SEARCH

SEARCH BY CITATION

Keywords:

  • children;
  • eczema;
  • haplotype;
  • single nucleotide polymorphism;
  • vitamin D

Abstract

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Background

Vitamin D is increasingly recognized to play crucial roles in cutaneous immunity, and vitamin D treatment improved eczema control in small clinical trials. Several vitamin D-related genes were associated with asthma, but there are no data for eczema.

Methods

Twenty-three single-nucleotide polymorphisms (SNPs) of five vitamin D-related genes (CYP27A1, CYP2R1, CYP27B1, GC and VDR) were genotyped in 1442 Chinese children with eczema and 1231 non-allergic controls. SNPs that followed Hardy–Weinberg equilibrium and yielded ≥95% genotyping call-rate were included. Haplotypic associations and SNP–SNP interactions for eczema diagnosis and subphenotypes were analysed.

Results

Atopic eczema was associated with rs4674343 of CYP27A1 (odds ratio 0.66, 95% confidence interval 0.53–0.83, P = 0.0004). Increased eosinophil percentage was associated with CYP2R1 rs2060793A (P = 0.001) and rs1933064A (P = 0.001). Two CYP2R1 haplotypes increased eczema risk whereas one VDR haplotype lowered eczema risk. GC rs7041 and CYP2R1 rs7935792 interacted to modulate total IgE (cross-validation consistency 10/10, P = 0.047). Specifically, high-risk eczema patients had higher log-transformed total IgE than low-risk patients (2.76 ± 0.76 vs 2.60 ± 0.80, P = 0.002).

Conclusion

A vitamin D-related SNP rs4674343 on CYP27A1 was found to be protective against atopic eczema. CYP2R1 and VDR haplotypes altered eczema susceptibility and eosinophil percentage, and GC and CYP2R1 interacted to determine total IgE among eczema patients.

Despite the well-known effect on bone health, vitamin D pathway emerged as a new regulatory mechanism for immune responses [1]. There are two sources of vitamin D, with the major one being the conversion of 7-dehydrocholesterol to provitamin D3 by UVB from sunlight exposure. Intake of vitamin D2 (ergocalciferol) or D3 from foods such as oily fish and cod liver oil is the other source. Inactive vitamin D is then metabolized to the major circulating form 25(OH)D, with a long half-life of more than 2 weeks, by CYP2R1 in microsomes and CYP27A1 in liver that encode 25-hydroxylase. The activated form of vitamin D (1α,25(OH)2D or calcitriol) is made following further hydroxylation step at C-1 position in the kidney by 1α-hydroxylase (CYP27B1), which is transported via circulation to the nuclei of target cells by the group-specific components of vitamin D-binding protein (GC). The binding of 1α,25(OH)2D to vitamin D receptor (VDR), a ligand-activated transcription factor and nuclear hormone receptor, results in the transcription of vitamin D-responsive genes [2].

Vitamin D was reported to be involved in inflammatory pathways and development of allergies [3]. VDR was found in most cells (T lymphocytes, neutrophils, macrophages, dendritic cells) of immune system. Vitamin D contributes to the increased production of cathelicidin and nitric oxide [4] in macrophages and inhibition of maturation and migration of dendritic cells. Vitamin D reduced the production of type 1 T helper cells by blocking IL-12 and induced type 2 T helper cell (Th2)-related cytokines IL-4, IL-5, and IL-10 [5]. Based on these in vitro findings, higher vitamin D levels are expected in eczema cases. Nonetheless, serum 25-hydroxyvitamin D levels showed inverse correlation with eczema severity [6], which might be due to reverse causation when patients with severe eczema were more likely to protect their skin from sun exposure. Thyssen et al. [7] reported that skin barrier abnormalities caused by filaggrin gene mutations were also associated with increased vitamin D levels. Association studies of vitamin D pathway genes will offer another approach to delineate the relationship between eczema and vitamin D.

Different research groups reported contradictory results of genetic variants encoding vitamin D pathway in the development of allergic conditions. VDR rs7975232 was associated with asthma in US children, but such result could not be replicated in a French Canadian family population. Nonetheless, the latter study reported six VDR single nucleotide polymorphisms (SNPs) to be weakly associated with asthma [8, 9]. A recent genome-wide association study (GWAS) of circulating vitamin D level found that rs2282679 of GC and rs10741657 of CYP2R1 were associated with 25(OH)D specifically among asthmatic children [10]. Two case–control studies between asthma and genes along the vitamin D pathway were carried out in northern Chinese. Two common GC functional SNPs, rs4588 and rs7041, and VDR rs7975232 were found to be associated with asthma [11, 12], whereas CYP2R1 showed no significant result. At present, there is only one report on possible relationship between eczema and vitamin D pathway genes. Heine and coworkers investigated the frequency of four common VDR polymorphisms in 265 adults with moderate-to-severe eczema and 265 healthy controls [13]. They found an association between severe eczema and VDRs rs1544410, rs7975232, and rs731236 as well as their haplotypes. This study investigated the association between eczema and SNPs of five genes (CYP27A1, CYP2R1, CYP27B1, GC, and VDR) along the vitamin D pathway in Hong Kong Chinese children.

Patients and methods

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Study population

This case–control study consisted of 1442 unrelated Chinese children with eczema and 1231 nonallergic controls who were recruited from several hospital-based and community studies conducted between January 2009 and December 2011 [14]. Patients with eczema were younger than 18 years, and their eczema was physician-diagnosed according to the widely accepted criteria [15]. All controls did not report any eczema, asthma, and allergic rhinitis by the Chinese International Study of Asthma and Allergies in Childhood questionnaire [16, 17], and they were recruited from (i) children attending a university teaching hospital for minor nonrespiratory and nonimmunologic complaints; and (ii) secondary schoolchildren who participated in an epidemiological study for obesity and diabetes [18]. All subjects were free from any infectious symptoms for 4 weeks before study. Subjects and their parents gave informed written consent, and Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee approved this study.

Eczema subphenotypes

Plasma total IgE concentration was measured by microparticle immunoassay (IMx analyser; Abbott Laboratories, Abbott Park, IL, USA), which was logarithm-transformed before analysis (LogIgE). Peripheral blood eosinophils were enumerated using a Coulter STKS counter (Beckman Coulter, Miami, FL, USA), which was expressed as percentage of total leukocytes (eos%). Skin prick test (SPT) was performed with normal saline as negative control, histamine 10 mg/ml as positive control, and standardized extracts of locally prevalent allergens (ALK Abelló, Round Rock, TX, USA) [19, 29]. Wheal ≥3 mm larger than negative control was considered positive. Plasma allergen-specific IgE levels to Dermatophagoides pteronyssinus, cat and cockroaches were measured by fluorescent enzyme immunoassay (AutoCAP, Phadia AB, Uppsala, Sweden), with ≥0.35 kIU/l being positive. Atopy was defined as ≥ one positive SPT or allergen-specific IgE.

SNP Selection and genotyping

Table S1 describes 24 SNPs genotyped in this study. Eleven SNPs were reported to have significant association with asthma or serum vitamin D level (2 for VDR, 6 for GC, and 3 for CYP2R1). For CYP27A1, CYP2R1, and CYP27B1, tagging SNPs were also selected based on genotypic data for Han Chinese in Beijing (CHB) in HapMap phase 3 database using TagSNP provided by National Institute of Environmental Health Sciences (NIEHS) (http://snpinfo.niehs.nih.gov/snptag.htm). The tagging strategy was applied for SNPs with MAF ≥0.05 and pairwise r2 ≥ 0.8 that were within 5–10 kb upstream and downstream of the five target genes. SNPs published to have significant associations with asthma or serum vitamin D level were forced to be included. Thirteen tagging SNPs were selected (5 for each of CYP2R1 and CYP27A1 and 3 for CYP27B1). Potential functions of SNPs were predicted using FuncPred (http://snpinfo.niehs.nih.gov/snpinfo/snpfunc.htm) provided by NIEHS.

Rs6716642 of CYP27A1 could not be genotyped due to failed probe design by both iPLEX and TaqMan assays. Thirteen and 10 SNPs were genotyped by iPLEX® Gold assay (Sequenom, San Diego, CA, USA) and TaqMan® SNP Genotyping assay (Applied Biosystems, Foster City, CA, USA), respectively (see Data S1 in the Supporting Information).

Statistical analysis

Results were expressed in percentage or mean and standard deviation. Allele frequencies were estimated by gene counting method. Hardy–Weinberg equilibrium (HWE) for SNPs among cases and controls was evaluated by chi-square exact test. Pairwise linkage disequilibrium (LD) coefficient was calculated for each SNP pair by HaploView software (Daly Lab, Cambridge, MA, USA) [21]. Because of markedly skewed data distribution, peripheral blood eosinophils were dichotomized at ≥5% of total leukocytes before analysis. The associations for single SNP with dichotomous variables were analysed by logistic regression and with eczema subphenotypes by linear regression, adjusted for age, sex, and physician-diagnosed asthma as covariates. All comparisons were made two-tailed using R version 2.15.1 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria). The level of significance was set at 0.0022 (0.05/23) by Bonferroni correction to account for 23 SNPs being analysed. Haplotypes were determined by HaploView 4.2 according to LD coefficients. Haplotypic associations were performed using R package haplo.stats, version 1.6.3 (http://www.R-project.org/). Epistatic interactions between SNPs were analysed by generalized multifactor dimensionality reduction (GMDR) software, version 0.7 [14, 22, 23], the details of which are provided under Data S1 in the Supporting Information. One-way ANOVA and Kruskal–Wallis test with post hoc tests were then used to compare, respectively, LogIgE and eos% among different GMDR-assigned risk groups.

Results

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Subjects

Table 1 summarizes characteristics of our subjects. Controls (13.7 ± 4.5 years) were older than eczema (11.1 ± 4.3 years) and had fewer males. Data for total IgE and eos% were available in 978 and 847 eczema patients and 699 and 654 controls, respectively, and 1152 (91.3%) eczema patients were atopic.

Table 1. Population characteristics of eczema patients and controls
ParameteraEczemaControls P
  1. LogIgE, log-transformed plasma total IgE concentration.

  2. a

    Results expressed as mean ± standard deviation, median (25–75% quartile), or percentage.

  3. b

    Age ranges for cases and controls were 0.3–18.0 years and 0.4–20.9 years, respectively.

  4. c

    Defined as eosinophils ≥ 5% of total leukocytes in peripheral blood.

  5. d

    By either positive skin prick test or presence of allergen-specific IgE.

n 14421231 
Age (year)b11.1 ± 4.313.7 ± 4.5<0.001
Male52.2%44.0%<0.001
n 978699 
LogIgE2.67 ± 0.781.86 ± 0.69<0.001
n 847654 
Eosinophil percentage (%)6 (3–10)2 (1–3)<0.001
Increased peripheral blood eosinophilsc63.6%15.1%<0.001
n 1152394 
Atopyd91.3%37.9%<0.001

Single SNP and haplotype analyses

Table S2 shows the 23 SNPs that were successfully genotyped in over 99% of samples. All of them were in HWE. Three SNPs, namely rs11819875 (odds ratio [OR] 1.18, 95% confidence interval [CI] 1.03–1.36; = 0.014) of CYP2R1 and rs222020 (OR 1.27, 95% CI 1.00–1.60; = 0.047) and rs2298849 (OR 1.37, 95% CI 1.08–1.72; = 0.008) of GC showed a trend for eczema diagnosis (Table S3). Two VDR SNPs rs7975232 (OR 0.81, 95% CI 0.68–0.96; = 0.017) and rs2239185 (OR 0.82, 95% CI 0.69–0.98; = 0.025) were protective against eczema, but none of them survived Bonferroni correction. Rs4674343 of CYP27A1 was protective against atopic eczema (OR 0.66, 95% CI 0.53–0.83; = 0.0004). Besides, two SNPs of CYP2R1 (rs2060793 and rs1553006) were associated with increased eos% among all subjects (Table 2). Among eczema patients, subjects homozygous for the minor A allele of rs2060793 had higher eos% than those with other genotypes (median: 8% vs 3%, = 0.002). Similarly, those homozygous for A allele of rs1553006 had higher eos% than other subjects (median: 8% vs 6%, = 0.008). None of the SNPs was associated with LogIgE (data not shown).

Table 2. Logistic regression analysis of the association between 23 SNPs and increased peripheral blood eosinophils (n = 1501)
GeneSNPAdditive modelDominant modelRecessive model
OR95% CI P a OR95% CI P a OR95% CI P a
  1. CI, confidence interval; OR, odds ratio, SNP, single nucleotide polymorphism.

  2. SNPs with P-value <0.05 are highlighted in bold.

  3. a

    Adjusted for age, gender, and eczema status as covariates; level of significance at 0.0022 after Bonferroni correction.

CYP2R1 rs79357920.770.531.120.1740.690.451.060.0901.200.393.770.750
rs127947140.960.721.290.7940.910.601.370.6351.040.591.850.885
rs2060793 1.371.011.87 0.043 1.170.781.760.4393.041.546.03 0.001
rs169306090.960.631.470.8481.090.681.750.7200.220.051.050.058
rs118198750.850.621.160.3000.760.511.140.1820.990.501.970.973
rs169306251.080.691.690.7371.220.751.980.4360.250.041.500.129
rs107661970.830.621.110.2020.840.561.280.4210.670.381.180.165
rs1553006 1.451.081.95 0.015 1.260.841.910.2682.691.494.84 0.001
CYP27A1 rs174702710.860.451.640.6470.870.441.720.6810.570.0213.610.725
rs46743430.830.551.260.3880.870.551.360.5290.350.062.010.240
rs6451631.260.951.670.1171.240.801.920.3271.520.922.510.102
rs134068310.680.411.130.1360.710.411.230.2180.170.021.620.123
CYP27B1 rs10486911.130.751.680.5651.160.731.820.5351.070.293.990.917
rs46465381.160.582.300.6821.050.512.160.89022.910.471125.70.115
rs46465361.190.911.560.2100.120.741.700.5991.540.932.560.092
GC rs22826791.080.781.500.6380.940.631.420.7882.000.904.460.090
rs45881.040.751.440.8020.920.621.390.7001.740.793.860.170
rs70410.940.701.270.7020.860.571.290.4601.120.592.110.728
rs2220201.110.831.490.4801.010.651.550.9781.400.822.390.212
rs11555630.960.711.310.8160.830.561.250.3771.390.712.710.341
rs22988491.140.861.530.3641.100.721.680.6721.360.802.320.254
VDR rs79752321.150.851.560.3731.270.851.910.2391.000.511.971.000
rs22391851.100.811.500.5311.180.791.770.4151.000.502.020.998

Figure S1 illustrates LD plots of SNPs from five target genes. CYP2R1 and GC formed three and two haplotypes, respectively, and each of CYP27A1 and VDR constructed one haplotype, but no haplotype was found for CYP27B1. Table 3 summarizes the haplotypic associations. Two CYP2R1 haplotypes increased eczema risk, with GG of rs10766197_rs1553006 having OR 1.25 (95% CI 1.07–1.47) for eczema when compared with AG haplotype (= 0.006). Eczema was also weakly associated with the other CYP2R1 haplotype, AG of rs16930609_rs11819875 (= 0.043), whereas AT haplotype of VDR protected against eczema (= 0.047).

Table 3. Summary of haplotypic associations with eczema diagnosis
GeneHaplotype blockHaplotypeFrequency in eczemaFrequency in controlsOR95% CI P a
  1. a

    Adjusted for age, gender, and asthma diagnosis as covariates.

  2. Haplotypes with statistical significance are highlighted in bold.

CYP2R1 rs7935792_rs12794714AG0.4500.452Ref.   
AA0.3640.3770.940.821.080.388
CG0.1850.1711.120.951.340.186
rs16930609_rs11819875AT0.6880.708Ref.   
AG 0.1810.1671.191.011.40 0.043
CG0.1290.1231.180.981.420.087
rs10766197_rs1553006AG0.3740.378Ref.   
GA0.3640.3901.000.871.160.963
GG 0.2620.2321.251.071.47 0.006
CYP27A1 rs645163_rs13406831CT0.4870.493Ref.   
TT0.4290.4211.000.881.130.957
CC0.0830.0840.980.771.240.838
GC rs2282679_rs4588_rs7041ACT0.4260.395Ref.   
ACG0.3120.3270.950.821.090.441
CAT0.2550.2720.890.771.040.156
rs222020_rs1155563CT0.4330.412Ref.   
TC0.3210.3420.910.781.050.187
TT0.2460.2460.930.791.090.364
VDR rs7975232_rs2239185CC0.7040.685Ref.   
AT 0.2820.3000.870.761.00 0.047
AC0.0130.0131.020.601.720.956

Gene–gene interactions for eczema traits

Table 4 shows the results of GMDR analysis for LogIgE. The model between GC rs7041 and CYP2R1 rs7935792 yielded borderline significant results with testing accuracy 0.54, cross-validation consistency 10/10, and P-value 0.047 (see the Supporting Information), suggesting epistatic interaction between these SNPs for LogIgE. Figure 1A shows the predicted risk of all genotype combinations, whereas Fig. 1B illustrates box plots for the distribution of LogIgE according to GMDR-assigned risk groups among patients and controls. High-risk eczema patients had higher LogIgE than low-risk patients (2.76 ± 0.76 vs 2.60 ± 0.80, = 0.002), whereas LogIgE did not differ in relation to risk assignment among controls (1.89 ± 0.67 vs 1.83 ± 0.70, = 0.609). No significant SNP–SNP interaction was found for eczema diagnosis and eos%.

Table 4. Gene–gene interaction analysis for LogIgE by generalized multifactor dimensionality reduction method
ModelTACVC P a
  1. CVC, cross-validation consistency; TA, testing accuracy.

  2. a

    Adjusted for age, gender, and asthma status as covariates and obtained from 5000 permutations.

rs22826790.509/100.186
rs7041_rs79357920.5410/100.047
rs2298849_rs7041_rs79357920.524/100.002
rs4646536_rs7041_rs7935792_rs79752320.533/100.043
image

Figure 1. Generalized multifactor dimensionality reduction (GMDR) analysis for LogIgE among our eczema patients and controls. (A) Risk groups of the best two-locus model for this subphenotype, with the numbers in brackets being subject counts of specific groups. Light gray squares are low-risk groups, whereas dark gray squares are high-risk groups. Left bar inside each square denotes the score of the control group processing respective genotypes, whereas right bar denotes eczema group. (B) The distribution of LogIgE among different GMDR-assigned risk groups. LogIgE level was significantly different between high- and low-risk eczema groups, but such difference was not found between control groups.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

This case–control association study investigated the relationship between eczema phenotypes and five candidate genes along the vitamin D pathway in southern Chinese children. A novel genetic variant rs4674343 of CYP27A1 was found to lower the risk of atopic eczema. Two CYP2R1 SNPs, rs2060793 and rs1553006, were associated with increased eos%. Two CYP2R1 haplotypes were associated with eczema susceptibility, whereas one VDR haplotype might protect against eczema. The two involved VDR SNPs were reported to increase asthma risk in previous studies [8, 9, 24]. Two SNPs of GC and CYP2R1 might interact to modulate total IgE levels among eczema patients.

Four human cytochrome P450s were able to convert vitamin D to its circulating form, but CYP2R1 and CYP27A1 are the most important ones [25]. Being the only mitochondrial enzyme exhibiting 25-hydroxylase activity, CYP27A1 was also studied for its 27-hydroxylation of sterol intermediates in bile acid synthesis. CYP27A1 SNPs have not been reported to be associated with eczema, asthma, or atopy. Our study was thus the first to suggest possible protective effect of CYP27A1 rs4674343 for atopic eczema (OR 0.66; Table S4). CYP2R1 was identified as a key microsomal 25-hydroxylase for vitamin D in 2003 [26]. Discordant association between CYP2R1 and asthma was found between Caucasian and Chinese populations. CYP2R1 SNPs had modest association with asthma [27] and total IgE [24] in Caucasian studies, but CYP2R1 was not associated with asthma in Chinese [12]. Our study first reported two SNPs (rs2060793 and rs1553006) of this gene to be associated with eos% among eczema children. Rs1553006 also formed a haplotype with rs107661697 to increase eczema risk. Another CYP2R1 haplotype from rs16930609 and rs11819875 also conferred eczema susceptibility. CYP2R1 functions to produce stable vitamin D, and this enzyme showed inverse correlation with eczema severity [6].

GC was best known by its two functional SNPs rs4588 and rs7041 [28]. These SNPs expressed three different GC globulin isoforms, GC1-F, GC1-S, and GC2 [29], which varied at positions 416 and 420 to alter the affinity (GC1F > GC1S > GC2) in carrying vitamin D metabolites to target cells [30]. GC2 genotype was associated with less chronic obstructive pulmonary disease in White people, whereas GC1F increased this disease among Asians [31]. Another SNP rs2282679 was found in GWAS to be associated with vitamin D level [32] and insufficiency [33] in Europeans. The latest GWAS in asthmatic children of Hispanic ancestry showed that GC affected circulating vitamin D levels [10]. Our study could not detect any major effect of GC on eczema phenotypes, with the exception that rs222020 and s2298849 were marginally associated with increased risk of atopic eczema and GC interacted with CYP2R1 to modulate total IgE among eczema patients.

VDRs rs7975232 and rs2239185 were the first variants along the vitamin D pathway reported to be associated with asthma. The former was also replicated in a Chinese case–control study [11]. The effect of VDR on eczema differed from its role for asthma, and the two SNPs showed modest protective effect. The AT haplotype was less prevalent in eczema patients (0.282) than controls (0.3). This discrepancy could be explained by interethnic difference between Asians and Caucasians of European ancestry, and AT was the major haplotype in the latter population (NCBI dbSNP at http://www.ncbi.nlm.nih.gov/projects/snp/). Our results found rs7975232C and rs2239185C to be risk alleles for allergies. Future studies should delineate the mechanisms whereby these intronic SNPs altered vitamin D binding or receptor function.

A number of studies suggested that targeting vitamin D metabolism and signaling might be an effective treatment for eczema. 1α,25(OH)2D increased in vitro cathelicidin expression and antimicrobial activity in keratinocytes [34]. Serum 25(OH)D levels were significantly higher in children with mild eczema than those with moderate or severe eczema [6]. Several small clinical trials reported daily vitamin D supplementation of 1000–6000 IU to be beneficial in eczema [35-37]. Larger randomized controlled trials are needed to confirm the therapeutic benefit of vitamin D on eczema.

The production of active vitamin D requires the concerted actions of many genes. Apart from our five selected genes, CYP24A1 is the other vitamin D pathway gene that inactivates vitamin D by 24-hydroxylation of the active form. We did not select this gene because it is not involved in vitamin D activation. Two Caucasian studies analysed possible associations between asthma and many genes encoding vitamin D metabolism and signaling [24, 27], which did not detect consistent association for CYP24A1.

This study has several limitations. Most patients were recruited from community studies that did not include the assessment of eczema severity. Only 418 cases had data for severity of their eczema. This study did not have sufficient power to analyse any effect of studied SNPs on eczema severity. Another drawback relates to the lack of vitamin D measurements in our subjects, which could serve as positive control to evaluate whether our studied SNPs alter serum vitamin D levels.

In conclusion, this case–control study investigating genetic association between childhood eczema and vitamin D pathway genes found rs4674343 in CYP27A1 to be associated with atopic eczema in southern Chinese. Peripheral blood eosinophil percentage was modulated by rs2060793 and rs1553006 of CYP2R1 as well as their haplotypes. CYP2R1 and VDR haplotypes altered eczema susceptibility, and GC rs7041 and CYP2R1 rs7935792 interacted to modulate total IgE among eczema patients. Replication of our findings in other ethnic populations is needed to confirm these vitamin D pathway genes as major eczema candidates.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

We thank E. Yung, F. Y. Y. Kwok, Y. Y. F. Ho, and W. C. Chan for helping with subject recruitment and sample processing. This work is funded by General Research Funds (469908, 470909 and 477110) of Hong Kong Research Grants Council; Research Committee Group Research Scheme (3110060 and 3110087); and Direct Grant for Research (2011.1.058) of The Chinese University of Hong Kong.

Author contributions

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Dr Ting F. Leung developed the idea, designed research protocol, and obtained funding for the study; Ms Susan S. Wang assessed subjects, collected study data, and performed genotyping experiments; Drs Kam L. Hon, Alice P. S. Kong, and Juliana C. N. Chan recruited and managed study participants; Mr Man F. Tang and Dr Hing Y. Sy collected study data and assisted genotyping experiments; Ms Susan S. Wang and Drs Hing Y. Sy and Ting F. Leung analysed data and interpreted study findings; Dr Ting F. Leung wrote the original draft of this manuscript; and all authors approved the final version for submission.

References

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Patients and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
all12337-sup-0001-SupportingInformation.docWord document242K

Data S1. Patients and methods (genotyping method and GMDR analysis).

Table S1. Selection criteria, function prediction, and genotyping method of 24 target SNPs of five genes along the vitamin D pathway.

Table S2. Genotyping efficiency of 23 SNPs and their respective Hardy–Weinberg equilibrium (HWE) among the controls.

Table S3. Logistic regression analysis of the association between 23 SNPs and eczema diagnosis.

Table S4. Case–control comparisons of minor allele frequencies of the studied SNPs.

Figure S1. Haplotype blocks of the five selected genes.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.