Original article: Polymorphisms in eosinophil pathway genes, asthma and atopy

Authors


Michael Kabesch MD
Children's University Hospital
Ludwig Maximilians University Munich
Lindwurmstrasse 4
D-80337 Munich
Germany

Abstract

Background:  Eosinophilic inflammation is considered to play an important role in the development as well as in the perpetuation of asthma. As eosinophil production and survival is under genetic control we investigated whether polymorphisms in eosinophil regulation pathway genes (IL-3, IL-5, GM-CSF and their respective enhancers and receptors) may influence the development of atopic diseases.

Methods:  In two large study populations of children, the German part of the International Study of Asthma and Allergy in Childhood (ISAAC II) and the German Multicentre Atopy Study (MAS), 3099 and 824 children, seven polymorphisms previously associated with the development of atopic diseases were genotyped: two in and around the GM-CSF gene (Ile117Thr and T3085G), one in IL-3 (Pro27Ser), in IL-5 (C-746T), and in the IL-5 high affinity receptor chain IL-5R (G-80A) and two in the common receptor chain CSFR2b for IL-3, IL-5, and GM-CSF (Asp312Asn and Glu249Gln). Statistical analyses were performed using chi-squared tests and variance analyses. Gene by gene interactions were evaluated in logistic regression models.

Results:  The T allele at position −746 in the IL-5 gene was significantly protective for atopy in the ISAAC II population (P = 0.006). Furthermore, the risk for atopic asthma was decreased in carriers of the T allele (P = 0.036) and evidence for interaction with the enhancer polymorphism 3085 bp 3′ of GM-CSF was detected.

Conclusions: IL-5 C-746T influenced atopic outcomes and showed evidence for gene by gene interaction. No significant associations were found with all other tested polymorphisms in the eosinophil regulation pathway after correction for multiple testing.

Eosinophilic inflammation has been identified as a hallmark of asthma as many inflammatory mechanisms in chronic asthma have been associated with eosinophil migration, maturation and activation. Mature eosinophils are important effectors of allergic inflammation releasing proinflammatory proteins from their granules such as major basic protein, neurotoxin, peroxidase, cationic protein (1). Eosinophils cooperate with mast cells and are a major source of leukotriens, directly affecting airway smooth muscle contraction and vascular permeability (1). Patients with asthma and allergy have higher peripheral eosinophil counts and eosinophils are increased in lung tissue and bronchioalveolar fluids of asthmatics (2, 3). As reviewed by Adamko et al.(4), GM-CSF, IL-3 and IL-5 are key cytokines in the induction and regulation of eosinophils. Eosinophils differentiate from the bone marrow precursor cells under the influence of growth factors such as IL-3 and GM-CSF. Local survival of eosinophils is mainly attributable to IL-5, which can also be expressed by epithelial cells and local fibroblasts in the respiratory tract (5). Furthermore, GM-CSF has also been shown to prolong the survival of eosinophils in asthmatics (6). All three key eosinophilia cytokines (IL-5, IL-3 and GM-CSF) signal through the receptor chain CSFR2β, which belongs to the type I cytokine receptor family, in combination with cytokine specific receptor chains such as IL-5R. Furthermore, for GM-CSF and IL-3 it has been shown that a common enhancer element located in between both cytokines, influences the expression of these two genes (7).

Eosinophilia seems to be under genetic control and linkage peaks have been mapped to the 5q31–33 locus (8, 9). For all components of the eosinophila regulating pathway, polymorphisms and mutations have been identified which lead to putative functional alterations as they change presumable transcription factor binding sites (IL-5 C-746T, IL-5R G-80A, GM-CSF T3085T, located 3085 bp 3′ of GM-CSF within an enhancer element) or the amino acid sequence of proteins (IL-3 Pro27Ser, GM-CSF Ile117Thr, CSFRB Glu249Gln and Asp312Asn) (7, 10–14).

Thus, it was hypothesized that polymorphisms which may influence the regulation of eosinophilia may also have an effect on the development of asthma and atopic disease. Up to now, studies have only focused on single genes of the pathway, but no overall picture on the effect of these polymorphisms in eosinophila regulation genes was available in any population. We included all previously described, putatively functional polymorphisms in eosinophilia regulating factors, their receptors and their genetic enhancer elements in our analyses. Thus, seven polymorphisms were investigated for their role in asthma, hay fever, atopic dermatitis, atopy status, lung function and bronchial hyperreactivity (BHR). A total of 3099 German children of the cross sectional International Study of Asthma and Allergy in Childhood, phase II (ISAAC II) and 824 children of the longitudinal German Multicenter Allergy Study (MAS) cohort were analysed.

Methods

Description of study populations

Informed written consent was obtained from all parents of children included in the cross sectional International Study of Asthma and Atopy in Childhood (ISAAC, phase II) and the German Multicentre Allergy Study (MAS) cohort. All study methods were approved by the local ethics committees, and as the populations and phenotyping methods have been described in detail before (15–17), only an overview of the methods pertaining to this analysis is given here.

Cross sectional International Study of Asthma and Allergy in Childhood

As part of the International Study of Asthma and Atopy in Childhood (ISAAC II), a cross sectional study was conducted in Munich and Dresden between 1995 and 1996 to assess the prevalence of asthma and allergies in schoolchildren age 9–11 years (15). Parental questionnaires for self-completion were sent through the schools to the families. In the ISAAC II population all children whose parents reported that a doctor diagnosed ‘asthma’ at least once or ‘asthmatic, spastic or obstructive bronchitis’ more than once were defined as having asthma. The definition of atopic dermatitis and hay fever relied on positive answers of the parents to questions on the doctor's diagnosis of these diseases in the questionnaire. Furthermore, children underwent skin prick testing for six common aeroallergens (Dermatophagoides pteronyssinus, D. farinae, Alternaria tenuis, cat dander, mixed grass and tree pollen) according to ISAAC phase II protocols. Atopy was defined as one positive reaction (wheal diameter of 3 mm or more) in at least one skin prick test. In the ISAAC II study pulmonary function and BHR were tested according to standard procedures previously described. Baseline lung function was measured and bronchial reactivity was assessed in a random 50% sub-sample of the study population by inhalation of nebulized, hyperosmolar saline (4.5%). Children with a drop in FEV1 of 15% or more from baseline were classified as positive for bronchial hyperresponsiveness (15).

Multicentre Allergy Study cohort

Initially, 1314 children born in five German cities in the year 1990 were followed up from birth to the age of 13 years. For 888 children DNA was available and genotyped. Of these, only children of German origin were included in this analysis (n = 824). Yearly follow-up visits included standardized interviews, questionnaires, and physical examinations. In the MAS study, asthma, hay fever and atopic dermatitis at age 10 were defined using the ISAAC II-core questions for children as described for the ISAAC II study population. Serum samples were obtained from the children at birth, and at 1, 2, 3, 5, 6, 7 and 10 years of age. Total IgE, specific IgE antibodies to food allergens and inhalant allergens (Dermatophagoides pteronyssinus, cat dander, mixed grass, birch pollen, as well as dog dander from age 3 years on) were determined by CAP-RAST FEIA (Pharmacia & Upjohn, Freiburg, Germany). In the MAS study, atopy was defined as a specific IgE level (CAP I) of ≥0.35 kU/l at age 7 or 10 years, respectively. Pulmonary function tests were performed at age 7, 10 and 13. To allow for a direct comparison with lung function measurements in ISAAC II, only lung function measurements at age 10 were considered for this analysis. Bronchial hyperresponsiveness was only assessed at age 7 in 610 individuals in the MAS study (18). Bronchial challenges in the MAS study were conducted after baseline spirometry using increasing concentrations of histamine (usually from 0.5–8.0 mg/ml) according to standard procedures until a 20% fall in FEV1 was observed (PC20FEV1).

Molecular Genetic Methods

Genomic DNA was successfully extracted from blood samples in from 3099 children in the ISAAC II study and blood for genetic testing was available in 888 MAS children. All polymerase chain reactions (PCR) were performed using standard thermocyclers (Eppendorf, Cologne, Germany). Polymerase chain reactions were carried out in a standard volume of 25 μl containing 40 ng of genomic DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 200 μM of each deoxynucleotide triphosphates, 0.8 unit of Taq DNA polymerase supplied by Promega Corp. (Mannheim, Germany) and 5 pmol of each primer.

For genotyping the ISAAC II population, the MassARRAY system (Sequenom, San Diego, CA, USA) was used as previously described in detail (19) and indicated in Table 1. All PCR reactions were performed using standard thermocyclers (MJ Research, Waltham, MA, USA). First, a PCR was carried out. To remove excessive dNTPs, shrimp alkaline phosphatase was added to the PCR products. The base-specific extension reaction was performed in 10 μl reactions by Thermosequenase (Amersham, Piscataway, NJ, USA). For the base extension reaction the denaturation was performed at 94°C for 2 min, followed by 94°C for 5 s, 52°C for 5 s, and 72°C for 10 s for 55 cycles. The final base extension products were treated with SpectroCLEAN resin to remove salts out of the reaction buffer, and 16 μl of water was added into each base extension reaction. After a quick centrifugation the reaction solution was dispensed onto a 384 format SpectroCHIP prespotted with a matrix of 3-hydroxypicolinic acid (3-HPA) by using a SpectroPoint nanodispenser. A modified Bruker Biflex matrix assisted laser desorption ionization-time-of-flight mass spectrometer was used for data acquisitions from the SpectroCHIP. Genotyping calls were made in real time with massarray rt software (Sequenom, San Diego, CA, USA).

Table 1.   Assay specifications for genotyping of 7 eosinophilia candidate genes in the ISAAC II population (n = 3099) using the massARRAY system
Polymorphisms*PrimersPrimer orientationExtension primerAllelesAllele frequency %
  1. * location of the first base pair of sequence variations in relation to the start of the open reading frame.

  2. † 3′ of the GM-CSF gene.

GM-CSF enhancer T3085G†GGC ATA GGA AAC TCC TTC CAG5′CCC CTT CCT ACA GAG ACA GGT58.4
GAC CAC GAT GGG TGG TAT GAC3′G41.6
CSF2Rb G8544A = CSF2RB Asp312AsnGTA TGT GTT TCT CTG CCC TCC5′CAC TGC CAG ATT CCC GTG CCCG97.7
GAG GCA CCA CTG CCA GAT TCC3′A2.3
CSF2Rb G8193C = CSF2RB Glu249GlnGTC CCA GGA GCA GCT GAG CAC5′CCC AGC CCC AGA ACC TGG94.8
GAG TAC ATG AGG ACC TGT CTC3′C5.2
GMCSF T2600C = GMCSF Ile117ThrGGC AGA AAG TCC TTC AGG TTC5′AAA CTT CCT GTG CAA CCC AGAT81.1
GTT AAA GGA AAC TTC CTG TGC3′C18.9
IL3 C-68T = IL3 Pro27SerGTG CTG CTG CAC ATA TAA GGC5′TCT GAA GAG TTG GCA ACAC76.6
GAG GCA CTC TGT CTT GTT CTG3′T23.4
IL5 C-746TGGC ACA GCT TGC TTT TTC CTG5′ATG TCC AGA CTC CTG GAT CTC70.5
GGG TCT CAA GAT GAT GTC CAG3′T29.5
IL5R G-80AGCA CTC TTT CAT CAT CAC GGC5′ATG AAG GAA GCT GCC TGA GAG73.7
GGA GAA AAC GTG TTG CCA TCC3′A26.3

In the MAS population, restriction endonuclease based assays and standard PCR techniques were used to genotype the study population for all specific polymorphisms as indicated in Table 1. PCR products were digested with restriction endonucleases (New England Biolabs, Frankfurt, Germany), electrophoresed on 3% agarose gels and visualized with ethidium bromide staining and ultraviolet illumination. To test for the reliability of genotyping calls, 5% of samples were genotyped twice and Hardy–Weinberg equilibrium (HWE) was calculated for every assay.

Statistical analysis

Hardy–Weinberg equilibrium was tested as a genotyping quality control procedure using the chi-squared statistic, with expected frequencies derived from allele frequencies. Univariate associations between SNPs and qualitative outcomes were investigated using chi-squared tests as well as Cochran–Armitage trend tests (20). Odds ratios and 95% confidence intervals of the chi-squared tests for allelic differences are reported. To test for differences between genotypes in lung function parameters, univariate variance analysis was performed. A nominal level of significance (α = 0.05) was used for all statistical tests. To correct for multiple testing we used a Bonferroni correction limited to the groups of tests for one phenotype in a population. Tests for pair-wise interaction of SNPs were performed using the interaction term of the dominant effects of two SNPs in different genes in a logistic regression model. To minimize the risk of false positive results, a stepwise procedure was used where pair-wise interactions were first performed in the MAS population and only the interaction found to be significant was replicated in the ISAAC II population. All calculations were carried out with sas (Version 9.1.3; Heidelberg, Germany) and SAS/Genetics.

Results

Genotyping was successful in 2901–3034 of German children in the ISAAC II population and 771–792 of the 824 German individuals in the MAS population, depending on the specific assay (Tables 1 and 2). All genotyping results were in HWE. Linkage disequilibrium was calculated for all studied SNPs located next to each other on chromosome 5q31 (GM-CSF, GM-CSF enhancer, IL-3, IL-5) but was not found to be significant (data not shown). Genotype frequencies were very similar between the ISAAC II (Table 1) and the MAS population (Table 2). The observed allele frequencies were in agreement with previously reported frequencies in Caucasian populations, wherever these data were available.

Table 2.   Assay specifications for genotyping of 7 eosinophilia candidate genes in the MAS population (n = 824) using allele specific polymerase chain reactions and restriction enzyme assays
Polymorphisms*PrimersPrimer orientationRestriction enzymeAllelesAllele specific fragmentsAllele frequency %
  1. * location of the first base pair of sequence variations in relation to the start of the open reading frame.

  2. † 3′ of the GM-CSF gene.

GM-CSF enhancer T3085G†GAC ACG CAT AGG AAA CTC CTT CCA GAG GGT TTG C5′58Hha1T16954.5
CAT GAC ACA GGC AGG CAT TCC TAG ATT GCG CTG3′G13445.5
CSF2Rb G8544A = CSF2RB Asp312AsnGCT CTG CTG TGC TCC TCA GGG AGG AAG AGT5′56Ahd IG10398.1
GGC TGA ACA GAG ACG ATG TAT TGG CCG TGG GAC G3′A1381.9
CSF2Rb G8193C = CSF2RB Glu249GlnCCA CTG AGA GCT ATG GGA GGG ATG AAT GAC C5′56Pst IG47794.0
GGC TGA ACA GAG ACG ATG TAT TGG CCG TGG GAC G3′C3916.0
GM-CSF T2600C = GMCSF Ile117ThrGGC AGC AGC GTC ACC TGA TCC5′58.5Dpn IIT16380.1
GTT CTC TTT GAA ACT TTC AAA GGT GAT G3′C19419.9
IL-3 C-68T = IL3 Pro27SerGGT AGT CCA GGT GAT GGC AGA TG5′62Hae IIIC33875.4
CTT GTT CTG GTC CTT CGT GGG ACT CTG AAG AGT TGG CAA CG3′T29624.6
IL-5 C-746TGCT CAT GAA CAG AAT ACG TA5′52Rsa IC14370.8
GAA GGT ATT GGC TCA TAG TAC3′T12429.2
IL-5R G-80ACTT GTT TAC CTT GTC ACC ATG AGT AAA AGT GAA5′55Apo IG27373.9
GAG GCG GTT CTT CAC TCT TTC ATC3′A22626.1

First, we assessed the relationship between genotypes of the seven tested polymorphisms and the primary atopic phenotypes asthma, hay fever, atopic dermatitis and atopy at age 10 in the MAS and ISAAC II populations. The polymorphic-G allele at the GM-CSF enhancer location 3085 bp 3′ of GM-CSF was associated with a protective effect for hay fever in the MAS population (OR 0.70, 95% CI 0.49–0.99, P = 0.042), but not in the much larger ISAAC II population. Also, no further association with any of the other investigated phenotypes was observed for this polymorphism (Table 3a). Furthermore, the polymorphic C allele at position 2600 in the GM-CSF gene increased the risk for hay fever in the ISAAC II population (OR 1.25, 95% CI 1.01–1.54, P = 0.042) and showed the same trend in the MAS population. However, none of these effects remained significant when the significance level was corrected to P < 0.007 for multiple testing using a Bonferroni approach based on tests for seven polymorphisms per phenotype. In contrast, the T allele at position −746 of the IL-5 gene showed a protective effect against atopy defined by skin prick test in the ISAAC II population (OR 0.83, 95% CI 0.73–0.95, P = 0.006), remaining significant even after correction for multiple testing. This association was not observed in the MAS study. However, the power for replicating this minor effect found in the ISAAC II population was low (power 0.21) in the MAS population. When asthma was stratified by atopy status, it was again the T allele at position −746 of the IL-5 gene which showed a protective effect against the development of atopic asthma in the ISAAC II population (OR 0.72, 95% CI 0.53–0.98, P = 0.036). A similar but not significant trend was observed in the MAS population (OR 0.78, 95% CI 0.48–1.26, P = 0.310) (Table 3b).

Table 3a.   Odds ratios for atopic diseases and atopy status by genotype in the ISAAC II population (n = 3099)
Gene/polymorphismSuccessfully genotyped NAsthma (n = 272/2901–3034)Hay fever (n = 280/2901–3034)Atopic dermatitis (n = 540/2901–3034)Atopy (SPT positive) (n = 777/2901–3034)
  1. Odds Ratios and 95%confidence intervals derived from allelic test. Maximal number of carriers of the respective phenotype in the successfully genotyped population is given in row 1. In column 2 the number of successfully genotyped individuals for each polymorphism is specified. P < 0.05 is indicated in bold.

  2. *Significant after correction for multiple testing; SPT, skin prick test.

GM-CSF enhancer T3085G29010.89 (0.74–1.07)0.98 (0.82–1.18)0.98 (0.85–1.12)0.96 (0.85–1.09)
CSF2Rb G8544A30020.90 (0.48–1.68)1.17 (0.67–2.04)1.33 (0.88–2.01)0.91 (0.61–1.35)
CSF2Rb G8193C30060.74 (0.47–1.17)0.84 (0.55–1.29)0.87 (0.64–1.20)0.95 (0.72–1.24)
GMCSF T2600C30341.22 (0.98–1.51)1.25 (1.01–1.54)1.08 (0.91–1.27)1.10 (0.95–1.28)
IL-3 C-68T29741.13 (0.92–1.39)1.15 (0.94–1.41)1.09 (0.93–1.27)1.04 (0.90–1.19)
IL-5 C-746T29850.91 (0.75–1.11)0.84 (0.69–1.02)0.96 (0.83–1.12)0.83 (0.73–0.95)*
IL-5R G-80A29831.14 (0.93–1.39)1.17 (0.96–1.42)0.88 (0.75–1.02)1.06 (0.93–1.22)
Table 3b.   Odds ratios for atopic diseases and atopy status by genotype at age 10 in the MAS cohort (n = 824)
Gene/polymorphismSuccessfully genotyped NAsthma (n = 84/771–792)Hay fever (n = 77/771–792)Atopic dermatitis (n = 128/771–792)Atopy (sens. >0.35 kU/l) (n = 215/771–792)
  1. Odds Ratios and 95%confidence intervals derived from allelic test. Maximal number of carriers of the respective phenotype in the successfully genotyped population is given in row one. In column 2 the number of successfully genotyped individuals for each polymorphism is specified. P < 0.05 is indicated in bold.

GM-CSF enhancer T3085G7710.89 (0.64–1.25)0.70 (0.49–0.99)0.96 (0.72–1.27)0.94 (0.73–1.21)
CSF2Rb G8544A7920.73 (0.22–2.45)1.86 (0.69–5.04)1.24 (0.49–3.13)1.24 (0.55–2.79)
CSF2Rb G8193C7851.11 (0.57–2.17)0.51 (0.20–1.28)1.25 (0.70–2.22)0.71 (0.41–1.22)
GM-CSF T2600C7720.91 (0.59–1.39)1.16 (0.76–1.77)0.71 (0.49–1.04)0.85 (0.62–1.18)
IL-3 C-68T7731.02 (0.70–1.51)1.04 (0.70–1.55)1.13 (0.82–1.56)0.99 (0.74–1.33)
IL-5 C-746T7820.84 (0.58–1.22)1.02 (0.70–1.48)0.80 (0.58–1.10)1.02 (0.77–1.33)
IL-5R G-80A7881.05 (0.72–1.53)1.01 (0.68–1.48)1.23 (0.90–1.68)0.94 (0.71–1.25)

When baseline levels of lung function using percent of reference values for FEV1 and FVC as outcomes were analysed, no consistent effect of any polymorphism was observed. Furthermore, no association with BHR was found, neither at age seven in the MAS population nor at age 9–11 in the ISAAC II population (data not shown).

In a further step, gene by gene interactions were analysed in the MAS population systematically between all seven polymorphisms, based on the hypothesis that genetic alterations in the same functional pathway may modify the effects of genetic variants in other components of the pathway. Indeed, a protective effect of the polymorphic T allele at position −746 of the IL-5 gene on atopy was observed in interaction with the GM-CSF enhancer wildtype allele in the MAS population OR 0.53 95% CI 0.26–1.11 (P for effect 0.092, P for interaction 0.027). When this interaction analysis was replicated in the ISAAC II population, the protective effect already observed with IL5 C-746T was further magnified by the same interaction OR 0.62, 95% CI 0.46–0.85 (P for effect 0.003, P for interaction 0.089).

Discussion

These data indicate that only the polymorphism IL-5 C-746T of all the tested polymorphisms in the eosinophilia pathway may play a role in the development of atopy in our study population comprising approximately 4000 children. Furthermore, the risk for atopic asthma was decreased in carriers of T allele at IL-5 C-746T and some evidence for interactions between these effects and the GM-CSF enhancer wildtype allele were detected.

The IL-5 polymorphism C-746T has been associated with decreased pulmonary function in Korean children with atopic asthma (11) and with eosinophilia levels in a Japanese study of patients with atopic dermatitis (21). Located in the promoter of the IL-5 gene, it was speculated that this polymorphism may influence levels of IL-5 expression. Indeed, increased expression of IL-5 mRNA has been observed in asthmatics (22) but has not yet been linked to promoter polymorphisms directly. In our analysis, the effect of the IL-5 C-746T polymorphism did not interact with a further promoter polymorphism in the IL-5 specific receptor (IL-5R) or the common CSF receptor chain (CSFRβ), although this may have been expected. In our study, a magnification of the IL-5 C-746T effect by the GM-CSF enhancer element was observed.

The finding that not only gene specific polymorphisms such as in the promoter of IL-5 are involved in eosinophil mediated atopic disease but that also a genetic variation in a superior regulation mechanism may play a role, is a novel finding. Both, IL-5 and the GM-CSF enhancer element are located in close vicinity on chromosome 5q31. The enhancer element is located approximately 3 kb upstream of the GM-CSF gene (towards the IL-3 gene) and in recent years, the function and role of this enhancer element has been studied extensively. It is responsible for cell line specific regulation of the GM-CSF locus and by the binding of the transcription factors NFAT and AP-1, GM-CSF promoter activity as well as chromatin remodeling in the region is controlled (23). Based on our results it seems that the single base pair change in the enhancer element is not sufficient to influence the development of atopic disease by itself. However, by interaction with a further polymorphism in the IL-5 promoter, the effect of this IL-5 promoter polymorphism increases significantly. How this interaction occurs, is unclear. It may be speculated that either both polymorphisms act independent of each other on the overall activity of the eosinophil activation pathway. However, it is also conceivable that a more direct interaction may occur in which the polymorphism in the enhancer element changes chromatin structures of the GM-CSF region, putatively also affecting the IL-5 gene locus (directly or indirectly).

For all other tested polymorphisms in the eosinophil regulation pathway, no consistent association with atopic phenotypes was found in our population. Our results are in contrast to a number of smaller previous studies where a major effect of the T2600C polymorphism in GM-CSF on the development of asthma in a case control population comprising 371 Swiss children (14) and on atopic dermatitis in another cohort of 370 children (24) was reported. For IL-3 it has been shown that two SNPs, one in the promoter at position IL-3 T-68C and one in the coding region of the gene at position C79T (leading to an amino acid change from proline to serine at position 27), are in almost complete linkage disequilibrium. These SNPs seem to influence proliferation and cytokine production in peripheral blood cells (25) and were associated with asthma and atopy in a Korean case control population (n = 721, 380 atopic and 170 nonatopic asthmatics) (13). Again, this association could not be replicated in our study populations.

Interpreting our results, it has to be taken into account, that in our study as in many other studies of these eosinophil pathway genes and atopic diseases, no data on eosinophil production, eosinophils in the peripheral blood or eosinophila in the lung was available. These measurements were performed neither in the ISAAC II nor the MAS population. Thus, it could be argued that some of the negatively tested polymorphisms may very well affect the production or survival of eosinophils but that these effects may not contribute significantly enough to the development of atopic diseases to be detected when only disease outcomes are studied. Based on our results, this may be true for all the tested polymorphisms in GM-CSF, IL-3, IL-5R and CSFRβ, which were previously reported to be associated with atopic diseases. A minor but interesting effect was detected with IL-5 C-746T which may protect from atopy only when studied in combination with further polymorphisms in a regulatory element for GM-CSF. While a major role of all other tested polymorphisms on asthma and atopy is unlikely, based on our analyses of approximately 4000 white children, further polymorphisms, not studied in these genes and polymorphisms in other genes contributing to eosinophil regulation (26), which have not be the focus of this study, may still contribute to the development of asthma.

Aknowledgments

This study was funded by the German ministry of education and research as part of the German national genome research network with grants GS 01 0122, GS 01 0172, GS 01 0002, GS 01 0429, IE S08T03, IE S08T06, by the Sonnenfeld-Foundation and by the research grant ‘United Airways’ by Aventis Germany.

Ancillary