Key Points
Question
Do FLG loss-of-function variants differ between African American and white children with atopic dermatitis?
Findings
In a registry cohort of 741 children with atopic dermatitis, FLG loss-of-function variants differed significantly by race, were associated with race, and overall differences were noted with respect to whether any variant was present as well as the frequency of individual variants.
Meaning
These findings suggest that using FLG loss-of-function variants to diagnose atopic dermatitis or to estimate the clinical course may be very difficult because very common variants in one race may be absent or very rare in another.
This registry cohort study describes and compares results of targeted sequencing of FLG loss-of-function variants in US children of African and European ancestry and the association of these variants with onset and persistence of atopic dermatitis.
Abstract
Importance
Atopic dermatitis (AD) is a common chronic illness that has been associated with variation in the filaggrin gene (FLG). Four variants are most often evaluated.
Objectives
To comprehensively describe and compare results from targeted sequencing of FLG loss-of-function (LoF) variants in children of African and European ancestry and the association of these variants with onset and persistence of AD.
Design, Setting, and Participants
This prospective US cohort study assessed the genetic subcohort of the Pediatric Eczema Elective Registry (PEER). Children with mild to moderate AD were included in the analysis. Massively parallel sequencing (MPS) was used to focus on FLG LoF variation in white and African American children. Patients were enrolled from June 2005 through July 2017. Data were analyzed from January 25 through May 10, 2019.
Main Outcomes and Measures
Associations of FLG LoF variation with white and African American ancestry and with the risk and persistence of AD.
Results
A total of 741 children were included in the analysis (394 [53.2%] female and 347 [46.8%] male; mean [SD] age at onset, 1.97 [2.72] years); of these, 394 (53.2%) were white, 326 (44.0%) were African American, and 21 (2.8%) were of other ancestries. Using MPS technology, 23 FLG LoF variants were found in children with AD. The prevalence of FLG LoF variants was 177 participants (23.9%) in the full cohort, 124 white participants (31.5%), and 50 African American participants (15.3%). The odds ratio for carrying any FLG LoF variant in a white child compared with an African American child with AD was 2.44 (95% CI, 1.76-3.39). Some FLG LoF variants are only found in children of a specific ancestry (eg, p.S3316* and p.R826* were not seen in white patients). Children with an FLG LoF were more likely to have persistent AD (odds ratio, 0.67; 95% CI, 0.56-0.80).
Conclusions and Relevance
The FLG LoF variants in a US cohort of children with mild to moderate AD differ significantly by race and their association with the persistence of AD. Conventional testing of the 4 frequently evaluated variants is inadequate. Any planned genetic diagnostic test for AD based on FLG LoF variants must be inclusive and not rely on the most frequently studied variants.
Introduction
Atopic dermatitis (AD) is a common chronic episodic ailment manifested by itchy red patches that most frequently occur in the flexural areas of the elbows and knees.1,2 The terms AD, eczema, and flexural dermatitis are often used interchangeably. Atopic dermatitis often has its onset in children younger than 2 years and was thought to be a childhood disease with the assumption that nearly all children were symptom free by adolescence.3,4 However, several studies3,4,5 show that AD is a life-long disease that often persists from childhood into adulthood and that the initial appearance of AD can occur in adulthood. In the United States, the yearly prevalence of AD is about 10% in children and adults, with an annual cost of about $4.2 billion.6,7,8,9
A family history of AD and other atopic illnesses, such as asthma and seasonal allergies, is associated with an increased risk of developing AD as a child.10,11 The heritability of AD is estimated to be as high as 84%.12 The most commonly reported genetic variants associated with skin barrier dysfunction are exon 3 loss-of-function (LoF) variants in the filaggrin gene (FLG [OMIM 135940]).13 FLG codes for a skin barrier protein, and FLG LoF variants are associated with the prevalence and persistence of AD.13,14,15,16,17,18,19 Although FLG LoF variants are found in approximately 25% to 30% of individuals of European and Asian ancestry with AD, except for 2 small case series, individuals of African ancestry with AD uncommonly exhibit FLG LoF variants.13,14,20,21,22 FLG LoF variants also are found in individuals without AD and, for instance, are present in about 8% of the general European population.13,23,24
Approximately 500 FLG LoF variants have been identified by the gnomAD (genome aggregation database), an international aggregate database of exome and genome sequencing data supported by the Broad Institute, Cambridge, Massachusetts.25 Several studies13,14,16,23,26,27 have shown that the prevalence of FLG LoF variants differs by race. Four FLG LoF variants have been consistently associated with AD in patients of European ancestry (p.R501*, c.2282del4[p.S761fs], p.S3247*, and p.R2447*), with allelic frequencies ranging from 1% to 10%, and studies often focus primarily on these variants.13,14,16,26 However, a study of individuals of East Asian ancestry with AD27 demonstrated that a larger number of variants should be assayed to fully evaluate the potential association of FLG LoF variants and AD disease prevalence. Older studies,14,23,24,28,29,30 using a range of genotyping and sequencing techniques and relying on an earlier bioinformatics pipeline, were not able to detect FLG LoF variants in individuals of African ancestry with AD at the frequencies noted in European and Asian populations. However, recent studies20 using newer analytic tools demonstrate that individual FLG LoF variants are present in those of African ancestry with AD, but their frequencies are less than 2.0%, suggesting that larger numbers of variants and participants need to be assessed. The goals of the present study are to comprehensively describe and compare results from fine sequencing of FLG LoF variants in US children of African and European ancestry and to evaluate the association of these variants with the risk of onset and the persistence of AD.
Methods
Population and Design
The Pediatric Eczema Elective Registry (PEER; https://www.thepeerprogram.org) is a US nationwide cohort of more than 8000 participants with pediatric-onset AD. The present prospective study represents the subcohort of PEER children who provided a genetic sample (PEER DNA cohort).14,31 Self-described race and ancestry informative markers were previously used to define race and were found in this cohort to be highly correlated.14 At the time of enrollment, children were aged 2 to 17 years, had a physician-confirmed diagnosis of AD, and had used pimecrolimus cream for at least 6 months.14 Participants were enrolled from June 2005 through July 2017. Participants were followed up for as long as 10 years (data cutoff, January 25, 2019) and during that time were not required to (and most did not) continue therapy with pimecrolimus.32 The study is still ongoing. The present study was approved by the institutional review board of the University of Pennsylvania, Philadelphia. All patients provided written informed consent. Full details of the PEER cohort have been previously reported.5,14,32 This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Genetic Analysis
We collected DNA using commercially available collection kits (Oragene; DNA Genotek) as previously reported.14 The FLG gene, including exon 3, was sequenced for all participants using massively parallel sequencing (MPS). The reliability of this technique with respect to FLG sequencing was previously reported.20
Raw sequencing data were aligned and mapped to the reference genome GRCh37 using the Burrows-Wheeler Aligner program and a standard 10 repeat in exon 3.33 Single-nucleotide variant and insertion/deletion calling was accomplished using the Genome Analysis Toolkit (GATK) HaplotypeCaller following GATK best practices realignment and recalibration.34,35,36,37 A set of hard filters was applied to the variant call set following recommendations described in the GATK user guide documentation. Briefly, the single-nucleotide variants were tested and filtered for quality by depth, strand bias, mapping quality bias, and read position bias. Similarly, insertion/deletion variants were excluded for quality by depth, strand bias, and read position bias criteria.
Statistical Analysis
Data were analyzed from January 25 through May 10, 2019. Variant frequency is reported for the full PEER cohort and by race as prevalence of any minor variant and as genome minor variant frequency per 100 individuals with 95% CI. The minor variant frequency was also determined for FLG LoF variants by race using data from gnomAD, version 2.11 (released March 2019; https://gnomad.broadinstitute.org/).25 As is the convention, a simple composite was created for all of the FLG LoF variants (ie, no variant, wild-type), a single variant (heterozygote), and 2 or more variants (including homozygote).13,14,16,26 Exon 3 of FLG contains 10 to 12 nearly identical repetitive sequences.38 Variation in the number of these sequences is called copy number variation. Most identified FLG LoF variants occur within the first 3 repeated sequences. A composite was also created depending on whether or not the FLG LoF variant occurred in the first 3 sequences. Variant frequency comparisons between races were conducted using the χ2 test, 2 × 2 table, or logistic regression. If the number of variants was less than 5, exact logistic regression was used.
Persistence was evaluated based on the self-reported outcome of whether or not a child’s skin was free of AD symptoms during the previous 6 months.14 Persistence was determined by survey responses to a series of questions, including: “During the last 6 months would you say that your child’s skin disease (AD) has shown: complete disease control, good disease control, limited disease control, or uncontrolled disease?” Symptom free was defined as an affirmative response of complete disease control. This response has been shown to correlate with other tools used to evaluate symptom control and is likely a marker of long-term disease severity.39 The association between these outcomes (multiple surveys recorded over time per participant) and the FLG variant composite were evaluated using generalized estimating equations for binary outcomes, assuming an independence working correlation structure with robust standard errors. All analyses were conducted with Stata, version 15.1 (StataCorp). Two-sided P < .05 was considered significant.
Results
The PEER DNA cohort had sufficient DNA for the MPS analysis from 741 children; of these, 394 (53.2%) were white, 326 (44.0%) were African American, and 21 (2.8%) were of other ancestries. For these children, the mean (SD) age at enrollment was 7.17 (3.77) years; the mean (SD) age at onset was 1.97 (2.72) years; 394 (53.2%) were female and 347 (46.8%) were male; and 421 (56.8%) completed the full 10 years of evaluations.
Minor variant frequencies for FLG LoF variants from PEER are presented in Table 1. Variant location is presented in Table 1 and the Figure. Variant frequency was found to be strongly associated with race (Table 1). For example, if we only evaluated the most frequently studied FLG LoF variants (ie, the classic p.R501*, p.S761fs [c.2282del4], p.S3247*, and p.R2447*),13,14,17,26 then the prevalence of FLG LoF variants was 134 (18.1%) in the full cohort, 115 (29.2%) in white individuals, and 18 (5.5%) among those of African ancestry. However, using the MPS technology, the prevalence of any FLG LoF variants was 177 (23.9%) in the full cohort, 124 (31.5%) in white individuals, and 50 (15.3%) in African American individuals. The 4 classic variants captured 92.6% of the prevalence of FLG LoF among white children (n = 9) but only 36.1% among children of African ancestry (n = 32). Using the results from MPS compared with interrogating only the classic 4 variants, 53 additional children (10 white, 42 African American, and 1 other) were noted to have at least 1 FLG LoF variant (P < .001). The odds ratio (OR) for carrying any FLG LoF variant in a white child compared with an African American child with AD was 2.44 (95% CI, 1.76-3.39). Some FLG LoF variants are only found in children of a specific ancestry (Table 1 and Figure). Furthermore, the OR for carrying a p.R501* variant in a white child compared with an African American child with AD was 5.92 (95% CI, 2.78-14.55), and p.S761fs, a more common variant in white children, was not found at all in PEER African American children (Table 1). Overall, the OR for an African American child having an FLG LoF variant compared with a white child was 0.44 (95% CI, 0.41-0.48).
Table 1. Loss-of-Function Variants of the FLG Gene.
| Variant | Location (GRCh37) | Reference Allele | Alternative Allele | Reference SNP Identifier | MVF per 100 Children (95% CI) | OR (95% CI)a | P Value | ||
|---|---|---|---|---|---|---|---|---|---|
| PEER | White | African American | |||||||
| p.Q3818* | 152275910 | G | A | rs148606936 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0.15 (0.00-0.85) | 0.78 (0.00-30.78) | .88 |
| p.S3707* | 152276242 | G | T | rs776800093 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0.15 (0.00-0.85) | 0.79 (0.00-30.78) | .88 |
| p.R3419* | 152277107 | G | A | rs143418984 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.R3409* | 152277137 | G | A | rs201356558 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0.15 (0.00-0.85) | 0.79 (0.00-30.78) | .88 |
| p.S3316* | 152277415 | G | C | rs149484917 | 0.61 (0.28-1.15) | 0 (0.00-0.47) | 1.38 (0.63-2.60) | 0.06 (0.00-0.39) | .001 |
| p.S3247*b | 152277622 | G | A | rs150597413 | 0.81 (0.42-1.41) | 1.14 (0.52-2.18) | 0.46 (0.09-1.34) | 2.40 (0.59-13.88) | .29 |
| p.R2447*b | 152280023 | G | A | rs138726443 | 1.35 (0.83-2.08) | 1.78 (0.97-2.96) | 0.77 (0.25-1.78) | 2.82 (0.88-11.90) | .09 |
| p.Y2119* | 152281005 | A | C | rs776681299 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.78 (0.02-∞) | >.99 |
| p.S2080* | 152281123 | G | T | rs147145234 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.S1906* | 152281645 | G | A | rs141784184 | 1.01 (0.57-1.16) | 0.26 (0.03-0.92) | 2.30 (1.29-3.77) | 0.12 (0.1-0.52) | .002 |
| p.K1872* | 152281748 | T | C | rs757686343 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.Q1807* | 152281943 | G | A | rs140164593 | 0.14 (0.02-0.49) | 0.13 (0.03-0.70) | 0.15 (0.00-0.85) | 0.79 (0.01-62.12) | >.99 |
| p.R1474* | 152282942 | G | A | rs146686141 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.R1388* | 152283200 | G | A | rs756594053 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.G1253* | 152283605 | C | A | rs199895224 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0 (0.00-0.56) | 0.79 (0.02-∞) | >.99 |
| p.R1150* | 152283914 | G | A | rs755996559 | 0.07 (0.00-0.38) | 0.13 (0.03-0.70) | 0.15 (0.00-0.85) | 0.79 (0.02-∞) | >.99 |
| p.Q977* | 152284433 | G | A | rs144419479 | 0.14 (0.02-0.49) | 0.13 (0.03-0.70) | 0.15 (0.00-0.85) | 0.79 (0.01-62.12) | >.99 |
| p.R826* | 152284886 | G | A | rs115746363 | 0.34 (0.10-0.78) | 0 (0.00-0.47) | 0.77 (0.25-1.78) | 0.12 (0.00-0.86) | .03 |
| p.R788* | 152285000 | G | A | rs183942200 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0 (0.00-0.56) | NC | NC |
| p.S761fsb | 152285076 | CACTG | C | rs138381300 | 3.98 (3.04-5.10) | 7.23 (5.52-9.27) | 0 (0.00-0.56) | NC | <.001 |
| p.Q570* | 152285654 | G | A | rs192402912 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0.15 (0.00-0.85) | 0.79 (0.00-30.78) | .88 |
| p.R501*b | 152285861 | G | A | rs61816761 | 4.45 (3.46-5.63) | 6.85 (5.19-8.85) | 1.84 (0.95-3.19) | 5.92 (2.78-14.55) | <.001 |
| p.H440fs | 152286063 | CTC | G | rs152286063 | 0.07 (0.00-0.38) | 0 (0.00-0.47) | 0.15 (0.00-0.85) | 0.78 (0.00-30.78) | .88 |
| Composite | |||||||||
| 4 Most frequent | NA | NA | NA | NA | 10.59 (9.07-12.27) | 17.51 (14.92-20.35) | 3.06 (1.88-4.70) | 7.05 (4.09-12.15) | <.001 |
| All | NA | NA | NA | NA | 14.04 (12.30-15.91) | 19.29 (16.59-22.22) | 8.59 (6.55-11.00) | 2.44 (1.76-3.39) | <.001 |
| First 3 repeats | NA | NA | NA | NA | 9.24 (7.82-10.83) | 15.61 (13.14-18.33) | 2.91 (1.76-4.51) | 5.76 (3.47-9.58) | <.001 |
Abbreviations: FLG, filaggrin; MVF, minor variant frequency; NA, not applicable; NC, not calculated; OR, odds ratio; PEER, Pediatric Eczema Elective Registry; SNP, single-nucleotide polymorphism.
Indicates odds of the variant found in a white or an African American child.
Indicates the 4 most frequently studied variants.
Figure. Diagram of Exon 3 of the Filaggrin Gene (FLG).
FLG loss-of-function variation assayed in this study was placed proportionally in the proper location based on GRCh37. The reference single-nucleotide polymorphim identifiers given in black were found in African American and white individuals; purple, in white individuals only; and red, in African American individuals only.
The frequency of specific FLG LoF variants in most cases was greater in children who had AD (PEER) vs a general population (gnomAD). Table 2 lists the 5 or 6 most frequently reported FLG LoF variants in gnomAD by race (European and African). Overall, the relative risk of variants for AD appears to be greater for individuals of European compared with African ancestry.
Table 2. Most Common FLG LoF Variants per gnomAD.
| Variant | Reference SNP Identifier | MVF Frequency per 100 Children (95% CI) | RRR (95% CI) | |
|---|---|---|---|---|
| gnomAD | PEER | |||
| White Individuals | ||||
| p.S761fs | rs138381300 | 2.16 (2.08-2.24) | 7.23 (5.52-9.27) | 3.48 (2.66-4.54)a |
| p.R501* | rs61816761 | 1.64 (1.57-1.71) | 6.85 (5.19-8.85) | 4.33 (3.30-5.69)a |
| p.R2447* | rs138726443 | 0.42 (0.38-0.46) | 1.78 (0.97-2.96) | 4.22 (2.50-7.11)a |
| p.S3247* | rs150597413 | 0.28 (0.25-0.31) | 1.14 (0.52-2.18) | 4.01 (2.10-7.68)a |
| p.R3419* | rs143418984 | 0.05 (0.04-0.07) | 0.13 (0.03-0.70) | 2.39 (0.34-16.72) |
| African American Individuals | ||||
| p.S1906* | rs141784184 | 1.16 (1.03-1.30) | 2.30 (1.29-3.77) | 1.96 (1.19-3.22)b |
| p.S3316* | rs149484917 | 0.78 (0.67-0.89) | 1.38 (0.63-2.60) | 1.75 (0.92-3.33) |
| p.R826* | rs115746363 | 0.76 (0.66-0.88) | 0.77 (0.25-1.78) | 1.00 (0.42-2.40) |
| p.R501* | rs61816761 | 0.42 (0.35-0.51) | 1.84 (0.95-3.19) | 4.04 (2.35-6.95)a |
| p.S761fsc | rs138381300 | 0.36 (0.29-0.44) | 0 (0.00-0.56) | NA |
| p.R2447* | rs138726443 | 0.06 (0.04-0.10) | 0.77 (0.25-1.78) | 9.40 (4.35-20.27)a |
Abbreviations: FLG, filaggrin gene; gnomAD; genome aggregation database; LoF, loss of function; MVF, minor variant frequency; NA, not applicable; PEER, Pediatric Eczema Elective Registry; RRR, relative rate ratio; SNP, single-nucleotide polymorphism.
Indicates P < .001.
Indicates P < .01 and P > .001.
A sixth variant was added because pS761fs was not noted in our sample.
The distribution of FLG LoF variation across FLG exon 3 also appears to differ by race. The frequency of a composite of FLG LoF variants from the first 3 repeats (Figure) is greater in white children than in African American children (Table 1). However, this is not true for the frequency of FLG LoF variants not found in the first 3 repeats (ORs, 4.19 [95% CI, 2.90-5.83] vs 5.67 [95% CI, 4.03-7.74], respectively; P = .28)
The associations between persistence of AD and children with FLG LoF variants as well as the FLG LoF composite with respect to the persistence of AD are presented in Table 3. During any survey period, children with any FLG LoF variants compared with children without an FLG LoF variant were more likely to have persistent AD (ie, the child was less likely to report disease-free skin) regardless of therapy (OR, 0.69; 95% CI, 0.58-0.83). In a composite of all variants, children with an FLG LoF were more likely to have persistent AD (odds ratio, 0.67; 95% CI, 0.56-0.80). A similar association was also noted for the composite that only included FLG LoF variants in the first 3 repeats (OR, 0.69; 95% CI, 0.58-0.83). The strength of the association did not vary by race. However, children with some variants, such as p.R501* (OR, 0.61; 95% CI, 0.44-0.84), appeared to have more persistent disease than other variants, such as p.S3247* (OR, 1.02; 95% CI, 0.47-2.19). In this cohort, regardless of FLG LoF variant status, African American compared with white children have more persistent disease (OR, 0.81; 95% CI, 0.67-0.99).
Table 3. Association of the Presence of FLG Loss-of-Function Variants With the Persistence of Symptoms of Atopic Dermatitis.
| Variant | Effect Estimate, OR (95% CI) | |||||
|---|---|---|---|---|---|---|
| Unadjusted | Adjusteda | |||||
| Full Cohort | White Individuals | African American Individuals | Full Cohort | White Individuals | African American Individuals | |
| Composite of 4 most frequent | 0.73 (0.60-0.89) | 0.69 (0.55-0.87) | 0.45 (0.26-0.78) | 0.67 (0.55-0.83) | 0.71 (0.56-0.90) | 0.42 (0.24-0.73) |
| P value | .002 | .002 | .004 | .0001 | .005 | .002 |
| Composite of first 3 repeats | 0.69 (0.58-0.83) | 0.62 (0.48-0.79) | 0.59 (0.34-1.03) | 0.63 (0.51-0.80) | 0.63 (0.49-0.82) | 0.57 (0.33-1.00) |
| P value | .001 | <.001 | .06 | <.001 | <.001 | .05 |
| Composite of all | 0.69 (0.58-0.83) | 0.67 (0.54-0.84) | 0.62 (0.44-0.86) | 0.67 (0.56-0.80) | 0.70 (0.56-0.87) | 0.62 (0.44-0.87) |
| P value | <.001 | .001 | .004 | <.001 | .002 | .006 |
| p.R501* | 0.61 (0.44-0.84) | 0.56 (0.39-0.81) | 0.44 (0.22-0.89) | 0.55 (0.40-0.76) | 0.57 (0.40-0.82) | 0.38 (0.20-0.78) |
| P value | .002 | .002 | .02 | <.001 | .003 | .008 |
| p.S761fs | 0.79 (0.56-1.11) | 0.70 (0.48-1.01) | NA | 0.78 (0.56-1.10) | 0.70 (0.49-1.02) | NA |
| P value | .17 | .06 | NA | .16 | .07 | NA |
| p.S3247* | 1.02 (0.47-2.19) | 1.21 (0.46-3.14) | 0.54 (0.13-2.29) | 0.94 (0.44-2.04) | 1.23 (0.47-3.19) | 0.44 (0.11-1.84) |
| P value | .97 | .70 | .41 | .89 | .65 | .26 |
| p.R2447* | 0.73 (0.40-1.30) | 0.70 (0.49-1.02) | 0.35 (0.11-1.16) | 0.78 (0.43-1.42) | 1.04 (0.49-2.32) | 0.49 (0.15-1.61) |
| P value | .28 | .07 | .10 | .43 | .90 | .24 |
| p.S1906* | 0.53 (0.26-1.03) | 0.57 (0.09-3.65) | 0.56 (0.29-1.10) | 0.61 (0.30-1.23) | 0.66 (0.10-4.18) | 0.61 (0.31-1.20) |
| P value | .06 | .55 | .95 | .17 | .66 | .15 |
| p.S3316* | 0.68 (0.28-1.62) | NA | 0.76 (0.33-1.76) | 0.78 (0.33-1.87) | NA | 0.79 (0.34-1.82) |
| P value | .38 | NA | .53 | .58 | NA | .58 |
| p.R826* | 0.74 (0.23-2.94) | NA | 0.83 (0.27-2.56) | 0.76 (0.24-2.47) | NA | 0.79 (0.26-2.41) |
| P value | .62 | NA | .75 | .65 | NA | .68 |
Abbreviations: FLG, filaggrin gene; NA, not applicable; OR, odds ratio.
Adjusted for age at onset, sex, race if appropriate, and presence of asthma or seasonal allergies.
Within the PEER cohort, the FLG LoF variant composite was associated with the age of onset of AD (mean [SD] age, 2.11 [2.80] years for wild-type, 1.56 [2.31] for composite heterozygote, and 1.27 [2.83] for compound composite; P = .02) as well as more frequent allergies to food (wild-type, 162 of 237 [68.4%]; composite heterozygote, 56 of 237 [23.6%]; and compound composite, 19 of 237 [8.0%]; P < .001) and animals (wild-type, 152 of 225 [67.6%]; composite heterozygote, 55 of 225 [24.4%]; and composite composite, 18 of 225 [8.0%]; P < .001) (Table 4). Fathers of a child with a variant were more likely to have AD. As opposed to age at onset, as might be expected for an inherited trait, age at study enrollment, sex, and household income were not associated with the prevalence of FLG LoF variants. Asthma and seasonal allergies, 2 atopic illnesses commonly associated with AD, also were not found to be associated with FLG LoF.
Table 4. Demographic and Associated Factors for Children With Atopic Dermatitis Based on Their FLG Loss-of-Function Status.
| Factor | FLG Loss-of-Function Status | |||
|---|---|---|---|---|
| Wild-Type (n = 564) | Composite Heterozygote (n = 146) | Compound Composite (n = 31) | P Valuea | |
| Age at enrollment, y | ||||
| Mean (SD) | 7.29 (3.81) | 6.60 (3.62) | 7.76 (3.59) | .40 |
| Median (range) | 6.59 (4.15-9.74) | 5.63 (3.89-8.45) | 6.90 (4.85-10.36) | |
| Age at onset, y | ||||
| Mean (SD) | 2.11 (2.80) | 1.56 (2.31) | 1.27 (2.83) | .02 |
| Median (range) | 0.75 (0.25-3.00) | 0.75 (0.25-2.00) | 0.25 (0.25-0.75) | |
| African American, No. (%) | 276 (84.7) | 44 (13.5) | 6 (1.8) | <.001 |
| Female, No. (%) | 297 (75.4) | 77 (19.5) | 20 (5.1) | .43 |
| Asthma, No. (%) | 294 (73.7) | 84 (21.1) | 21 (5.3) | .14 |
| Allergies, No. (%) | ||||
| Seasonal | 390 (75.1) | 106 (20.4) | 23 (4.4) | .67 |
| Food | 162 (68.4) | 56 (23.6) | 19 (8.0) | <.001 |
| Medication | 80 (69.0) | 30 (25.9) | 6 (5.2) | .14 |
| Animal | 152 (67.6) | 55 (24.4) | 18 (8.0) | <.001 |
| Family history of eczema, No. (%) | ||||
| Mother | 141 (78.8) | 31 (17.3) | 7 (3.9) | .62 |
| Father | 94 (81.7) | 13 (11.3) | 8 (7.0) | .02 |
| Annual income <$25 000, No. (%) | 182 (81.3) | 37 (16.5) | 5 (2.2) | .28 |
Abbreviation: FLG, filaggrin gene.
Calculated using analysis of variance or χ2 test.
Discussion
As genotyping technology has improved, more FLG LoF variants have been identified.20 Margolis et al20 first reported the use of MPS in 262 African American children from the PEER DNA cohort. In the present study, we expand those observations to the full PEER DNA cohort of 741 children consisting of white and African American participants. In total, we observed 23 FLG LoF variants. The minor variant frequency for these variants differs greatly and is associated with race in that many are exclusively found in one race but not another. African American individuals also tend to have a similar frequency of variants throughout exon 3, whereas white individuals tend to have a higher frequency of FLG LoF variants in the first 3 repeats. In most cases, the frequency of these variants was greater in our PEER cohort children, all of whom have AD, than in the more general population in gnomAD. As has been previously noted, those children with FLG LoF variants are more likely to have AD than those who do not, and those with an FLG LoF variant are more likely to have more persistent AD.13,14,20,21,40 However, all FLG LoF variants might not confer an increased risk of AD, and further, they may not all have the same effect on the persistence of AD over time. For example, p.R826*, which is not observed in Europeans, might not be more common in African American individuals with AD when compared with a general population (relative risk ratio, 1.00; 95% CI, 0.42-2.40).
As genetic testing becomes more prevalent and is used to determine risk and severity of an illness, it is important to examine how testing is conducted. Testing for FLG LoF variants should not be limited to the classic 4 European FLG LoF variants. First, as previous studies20,21,27 have shown, the classic European variants do not fully capture variants seen in African American and Asian individuals. Second, all variants may not confer increased risk.14,31 Third, FLG is not well characterized by all sequencing and genotyping techniques, so care must be taken when creating assays.20,27 Finally, although many studies have shown increased risk of AD in association with FLG LoF variants, very few have translated these findings into understanding how a child experiences AD over time (eg, persistence). In fact, for many patients, by the time genetic testing is complete, the child already has clinical manifestations of the disease, and parents or patients want to know not only if their child is at increased risk of developing the disease but also information about disease course.
The FLG LoF variation, with respect to the actual variant and frequency, is highly associated with race.13,20,27 Overall, white children have a higher frequency of FLG LoF variants, and a small number of these variants account for most of the FLG LoF variant prevalence. In children of African or Asian ancestry, no variants are seen as commonly as the 2 most common European variants, and overall there seem to be a large number of variants.20,27 In fact, 1 of the most prevalent variants in Europeans, p.S671fs, is rarely seen in individuals of African or Asian ancestry.14,20,27 The variant p.R501*, also common in individuals of European ancestry, is seen in those of African and Asian ancestry at a much lower frequency.20,27 In addition, the location of variants seems to differ by race, and the higher frequency of FLG LoF variants in white individuals is mainly owing to variants located within the first 3 repeats in exon 3. In this study, regardless of FLG LoF variant status, African American children compared with white children have more persistent disease. The skin barrier defect that results from filaggrin protein alteration and that is associated with AD may also be associated with a currently unexplained evolutionary advantage that seems to predominate in white populations.41,42
Limitations
Similar to other studies of AD and FLG, we rely on the convention that FLG LoF variants are all stop-gain variants of exon 3 and that these variants alter filaggrin protein function.17 However, limited direct evidence indicates that most of the described variants alter FLG protein function. The PEER subgroup of participants who agreed to provide DNA was not a random sample of PEER, and PEER is not a random sample of the US population, so our cohorts may not generalize to the US population. Analysis of copy number variation was not performed for repeats 8 and 10 of exon 3, the number of which has been shown to be associated with AD risk.13,21,27 Finally, owing to the low frequency of many of the variants, our study may not be large enough to demonstrate statistical associations or to estimate their frequency precisely.
Conclusions
We have presented a comprehensive listing of FLG LoF variants in a US cohort of children with mild to moderate AD. These variants differ greatly by race and their influence on the persistence of AD. This difference can be profound in that very common variants in one race may be absent or very rare in another, and their location within exon 3 may also vary widely. Overall, white children with AD are more than twice as likely to have an FLG LoF variant as African American children, and African American children have more prevalent and persistent AD. Relying on the classic 4 FLG LoF variants would result in approximately 8% of white children and 64% of African American children with an FLG LoF variant being improperly classified. Any planned genetic diagnostic test for AD based on FLG LoF variants must be inclusive and not rely on the most frequently studied variants. Furthermore, use of testing needs to carefully differentiate between disease diagnosis and disease prognostication.
References
- 1.Akdis CA, Akdis M, Bieber T, et al. ; European Academy of Allergology; Clinical Immunology/American Academy of Allergy, Asthma and Immunology/PRACTALL Consensus Group . Diagnosis and treatment of atopic dermatitis in children and adults: European Academy of Allergology and Clinical Immunology/American Academy of Allergy, Asthma and Immunology/PRACTALL Consensus Report. Allergy. 2006;61(8):969-987. doi: 10.1111/j.1398-9995.2006.01153.x [DOI] [PubMed] [Google Scholar]
- 2.Charman CR, Williams HC. Epidemiology In: Bieber T, Leung DYM, eds. Atopic Dermatitis. Vol 1 New York, NY: Marcel Dekker Inc; 2002:21-42. doi: 10.3109/9780203908877-3 [DOI] [Google Scholar]
- 3.Abuabara K, Magyari A, McCulloch CE, Linos E, Margolis DJ, Langan SM. Prevalence of atopic eczema among patients seen in primary care: data from the Health Improvement Network. Ann Intern Med. 2019;170(5):354-356. doi: 10.7326/M18-2246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Abuabara K, Margolis DJ. Do children really outgrow their eczema, or is there more than one eczema? J Allergy Clin Immunol. 2013;132(5):1139-1140. doi: 10.1016/j.jaci.2013.08.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Margolis JS, Abuabara K, Bilker W, Hoffstad O, Margolis DJ. Persistence of mild to moderate atopic dermatitis. JAMA Dermatol. 2014;150(6):593-600. doi: 10.1001/jamadermatol.2013.10271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Silverberg JI, Hanifin JM. Adult eczema prevalence and associations with asthma and other variables: a US population-based study. J Allergy Clin Immunol. 2013;132(5):1132-1138. doi: 10.1016/j.jaci.2013.08.031 [DOI] [PubMed] [Google Scholar]
- 7.Silverberg JI, Simpson EL, Durkin HG, Joks R. Prevalence of allergic disease in foreign-born American children. JAMA Pediatr. 2013;167(6):554-560. doi: 10.1001/jamapediatrics.2013.1319 [DOI] [PubMed] [Google Scholar]
- 8.Shaw TE, Currie GP, Koudelka CW, Simpson EL. Eczema prevalence in the United States: data from the 2003 National Survey of Children’s Health. J Invest Dermatol. 2011;131(1):67-73. doi: 10.1038/jid.2010.251 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bickers DR, Lim HW, Margolis D, et al. ; American Academy of Dermatology Association; Society for Investigative Dermatology . The burden of skin diseases: 2004: a joint project of the American Academy of Dermatology Association and the Society for Investigative Dermatology. J Am Acad Dermatol. 2006;55(3):490-500. doi: 10.1016/j.jaad.2006.05.048 [DOI] [PubMed] [Google Scholar]
- 10.Eichenfield LF, Tom WL, Chamlin SL, et al. Guidelines of care for the management of atopic dermatitis, section 1: diagnosis and assessment of atopic dermatitis. J Am Acad Dermatol. 2014;70(2):338-351. doi: 10.1016/j.jaad.2013.10.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Leung DY, Bieber T. Atopic dermatitis. Lancet. 2003;361(9352):151-160. doi: 10.1016/S0140-6736(03)12193-9 [DOI] [PubMed] [Google Scholar]
- 12.Ober C, Yao TC. The genetics of asthma and allergic disease: a 21st century perspective. Immunol Rev. 2011;242(1):10-30. doi: 10.1111/j.1600-065X.2011.01029.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Brown SJ, McLean WH. One remarkable molecule: filaggrin. J Invest Dermatol. 2012;132(3, pt 2):751-762. doi: 10.1038/jid.2011.393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Margolis DJ, Apter AJ, Gupta J, et al. The persistence of atopic dermatitis and filaggrin (FLG) mutations in a US longitudinal cohort. J Allergy Clin Immunol. 2012;130(4):912-917. doi: 10.1016/j.jaci.2012.07.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Margolis DJ, Hoffstad O, Bilker W. Lack of association between exposure to topical calcineurin inhibitors and skin cancer in adults. Dermatology. 2007;214(4):289-295. doi: 10.1159/000100879 [DOI] [PubMed] [Google Scholar]
- 16.Palmer CN, Irvine AD, Terron-Kwiatkowski A, et al. Common loss-of-function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis [see comment]. Nat Genet. 2006;38(4):441-446. doi: 10.1038/ng1767 [DOI] [PubMed] [Google Scholar]
- 17.Palmer CN, Ismail T, Lee SP, et al. Filaggrin null mutations are associated with increased asthma severity in children and young adults. J Allergy Clin Immunol. 2007;120(1):64-68. doi: 10.1016/j.jaci.2007.04.001 [DOI] [PubMed] [Google Scholar]
- 18.Brown SJ, Relton CL, Liao H, et al. Filaggrin haploinsufficiency is highly penetrant and is associated with increased severity of eczema: further delineation of the skin phenotype in a prospective epidemiological study of 792 school children. Br J Dermatol. 2009;161(4):884-889. doi: 10.1111/j.1365-2133.2009.09339.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Brown SJ, Relton CL, Liao H, et al. Filaggrin null mutations and childhood atopic eczema: a population-based case-control study. J Allergy Clin Immunol. 2008;121(4):940-946.e3. doi: 10.1016/j.jaci.2008.01.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Margolis DJ, Mitra N, Gochnauer H, et al. Uncommon filaggrin variants are associated with persistent atopic dermatitis in African Americans. J Invest Dermatol. 2018;138(7):1501-1506. doi: 10.1016/j.jid.2018.01.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mathyer ME, Quiggle AM, Wong XFCC, et al. Tiled array-based sequencing identifies enrichment of loss-of-function variants in the highly homologous filaggrin gene in African-American children with severe atopic dermatitis. Exp Dermatol. 2018;27(9):989-992. doi: 10.1111/exd.13691 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Polcari I, Becker L, Stein SL, Smith MS, Paller AS. Filaggrin gene mutations in African Americans with both ichthyosis vulgaris and atopic dermatitis. Pediatr Dermatol. 2014;31(4):489-492. doi: 10.1111/pde.12355 [DOI] [PubMed] [Google Scholar]
- 23.Margolis DJ, Gupta J, Apter AJ, et al. Exome sequencing of filaggrin and related genes in African-American children with atopic dermatitis. J Invest Dermatol. 2014;134(8):2272-2274. doi: 10.1038/jid.2014.126 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Winge MC, Bilcha KD, Liedén A, et al. Novel filaggrin mutation but no other loss-of-function variants found in Ethiopian patients with atopic dermatitis. Br J Dermatol. 2011;165(5):1074-1080. doi: 10.1111/j.1365-2133.2011.10475.x [DOI] [PubMed] [Google Scholar]
- 25.Lek M, Karczewski KJ, Minikel EV, et al. ; Exome Aggregation Consortium . Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536(7616):285-291. doi: 10.1038/nature19057 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sandilands A, Terron-Kwiatkowski A, Hull PR, et al. Comprehensive analysis of the gene encoding filaggrin uncovers prevalent and rare mutations in ichthyosis vulgaris and atopic eczema. Nat Genet. 2007;39(5):650-654. doi: 10.1038/ng2020 [DOI] [PubMed] [Google Scholar]
- 27.Wong XFCC, Denil SLIJ, Foo JN, et al. Array-based sequencing of filaggrin gene for comprehensive detection of disease-associated variants. J Allergy Clin Immunol. 2018;141(2):814-816. doi: 10.1016/j.jaci.2017.10.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Taylan F, Nilsson D, Asad S, et al. Whole-exome sequencing of Ethiopian patients with ichthyosis vulgaris and atopic dermatitis. J Allergy Clin Immunol. 2015;136(2):507-509.e19. doi: 10.1016/j.jaci.2015.02.010 [DOI] [PubMed] [Google Scholar]
- 29.Thawer-Esmail F, Jakasa I, Todd G, et al. South African amaXhosa patients with atopic dermatitis have decreased levels of filaggrin breakdown products but no loss-of-function mutations in filaggrin. J Allergy Clin Immunol. 2014;133(1):280-282.e2. doi: 10.1016/j.jaci.2013.09.053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Pigors M, Common JEA, Wong XFCC, et al. Exome sequencing and rare variant analysis reveals multiple filaggrin mutations in Bangladeshi families with atopic eczema and additional risk genes. J Invest Dermatol. 2018;138(12):2674-2677. doi: 10.1016/j.jid.2018.05.013 [DOI] [PubMed] [Google Scholar]
- 31.Chang J, Mitra N, Hoffstad O, Margolis DJ. Association of filaggrin loss of function and thymic stromal lymphopoietin variation with treatment use in pediatric atopic dermatitis. JAMA Dermatol. 2017;153(3):275-281. doi: 10.1001/jamadermatol.2016.4467 [DOI] [PubMed] [Google Scholar]
- 32.Margolis DJ, Abuabara K, Hoffstad OJ, Wan J, Raimondo D, Bilker WB. Association between malignancy and topical use of pimecrolimus. JAMA Dermatol. 2015;151(6):594-599. doi: 10.1001/jamadermatol.2014.4305 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Li H, Durbin R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics. 2010;26(5):589-595. doi: 10.1093/bioinformatics/btp698 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43(5):491-498. doi: 10.1038/ng.806 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.McKenna A, Hanna M, Banks E, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297-1303. doi: 10.1101/gr.107524.110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Van der Auwera GA, Carneiro MO, Hartl C, et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Curr Protoc Bioinformatics. 2013;43:1-33. doi: 10.1002/0471250953.bi1110s43 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Poplin R, Ruano-Rubio V, DePristo MA, et al. Scaling accurate genetic variant discovery to tens of thousands of samples [published online July 24, 2018]. bioRxiv. doi: 10.1101/201178 [DOI] [Google Scholar]
- 38.Brown SJ, Kroboth K, Sandilands A, et al. Intragenic copy number variation within filaggrin contributes to the risk of atopic dermatitis with a dose-dependent effect. J Invest Dermatol. 2012;132(1):98-104. doi: 10.1038/jid.2011.342 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Chang J, Bilker WB, Hoffstad O, Margolis DJ. Cross-sectional comparisons of patient-reported disease control, disease severity and symptom frequency in children with atopic dermatitis. Br J Dermatol. 2017;177(4):e114-e115. doi: 10.1111/bjd.15403 [DOI] [PubMed] [Google Scholar]
- 40.Irvine AD, McLean WH, Leung DY. Filaggrin mutations associated with skin and allergic diseases [review]. N Engl J Med. 2011;365(14):1315-1327. doi: 10.1056/NEJMra1011040 [DOI] [PubMed] [Google Scholar]
- 41.Eaaswarkhanth M, Xu D, Flanagan C, et al. Atopic dermatitis susceptibility variants in filaggrin hitchhike hornerin selective sweep. Genome Biol Evol. 2016;8(10):3240-3255. doi: 10.1093/gbe/evw242 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Amos W, Hoffman JI. Evidence that two main bottleneck events shaped modern human genetic diversity. Proc Biol Sci. 2010;277(1678):131-137. doi: 10.1098/rspb.2009.1473 [DOI] [PMC free article] [PubMed] [Google Scholar]