) Google Scholar Several genetic linkage and association studies have sought the underlying genetic causes of AD, and a number of different chromosomal regions have been implicated by these studies.2Cookson WO Ubhi B Lawrence R Abecasis GR Walley AJ Cox HE Coleman R Leaves NI Trembath RC Moffatt MF Harper JI Genetic linkage of childhood atopic dermatitis to psoriasis susceptibility loci.Nat Genet. 2001; 27: 372-373Crossref PubMed Scopus (326) Google Scholar, 3Lee YA Wahn U Kehrt R Tarani L Businco L Gustafsson D Andersson F Oranje AP Wolkertstorfer A v Berg A Hoffmann U Kuster W Wienker T Ruschendorf F Reis A A major susceptibility locus for atopic dermatitis maps to chromosome 3q21.Nat Genet. 2000; 26: 470-473Crossref PubMed Scopus (230) Google Scholar, 4Bradley M Soderhall C Luthman H Wahlgren CF Kockum I Nordenskjold M Susceptibility loci for atopic dermatitis on chromosomes 3, 13, 15, 17, and 18 in a Swedish population.Hum Mol Genet. 2002; 11: 1539-1548Crossref PubMed Scopus (88) Google Scholar, 5Haagerup A Bjerke T Schiotz PO Dahl R Binderup HG Tan Q Kruse TA Atopic dermatitis—a total genome-scan for susceptibility genes.Acta Derm Venereol. 2004; 84: 346-352Crossref PubMed Scopus (69) Google Scholar In a Swedish population with AD, a previous genomewide linkage analysis4Bradley M Soderhall C Luthman H Wahlgren CF Kockum I Nordenskjold M Susceptibility loci for atopic dermatitis on chromosomes 3, 13, 15, 17, and 18 in a Swedish population.Hum Mol Genet. 2002; 11: 1539-1548Crossref PubMed Scopus (88) Google Scholar showed suggestive linkage (P<7.4×10−4) to susceptibility loci on chromosomes 3, 13, 15, 17, and 18. Yet very little is known about specific genes involved in the pathogenesis of AD. To search systematically for gene-expression patterns associated with AD, we recently adopted a genomewide approach, using human cDNA microarrays (M. Bradley, P. O. Brown, H. Y. Chang, M. Nordenskjöld, A. Scheynius, A. Sääf, and M. Tengvall-Linder, unpublished data). Nine adult patients with AD were collected at the Dermatology Unit, Department of Medicine, Karolinska University Hospital Solna, Stockholm. From each individual, two punch biopsy specimens were taken: one from lesional skin and one from noninvolved PBS-patch-tested skin on the back. AD was diagnosed using the U.K. Working Party's diagnostic criteria.6Williams HC Burney PG Hay RJ Archer CB Shipley MJ Hunter JJ Bingham EA Finlay AY Pembroke AC Graham-Brown RA Atherton DA Lewis-Jones MS Holden CA Harper JI Champion RH Poyner TF Launer J David TJ The U.K. Working Party's diagnostic criteria for atopic dermatitis. I. Derivation of a minimum set of discriminators for atopic dermatitis.Br J Dermatol. 1994; 131: 383-396Crossref PubMed Scopus (767) Google Scholar The patients had not received treatment for the previous 2 mo and were considered by a dermatologist to be moderately to severely affected with AD. For controls, we used four healthy individuals from the Stockholm area. Two punch biopsy specimens were taken from the back: one from normal skin and one from PBS-patch-tested skin. Each punch biopsy specimen provided 10–30 μg of total RNA. All patients with AD had allergen-specific IgE antibodies, which were not present in any of the controls. We used human cDNA microarrays containing 41,792 elements representing ∼24,500 unique genes (based on UniGene clusters) that were manufactured by the Stanford Functional Genomics Facility. Fluorescently labeled cDNA prepared from amplified RNA was hybridized to the array in a two-color comparative format, with samples from patients with AD and healthy controls labeled with Cy5 (red) and with a reference pool of human mRNAs (Stratagene) derived from 10 cell lines labeled with Cy3 (green). Data were normalized and retrieved as the log2 ratio of fluorescence intensities of the sample and the reference. In analyzing the microarray data by use of the significance-analysis-of-microarrays (SAM) algorithm,7Tusher VG Tibshirani R Chu G Significance analysis of microarrays applied to the ionizing radiation response.Proc Natl Acad Sci USA. 2001; 98: 5116-5121Crossref PubMed Scopus (9447) Google Scholar we found ∼5,000 genes consistently differentially expressed between skin from patients with AD and skin from control individuals, with a false-discovery rate <0.7%. The genomewide portrait of AD will be presented elsewhere (M. Bradley, P. O. Brown, H. Y. Chang, M. Nordenskjöld, A. Scheynius, A. Sääf, and M. Tengvall-Linder, unpublished data). Among all the differentially expressed genes, we focused specifically on genes located within chromosomal regions previously linked to AD. The SOCS3 gene (Locus Link ID 9021; NCBI accession number NM_003955; UniGene cluster ID Hs.527973), located in chromosome region 17q25, was found to be significantly more highly expressed in skin from patients with AD than in skin from healthy individuals (P<.03, corrected for multiple testing) (table 1 and fig. 1).Table 1SOCS3 Microarray DataSOCS3Sequence GenBank Accession NumberSkin SampleAA001218AI922872T72915AD lesional skin: ADL1.87.01−.98 ADL22.432.422.18 ADL32.482.322.43 ADL4.77.74.72 ADL5.36.42.61 ADL61.161.362.13 ADL7.47.54.52AD PBS skinaPBS-patch-tested skin.: ADP1.84.04−.12 ADP2.62.37−.58 ADP3NA−.38−.45 ADP4.79.33.47 ADP5.5.41−.07 ADP6.88.64.69Control normal skin: CN1−.43−.85−.72 CN2NA−.95−1.32 CN3NA−.95−1.14 CN4NA−.41−.78Control PBS skinaPBS-patch-tested skin.: CP1−.91−1.04−.86 CP2−.73−1.14−1.16 CP3−.21−.72−.77 CP4−.42−.75−.62Note.—Data are log2 of red/green normalized ratio. NA = not available.a PBS-patch-tested skin. Open table in a new tab Note.— Data are log2 of red/green normalized ratio. NA = not available. SOCS3 is a member of the suppressors of cytokine signaling (SOCS) family of proteins that act as negative regulators of cytokine signaling. SOCS proteins interfere with the binding of cytokine receptors and with intracellular molecules that act downstream. Several studies have pointed to SOCS proteins as an effective block of the Janus kinase–signal transducer and activator of transcription (JAK-STAT) pathway in the pathogenesis of various inflammatory diseases.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar Furthermore, Seki et al.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar have reported that patients with AD show significantly increased SOCS3 expression (P<.001) in peripheral CD3+ T cells. On the basis of the microarray results and the previously described role of SOCS3 in various inflammatory diseases, we considered SOCS3 to be a candidate gene for AD. To explore the link between mRNA levels and protein expression of SOCS3 in AD, we used immunohistochemistry. For this assessment, 15 adult patients with AD were collected at the Dermatology Unit, Department of Medicine, Karolinska University Hospital Solna, Stockholm. They were classified as having AD on the basis of the U.K. Working Party's diagnostic criteria.6Williams HC Burney PG Hay RJ Archer CB Shipley MJ Hunter JJ Bingham EA Finlay AY Pembroke AC Graham-Brown RA Atherton DA Lewis-Jones MS Holden CA Harper JI Champion RH Poyner TF Launer J David TJ The U.K. Working Party's diagnostic criteria for atopic dermatitis. I. Derivation of a minimum set of discriminators for atopic dermatitis.Br J Dermatol. 1994; 131: 383-396Crossref PubMed Scopus (767) Google Scholar Two punch biopsy specimen from the trunk were taken from each individual: one from lesional skin and one from noninvolved skin. Eight healthy adult controls were collected from the Stockholm area, and biopsy specimens were taken from normal skin. Biopsy specimens were snap frozen on dry ice and were stored at −80°C. Six-micrometers-thick cryo sections were prepared and put on glass slides, fixed in acetone, and were used in the ABC-ELITE (Vector Laboratories) immunohistochemical staining method. The sections were treated with 0.3% H2O2 to block endogenous peroxidase activity, were incubated with normal goat serum (dilution 1/10), and were blocked with avidin and biotin (Vector Laboratories). A mouse monoclonal antibody against human SOCS3, dilution 1/250 (obtained from Assay Designs), was added to incubate the specimens for 30 min. A biotinylated, rabbit anti-mouse secondary antibody (dilution 1/200 [Vector Laboratories]) was also added. The specimens were allowed to react with preformed avidin-biotin-enzyme complex (ABC-ELITE reagent [Vector Laboratories]) for 30 min, followed by incubation with 3-amino-9-ethylcarbazole substrate for 15 min to develop a staining pattern. Counterstaining was done with Mayer's hematoxylin. As a control, irrelevant mouse IgG1 was used. From the immunohistochemical staining, we found a higher expression of SOCS3 protein in dendritic cells and an increased number of SOCS3-positive cells in epidermis from lesional AD skin compared with nonlesional AD skin and normal skin from healthy controls (fig. 2). The altered SOCS3 protein expression in AD skin and the fact that SOCS3 maps to a region that has been reported to be linked to AD supported the idea that SOCS3 might be critically involved in AD and prompted us to investigate genetic association between SOCS3 and AD. We first analyzed a set of 406 Swedish families with at least one sibling affected with AD (1,334 total affected individuals). The families were recruited at the Dermatology Unit, Department of Medicine, Karolinska University Hospital Solna, Stockholm, during 1995–1997. The patients were all classified as having AD on the basis of clinical examination by dermatologists, who applied the U.K. Working Party's diagnostic criteria.6Williams HC Burney PG Hay RJ Archer CB Shipley MJ Hunter JJ Bingham EA Finlay AY Pembroke AC Graham-Brown RA Atherton DA Lewis-Jones MS Holden CA Harper JI Champion RH Poyner TF Launer J David TJ The U.K. Working Party's diagnostic criteria for atopic dermatitis. I. Derivation of a minimum set of discriminators for atopic dermatitis.Br J Dermatol. 1994; 131: 383-396Crossref PubMed Scopus (767) Google Scholar Mean age was 29 years. Allergen-specific serum IgE antibodies to inhalant and food allergens were measured in all siblings with AD.9Bradley M Kockum I Soderhall C Van Hage-Hamsten M Luthman H Nordenskjold M Wahlgren CF Characterization by phenotype of families with atopic dermatitis.Acta Derm Venereol. 2000; 80: 106-110PubMed Google Scholar Subjects were classified as having allergic asthma or allergic rhinoconjunctivitis on the basis of an interview. Of these 406 families, 109 were studied in the previously published genomewide linkage analysis.4Bradley M Soderhall C Luthman H Wahlgren CF Kockum I Nordenskjold M Susceptibility loci for atopic dermatitis on chromosomes 3, 13, 15, 17, and 18 in a Swedish population.Hum Mol Genet. 2002; 11: 1539-1548Crossref PubMed Scopus (88) Google Scholar Ten SNPs were selected from the NCBI dbSNP Home Page from ∼10 kb upstream of the SOCS3 gene and throughout the gene. SNPs that have been been validated by several methods according to dbSNP and that had a minor-allele frequency >5% were chosen with priority. The SNPs were rs8074003, rs12952093, rs9914196, rs4969170, rs12949584, rs11868378, C−920→A, rs1061489, rs2280148, and rs4969168. Genotyping was performed with TaqMan SNP Genotyping Assay (Applied Biosystems [ABI]). Primers and probes were obtained from ABI. One allelic TaqMan probe was labeled with fluorescent FAM dye, and the other with VIC dye. PCRs were run in 2× TaqMan Universal PCR Master Mix and No AmpErase UNG, with primer concentrations of 225 nM and probe concentration of 50 nM. The total reaction volume was 5 μl of 1-ng/μl genomic DNA. All SNPs were amplified using TaqMan Universal PCR protocol (95°C for 10 min, followed by 40 cycles at 92°C for 15 s, and 60°C for 1 min). The plates were then placed in the 7900HT Fast Real Time PCR System, where the fluorescence intensity in each well of the plate was read using the program SDS 2.2.1. Three of the genotyped SNPs were not found to be polymorphic when tested in 380 individuals (rs12949584, C-920→A, and rs1061489) and a fourth (rs11868378) failed to pass ABI's quality control. Of 10 tested SNPs, 6 were successfully genotyped in the 3.5-kb SOCS3 gene and in a region ∼10 kb upstream of the gene (fig. 3). Population Hardy-Weinberg equilibrium was evaluated for each SNP by use of a χ2 test as implemented in the zGenStat 1.128 software (Henric Zazzi), with a cutoff value of P>.01. Two of the SNPs were significantly associated with the AD phenotype (P<.03 for rs12952093; P<.02 for rs4969170) (table 2). When analyzing the linkage disequilibrium (LD) structure in the gene, using the program Unphased,10Dudbridge F Pedigree disequilibrium tests for multilocus haplotypes.Genet Epidemiol. 2003; 25: 115-121Crossref PubMed Scopus (1044) Google Scholar we found that all analyzed SNPs fell into two haplotype blocks. We then identified three haplotype-tagging SNPs as representative of these haplotype blocks: rs12952093 and rs4969170 in block 1 and rs4969168 in block 2. Two of these were in the promoter region of the gene (rs12952093 and rs4969170), and the third was in the 3′ UTR (rs4969168). One haplotype from the first block was positively associated with AD (A-A; P<.02), and one was negatively associated (C-G; P<.04) (table 2). In addition, the SOCS3 gene was linked to AD in these families (LOD=1.63; P<.04) in a nonparametric linkage analysis performed using the program Allegro (exponential model with power weighting of families).13Gudbjartsson DF Jonasson K Frigge ML Kong A Allegro, a new computer program for multipoint linkage analysis.Nat Genet. 2000; 25: 12-13Crossref PubMed Scopus (650) Google Scholar However, the linkage did not reach the previously reported suggestive linkage level (P 5%; n=4). NS = not significant.a Transmissions to affected individuals. Open table in a new tab Note.— The data shown are from 406 families, including 1,334 individuals with AD. Mean age was 29 years. Statistical analysis was performed using Unphased,10Dudbridge F Pedigree disequilibrium tests for multilocus haplotypes.Genet Epidemiol. 2003; 25: 115-121Crossref PubMed Scopus (1044) Google Scholar with the pdtphase command for multiplex families. Transmission frequencies were estimated with the program PDT11Martin ER Monks SA Warren LL Kaplan NL A test for linkage and association in general pedigrees: the pedigree disequilibrium test.Am J Hum Genet. 2000; 67: 146-154Abstract Full Text Full Text PDF PubMed Scopus (518) Google Scholar for alleles and with the program Transmit12Clayton D A generalization of the transmission/disequilibrium test for uncertain-haplotype transmission.Am J Hum Genet. 1999; 65: 1170-1177Abstract Full Text Full Text PDF PubMed Scopus (565) Google Scholar for haplotypes. P values were corrected with either the number of tested SNPs (n=3) or the number of tested haplotypes (haplotypes with frequency >5%; n=4). NS = not significant. To confirm the genetic association, we analyzed the three haplotype-tagging SNPs in two independent sets of patients and controls. The first was a nested case-control sample consisting of 555 children from the population-based birth cohort BAMSE, which comprises 4,089 children who were recruited in the Stockholm area between 1994 and 1996.14Bohme M Lannero E Wickman M Nordvall SL Wahlgren CF Atopic dermatitis and concomitant disease patterns in children up to two years of age.Acta Derm Venereol. 2002; 82: 98-103Crossref PubMed Scopus (79) Google Scholar, 15Melen E Gullsten H Zucchelli M Lindstedt A Nyberg F Wickman M Pershagen G Kere J Sex specific protective effects of interleukin-9 receptor haplotypes on childhood wheezing and sensitisation.J Med Genet. 2004; 41: e123Crossref PubMed Scopus (16) Google Scholar, 16Wickman M Kull I Pershagen G Nordvall SL The BAMSE project: presentation of a prospective longitudinal birth cohort study.Pediatr Allergy Immunol. 2002; 13: 11-13Crossref PubMed Scopus (197) Google Scholar When the children were 1, 2, and 4 years old, questionnaires that included questions on symptoms related to wheezing and allergic diseases were completed. The children with reported symptoms of dry skin in combination with itchy rash for at least 2 wk with typical localization and/or who received a diagnosis of eczema from a doctor were classified as having AD.14Bohme M Lannero E Wickman M Nordvall SL Wahlgren CF Atopic dermatitis and concomitant disease patterns in children up to two years of age.Acta Derm Venereol. 2002; 82: 98-103Crossref PubMed Scopus (79) Google Scholar Allergen-specific serum IgE antibodies to inhalant and food allergens were measured in all children. At 2 and/or 4 years of age, 328 children received the diagnosis of AD. Children with no reported eczema, asthma, or allergic rhinoconjunctivitis and with no allergen-specific IgE were used as healthy controls. Of the 555 total children, 328 were affected with AD and 227 were controls. Mean age was 4.3 years. The second was a case-control study comprising 187 adult patients with AD (102 females and 85 males) and 230 age- and sex-matched controls (162 females and 68 males) from the United Kingdom.17Veal CD Reynolds NJ Meggitt SJ Allen MH Lindgren CM Kere J Trembath RC Barker JN Absence of association between asthma and high serum immunoglobulin E associated GPRA haplotypes and adult atopic dermatitis.J Invest Dermatol. 2005; 125: 399-401PubMed Scopus (21) Google Scholar For all patients, AD was diagnosed by a dermatologist using the standard clinical criteria of Hanifin and Rajka.18Hanifin JM Atopic dermatitis.in: Moschella SL Hurley HJ Dermatology. 3rd edition. W. B. Saunders, Philadelphia1992Google Scholar The minimum age at study was 18 years, and the mean age was 45 years. Total serum IgE levels were elevated in 127 of the patients with AD, and 164 had other atopic manifestations. The controls did not have AD but were not checked for IgE levels or other atopic manifestations. When analyzing the Swedish case-control samples, we found a highly significant genetic association between AD and both the positively associated haplotype (P<.03) and the negatively associated haplotype (P<.02) (table 3). However, we could not confirm these associations in the U.K. study (table 3). The power to detect association for the two associated haplotypes in the U.K. study was 83% for the negatively associated haplotype (P<.05; odds ratio [OR] 0.62) and 80% for the positively associated haplotype (P<.05; OR 1.72), with the assumption of the same haplotype frequencies as observed in the Swedish case-control study. Power analysis was performed using the PS software,19Dupont WD Plummer WD PS power and sample size program available for free on the Internet.Controlled Clin Trials. 1997; 18: 274Abstract Full Text PDF Google Scholar version 2.1.31.Table 3Genetic Association of SOCS3 with AD in Two Independent Case-Control StudiesSwedish Nested Case-Control StudyU.K. Case-Control StudyAllele or HaplotypeFrequency (%) inFrequency (%) inrs12952093rs4969170rs4969168ControlsPatientsOR (95% CI)PControlsPatientsOR (95% CI)PA26331.45 (1.10–1.92).032726.95 (.69–1.31)NSA31411.51 (1.16–1.97).0093431.87 (.64–1.18)NSG8584.90 (.62–1.31)NS85901.53 (.99–2.36)NSCG6858.67 (.51–.87).0266681.12 (.82–1.53)NSAA33241.50 (1.12–2.00).032625.94 (.64–1.38)NSCGG6958.63 (.48–.82).00466691.15 (.85–1.56)NSAAG15231.72 (1.22–2.43).0119201.07 (.70–1.63)NSNote.—Shown are the association results from the Swedish nested case-control study (328 individuals with AD; 227 controls; mean age 4.3 years) and from the U.K. case-control study (187 individuals with AD; 230 controls; mean age 45 years). Statistical analysis was performed using Unphased,10Dudbridge F Pedigree disequilibrium tests for multilocus haplotypes.Genet Epidemiol. 2003; 25: 115-121Crossref PubMed Scopus (1044) Google Scholar with the cocaphase command for case-control studies. The significance of associations of individual haplotypes in the case-control studies was estimated using standard χ2 tests. P values have been corrected with either the number of tested SNPs (n=3) or the number of tested haplotypes (haplotypes with frequency >5%; n=4). NS = not significant. Open table in a new tab Note.— Shown are the association results from the Swedish nested case-control study (328 individuals with AD; 227 controls; mean age 4.3 years) and from the U.K. case-control study (187 individuals with AD; 230 controls; mean age 45 years). Statistical analysis was performed using Unphased,10Dudbridge F Pedigree disequilibrium tests for multilocus haplotypes.Genet Epidemiol. 2003; 25: 115-121Crossref PubMed Scopus (1044) Google Scholar with the cocaphase command for case-control studies. The significance of associations of individual haplotypes in the case-control studies was estimated using standard χ2 tests. P values have been corrected with either the number of tested SNPs (n=3) or the number of tested haplotypes (haplotypes with frequency >5%; n=4). NS = not significant. The lack of association in the U.K. study could be because of differences between the two populations in the relative contribution of SOCS3 as a genetic risk factor. Another potential explanation for the discrepancy is that the importance of SOCS3 as a risk factor might be age dependent, since the most significant association of SOCS3 was seen in children with AD. It is also possible that the selection of controls in the U.K. study influenced the results. Controls in the U.K. study were not checked for other atopic manifestations as they were in the Swedish case-control study, which is a limitation because the potential role of SOCS3 in other atopic manifestations is unknown.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar However, in a meta-analysis20Woolf B On estimating the relation between blood group and disease.Ann Hum Genet. 1955; 19: 251-253Crossref PubMed Scopus (2238) Google Scholar of the two case-control sets, the risk haplotype was significantly associated (P<.02; combined OR 1.37; 95% CI 1.07–1.77) and a test of heterogeneity between the studies was not significant. The meta-analysis was performed in accordance with the method of Woolf.20Woolf B On estimating the relation between blood group and disease.Ann Hum Genet. 1955; 19: 251-253Crossref PubMed Scopus (2238) Google Scholar The consistently elevated expression of SOCS3 in skin of patients with AD, together with evidence for genetic association in two independent patient data sets, makes SOCS3 a strong candidate gene for AD susceptibility. This evidence is reinforced by the potential pathophysiological link between the activity of SOCS3 and the pathogenesis of AD. SOCS3 is an SH2-containing protein that binds to the activation loop of Janus kinases, inhibiting kinase activity and thereby suppressing cytokine signaling. The promoter of the SOCS3 gene contains two STAT-responsive elements and one GC-rich element.21Ehlting C Haussinger D Bode JG Sp3 is involved in the regulation of SOCS3 gene expression.Biochem J. 2005; 387: 737-745Crossref PubMed Scopus (38) Google Scholar SOCS3 is expressed predominantly in T-helper type 2 (TH2) cells and has an important role in regulating the onset and maintenance of TH2-mediated responses.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar Patients with AD have increased numbers of TH2-expressing cells in peripheral blood and elevated interleukin-13 expression in the skin, which indicates a systemic TH2 environment.22Leung DY Molecular basis of allergic diseases.Mol Genet Metab. 1998; 63: 157-167Crossref PubMed Scopus (79) Google ScholarSocs3 transgenic mice showed increased TH2 responses.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar These mice, constitutively expressing the Socs3 protein, had about two-to-three times higher expression of Myc-tagged Socs3 than of endogenous Socs3 in TH2 cells. However, the presence or absence of AD-like features in the Socs3 transgenic mice and the expression of Socs3 protein in cell types other than T cells were not reported. In contrast, dominant-negative mutant Socs3 transgenic mice and mice with a heterozygous deletion of Socs3 had decreased TH2 development.8Seki Y Inoue H Nagata N Hayashi K Fukuyama S Matsumoto K Komine O Hamano S Himeno K Inagaki-Ohara K Cacalano N O'Garra A Oshida T Saito H Johnston JA Yoshimura A Kubo M SOCS-3 regulates onset and maintenance of TH2-mediated allergic responses.Nat Med. 2003; 9: 1047-1054Crossref PubMed Scopus (312) Google Scholar Despite the connection between SOCS3 and AD, it is still unclear how the polymorphisms in the SOCS3 gene might affect development of AD, but there may be a direct effect on SOCS3 transcription, since two of the associated SNPs are located in predicted transcription-factor binding sites (RAVEN). In conclusion, we found that SOCS3 is consistently more highly expressed in skin from patients with AD than in normal control skin and that specific haplotypes of the SOCS3 gene are significantly associated with the disease. We suggest that SOCS3 may play an important role in AD pathogenesis, at least for a Swedish population with AD. We also suggest that SOCS3 and the pathways that it regulates may be important targets for future investigations of the molecular pathogenesis of AD and for therapeutic developments. We gratefully acknowledge financial assistance from the Swedish Asthma and Allergy Association's Research Foundation, the Edward Welander-Finsen Foundation, the Swedish Society for Medical Research, the Juvenile Diabetes Foundation International, and the U.S. National Eczema Association for Science and Education.