INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is a complex genetic disease characterized by increased airway responsiveness to a variety of stimuli, reversible airway obstruction, and airway inflammation. Eosinophils, mast cells, and lymphocytes together with a multitude of cytokines have important pathogenic roles in this inflammation (1). Atopy is a common familial trait characterized by increased specific or total serum IgE as well as skin prick test responsiveness to common allergens and underlies allergic asthma, rhinitis, and atopic dermatitis. As much as 94% of childhood asthma is associated with atopy (2, 3). Genetic susceptibility, environmental exposures, and their interaction are the major components contributing to the phenotypic expression of asthma. Genetic studies of asthma using genome screens and candidate gene approaches have suggested that multiple genes are involved in the development of atopy and asthma (4).

Several candidate genes have been identified in the pathway leading to atopy and asthma. Tumor necrosis factor-alpha (TNF-) has been tested as a candidate gene for asthma (5). TNF is a potent proinflammatory cytokine that is found in increased concentration in bronchoalveolar lavage fluid (BALF) from symptomatic versus asymptomatic asthmatics (6) and is increased in BALF from asthmatics after allergen challenge (7). The genes for TNF- and lymphotoxin- (LT or TNF-) are located within the class III region of the MHC complex on chromosome 6p. It has been shown that allele 2 of a TNF- promoter polymorphism (TNF-308*2) is associated with higher TNF gene expression (8) and secretion (9) and with a NcoI polymorphism (LTNcoI*1) in the LT gene (5). However, the associations with gene expression were not found by others (10) and therefore remain controversial. More recently, the results of one study have shown that asthma was significantly more common in subjects possessing the LTNcoI*1 and TNF-308*2 alleles (5). The association was confined to the LTNcoI*1/TNF-308*2 haplotype. In another study, Trabetti and colleagues (11) found that the LTNcoI polymorphism and LTNcoI/TNF-308 haplotypes showed significant linkage to atopy by sib-pair analysis (11). However, in contrast to the results of Moffatt and Cookson (5) there was an association between increased total serum IgE levels and the LTNcoI*2 allele. A third group produced results similar to those of Trabetti and colleagues (11) when they showed that the LTNcoI*2/TNF-308*1 haplotype was associated with childhood asthma (12).

The interleukin-4 (IL-4) gene is an attractive candidate gene for atopy. The IL-4 gene is of great interest because it causes B-cell isotype switching from IgM to IgE and stimulates IgE production in allergic sensitization. Evidence for linkage of total serum IgE levels to the IL-4 gene locus has been reported in two different populations (13, 14). Further investigation obtained evidence that a C T transition at position -589 in the IL-4 promoter was associated with elevated total IgE levels in asthmatic families and increased reporter gene expression in vitro (15). However, others found that the IL4-589 C T polymorphism was only weakly associated with certain measures of asthma and atopy in some subjects (16). A more recent study has shown that IL-4-589*T was associated with asthma, but not with total or specific IgE levels, in Japanese children (17).

Another candidate gene for atopy is the beta subunit of the high-affinity IgE receptor (Fc RI). This gene is located in the 11q13 region, which has shown linkage to atopy and asthma (18). The high affinity IgE receptor is found on the surface of mast cells and basophils. Cross-linking of the receptor after binding to specific antigens leads to degranulation of these cells and increased IL-4 production. Therefore, variants of the Fc RI gene may influence the production of IL-4. An amino acid change at residue 237 from glutamic acid to glycine (E237G) has been shown to be associated with total and specific IgE levels and with childhood atopic asthma (21, 22). Other investigators have reported the lack of association of this polymorphism with atopy and asthma (23, 24).

The previous studies of these polymorphisms have shown a lack of consistency with regard to the associated phenotypes. Some of these associations are still considered controversial. In addition, the previous studies were linkage or case-control studies and therefore the relative risks associated with these polymorphisms have not been investigated. Therefore, a prospective study of an infant cohort at high risk for atopy and other allergic diseases was designed. Techniques based on the polymerase chain reaction (PCR) were used to genotype the infants and their parents. We determined whether these polymorphisms were associated with the development of childhood atopy and other allergic diseases. The presence of disease associations with these polymorphisms was also assessed using the transmission disequilibrium test (TDT).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Infants at high risk for asthma were recruited from two centers: Vancouver and Winnipeg. The definition of high risk infants was those who had at least one first-degree relative with a history of asthma or two first-degree relatives with a history of other allergic disorders such as allergic rhinitis, atopic dermatitis, or food allergy. The infants were recruited for a study designed to test whether an intervention program instituted during the first 12 mo of life protected against the development of asthma and other atopic disorders (25, 26).

We initially recruited 545 women into the study, of whom 271 came from Vancouver and 274 from Winnipeg. Fifty-six families dropped out from the study for various reasons. The history of allergic diseases (previously diagnosed asthma, atopic eczema, allergic rhinitis, or food allergy) and smoking habits in the mothers, their partners, or their children were documented. Ethnicity was recorded by questioning of the parents as to their ethnic origin.

Peripheral blood was taken from parents, and cord blood was collected for each infant. Home visits were carried out before delivery and every 4 mo for 12 mo. A questionnaire was administered to both parents. Assessment of the infants at 12 mo for respiratory symptoms and other allergic symptoms such as rhinitis was done by pediatric allergists. Because of the difficulty in diagnosing asthma at 12 mo of age, we used a conservative definition of "probable" asthma for these infants.

"Probable" asthma was defined as (1) at least two distinct episodes of cough, each lasting for 2 wk or more; or (2) at least two distinct episodes of wheeze, each lasting for 1 wk or more, plus at least one of the following: nocturnal cough at least once a week in the absence of a cold, hyperpnea-induced cough or exercise-induced cough or wheeze at any time, response to treatment with beta-agonist and/or anti-inflammatory drugs. The definition of rhinitis without colds ("noninfectious" rhinitis) was two or more episodes of runny nose and sneezing without apparent colds. Atopy was defined by the presence of a positive epicutaneous (prick) skin test reaction (3 mm or greater mean wheal diameter than the negative control) to one or more of the following inhalant allergens: cat, dog, house dust mites (extracts of Dermatophagoides pteronyssinus and Dermatophagoides farinae), cockroach, Cladosporium, Alternaria, and to the following ingestant allergens: cow's milk, egg, soy, wheat, and peanut.

Of the 489 families, there were four pairs of twins, so completed phenotypic data at the age of 12 mo were obtained for 493 infants. Cord blood was collected for 373 infants. Parents of 283 of these infants also had blood taken. Of the 373 infants, 281 were of white ancestry and 22 were of Asian ancestry, the rest were of other ancestry including First Nation, East Indian, black, and mixed ethnic background.

Genotyping

Genomic DNA was extracted from peripheral blood or cord blood by standard techniques (27). The 373 infants and 572 parents were genotyped for the Fc RI 237, IL-4-589, and TNF--308 polymorphisms. All of the genotyping was done by PCR-based techniques. Genotyping of the Fc RI polymorphism was by allele-specific PCR as described previously (21).

The TNF--308 polymorphism was amplified using an upstream primer with a mismatch that introduced an artificial Sty I restriction site into the wild-type allele (allele 1), but not into the variant allele (allele 2). The sequences of the primers were 5'-AGG CAA TAG GTT TTG AGG GCC ATG-3' and 5'-ACA CAC AAG CAT CAA GGA TAC C-3'. Amplification with these primers produced a 143 bp product. The reaction conditions used were: 0.1 µg of genomic DNA in a total volume of 20 µl of reaction mixture containing 0.1 µM of each primer, 200 µM of each dNTP, 20 mM TRIS-HCl (pH, 8.4), 50 mM KCl, 1.5 mM MgCl₂, and 0.5 U of Taq DNA polymerase. The amplification was accomplished by 35 cycles at 94° C for 30 s, 59° C for 30 s, 72° C for 10 s, and a final 10-min extension at 72° C. After amplification, a 20-µl aliquot of digestion mixture containing 10 U of Sty I (New England BioLabs, Beverly, MA) was added to 20 µl of PCR product and incubated at 37° C for at least 4 h. Resultant products were analyzed on 3% agarose gels. TNF--308 allele 1 was identified by 123 bp and 20 bp fragments and allele 2 by a single 143 bp fragment (Figure 1).

View larger version (27K): [in this window] [in a new window]   Figure 1.   Sty I digestion for the TNF--308 polymorphism. M indicates 100 bp DNA molecular size marker; -ve represents template-free negative control. The homozygous mutant is labeled by an asterisk; heterozygotes are labeled by arrows. Other lanes are homozygous wild types.

Genotyping of the IL-4-589 polymorphism was also by an allele-specific PCR. However, two separate reactions were used for each sample. Primers for these reactions were as follows: control primers: 5'-TAG ACC TAC CTT GCC AAG GGC -3' (P3F), 5'-TGC ATA GAA GGG AGA GGC CAC -3' (P3R); PW2: 5'-CTC AAA ACA CTA AAC TTG GGA GAA CAT TCT C -3' (specific to C allele); PM2: 5'-AGA CTC TCC TAC CCC AGC ACT GGC GA-3' (specific to T allele). P3F and P3R produced a 350 bp control band irrespective of the genotype at the 590 position. PW2 produced a 258 bp fragment with P3R only when the C allele was present. PM2 produced a 148 bp band with P3F only when the T allele was present. All subjects were first screened for the T allele. P3F, P3R and PM2 were needed for this reaction. The optimized PCR was performed in a 40-µl volume containing the same conditions as stated above except 0.32 µM of each primer was used. The cycling parameters were 35 cycles at 94° C for 30 s, 60° C for 30 s, 72° C for 1 s, and final 10-min incubation at 72° C. The reaction products were analyzed by electrophoresis in 2% agarose stained with ethidium bromide.

To differentiate the homozygous mutants and the heterozygotes a second PCR was performed for subjects who had the T allele. P3F, P3R, and PW2 were required for this PCR. The reaction was carried out as described above except a touchdown PCR (28) was used as follows: the annealing temperature started at 63° C, then progressively decreased every second cycle by 1° C to 57° C, followed by 30 cycles at 56° C. Each annealing temperature was set for 30 s. A denaturation step (94° C for 30 s) and extension step (72° C for 30 s) were also included for each cycle. PCR products were analyzed by electrophoresis in 2% agarose gels. If only the control fragment was present, this was indicative of a homozygous mutant.

Template-free and known genotype controls were included in each experiment. Genotypes were scored without knowledge of the phenotypes. Another member of our laboratory independently checked all the genotypes. The samples were regenotyped if there was any disagreement concerning the genotyping.

Statistical Analysis

The frequencies of the alleles and genotypes between groups were compared by chi-square analyses for 2 × 2 contingency tables. Because intervention group is a potential confounding factor, the associations of asthma, atopy, and rhinitis with the TNF--308, IL-4-589, and Fc RI polymorphisms were analyzed by multiple logistic regression. The statistical significance of the TDT was tested by a ² (McNemar) test (29).

Power Analysis

A power analysis was performed for the study. Given the numbers of infants in the study the minimum relative risk (RR) that could be detected with 80% power for the TNF--308*2, IL-4-589*T, and Fc RI*G alleles for a variety of phenotypic frequencies is shown in Table 1.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Allergic disorders had developed in some of the infants by 12 mo of age. The most prevalent positive skin test was to food allergens. Asian infants had a higher prevalence of positive skin tests than did white infants (Table 2). No association was found between "noninfectious" rhinitis, "probable" asthma, and positive skin test at 1 yr of age. The allele frequency of TNF-308*2 was not different between whites and Asians. However, the allele frequencies of Fc RI*G and IL-4-589*T in white infants were significantly less than in Asian infants (Table 3). Because of these differences in allele frequencies and phenotypes between ethnic groups, we limited our association study to whites, of whom we had the largest number of subjects. The prevalence of "probable" asthma and allergic diseases was compared in infants with and without these polymorphisms. No significant difference in phenotype frequency was found between the white infants with TNF-308*2 or Fc RI*G and those without these alleles (Table 4). However, significant associations were found for the IL-4-589 polymorphism with rhinitis and with "probable asthma." The relative risk of TT (homozygous mutant) and CT (heterozygous) compared with the CC (homozygous wild type) genotype for "probable asthma" was 4.1. The relative risk of TT compared with other genotypes for rhinitis was 2.4. Therefore, TNF-308*2 and Fc RI*G were not risk factors for the development of "asthma" or atopy in white infants by 1 yr of age. IL-4-589*T was associated with rhinitis and "probable asthma" in this population. The association of IL-4-589*T remained significant after correction for intervention group.

The proportion of Asian CT and TT infants who developed atopy was 62%. Only one Asian CC infant was found; therefore, the relative risk for the IL-4-589 polymorphism for the development of atopy in Asians could not be accurately determined.

To further test for association of this polymorphism with atopy and to control for ethnic diversity, the TDT was performed. Thirty families were selected. All infants had at least one positive skin test and at least one of their parents was heterozygous for the IL-4-589 polymorphism. In the absence of an association, the expected number of transmissions of T and C alleles from the heterozygous parents would be equal. We found that allele T was transmitted more often than nontransmitted (26 transmitted versus 13 nontransmitted, p = 0.037).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previous studies have shown that IL-4-589*T was associated with atopy and asthma (15). We found association of IL-4-589*T with "probable" asthma, rhinitis, and atopy. Our results have shown that IL-4-589*T may be a risk factor for the development of asthma and atopy in the white infant population at 1 yr of age. First, the prevalence of "probable" asthma in white infants with IL-4-589*T was significantly higher than that in those with the wild-type allele. However, it is very difficult to define asthma phenotypes at 1 yr of age, and further follow-up for the development of atopy and asthma in this cohort is essential. Second, the prevalence of rhinitis was significantly higher in white infants with TT than in those with the other genotypes. This is the first report of this association. Third, an association of the IL-4-589*T allele for atopy (p = 0.037) was detected using the TDT. These findings support previous association studies of this polymorphism with asthma (17) and atopy (15, 16). Previous studies have found associations between IL-4-589*T and total serum IgE in American whites (15) but not in Australian and British whites (16). In the latter study, only a weak association was detected between IL-4-589*T and specific IgE to house dust mite and wheeze (RR = 1.33). In a more recent study, it was shown that IL-4-589*T was associated with asthma but not with total and specific IgE levels in Japanese children (17). All previous association studies have been retrospective. This is the first prospective study in which genetic markers have been predictive of disease development. Our results suggest that IL-4-589*T is a predictor for atopy, asthma, and rhinitis. It is possible that this variant sequence in the promoter region of the IL-4 gene increases expression of IL-4 and this may explain these associations. Alternatively the polymorphism could be in linkage disequilibrium with other causal mutations.

The relative risks associated with IL-4-589*T for "probable" asthma and rhinitis are specific to this cohort and may not be applicable to the general population. This is because the infants were chosen on the basis of a family history of asthma and allergic disease. Therefore, the unaffected children in this study may have a higher prevalence of risk alleles than unaffected subjects from the general population. Thus, the relative risks found in this study may be underestimates of the true risks associated with these alleles.

The occurrence of high levels of atopy (18 of 36) and increased allele frequency of the IL-4-589*T in the Asian infants is intriguing and suggests that the two may be causally linked. However, it was not possible to test this hypothesis because of the low prevalence of the IL-4-589*C allele in this group.

The data presented here for the infants based on the first year assessment indicated that TNF--308*2 and FcRI*G were not risk factors for the development of atopy or asthma in the infant population at 1 yr of age. This conclusion was drawn from the evidence that the prevalence of "probable asthma" and atopy were not significantly different between infants with and those without TNF--308*2 and FcRI*G. However, an important consideration in studies such as this is the issue of power. We calculated the power of our cohort given the genotype frequencies present in the infants (Table 1). A previous study of 413 subjects (age range, 5 to 51 yr) from a general white population found that asthma was significantly more common in subjects with the TNF--308*2 and LT NcoI*1 alleles (5). The odds ratio (OR) for asthma associated the presence of the TNF--308*2 allele was 2.1. A study of 1,004 Australian whites (age range, 5 to 55 yr) found that subjects with Fc RI*G had increased skin test responses to aeroallergens and bronchial reactivity to methacholine (21). The RR of subjects with Fc RI*G having asthma compared with subjects without the variant was 2.3. A study of a Japanese population showed that the prevalence of Fc RI*G was significantly increased in atopic asthmatics, particularly childhood atopic asthmatics, compared with nonasthmatic, nonatopic control subjects (OR = 3.9) (22). Therefore, given the prevalence of asthma in this cohort the lack of association of these polymorphisms with asthma could be due to lack of power.

The lack of association of the TNF--308 and Fc RI E237G polymorphisms and asthma/atopy phenotypes in the white infant population has several other possible explanations. First, it is difficult to define asthma at 12 mo of age, since most wheezing in infants is transient and is not associated with increased risk of asthma or allergies later in life. Only those infants who have persistent wheezing are more likely to have rhinitis apart from colds, reduced lung function, sensitization to common aeroallergens, and elevated serum IgE levels (3, 30). The diagnosis of asthma at 1 yr of age may not be an accurate predictor of later childhood or adult asthma. It is likely that for some of the infants who have the "probable asthma" phenotype the symptoms are the result of respiratory viral infections. The wheezing in these infants will be transient and they will not be at risk for asthma later in life. Furthermore, other infants may develop asthma in later life even though they did not have the "probable asthma" phenotype at 1 yr of age. Therefore, a follow-up study is planned for this cohort.

Second, because atopy and asthma are complex genetic traits, it is unlikely that every study will demonstrate the same associations. The lack of associations may reflect genetic heterogeneity in the pathogenesis of atopy and asthma.

The third reason for the lack of association in this study may be that the frequency of positive skin tests increases with age (31). Therefore, negative skin tests in children younger than 5 yr of age may not fully reflect their atopic status. The peak age of sensitization is between 15 and 25 yr (32, 33). Previous association studies used positive skin tests to aeroallergens to define atopy. In our study, most of the positive skin tests were to food allergens.

In summary, the results of this study suggest that TNF--308*2 and Fc RI*G are not risk factors for the development of atopy and asthma in both white and Asian infants. The IL-4-589*T was associated with rhinitis and "probable asthma" in white infants. The allele frequencies of IL-4-589*T and Fc RI*G were more prevalent in other ethnic groups than in whites.