1. ¹Department of Dermatology and Allergy Biederstein, Technical University Munich (TUM), Munich, Germany
2. ²Division of Environmental Dermatology and Allergy, GSF National Research Center for Environment and Health/Technical University Munich, Munich, Germany
3. ³Institute for Medical Statistics and Epidemiology IMSE, Technical University Munich (TUM), Munich, Germany
4. ⁴Department of Epidemiology, GSF-National Research Center for Environment and Health, Neuherberg, Germany
5. ⁵Department of Dermatology, University of Bonn, Bonn, Germany

Correspondence: Stephan Weidinger, E-mail: weidinger@lrz.tum.de

TO THE EDITOR

Atopic dermatitis (AD) arises from the interaction between strong genetic and environmental factors (Morar et al., 2006). As it commonly occurs in concert with asthma and/or hay fever, and the majority of patients exhibits elevated serum IgE levels and/or sensitizations against aero- and food allergens, most pathophysiological concepts are dominated by atopic mechanisms and regard IgE-mediated sensitization as central cornerstone in the development of AD (Akdis et al., 2006). However, there is accumulating evidence from epidemiological studies, suggesting that sensitization is not essential for the development of eczema and that atopy may be secondary rather than causative (Williams and Flohr, 2006). In addition, genome-wide screens and candidate gene studies have shown only partial overlap of AD with other atopic phenotypes, but a closer relationship with other epithelial diseases such as psoriasis (Cookson, 2004). Thus, attention has been drawn to the impaired epidermal barrier function, which is a hallmark feature of AD. In this context, the epidermal differentiation complex on chromosome 1q21, for which linkage to both AD and psoriasis has been reported, is of particular interest, as it harbors a number of genes and gene families expressed in the terminally differentiating epithelium (Cookson et al., 2001). Recently, it was demonstrated that two functional mutations within the epidermal differentiation complex gene filaggrin (FLG), underlie ichthyosis vulgaris (Smith et al., 2006) and are strong risk factors for AD and asthma in the context of AD in European populations (Palmer et al., 2006). In a subsequent family-based association study, we could show that these mutations particularly predispose to the extrinsic subtype of AD, but not intrinsic AD, suggesting that the skin barrier defect in AD is a key early event that precedes the development of allergic sensitization and respiratory atopy (Weidinger et al., 2006). To replicate these associations, we investigated an independent collection of 274 unrelated adult AD cases (average age 35.910.8 years; 61.2% female subjects and 38.8% male subjects) and 252 population-based controls (average age 39.416.1 years; 61.5% female subjects and 38.5% male subjects). All patients were recruited in the dermatologic outpatient departments of the University of Bonn and the Technical University Munich, Germany. AD was diagnosed on the basis of a skin examination by experienced dermatologists using the UK diagnostic criteria for AD. The severity of eczema was assessed using the SCORAD (scoring atopic dermatitis index). Subjects were classified as having asthma or allergic rhinitis when they reported a physician's diagnosis of asthma or rhinoconjunctivitis. Hypernormal control individuals without AD or allergic diseases and IgE levels <100 kU/l were selected from the KORA S4 (Cooperative Health Research in the Region of Augsburg Survey 4: 1999–2001) population-based cross-sectional study, which was carried out 1999–2001 in the city and region of Augsburg, South Germany (Weidinger et al., 2005). Total as well as specific IgE antibodies against common environmental allergens were measured with the help of an enzyme-immuno assay (CAP-FEIA, Pharmacia, Uppsala, Sweden). Specific sensitization was defined to be present if at least one of the specific IgE antibodies was positive (CAP-RAST class 1, corresponding to 0.35 kU/l). A raised total serum IgE was taken to be greater than 100 kU/l. Extrinsic AD was defined as AD with concomitant sensitization and/or IgE levels 100 kU/l. Severity of AD was regarded as binary trait using the median value of the objective SCORAD components (31 points) as cutoff.

All study methods were approved by the ethics committee of the "Bayerische Landesärtzekammer" Munich and the ethics committee of the University of Bonn and a written and informed consent that complies with all the Declaration of Helsinki Principles was obtained from the participants before the beginning of the study.

Out of the 274 AD cases, 71.7% were extrinsic, 62.6% had allergic rhinitis, 34.8% suffered from asthma, and 41.6% showed an early onset of the disease (<2 years of age). The mean total IgE of AD cases was 1,024 kU/l.

Genotyping for R501X and 2282del4 was performed using the MassARRAY system (Sequenom, San Diego, CA) as described previously (Weidinger et al., 2005). To test for associations with binary traits, logistic regression models were applied with age, sex, and genotype as independent variables. In addition, associations with IgE levels were evaluated using a linear regression model on log-transformed IgE values with the same covariates. For either regression analysis backward variable selection was applied. Results are given for the estimates of genotype effects after variable selection. The P-values are two-sided and subject to a significance level of 0.05. The issue of multiple testing does not affect the overall results and interpretation. Therefore P-values were not adjusted.

In our cohort, we did not observe any homozygotes for either mutation, but seven compound heterozygotes. The FLG variants were greatly overrepresented in AD cases with a combined carrier frequency of 57 (21.1%) vs 15 (6.0%) in controls (odds ratio 3.53, 95% confidence interval 1.92–6.48, P=4.9 10^-5, Tables 1 and 2). The lower than expected carrier frequency for heterozygote FLG status in controls (6 vs 10% for other comparable populations) is probably due to the fact that our control population had been screened so as to not have AD, allergic diseases or raised IgE levels and therefore represents a hypernormal type of control population.

Table 1 - Frequency of FLG null alleles in cases (AD) and controls.

Full table

Table 2 - Results of the logistic regression models adjusted for age and sex using backward variable selection.

Full table

The combined carrier frequency among patients with extrinsic AD was 24.6%. The highest carrier frequency was seen in patients with early onset of disease (33 out of 111, 29.7%). In addition, logistic regression analysis indicated that the 2282del4 mutation was associated with a more severe phenotype (odds ratio 3.16, 95% confidence interval 1.49–6.72, P=0.0027). The lack of association with severity for R501X may be due to the age at which SCORAD was determined and due to the fact that the SCORAD is a "snap shot" variable that does not reflect the overall disease activity. Both FLG variants showed significant associations with total serum IgE levels with geometric means for serum IgE (kU/l) 135.6 in wild type/wild type versus 317.4 in carriers of at least one mutation (P=0.010). This association is confirmed by linear regression using log-transformed values of observed IgE levels as quantitative dependent variable (P=0.011). In addition, both asthma and rhinitis in the context of AD were significantly associated (P=2.5 10^-6 and P=2.4 10^-5 for combined genotype). Consistent with previous findings (Weidinger et al., 2006), no association was seen with the intrinsic subtype of AD. Results of association analysis are summarized in Table 2.

Our present findings confirm that FLG mutations are strong risk factors for AD, in particular for the extrinsic subtype of AD, and with concomitant respiratory allergic diseases. Our results further indicate that these variations predispose to an early onset and persistent course of the disease. It seems conceivable that a genetically determined primary epidermal barrier disruption is the key-event in AD that predisposes to secondary allergic manifestations by causing a reduced resistance to environmental agents such as allergens, irritants, and microbes. It will be highly interesting to further evaluate the epidermal differentiation complex for additional candidate genes modulating the skin barrier function.