Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

Interplay of Filaggrin Loss-of-Function Variants, Allergic Sensitization, and Eczema in a Longitudinal Study Covering Infancy to 18 Years of Age

  • Ali H. Ziyab,

    Affiliations: Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, South Carolina, United States of America, Department of Community Medicine and Behavioral Sciences, Faculty of Medicine, Kuwait University, Kuwait City, Kuwait

  • Wilfried Karmaus ,

    karmaus@sc.edu

    Affiliation: Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, South Carolina, United States of America

  • Mitra Yousefi,

    Affiliation: Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, South Carolina, United States of America

  • Susan Ewart,

    Affiliation: College of Veterinary Medicine, Michigan State University, East Lansing, Michigan, United States of America

  • Eric Schauberger,

    Affiliations: Genetics Graduate Program, Michigan State University, East Lansing, Michigan, United States of America, College of Osteopathic Medicine, Michigan State University, East Lansing, Michigan, United States of America

  • John W. Holloway,

    Affiliations: Academic Units of Clinical and Experimental Medicine, Faculty of Medicine, University of Southampton, Southampton, United Kingdom, Academic Units of Human Genetics, Faculty of Medicine, University of Southampton, Southampton, United Kingdom

  • Hongmei Zhang,

    Affiliation: Department of Epidemiology and Biostatistics, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, South Carolina, United States of America

  • Syed Hasan Arshad

    Affiliations: Academic Units of Clinical and Experimental Medicine, Faculty of Medicine, University of Southampton, Southampton, United Kingdom, David Hide Asthma and Allergy Research Centre, Isle of Wight, United Kingdom

Interplay of Filaggrin Loss-of-Function Variants, Allergic Sensitization, and Eczema in a Longitudinal Study Covering Infancy to 18 Years of Age

  • Ali H. Ziyab, 
  • Wilfried Karmaus, 
  • Mitra Yousefi, 
  • Susan Ewart, 
  • Eric Schauberger, 
  • John W. Holloway, 
  • Hongmei Zhang, 
  • Syed Hasan Arshad
PLOS
x
  • Published: March 5, 2012
  • DOI: 10.1371/journal.pone.0032721

Abstract

Background

Immune specific genes as well as genes regulating the formation of skin barrier are major determinants for eczema manifestation. There is a debate as to whether allergic sensitization and filaggrin gene (FLG) variants lead to eczema or FLG variants and eczema increase the risk of allergic sensitization. To investigate the time-order between eczema and allergic sensitization with respect to FLG variants, data from a large prospective study covering infancy to late adolescence were analyzed.

Methodology/Principal Findings

Repeated measurements of eczema and allergic sensitization (documented by skin prick tests) at ages 1, 2, 4, 10, and 18 years were ascertained in the Isle of Wight birth cohort (n = 1,456). Three transition periods were analyzed: age 1-or-2 to 4, 4 to 10, and 10 to 18 years. FLG variants were genotyped in 1,150 participants. Over the three transition periods, in temporal sequence analyses of initially eczema-free participants, the combined effect of FLG variants and allergic sensitization showed a 2.92-fold (95% CI: 1.47–5.77) increased risk ratio (RR) of eczema in subsequent examinations. This overall risk was more pronounced at a younger age (transition period 1-or-2 to 4, RR = 6.47, 95% CI: 1.96–21.33). In contrast, FLG variants in combination with eczema showed a weaker, but significant, risk ratio for subsequent allergic sensitization only up to 10 years of age.

Conclusions/Significance

Taking the time order into account, this prospective study demonstrates for the first time, that a combination of FLG variants and allergic sensitization increased the risk of eczema in subsequent years. Also FLG variants interacted with eczema and increased the risk of subsequent allergic sensitization, which, was limited to the younger age. Hence, early restoration of defective skin barrier could prevent allergic sensitization and subsequently reduce the risk of eczema development.

Introduction

The prevalence of common allergic disorders including eczema has increased in recent decades among children in westernized societies [1]. Eczema is a common chronic inflammatory skin disorder characterized by relapsing-remitting combination of skin dryness, pruritus, and rash starting in childhood and often persisting into adulthood [1], [2]. Allergic sensitization, characterized by increased immunoglobulin-E (IgE) antibodies in response to specific allergens, has been linked to the increased risk of allergic disorders and is believed to account for 30% to 40% of cases of eczema, asthma, and allergic rhinitis [3], [4]. In past decades, most of the clinical etiology of eczema was attributed to genetic predisposition to sensitization, thus relating eczema to immune specific genes. However, the recently recognized role of a dysfunctional skin barrier in the pathogenesis of eczema suggests that a functional skin barrier could prevent the development of allergic sensitization and consequently eczema [5].

One important gene involved in the formation of a normal skin barrier is filaggrin (FLG). This gene encodes filament-aggregating protein, which serves as a primary shield against increased transepidermal water loss and entry of microbes, allergens, and irritants [5], [6]. The FLG gene is situated within the epidermal differentiation complex on chromosome 1q21. Null (Loss-of-function) variants in the FLG gene have been linked to disrupted skin barrier and are strong predisposing factors in the development of eczema [7]. In European populations, R501X and 2282del4 are the most prevalent FLG variants; their association with eczema has been replicated in several investigations, summarized in meta-analysis [8]. Three other less common FLG variants (3702delG, R2447X, S3247X) have also been associated with eczema [9], [10], [11]. It is estimated that about 10% of European populations carry at least one of the FLG variants [12]. Several studies have shown that FLG variants are major risk factors for eczema in general as well as for the subgroups of atopic and non-atopic eczema [13], [14], [15]. In addition, a meta-analysis study summarized findings of several investigations that FLG variants also increase the risk for allergic sensitization [16].

A complex interplay between genetic, environmental, and immunological factors is considered to contribute to the pathogenesis of eczema [17]. However, the time-order of the associations between allergic sensitization and eczema is highly disputed in the literature [18] without consensus on which comes first [19]. To our knowledge, analyses of longitudinal studies investigating the temporal sequence of the occurrence of allergic sensitization and eczema while accounting for the effect of FLG variants are lacking. One position is that allergic sensitization interacts with FLG variants leading to increased risk of eczema. The other position is that eczema interacts with FLG variants leading to higher risk of allergic sensitization [19]. This dispute is more than an academic discourse, since prophylaxis can either focus on avoiding allergic sensitization or preventing eczema. To solve this dispute, longitudinal studies are needed that can address the critical time order. Therefore, we assessed the interplay of FLG variants, allergic sensitization, and eczema in a longitudinal study.

The determination of FLG variants and repeated assessments of eczema and allergic sensitization at ages 1, 2, 4, 10 and 18 years in the Isle of Wight birth cohort facilitate the investigation of the combined effects. First, we investigated the individual and combined effects of allergic sensitization and FLG variants on the occurrence of eczema in a concurrent model (all manifestations were detected at the same age). Second, to elucidate the temporal sequence of events we tested two hypotheses: (i) the combined effect of eczema and FLG variants increases the risk of subsequent allergic sensitization and (ii) the combined effect of allergic sensitization and FLG variants increases the risk of subsequent eczema.

Methods

The Isle of Wight birth cohort

A whole population birth cohort was established on the Isle of Wight, UK, in 1989 to prospectively study the natural history of allergies from birth to 18 years of age. The island is close to the British mainland, semi-rural, without heavy industry. Both the Isle of Wight and the study populations are 99% Caucasian. Ethics approvals were obtained from the Isle of Wight Local Research Ethics Committee (now named the National Research Ethics Service, NRES Committee South Central – Southampton B) at recruitment and for the 1, 2, 4, 10 and 18 years follow-up (06/Q1701/34). Of the 1,536 children born between January 1, 1989, and February 28, 1990, written informed consent was obtained from parents to enrol 1,456 newborns. Children were followed up at the ages of 1 (n = 1,167), 2 (n = 1,174), 4 (n = 1,218), 10 (n = 1,373), and 18 years (n = 1,313). Detailed questionnaires were completed for each child at each follow-up. When a visit was not possible, a telephone questionnaire was completed or a postal questionnaire sent for completion and return.

Phenotypes

In all assessments of the Isle of Wight birth cohort, eczema was defined as chronic or chronically relapsing, itchy dermatitis lasting more than 6 weeks with characteristic morphology and distribution [20], following Hanifin and Rajka criteria [21]. Since the 1-year and 2-year follow-up data on eczema were collected in a relatively small time window, we combined them for analytic purposes (reported as 1-or-2 years).

To determine allergic sensitization status, skin prick testing (SPT) at ages 1 and 2 years was performed on children with any symptoms of eczema, asthma, or rhinitis. We combined SPT results for ages 1 and 2 years, since they occurred within a short time period and will henceforth refer to this as SPT at 1-or-2 years. At 4, 10 and 18 years, regardless of symptoms, SPT was performed on most children attending the research center to a standard battery of common allergens (ALK-Albello, Horsholm, Denmark). Inhalant allergens tested were house dust mite, cat, dog, Alternaria alternata, Cladosporium herbarium, grass pollen mix, and tree pollen mix. Food allergens tested were cows' milk, soya, hens' egg, peanut and cod. Positive and negative controls were included. Allergic sensitisation was defined by having a SPT to at least one allergen test with mean wheal diameter of 3 mm greater than the negative control. Since allergic sensitization is a dynamic rather than a completely stable phenotype, we used the concurrent status and thus allowed the risk to change over time. In this manuscript we use the terms allergic sensitization and positive SPT interchangeably.

FLG genotyping

We approached the participants in 1999 and ask them to provide written informed consent to genetic analyses, when blood samples were collected in the vast majority of children (n = 907). In 2006/2007, DNA samples were collected from an additional 304 children who did not provide the samples in 1999. DNA was extracted from blood or saliva samples from cohort subjects (n = 1,211). Five variants in the FLG gene that result in loss of function and are reported to be common in populations of European ancestry were selected for genotyping. DNA samples were interrogated using GoldenGate Genotyping Assays (Illumina, Inc, SanDiego, CA) on the BeadXpressVeracode platform (Illumina, Inc, SanDiego, CA) per Illumina's protocol. In brief, samples were fragmented and hybridized to the pool of allele-specific primer sets. Following an extension/ligation reaction the samples were then hybridized to the Veracode bead pool and processed on the BeadXpress reader. Data were analyzed using the genotyping module of the GenomeStudio Software package (Illumina, Inc, SanDiego, CA). DNA from each subject plus 37 replicate samples were analyzed for a total of 1,248 samples. The quality threshold for allele determination was set at a GenCall score >0.25 (scores ≤0.25 were “no calls”) with n = 1,227 samples (98.3%) retained for further analysis. Analysis of each locus included reclustering of genotyping data using our project data to define genotype cluster positions with additional manual reclustering to maximize both cluster separation and the 50th percentile of the distribution of the GenCall scores across all genotypes (50% GC score). Children were classified as having FLG loss-of-function defect if they carry the minor allele for at least one of the following FLG null variants: R501X, 2282del, or S3247X.

Statistical analysis

The study represents a dynamic cohort; some children did not participate in one assessment but re-joined the next. For each follow-up, the 12-month (24-month for 1-or-2 years) period prevalence of eczema was stratified by allergic sensitization and FLG variants. Hence, the number of children with eczema was divided by the number of children in the respective stratum.

Since period prevalence and transitions of eczema and allergic sensitizations are not rare events (>10%), odds ratios for explanatory variables are likely to overestimate risk ratios [22]. To estimate risk ratios, we applied log-binomial regression using log link function describing the distribution of a binary variable (i.e. dichotomous outcomes). In this type of regression, the log-linear function describes the association between the mean number of “successes” in the binomial distribution on a log-scale and the explanatory variables. Thus, this approach enabled us to directly estimate risk ratios, which is the exponential of the estimated coefficients in the log-linear function. To assess long-term development in individual children, we had to consider that repeated measurements within one child represent correlated observations. The within-child effect was taken into consideration by employing a covariance matrix solved through an iterative estimating process based on a working correlation matrix. This is the idea of the generalized estimating equation (GEE) approach for estimating the parameters [23]. The appropriate working correlation matrix was determined by using the Akaike information criterion. Applying GEE analysis using an autoregressive working correlation matrix to take the within-child effect into account, we estimated marginal probabilities for the risk factors and their interaction using the GENMOD procedure in SAS 9.2 (SAS, Gary, NC, USA). For all GEE models we included age at follow-up as categorical variables (1-or-2, 4, 10, and 18 years) and gender as potential confounders. In this manuscript we use the terms repeated measurement analysis and longitudinal analysis interchangeably. A p-value<0.05 was taken to indicate statistical significance.

First for descriptive purposes, we analyzed concurrent associations between eczema and FLG variants, eczema and SPT, and FLG variants and SPT. Focusing on the overall cohort we calculated risk ratios (RR) using log-binomial regression. Within this approach, we then restricted the cohort to children with a negative SPT and estimated the association between eczema and FLG variants to determine the effect of FLG variants in the absence of allergic sensitization [24]. In contrast, to determine the effect of allergic sensitization in the absence of FLG variants, we focused on the cohort of children without FLG variants and evaluated that association between eczema and SPT. Then, to determine the combined effect of both risk factors, we compared the risk of eczema in children who have both risk factors (FLG variants and positive SPT) to the risk of children who have neither (no FLG variants and negative SPT). We also applied an alternate approach using the whole sample and treating FLG variants and allergic sensitization as main effects and estimated the combined effect through log-binomial model {RR = exp(β1×SPT+β2×FLG variants+β3×[SPT×FLG variants])} controlling for age and gender of the child. The interaction term (β3×[SPT×FLG variants]) was used to estimate the additional effect of two co-occurring risk factors on the health outcome above and beyond their individual effects. The term “combined effect” was used to describe the joint impact of two individual risk factors plus their interaction on the occurrence of the outcome. To estimate concurrent age-specific effects, we analyzed each examination separately. To evaluate whether there is an overall concurrent effect covering all ages, we used repeated measurements analyses.

To address temporality and to decipher whether allergic sensitization leads to eczema or vice versa, while accounting for the FLG variants effect, we structured the analysis into three transition periods (Figure 1). We have three transitions (1-or-2 to 4, 4 to 10, and 10 to 18 years) between four follow-ups (ages 1-or-2, 4, 10, and 18 years). For each transition period, at baseline children were without the outcome (allergic sensitization or eczema, respectively) but had either allergic sensitization or eczema (alternate risk factors). The value of the risk factors could change for each transition (time-dependent covariate), since eczema and SPT were conducted at each examination. This setting provides three repeated transition periods. In each we assessed the interaction between the risk factors (time-dependent) and FLG variants (time-independent) using log-binomial models. To address the overall effect of all three transition periods, we applied repeated measurements analyses. Hence, the alternate models were: 1) eczema model: FLG variants × preceding allergic sensitization → eczema; and 2) allergic sensitization model: FLG variants × preceding eczema → allergic sensitization.

thumbnail
Figure 1. Graphical presentation of temporal (time-order) analyses.

“+ SPT” refers to positive skin prick test. Between four follow-ups (ages 1-or-2, 4, 10, and 18 years) three transition periods (1-or-2 to 4, 4 to 10, and 10 to 18) were analyzed. The combined effect of eczema and FLG variants on the development of subsequent allergic sensitization (defined by positive SPT) was determined for the three transition periods separately and in repeated measurement analysis. Similarly, the combined effect of allergic sensitization and FLG variants on the development of subsequent eczema was determined. Dashed arrows refer to the combined effect of eczema and FLG variants on the development of + SPT. Solid arrows represent the combined effect of + SPT and FLG variants on the development of eczema.

doi:10.1371/journal.pone.0032721.g001

Results

A total of 1,377, 1,214, 1,359, and 1,307 children had available information on eczema diagnosis at ages 1-or-2, 4, 10, and 18 years, respectively. Skin prick tests (SPTs) were performed at ages 1-or-2 (n = 515, symptomatic children), 4 (n = 982), 10 (n = 1,036), and 18 (n = 853) years. Genotype information on the FLG variants was available for 1,150 children. No significant group differences were detected between the total cohort and the genotyped children (Table 1).

thumbnail
Table 1. Characteristics of total cohort and subpopulation of children with genotype information for FLG variants.

doi:10.1371/journal.pone.0032721.t001

FLG variants

Participants were genotyped for FLG variants R501X, 2282del4, S3247X, 3702delG, and R2447X. The polymorphic FLG variants were in Hardy-Weinberg equilibrium. Variants 3702delG and R2447X were not informative in our population due to minor allele frequencies <0.1% (Table S1). Therefore, FLG loss-of-function status was determined using R501X, 2282del4, and S3247X variants. Minor allele frequencies of R501X, 2282del4 and S3247X were 2.0%, 2.3% and 0.8%, respectively. The proportion of carriers (heterozygous) of FLG variants R501X, 2282del4 and S3247X were 4.1%, 4.6% and 1.6%, respectively, resulting in a combined carrier frequency of 10.3%. No individuals were homozygous for the minor allele of any FLG variants. One person had a compound heterozygous genotype for R501X and 2282del4, and this person was analyzed with the heterozygotes.

Effect of FLG variants and allergic sensitization on eczema

The combined genotype of FLG variants was strongly associated with eczema at all time points except at 10 years of age (Table 2). This finding is further supported by repeated measurement analysis (age 1-or-2 to 18 years), which showed an overall effect of FLG variants on eczema ([Risk Ratio] RR = 1.56; 95% CI: 1.17–2.08). Likewise, allergic sensitization was associated with eczema at 1-or-2, 4, 10, and 18 years of age, with the strongest association observed at age 4 years (RR = 3.05; 95% CI: 2.21–4.2; Table 2).

thumbnail
Table 2. Risk ratios of FLG variants and allergic sensitization on the occurrence of concurrent eczema in the course of childhood and adolescence.

doi:10.1371/journal.pone.0032721.t002

Effect of FLG variants on allergic sensitization

Association analyses between FLG variants and allergic sensitization were conducted at 1-or-2, 4, 10, and 18 years of age, and to determine the overall association a repeated measurement approach was applied. FLG variants increased the risk of allergic sensitization at age 1-or-2 years (RR = 1.57; 95% CI: 1.02–2.4; Table S2) and at age 10 years (RR = 1.53; 95% CI: 1.18–1.99; Table S2). At other ages (4 and 18 years) and in the longitudinal analysis the associations were weaker.

Single and combined associations of FLG variants and allergic sensitization with eczema

To examine the individual and combined effect of FLG variants and allergic sensitization on the concurrent risk of eczema, we used stratified analyses. Associations between FLG variants and eczema were conducted among a sub-group of children who tested negative in SPT assessments. This stratification allowed us to determine the effect of FLG variants on the development of eczema independent of allergic sensitization. FLG variants did not increase the risk of eczema in the sub-group of non-allergic children at any age or in the longitudinal analysis (RR = 1.12; 95% CI: 0.77–1.62; Table 3). Next, to determine the effect of allergic sensitization in the absence of FLG variants, we evaluated the association between positive SPT and concurrent eczema among children who do not carry FLG variants. At 1-or-2, 4, and 10 years (not at 18 years) of age a positive SPT increased risk of eczema (longitudinal analysis, RR = 1.48; 95% CI: 1.22–1.8) among children without FLG variants. Then we assessed the combined effect of FLG variants and allergic sensitization by comparing the risk of eczema in children with both FLG variants and positive SPT to the risk of eczema among children without FLG variants and who had a negative SPT. The combined effect of FLG variants and allergic sensitization was significant at all ages (repeated measurement analysis: RR = 2.82; 95% CI: 2.15–3.69) with a maximum effect at 4 years of age (RR = 8.17; 95% CI: 5.55–12.02; Table 3). Finally, alternative to stratification, we applied a log-binomial regression model to estimate the interaction between FLG variants and allergic sensitization using repeated measurements. This approach, controlling for the main effect of FLG variants and SPT, showed that FLG variants and positive SPT have an increased combined effect on the concurrent risk of eczema (RR = 2.51; 95% CI: 1.86–3.38; Table 3). Thus, stratification and modelling approaches yielded comparable results.

thumbnail
Table 3. Risk ratios of FLG variants, allergic sensitization, and their combined effect on the occurrence of concurrent eczema at different ages.

doi:10.1371/journal.pone.0032721.t003

Association analyses accounting for temporal sequence of events

To address temporality of allergic sensitization and eczema we analyzed three transition periods (1-or-2 to 4, 4 to 10, and 10 to 18 years) between four follow-ups (ages 1-or-2, 4, 10, and 18). At the baseline of the transition period, children were without the outcome but had developed eczema in the next exam: 6.2%, 9.4%, and 6.7%, respectively, between 1-or-2 and 4 years, 4 and 10 years, and 10 and 18 years, or had developed allergic sensitization: 14.2%, 14.8%, and 23.1%, respectively (Table 4). The combined effect of allergic sensitization in the preceding exam and FLG variants on the development of eczema (eczema model) showed a 2.92-fold increased risk (95% CI: 1.47–5.77) in the longitudinal analysis. Alternatively, we tested whether the combined effect of eczema in the preceding exam and FLG variants affected the development of positive SPT (allergic sensitization model), but did not find a significant association in the repeated measurements analysis.

thumbnail
Table 4. Risk ratios of FLG variants interacting with allergic sensitization and eczema on the subsequent occurrence of eczema (Eczema model) and allergic sensitization (Allergic sensitization model).

doi:10.1371/journal.pone.0032721.t004

To decipher whether time (transition-period) specific effects are present in the analysis of the temporal sequence, we determined the risk associated with the combined effect for each transition separately. The combined effect of FLG variants and allergic sensitization statistically significantly increased the risk of subsequent eczema for the transition periods 1-or-2 to 4 years (RR = 6.47; 95% CI: 1.96–21.33, p-value = 0.002; Figure 2A) and 10 to 18 years (RR = 2.68; 95% CI: 1.04–6.9, p-value = 0.041), which was further supported by repeated measurements analysis that suggested the overall effect of all transition periods is statistically significant (Table 4). In contrast, the combined effect of FLG variants and eczema statistically significantly increased the risk of allergic sensitization only for the transition periods 1-or-2 to 4 years (RR = 3.5; 95% CI: 1.46–8.39, p-value = 0.005) and 4 to 10 years (RR = 2.88; 95% CI: 1.05–7.95, p-value = 0.041; Figure 2B). The overall effect of the three transitions periods for this model (allergic sensitization model) was not statistically significant (Table 4).

thumbnail
Figure 2. Overall and time-specific risk ratios in the temporal (time-order) analyses.

(A) Risks ratios associated with the combined effect of FLG variants and allergic sensitization on the development of subsequent eczema for three transition periods and repeated measurements. (B) Risk ratios associated with the combined effect of FLG variants and eczema on the development of subsequent allergic sensitization for three transition periods and repeated measurements. “RM” refers to repeated measurements.

doi:10.1371/journal.pone.0032721.g002

Discussion

Using data of a large follow-up study from infancy to age 18 years, our study explores two assertions: first, the combined occurrence of eczema and FLG variants leads to allergic sensitization, second, the combined effect of allergic sensitization and FLG variants leads to eczema. We applied two explanatory models, a concurrent and a delayed effect (time order) model (Figure 1), and investigated both the overall effect (repeated measurements) of eczema and allergic sensitization and period-specific effects (ages 1-or-2, 4, 10 and 18 years). In the repeated measurement analyses of concurrent effects, FLG variants, in the absence of a positive SPT, did not significantly increase the risk of the concurrent eczema (RR = 1.12). In contrast, among individuals without FLG variants, allergic sensitization was associated with concurrent risk of eczema (RR = 1.48). In the presence of both FLG variants and allergic sensitization (combined effect), a considerable increase in the risk ratio was observed (RR = 2.82). Focussing on the temporal sequence in repeated measurement analyses, we found that among eczema-free children, a positive SPT combined with FLG variants showed a significant interaction (combined effect RR = 2.92) on the occurrence of eczema in subsequent follow-up examinations up to 18 years of age. In contrast, in non-sensitized children the combined effect of eczema and FLG variants posed no overall increased risk for subsequent allergic sensitization, but only for the transition periods 1-or-2 to 4 years and 4 to 10 years.

Other birth cohort studies, the Copenhagen Study on Asthma in Childhood (COPSAC), the Manchester Asthma and Allergy Study (MAAS), the Prevention and Incidence of Asthma and Mite Allergy (PIAMA) in the Netherland, and the Avon Longitudinal Study of Parents and Children (ALSPAC), have ascertained FLG variants and eczema longitudinally up to the age of 5, 8, and 11 years, respectively, with several repeated measurements [13], [25], [26]. However, the temporal sequence of eczema and allergic sensitization in interaction with FLG variants was not tested. Nonetheless, when comparing the results of our concurrent models with these studies, the magnitude of the associations between FLG variants and eczema is comparable. In our cohort, at 4 years of age FLG variants increased the risk of eczema by approximately 2-fold (RR = 1.92; 95% CI: 1.27–2.91). Similarly, results from the COPSAC cohort study demonstrated that FLG variants increase the risk of eczema during the first five years of life by almost the same amount (OR = 1.75; 95% CI: 1.29–2.37) [27]. Furthermore, results from the PIAMA cohort study showed that FLG variants were associated with eczema during the first eight years of life (OR = 2.0; 95% CI: 1.4–2.8) [25]. A slightly higher risk of eczema due to FLG variants during the first 11 years of life (OR = 2.46; 95% CI: 2.02–2.99) was reported by the ALSPAC cohort study [13].

A report of the German Multicentre Allergy Study (MAS) cohort show an interaction effect of FLG variants and food sensitization (RR = 4.07; 95% CI: 2.94–5.65) on the development of concurrent “eczema-associated asthma” from birth to 13 years of age [28]. In our study we investigated the combined effects of FLG variants and allergic sensitization (instead of food sensitization only) on the development of eczema (not “eczema-associated asthma”). Using repeated measurements from age 1-or-2 to age 18, we found that allergic sensitization interacts with FLG variants and increase the risk of concurrent eczema (RR = 2.82).

Without considering FLG variants, previous longitudinal investigations showed inconsistent results as to whether manifestation of eczema occurs before or after allergic sensitization development [18]. A report based on Melbourne Atopy Cohort Study (MACS, investigating the period from 6 months to 7 years of age) suggested that allergic sensitization may precede eczema development or eczema develops before allergic sensitization [29]. In contrast, results from the Childhood Asthma Prevention Study (CAPS) cohort in Sydney showed that eczema preceded and increased the risk of subsequent allergic sensitization between ages 18 months to 5 years [30]. Both studies did not account for the effect of FLG variants and did not cover the period of adolescence. When investigating the temporal sequence of events in relation to FLG variants effect, we found that both directions coexist. The combined effect of FLG variants and eczema leading to later allergic sensitization was only seen in the first decade of life. This finding speaks in favour of the model proposed by Cork et al., suggesting that the skin barrier is less developed at younger ages, which might explain the increased risk of allergic sensitization associated with FLG variants and eczema only in infants and young children [5]. For the transition period 10 to 18 years, the combined effect of FLG variants and eczema had no effect on the development of subsequent allergic sensitization. This observation is further supported by a previous study which showed that the majority of children carrying FLG variants outgrow eczema by 12 years of age [13]. Hence, in the transition from 10 to 18 years, the eczema × FLG variants interaction may no longer add to the risk of allergic sensitization. Overall, compared to the combined effect of eczema and FLG variants on allergic sensitization, the combined effect of FLG variants and allergic sensitization leading to eczema was stronger, particularly in early childhood, and covered the first two decades of life (Figure 2). This novel finding suggests strongly that allergic sensitization is not an epiphenomenon of eczema or its sequel, but a risk factor for subsequent eczema.

The longitudinal design of the Isle of Wight birth cohort study that covered the first 18 years of life is a major strength of our study. Throughout all assessments, the proportion of participation remained high, which ruled-out a major bias due to loss-to-follow up. Moreover, SPTs were repeatedly performed, thus allowing us to treat allergic sensitization as a time-dependant covariate. The prevalence of eczema and allergic sensitization did not differ between the total cohort and the sample that had available FLG variants genotypes (Table 1). Thus, there is no indication of a selection bias. Misclassification of eczema cases is minimal since a high proportion of subjects showed typical manifestation of eczema in usual locations (antecubital or popliteal fossae, ankles, face or neck for 97% at 1 year, 91% at 2 years, 75% at 4 years, 86% at 10 years and 76% at 18 years) [31]. In addition, the prevalence of eczema in our cohort (13.7% at age 10 years and 12.3% at age 18 years) is comparable to results from the International Study of Asthma and Allergies in Childhood aged 13–14 years in the United Kingdom (UK) (14.7% phase one and 10.6% phase three) and other studies conducted in the UK [32], [33], [34]. The proportion of FLG variants carriers in our study population (10.3%) is similar to the estimated heterozygous carries (10%) for individuals of European ancestry [12]. In line with previous studies, individuals homozygous for the minor allele of FLG variants were rarely identified due to the low minor allele frequencies of these variants [13], [25], [35].

A limitation of this study is that we have information on SPTs for the exams at age 1 and 2 years only for symptomatic children. However, there is no suspicion that the inclusion of the SPT at age 1 and 2 resulted in a selection bias: First, in both the eczema and the control children at age 1-or-2 the proportion of positive SPT is not higher than at age 4 when all available children were skin prick tested (21.9 vs. 20.4, Table 1). Second, there is no increased risk ratio of allergic sensitization at age 1-or-2 for eczema compare to other ages (Table 2). Statistical power could be another potential limitation due to the low sample sizes in some of the analyzed subgroups. However, statistical significance was demonstrated for most of the clinically important effects (i.e. associations with RR considerably higher than the null). Therefore, we believe that our study was not underpowered to detect statistically significant and clinically important associations.

Our results suggest that a defective skin barrier and altered immune responses work synergistically to increase susceptibility to eczema, a concept recently proposed by Biagini Myers et al. [36]. Null variants in FLG result in the formation of a disrupted skin barrier that predisposes to higher risk of eczema due to higher penetration of environmental factors (outside-inside barrier) and excessive water loss (inside-outside barrier) [17]. Our findings, that allergic sensitization is associated with eczema even in the absence of FLG variants, further support the concept that penetration of allergens leading to manifestation of eczema may also be related to other factors that regulate the integrity of the skin barrier. Therefore, factors regulating the immune response as well as genes responsible for the formation of functional skin barrier, may play a role in the pathophysiology of eczema [37].

A recent study showed that the skin barrier is compromised even in non-FLG variants carriers, supporting the notion that other factors may be important for the formation of a functional skin barrier [38]. Recent reports suggested that a single nucleotide polymorphism within the HRNR (hornerin) gene, located 78 kb away from FLG, conferred susceptibility to eczema [39], [40]. Hence comprehensive tests of genetic variants in the epidermal differentiation complex will improve the sensitivity of identifying susceptible children. Improved knowledge of skin susceptibility factors would then enhance randomization of susceptible groups to test treatment effects of possible skin treatments.

In conclusion, no prior study has investigated the long-term development of eczema and allergic sensitization with respect to interactions with FLG variants. The results of our longitudinal study covering the first 18 years of life corroborate prior findings that a higher risk of eczema is associated with FLG variants. For the first time, taking the time-order into account, we estimated the risk ratio of interaction of FLG variants with allergic sensitization and eczema on later occurrence of eczema or allergic sensitization, respectively. We demonstrated that both directions coexist; however, the interaction between allergic sensitization and FLG variants on subsequent eczema was more important: First, this interaction was detected in both childhood and adolescence, whereas the effect of the FLG variants × eczema interaction was only seen up to the age of 10 years. Second, the interaction FLG variants × allergic sensitization was stronger particularly in early childhood. The findings suggest that improvements of the skin barrier, for example by using barrier creams or anti-inflammatory treatments, at an early age may reduce the risk of eczema associated with the combination of FLG variants and allergic sensitization.

Supporting Information

Table S1.

Summary of FLG variants genotype data.

doi:10.1371/journal.pone.0032721.s001

(PDF)

Table S2.

Risk ratios of FLG variants for allergic sensitization in the course of childhood and adolescence.

doi:10.1371/journal.pone.0032721.s002

(PDF)

Acknowledgments

We would like to acknowledge the help of all the staff at The David Hide Asthma and Allergy Research Centre in undertaking the 18 year and previous assessments of 1989 Isle of Wight birth cohort. We would also like to acknowledge the help of the participants and their families who have helped us with this project over the last two decades.

Author Contributions

Analyzed the data: AHZ WK MY HZ. Wrote the paper: AHZ WK MY. Conceived and designed the epidemiological study: AHZ SHA WK ES Performed the examinations and measurements: AHZ SE ES. Revised the manuscript: HZ SE JWH WK.

References

  1. 1. Bieber T (2010) Atopic dermatitis. Ann Dermatol 22: 125–137.
  2. 2. Peters AS, Kellberger J, Vogelberg C, Dressel H, Windstetter D, et al. (2010) Prediction of the incidence, recurrence, and persistence of atopic dermatitis in adolescence: a prospective cohort study. J Allergy Clin Immunol 126: 590–595 e591–593.
  3. 3. Gold MS, Kemp AS (2005) Atopic disease in childhood. Med J Aust 182: 298–304.
  4. 4. Arshad SH, Tariq SM, Matthews S, Hakim E (2001) Sensitization to common allergens and its association with allergic disorders at age 4 years: a whole population birth cohort study. Pediatrics 108: E33.
  5. 5. Cork MJ, Danby SG, Vasilopoulos Y, Hadgraft J, Lane ME, et al. (2009) Epidermal barrier dysfunction in atopic dermatitis. J Invest Dermatol 129: 1892–1908.
  6. 6. Akiyama M (2010) FLG mutations in ichthyosis vulgaris and atopic eczema: spectrum of mutations and population genetics. Br J Dermatol 162: 472–477.
  7. 7. Palmer CN, Irvine AD, Terron-Kwiatkowski A, Zhao Y, Liao H, et al. (2006) Common loss-of-function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat Genet 38: 441–446.
  8. 8. Rodriguez E, Baurecht H, Herberich E, Wagenpfeil S, Brown SJ, et al. (2009) Meta-analysis of filaggrin polymorphisms in eczema and asthma: robust risk factors in atopic disease. J Allergy Clin Immunol 123: 1361–1370 e1367.
  9. 9. Brown SJ, Sandilands A, Zhao Y, Liao H, Relton CL, et al. (2008) Prevalent and low-frequency null mutations in the filaggrin gene are associated with early-onset and persistent atopic eczema. J Invest Dermatol 128: 1591–1594.
  10. 10. Greisenegger E, Novak N, Maintz L, Bieber T, Zimprich F, et al. (2009) Analysis of four prevalent filaggrin mutations (R501X, 2282del4, R2447X and S3247X) in Austrian and German patients with atopic dermatitis. J Eur Acad Dermatol Venereol 24: 607–610.
  11. 11. Sandilands A, Terron-Kwiatkowski A, Hull PR, O'Regan GM, Clayton TH, et al. (2007) Comprehensive analysis of the gene encoding filaggrin uncovers prevalent and rare mutations in ichthyosis vulgaris and atopic eczema. Nat Genet 39: 650–654.
  12. 12. Irvine AD, McLean WH, Leung DY (2011) Filaggrin mutations associated with skin and allergic diseases. N Engl J Med 365: 1315–1327.
  13. 13. Henderson J, Northstone K, Lee SP, Liao H, Zhao Y, et al. (2008) The burden of disease associated with filaggrin mutations: a population-based, longitudinal birth cohort study. J Allergy Clin Immunol 121: 872–877 e879.
  14. 14. Marenholz I, Nickel R, Ruschendorf F, Schulz F, Esparza-Gordillo J, et al. (2006) Filaggrin loss-of-function mutations predispose to phenotypes involved in the atopic march. J Allergy Clin Immunol 118: 866–871.
  15. 15. Weidinger S, Rodriguez E, Stahl C, Wagenpfeil S, Klopp N, et al. (2007) Filaggrin mutations strongly predispose to early-onset and extrinsic atopic dermatitis. J Invest Dermatol 127: 724–726.
  16. 16. van den Oord RA, Sheikh A (2009) Filaggrin gene defects and risk of developing allergic sensitisation and allergic disorders: systematic review and meta-analysis. BMJ 339: b2433.
  17. 17. Proksch E, Folster-Holst R, Brautigam M, Sepehrmanesh M, Pfeiffer S, et al. (2009) Role of the epidermal barrier in atopic dermatitis. J Dtsch Dermatol Ges 7: 899–910.
  18. 18. Simpson E (2010) Are epidermal defects the key initiating factors in the development of atopic dermatitis? Br J Dermatol 163: 1147–1148.
  19. 19. Boguniewicz M, Leung DY (2011) Atopic dermatitis: a disease of altered skin barrier and immune dysregulation. Immunol Rev 242: 233–246.
  20. 20. Arshad SH, Karmaus W, Kurukulaaratchy R, Sadeghnejad A, Huebner M, et al. (2008) Polymorphisms in the interleukin 13 and GATA binding protein 3 genes and the development of eczema during childhood. Br J Dermatol 158: 1315–1322.
  21. 21. Hanifin JM, Rajka G (1980) Diagnostic features of atopic dermatitis. Acta Derm Venereol Suppl (Stockh) 92: 44–47.
  22. 22. Zhang J, Yu KF (1998) What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 280: 1690–1691.
  23. 23. Zeger SL, Liang KY (1986) Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42: 121–130.
  24. 24. Ottman R (1996) Gene-environment interaction: definitions and study designs. Prev Med 25: 764–770.
  25. 25. Schuttelaar ML, Kerkhof M, Jonkman MF, Koppelman GH, Brunekreef B, et al. (2009) Filaggrin mutations in the onset of eczema, sensitization, asthma, hay fever and the interaction with cat exposure. Allergy 64: 1758–1765.
  26. 26. Bisgaard H, Simpson A, Palmer CN, Bonnelykke K, McLean I, et al. (2008) Gene-environment interaction in the onset of eczema in infancy: filaggrin loss-of-function mutations enhanced by neonatal cat exposure. PLoS Med 5: e131.
  27. 27. Bonnelykke K, Pipper CB, Tavendale R, Palmer CN, Bisgaard H (2010) Filaggrin gene variants and atopic diseases in early childhood assessed longitudinally from birth. Pediatr Allergy Immunol 21: 954–961.
  28. 28. Marenholz I, Kerscher T, Bauerfeind A, Esparza-Gordillo J, Nickel R, et al. (2009) An interaction between filaggrin mutations and early food sensitization improves the prediction of childhood asthma. J Allergy Clin Immunol 123: 911–916.
  29. 29. Lowe AJ, Abramson MJ, Hosking CS, Carlin JB, Bennett CM, et al. (2007) The temporal sequence of allergic sensitization and onset of infantile eczema. Clin Exp Allergy 37: 536–542.
  30. 30. Almqvist C, Li Q, Britton WJ, Kemp AS, Xuan W, et al. (2007) Early predictors for developing allergic disease and asthma: examining separate steps in the ‘allergic march’. Clin Exp Allergy 37: 1296–1302.
  31. 31. Ziyab AH, Raza A, Karmaus W, Tongue N, Zhang H, et al. (2010) Trends in eczema in the first 18 years of life: results from the Isle of Wight 1989 birth cohort study. Clin Exp Allergy 40: 1776–1784.
  32. 32. Asher MI, Montefort S, Bjorksten B, Lai CK, Strachan DP, et al. (2006) Worldwide time trends in the prevalence of symptoms of asthma, allergic rhinoconjunctivitis, and eczema in childhood: ISAAC Phases One and Three repeat multicountry cross-sectional surveys. Lancet 368: 733–743.
  33. 33. Marinho S, Simpson A, Lowe L, Kissen P, Murray C, et al. (2007) Rhinoconjunctivitis in 5-year-old children: a population-based birth cohort study. Allergy 62: 385–393.
  34. 34. Osman M, Tagiyeva N, Wassall HJ, Ninan TK, Devenny AM, et al. (2007) Changing trends in sex specific prevalence rates for childhood asthma, eczema, and hay fever. Pediatr Pulmonol 42: 60–65.
  35. 35. Weidinger S, O'Sullivan M, Illig T, Baurecht H, Depner M, et al. (2008) Filaggrin mutations, atopic eczema, hay fever, and asthma in children. J Allergy Clin Immunol 121: 1203–1209 e1201.
  36. 36. Biagini Myers JM, Khurana Hershey GK (2010) Eczema in early life: genetics, the skin barrier, and lessons learned from birth cohort studies. J Pediatr 157: 704–714.
  37. 37. O'Regan GM, Irvine AD (2010) The role of filaggrin in the atopic diathesis. Clin Exp Allergy 40: 965–972.
  38. 38. Jakasa I, Koster ES, Calkoen F, McLean WH, Campbell LE, et al. (2011) Skin barrier function in healthy subjects and patients with atopic dermatitis in relation to filaggrin loss-of-function mutations. J Invest Dermatol 131: 540–542.
  39. 39. Henry J, Hsu CY, Haftek M, Nachat R, de Koning HD, et al. (2011) Hornerin is a component of the epidermal cornified cell envelopes. Faseb J 25: 1567–1576.
  40. 40. Esparza-Gordillo J, Weidinger S, Folster-Holst R, Bauerfeind A, Ruschendorf F, et al. (2009) A common variant on chromosome 11q13 is associated with atopic dermatitis. Nat Genet 41: 596–601.