|Home | About | Journals | Submit | Contact Us | Français|
Small proline rich protein 2B (SPRR2B) is a skin and lung epithelial protein associated with allergic inflammation in mice that has not been evaluated in human atopic diseases.
To determine whether single-nucleotide polymorphisms (SNPs) in SPRR2B are associated with childhood eczema and with the phenotype of childhood eczema combined with asthma.
Genotyping for SPRR2B and filaggrin (FLG) was performed in 2 independent populations: the Cincinnati Childhood Allergy & Air Pollution Study (CCAAPS; N = 762; birth-age, 4 years) and the Greater Cincinnati Pediatric Clinical Repository (GCPCR;N = 1152; ages 5–10 years). Eczema and eczema plus asthma were clinical outcomes based on parental report and clinician’s diagnosis. Genetic analyses were restricted to whites and adjusted for sex in both cohorts and adjusted for environmental covariates in CCAAPS.
Variants in SPRR2B were not significantly associated with eczema in either cohort after Bonferroni adjustment. Children from both cohorts with the CC genotype of the SPRR2B rs6693927 SNP were at 4 times the risk for eczema plus asthma (adjusted odds ratio, 4.1; 95% confidence interval, 1.5– 10.9; P = .005 in CCAAPS; and adjusted odds ratio, 4.0; 95% confidence interval, 1.8 –9.1; P <.001 in the GCPCR), however. SNPs in SPRR2B were not in strong linkage disequilibrium with the R501X and del2282 FLG mutations, and these findings were independent of FLG.
An SNP in SPRR2B was predictive of asthma among white children with eczema from 2 independent populations. SPRR2B polymorphisms may serve as important predictive markers for the combined eczema plus asthma phenotype.
Eczema is a chronic inflammatory skin condition commonly associated with other atopic diseases, such as asthma.1,2 Genetic variants in epidermal proteins, such as filaggrin (FLG), are thought to play a critical role in the development of eczema and have been associated with the combination of eczema plus asthma.2–6 The FLG gene is located within the epidermal differentiation complex, a region on chromosome 1q21 that also encodes for numerous other proteins involved in maintaining the skin barrier.7–10
Linkage and expression studies indicate that FLG alone does not account for the strong association between region 1q21 and eczema nor for the possible association between this region and the combined phenotype of eczema plus asthma.4,11,12 Small proline rich proteins (SPRRs) are cross-linking components of the keratinocyte-cornified envelope encoded on 1q21 that may be critically important to the formation of the skin barrier.13,14 SPRR2B in particular has been shown to be induced by allergens and the TH2 cytokine, interleukin 13, in mouse models of asthma and allergic gastrointestinal disease.15,16 Moreover, SPRR2B messenger RNA expression is increased during asthma exacerbations, and large allele frequency differences exist between human populations with high rates of asthma vs those with a lower prevalence of the disease.17,18 Together, these findings suggest that SPRR2B may play a role in both eczema and asthma.
We hypothesized that polymorphisms in SPRR2B would be significantly associated with childhood eczema and with the eczema plus asthma phenotype. To test this hypothesis, we examined differences in the frequency of single-nucleotide polymorphisms (SNPs) in SPRR2B between children with and without eczema, and eczema plus asthma who were participating in the Cincinnati Childhood Allergy & Air Pollution Study (CCAAPS) and the Greater Cincinnati Pediatric Clinic Repository (GCPCR) at Cincinnati Children’s Hospital Medical Center.18,19
CCAAPS is a longitudinal birth cohort at high risk for atopy.19 Newborns in the Cincinnati metropolitan area were identified by public birth records from 2001 to 2003. All 762 infants had a parent with symptoms of asthma, allergic rhinitis, or eczema and with at least one positive skin prick test result to a panel of 15 aeroallergens. Parents signed an informed consent approved by the institutional review board of the University of Cincinnati. The GCPCR (N = 1152) includes children 5 to 17 years of age from subspecialty clinics at Cincinnati Children’s Hospital and control subjects from the general population.18 Parents signed an informed consent approved by the Cincinnati Children’s Hospital institutional review board. (Further details are provided in Table 1 and in the eMethods.)
The definition of eczema in CCAAPS was adapted from a validated questionnaire (International Study of Asthma and Allergies in Childhood) and included a parental report of the child’s scratching and redness, “raised bumps,” or dry skin/scaling for at least 6 of the last 12 months .20 Children who tested positive for eczema at any point from 1 through 4 years of age were considered cases. This definition was established a priori and was previously used in CCAAPS.21 For the GCPCR, cases with eczema included children with a physician’s diagnosis or a parental report of eczema. A separate analysis with eczema based on the clinician’s diagnosis at 1, 2, 3, or 4 years of age was performed in CCAAPS. Physical examination findings considered to be consistent with eczema included erythema, papulation, excoriations, and/or lichenification.
Having eczema at any point from 1 through 4 years of age combined with asthma at the age of 4 years served as a secondary outcome measure for CCAAPS.2 Given that objective testing for asthma could not be performed in CCAAPS due to the age of the children, a case definition for asthma at the age of 4 years was developed; this case definition included a parental report of persistent wheezing (≥2 wheezing episodes not associated with a cold or upper respiratory tract infection) at both 3 and 4 years of age or a diagnosis of asthma by a child’s physician by 4 years of age.22 Although this case definition is not asthma per se, for the sake of simplicity it will be referred to as asthma for the remainder of the article. The reference group for the eczema plus asthma outcome included children with neither condition or just eczema or asthma alone (but not both). For the GCPCR, cases included children with both eczema and asthma, whereas the reference group had neither condition or just eczema or asthma alone. In both CCAAPS and the GCPCR, a separate analysis was performed in which only children with neither eczema nor asthma were included in the reference group. (Additional details are provided in the eMethods.)
A priori, we decided to use the GCPCR as a replication cohort. Therefore, SNPs evaluated in CCAAPS were the same as those analyzed concurrently in the GCPCR. The SPRR2B SNPs chosen included 1 intronic tagging SNP (rs6693927), 2 SNPs in the 5′ and 3′ untranslated regions (rs7525198 and rs6673356, respectively), and a missense mutation (rs404408).18 Tagging SNPs and SNPs in regulatory regions with minor allele frequency greater than 5% across the SPRR2B gene were chosen to adequately represent the gene, as previously described.23
SPRR2B genotyping for CCAAPS and the GCPCR was accomplished using a LightTyper platform and a custom Illumina Golden Assay, respectively.18,24,25 The rs6693927 and rs7525198 SNPs were genotyped with error rates of less than 2% and were in Hardy Weinberg equilibrium (eTable 1). The rs6673356 and rs404408 SNPs were not included in further analyses because genotyping results did not meet quality controls. Genotyping of FLG R501X was accomplished by polymerase chain reaction and restriction fragment length polymorphism in CCAAPS and the GCPCR as previously described (eMethods).6,26 Primer sequences and genotyping methods for FLG 2282del4 in the GCPCR, as well as further information regarding genotyping of SPRR2B, are provided in the online supplement.
SAS statistical software (SAS Institute Inc, Cary, North Carolina) was used to perform multivariate logistic regression to adjust findings in CCAAPS for host and environmental covariates, including income under $50,000, breastfeeding (yes/no), dog and/or cat ownership (yes/no based on home visit), and house dust endotoxin exposure at 1 year of age (based on home visit). All potential covariates listed in eTable 2 were initially included in the maximum likelihood estimated model and removed if P > .15. Analyses in the GCPCR were adjusted for sex. Environmental covariates from the GCPCR were not available at the time of this analysis. Analyses in both cohorts were restricted to whites given the small sample size of African Americans. A priori, children older than 10 years were excluded from the analysis in the GCPCR to improve comparability with CCAAPS and to decrease recall bias related to transient eczema occurring at younger ages.27
A genotypic analysis for SPRR2B was performed. This model was chosen given that the mode of inheritance is not known and an additive model may be less likely to detect an association with relevant outcomes in the case of recessive traits.28 To perform a genotypic analysis, we used logistic regression to calculate adjusted odds ratios (aORs), 95% confidence intervals (CIs), and P values that represented the relationship between genotypes of rs6693927 and rs7525198 (eg, TT vs all others; CC vs all others; CT vs all others) and the outcomes of eczema and eczema plus asthma. For rs6693927, TT represents a homozygous recessive with 2 TT alleles, CT represents a heterozygote, and CC represents a homozygous recessive with 2 CC alleles. A haplotype analysis using PLINK software was also conducted to determine whether there were additional effects from combined genotypes of rs6693927 and rs7525198.29 Linkage disequilibrium (LD) (D′ and r2) between SPRR2B and FLG variants was estimated using Haploview.30 A LD-adjusted Bonferroni calculation (0.03) was applied to genetic findings to account for multiple comparisons.
In the CCAAPS cohort, 161 of 396 whites (41%) had eczema for at least 1 year from 1 through 4 years of age, whereas 59 of 328 whites (18%) ages 5 to 10 years in the GCPCR had the disease (Table 2). Among whites with eczema from CCAAPS, 144 of 161 (89%) were atopic based on 1 positive skin prick test result to a panel of 15 aeroallergens at any point from 1 through 4 years of age; skin prick test results were not available for the GCPCR. Findings related to host and environmental covariates in CCAAPS are presented in eTable 3.
Among whites from CCAAPS, the TT genotype of the rs6693927 SNP (TT_rs6693927) in SPRR2B was protective against eczema after adjustment for sex (aOR, 0.5; 95% CI, 0.2–0.9; P = .03); however, this finding did not meet criteria for significance based on LD-adjusted Bonferroni correction (P < .03 was significant). There was a trend toward an association with eczema after adjustment for environmental covariates (aOR, 0.5; 95% CI, 0.3–1.1; P = .09; Fig 1). Findings for TT_rs6693927 in the GCPCR study population were not statistically significant (Fig 1 and eTable 4). The SPRR2B rs7525198 SNP was not associated with eczema in either CCAAPS or the GCPCR. In addition, the combination of rs7525198 and rs6693927 was not more predictive of eczema than rs6693927 alone, based on findings from combined genotype (diplotype) and haplotype analyses (data not shown).29
We next evaluated the association between polymorphisms in SPRR2B and the presence of both eczema and asthma.2 As indicated in Table 2, 25 whites (7%) from CCAAPS and 28 whites (9%) from the GCPCR had both conditions. An odds ratio for TT_rs6693927 could not be calculated for CCAAPS because none of the children with this genotype had the phenotype of eczema plus asthma; the odds ratio for the CC genotype of the rs6693927 SNP (CC_rs6693927) was therefore calculated instead. Among whites from both CCAAPS and the GCPCR, CC_rs6693927 conferred a 4-fold increased risk for the eczema plus asthma phenotype (aOR, 4.1; 95% CI, 1.5–10.9; P = .005 after adjustment for host and environmental covariates from the logistic regression model in CCAAPS; and aOR, 4.0; 95% CI, 1.8 –9.1; P < .001 after adjustment for sex in the GCPCR; Fig 2 and eTable 5). These findings met the LD-adjusted Bonferroni criteria for significance (P < .03). As was the case with eczema, rs7525198 did not confer significant effects for eczema plus asthma by itself, and the combination of rs7525198 and rs6693927 was not more predictive of eczema plus asthma than the latter SNP alone (data not shown).29
As a secondary analysis, we also evaluated eczema defined solely by clinician’s diagnosis at 1, 2, 3, or 4 years of age in CCAAPS. Findings were similar to those seen with the survey definition of eczema in CCAAPS. Among whites ages 1 to 4 years, 117 (44%) had clinician-diagnosed eczema and 149 (56%) did not. There was a possible protective effect from clinician-diagnosed eczema conferred by the TT_6693927 genotype after adjustment for host and environmental covariates (aOR, 0.6; 95% CI, 0.4 –1.1; P = .10; Fig 3); however, this finding was not statistically significant.
We also evaluated the relationship between the eczema plus asthma phenotype using the clinician’s diagnosis of eczema definition in CCAAPS. Findings were strikingly similar to those seen with the survey definition of eczema. There were 16 children (7%) with the clinician’s diagnosis of eczema plus asthma and 217 children (93%) in the reference group. A strong association was found between the CC_6693927 genotype and the combination of clinician-diagnosed eczema plus asthma (aOR, 5.3; 95% CI, 1.6–17.4; P = .005 after adjustment for host and environmental covariates from the logistic regression model; Fig 4). These findings met the LD-adjusted Bonferroni criteria for significance (P < .03).
To further clarify the relationship between polymorphisms in SPRR2B and the eczema plus asthma phenotype, an additional analysis was conducted in which the reference group included only children with neither eczema nor asthma (n = 195 in CCAAPS and n = 100 in the GCPCR). Findings from this analysis were nearly identical to those previously reported. That is, CC_rs6693927 conferred increased risk for eczema plus asthma (aOR, 4.5; 95% CI, 1.6–13.1; P = .005 after adjustment for host and environmental covariates from the logistic regression model in CCAAPS; and aOR, 4.6; 95% CI, 1.9–11.3; P < .001 in the GCPCR). These findings met the LD-adjusted Bonferroni criteria for significance (P < .03).
To better understand the relationship between the rs6693927 SNP and the eczema plus asthma phenotype, we also evaluated the relationship between this SNP and asthma alone (without eczema). Among whites from CCAAPS, there was a trend toward a protective effect from asthma conferred by the TT_6693927 genotype after adjustment for host and environmental covariates (aOR, 0.3; 95% CI, 0.09–1.1; P = .06; Fig 5). Among whites 5 to 10 years of age from the GCPCR, there was also a trend toward increased risk for asthma alone conferred by the CC_6693927 genotype (aOR, 1.6; 95% CI, 1.0 –2.7; P = .06).
Given the previously reported association between FLG null mutations and the eczema plus asthma phenotype and that both FLG and SPRR2B are located on chromosome 1q21, we next determined whether the effects of SPRR2B were related to LD with FLG. FLG null mutations were not in strong LD with SPRR2B polymorphisms (r2 < .05, D′ < .68 for R501X and 2282del4 in the GCPCR, with similar findings for R501X in CCAAPS).
Evidence from the literature has established that epidermal barrier function plays a critical role in eczema and possibly in the phenotype of eczema plus asthma.1,2,5 The combination of eczema plus asthma may represent a distinct endophenotype with unique pathogenic and prognostic implications.5,31 Identification of risk factors that may modify the risk of asthma among children with eczema is important.31,32 In addition, improved understanding of the relationship between eczema and asthma may provide mechanistic insights into the pathogenesis of asthma.5,32 FLG mutations alone are unlikely to account for the strong genetic linkage between the epidermal differentiation complex on chromosome 1q21 and the eczema plus asthma phenotype.1,2 Given this premise and based on preliminary data from RNA expression studies and allele frequency differences in asthmatic populations, we explored the relationship among childhood eczema, the combination of eczema plus asthma, and SPRR2B, an epithelial protein encoded on 1q21.17,18 We demonstrated a possible association between an SNP in SPRR2B and eczema among whites 4 years and younger from the CCAAPS cohort; however, this finding was not statistically significant after Bonferroni correction. A nonsignificant trend toward an association between this SNP and asthma was also seen in both CCAAPS and the GCPCR. Interestingly, in both cohorts, the CC genotype of the rs6693927 SNP conferred strong, statistically significant effects among children who had the eczema plus asthma phenotype. That is, the rs6693927 CC genotype conferred an approximately 4-fold increased risk for eczema plus asthma in both CCAAPS and the GCPCR. Findings for this SNP did not appear to be related to FLG polymorphisms. Thus, polymorphisms in SPRR2B may independently serve as useful markers to predict the development of asthma among children with eczema.
Mechanistic studies to explain the reasons for an association between SPRR2B polymorphisms and the eczema plus asthma phenotype have yet to be conducted. Alterations in SPRR2B may potentially result in a defective skin barrier. Small proline rich proteins, including SPRR2B, play an important role in the formation of the keratinocyte-cornified envelope.33 After cross-linking by transglutaminases, they serve as the scaffold on which other epithelial proteins are added.34 It is not clear, however, whether the link between SPRR2B and asthma is entirely mediated by effects on the skin barrier. In contrast to FLG, which is not expressed in the lungs, SPRRs are expressed in multiple tissues throughout the body. SPRR2B in particular is expressed in the lungs of mice, and its expression is specifically altered in the nasal epithelium of asthmatic patients.15,17,34 It has also been speculated that the altered expression of SPRR genes as part of a stress response to UV light, infection, or allergic inflammatory signals may reflect a role in tissue repair.34–36 This implies that quantitative or functional abnormalities involving SPRRs might lead to a dysfunctional repair response with the potential for pathologically adverse outcomes. Alternatively, up-regulation of SPRR2B during asthma exacerbations may reflect another undiscovered role of this protein in allergic inflammation.17,34
Although there is a possible association between the rs6693927 polymorphism in SPRR2B and eczema and a strong relationship with the combination of eczema and asthma, there are limitations to what may be concluded from these findings. It is possible that this association reflects genetic linkage with other epidermal barrier proteins encoded on region 1q21, besides FLG, and/or that SPRR2B is coexpressed with other SPRRs that are responsible for clinically meaningful responses. The rs6693927 SNP has not been clearly related to changes in SPRR2B protein function or expression, and the possibility exists that this SNP may be a marker of disease causing mutations rather than disease causing itself. Another possible limitation is that there were relatively small numbers of children with the eczema plus asthma phenotype from both cohorts. There were also differences in the cohorts studied, including that they enrolled children of different ages and that CCAAPS is a high-risk birth cohort, whereas the GCPCR includes both atopic and control children. Differences in the prevalence of eczema between the cohorts are reflective of these factors. Findings related to asthma in the CCAAPS cohort will be validated once the children are old enough to undergo objective testing. Despite these differences, the prevalence of the eczema plus asthma phenotype was similar between cohorts; further, odds ratios for the effects of SPRR2B polymorphisms were almost identical. In addition, findings were similar despite different outcome definitions between cohorts and were comparable in CCAAPS regardless of the outcome definition used.
In conclusion, the rs6693927 SNP in SPRR2B was significantly associated with the eczema plus asthma phenotype among whites from both cohorts. It was not significantly associated with eczema in CCAAPs or in the GCPCR. Findings were independent of FLG polymorphisms and were robust despite potential differences in the cohorts.
Funding Sources: This study was supported by grants U19AI70235 from the National Institutes of Health, ES11170 from the National Institute of Environmental Health Sciences, T32 AI60515 from the National Institute of Allergy and Infectious Diseases, and P30ES006096 from the National Institutes of Health.
Disclosures: Authors have nothing to disclose.
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.anai.2012.01.004.