1. ¹Department of Pediatric Dermatology, Our Lady's Children's Hospital, Dublin, Ireland
2. ²Department of Clinical Medicine, Trinity College, Dublin, Irel

Correspondence: Dr Alan D. Irvine, Department of Pediatric Dermatology, Our Lady's Hospital for Sick Children, Crumlin, Dublin 12, Ireland. E-mail: irvinea@tcd.ie, alanirvine@eircom.net

Abstract

The association of filaggrin null alleles with eczema has been replicated in several European populations. Three large, well-conducted studies confirm this association and offer insights into the phenotypic nature of eczema associated with these alleles. Early data suggest that FLG-associated eczema may be more persistent, more likely to have palmar hyperlinearity, and more likely to be associated with asthma. These initial hints will require further confirmation in cohort studies.

Earlier this year, two recurrent null alleles in filaggrin (FLG) were observed to confer substantial risk of atopic dermatitis (AD; henceforth called eczema, according to the World Allergy Organization nosology (Johansson et al., 2004)) and of asthma occurring in the context of eczema (Palmer et al., 2006). Within a short time these two null alleles in FLG (R501X and 2282del4) have become the most widely and independently replicated and most strongly associated genetic factor conferring susceptibility to eczema in European populations to date, with odds ratios of between 3.73 and 7.1 reported in the larger cohorts (Marenholz et al., 2006; Ruether et al., 2006). In this issue of the Journal of Investigative Dermatology, three additional replication studies are reported (Barker et al., 2007; Stemmler et al., 2007; Weidinger et al., 2007). Each of these studies reports strong association between eczema and FLG null alleles, and each allows further exploration of the genetic contribution of these common recurrent null alleles to the pathogenesis of eczema. A summary of the populations studied to date and relevant findings is presented in Table 1. Significantly, all studies have shown strong replication of association with eczema, and no negative or equivocal study in Europeans, in whom these mutations are prevalent, has been reported. The position of FLG as a major gene in AD is further reinforced by the striking finding that these two common recurrent mutations have a penetrance of 42% in the German Multicenter Allergy Study and that the population attributable risk factor (the proportion of eczema that is directly attributable to these two recurrent null alleles at population level assuming a causal association) for eczema in this cohort was 11% (Marenholz et al., 2006). These calculations of the total cumulative effect of FLG null alleles in eczema are likely to be an underestimate, as several additional recurrent FLG null alleles are present in European populations at lower levels with intra-European, regionally distinct allele frequencies (A. Sandilands, W.H.I. McLean, and A.D. Irvine, unpublished data).

Table 1 - Summary of genetic studies of association of FLG null alleles with eczema and related phenotypes.

Full table

Given such a strong gene effect, and given that these same alleles cause the distinct phenotype of ichthyosis vulgaris in a semidominant inheritance pattern (Smith et al., 2006), it is reasonable to ask whether the phenotypic characteristics of eczema related to FLG null alleles can in any way be dissected out from the broad and inclusive concept of eczema suggested by current guidelines. Subcategorization of eczema based on elevated circulating specific IgE levels (into atopic and nonatopic, or extrinsic and intrinsic forms) has been widely suggested, but it has failed to increase our predictive ability in terms of phenotype or course of disease (Flohr et al., 2004). In particular, elevated IgE levels are less relevant in community-based studies than in hospital-based studies (Flohr et al., 2004). Could the presence or absence of FLG null alleles help define sub-phenotypes and predict the course of eczema? Examination of the early data allows a degree of cautious interpretation as to the precise contribution of FLG null alleles to eczema phenotypes and to the consideration of possible genotype–phenotype correlates. Within cohorts that have been ascertained through attendance of people with eczema at dermatology clinics, some recurrent themes emerge. Two large series of adult patients with AD presenting to hospital clinics show a high frequency of FLG null carrier rates in adults who have persistent AD with onset within the first 2years (29.7% in a German cohort (Weidinger et al., 2007), and 42% in an English cohort (Barker et al., 2007)). These observations suggest that individuals with eczema who carry FLG null alleles could be more likely to have early onset and persistent disease, although controls with late-onset eczema and cohort studies are needed for further proof of such a notion. In well-characterized eczema cohorts from Munich, these alleles predispose to elevated IgE as a secondary trait (Weidinger et al., 2006; Weidinger et al., 2007). This predisposition was also seen in the extended Genetic Studies in Nuclear Families with Atopic Dermatitis cohort (Marenholz et al., 2006) and in relation to "extrinsic" AD (AD plus sensitization determined by positive skin-prick testing or specific IgE elevation) (Weidinger et al., 2007). These findings may reflect the more severe spectrum of disease seen in hospital studies, where elevated IgE is more commonly found (Flohr et al., 2004), and it is interesting that in a birth-cohort study, FLG-null status was also strongly associated with nonatopic eczema (Marenholz et al., 2006). Thus, elevated circulating IgE may be primarily an epiphenomenon that is a proxy for severity in eczema and is therefore more represented in hospital-ascertained series (Flohr et al., 2004).

Although the relationship between FLG null alleles and asthma remains to be fully understood, in eczema cohorts where asthma status has been ascertained there is a consistent additional association with asthma as a secondary trait (Marenholz et al., 2006; Weidinger et al., 2006; Weidinger et al., 2007). The more informative cohorts in this regard are those ascertained through a diagnosis of asthma, and population-based cohorts. In the Dundee BREATHE cohort, the association of FLG with asthma was entirely limited to the subgroup with a coincident history of eczema, with no association between FLG alleles and asthma in the absence of eczema (Palmer et al., 2006). Similarly, in the German Multicenter Allergy Study cohort, FLG alleles had no association with asthma or allergic sensitization in the absence of prior eczema (Marenholz et al., 2006). Importantly, this study also observed an additional risk of asthma in patients homozygous for FLG null alleles and calculated a striking population attributable risk factor of 20.6% for FLG null alleles for the phenotype eczema plus asthma (Marenholz et al., 2006). These data suggest a sequence of phenotypic manifestations rather than individual susceptibilities to each of these manifestations of atopy and also allow speculation that early control of eczema may have an impact on future development of asthma, a concept that has already begun to be explored in studies such as the Early Treatment of the Atopic Child study (Warner, 2001).

Taken together, these early data hint at possible discriminating features of eczema associated with FLG null alleles: that the eczema is more strongly associated with asthma, has associated hyperlinearity of the palms, and is more likely to be persistent into adulthood. An additional association with elevated IgE is also possible but not yet clear. This tentative analysis will doubtless be tempered as additional data become available, but it suggests a starting point to explore genotype–phenotype correlations. It should also be noted that although these carefully phenotyped cohorts where participants have been recruited through a diagnosis of eczema are very informative, they are unable to fully answer the question of what (if anything) is different about FLG-related eczema versus non-FLG-related eczema, nor can they determine the relative risk of FLG null alleles to asthma, atopy, and rhinitis (if any) per se. Large longitudinal or cross-sectional population-based cohorts that have been typed for all FLG variants will facilitate the dissection of these phenotypic characteristics and will enable researchers to compare and contrast FLG phenotypes versus others to test these speculations. Given the xerotic phenotype of ichthyosis vulgaris (Smith et al., 2006), which is seen in many eczema sufferers, future studies observing additional physical parameters of skin barrier function, such as the degree of skin hydration, the rate of transepidermal water loss, and susceptibility to skin irritation, may help to test the hypothesis of a "dry or defective barrier" type of FLG-related eczema. The pharmacogenetic significance of filaggrin status in predicting possible treatment responses to interventions such as emollients, soap substitution, water softening, or pharmacological intervention, and whether such interventions alter the disease trajectory, will also be worth examining.

To date, filaggrin has not been shown to be a susceptibility factor for eczema in non-European populations. There are a few possible explanations for this, including the fact that the currently known recurrent mutations either are not present or are present at very low levels in many non-European populations (Palmer et al., 2006). Also, except in Japan, few research groups have established data sets of non-European eczema patients for genetic analysis. The technical difficulties of sequencing filaggrin may mean that it will take some time to determine the role of FLG mutations in populations outside of Europe. The incomplete penetrance of these alleles also merits further attention and is presumably explained by both genetic and environmental modifiers. Possible genetic modifiers of FLG null-allele penetrance include SPINK5 (Walley et al., 2001) and SCCE (Vasilopoulos et al., 2004), both of which have been associated with eczema and both of which are plausibly involved in controlling filaggrin processing. Much has been learned in a short time about the role of FLG status in eczema. Much more needs to be understood about its function before this knowledge can be translated directly into patient benefit.