Introduction

Asthma is a complex disease with several intermediate phenotypes, notably atopy and chronic airway inflammation.1 A potential risk factor for asthma is tumour necrosis factor (TNF) alpha, a pleiotropic cytokine involved with immune inflammatory responses.2,3 High levels of TNFα have been observed in the bronchoalveolar lavage fluid, serum, and bronchial submucosa of asthmatics.4,5 In addition, following an allergen challenge, the secretion of TNFα is higher among atopic than non-atopic individuals.6,7,8 Therefore, TNFα appears to be implicated in multiple characteristics of asthma: airway inflammation, increased bronchial hyper-responsiveness, and atopy.2,8,9,10,11,12

Genetic factors may affect TNFα levels, and the TNF locus is within the Class III region of the human major histocompatibility complex (MHC) on chromosome 6p21.13 This locus includes two closely linked genes that encode the cytokines TNFα and lymphotoxin alpha (LTα) (also known as TNFβ). A TNFα promoter SNP (guanine (G, allele ‘1’) to adenosine (A, allele ‘2’) at position -308, TNFα-308) and a LTα-NcoI SNP (A to G in the first intron) have been positively associated with clinical symptoms of asthma and self-reported asthma.14,15,16,17,18,19,20 These SNPs have also been shown to influence the rate of transcription and protein production of TNFα and LTα, respectively.21,22,23 However, other studies have observed equivocal or inverse associations between the TNFα-308*2 and LTα-NcoI*1 alleles and asthma.19,24,25,26 In light of these conflicting results, we have undertaken a case-control study of 511 individuals to further investigate the potential association between these alleles and asthma, and whether this relation is stronger among individuals with acute asthma and/or a family history of asthma.

Materials and methods

All subjects were recruited from an outpatient clinic at the Sharp Rees-Stealy Medical Group in San Diego, California, USA. Individuals diagnosed by a physician with asthma (coded as ICD-9 493) were included as cases in the present study. Almost all diagnoses of asthma were confirmed by bronchodilatory response to albuterol. Controls were those individuals confirmed as non-asthmatic, with no previous history of asthma, hayfever, allergic rhinitis, or eczema. A total of 236 cases and 275 controls were recruited into the study. Blood was taken by venipuncture from these subjects, and a questionnaire asked about family history of asthma, basic demographic information, and for the cases, how often they had acute asthma attacks. The latter question was used to distinguish cases with sufficiently severe symptoms to be considered an attack (not just a sense of shortness of breath). Institutional Review Board (IRB) review and approval was obtained for the study protocol from Western IRB (Olympia, Washington, USA), and all study subjects gave informed consent.

DNA was extracted from whole blood using a kit from Gentra Systems, Inc. (Minnesota, USA). SNP genotypes were determined using the TaqMan assay.27 Samples were assayed in triplicate in a Robbins 96-well plate. The primers for TNFα-308 and LTα-NcoI were derived from previously published studies.24,25 Specifically, TNFα-308 detection was carried out by PCR with the following oligonucleotide primers: 5′-AAACAGACCACAGACCTGGTCC-3′ and 5′-CCATCCTCCCTGCTCCGATTCCG-3′. LTα-NcoI detection was carried out by PCR with the following oligonucleotide primers: 5′-CCGTGCTTCGTGCTTTGGACTA-3′ and 5′-AGAGCTGGTGGGGACATGTCTG-3′. Fragments were amplified in PCR reactions containing 20 ng genomic DNA, 900 nM forward unlabelled inner primer, 900 nM reverse unlabelled inner primer, 200 nM FAM labelled probe, 200 nM TET labelled probe and 1× Perkin-Elmer TaqMan Reagent Mix #43C4447. PCR reactions were preincubated at 50°C for 2 min, then 95°C for 10 min. Two-step thermocycling was performed for 45 cycles: denaturation at 95°C for 30 s and annealing at 64°C for 30 s. Upon completion of thermocycling, the fluorescence was read on an ABI 7700 Sequence Detector using the allelic discrimination software. FAM to TET ratios for each sample DNA, normalised against the TAMRA signal, indicated the genotype of each patient and was further confirmed by similar signals from known control DNAs. The TNF-308 and LTα-NcoI assays did not amplify for three and eight subjects, respectively. These individuals were excluded from the analyses of the corresponding SNP.

Basic descriptive statistics, stratified by asthma status, were calculated for demographic factors considered potential effect modifiers or confounders (eg, age, ethnicity). To estimate the relative risks of asthma for carriers of the TNF-308*2 or LTα-NcoI*1 variants, odds ratios (ORs) were calculated using logistic regression. The genotypes with one or two variants were combined to reflect a dominant model (eg, for TNF-308, 2,2 and 1,2 combined vs 1,1). We also evaluated the relation of asthma risk and TNF-308–LTα-NcoI haplotypes (estimated with the EM algorithm), and combinations of their genotypes. We statistically adjusted for or stratified on the following potential confounders in the logistic regression models: age at diagnosis, sex, ethnicity, physical activity, and family history of asthma. (Adjusting for body mass index and smoking had no appreciable statistical impact on our results, and so these covariates are not included in our final models.) An investigation of the need for interaction or polynomial terms resulted in our inclusion of an age×sex interaction term in the final model. All statistical analyses were undertaken with SAS software (SAS Institute Inc., Cary, North Carolina, 2000).

Results

The average age of study subjects included here was 42 years (standard deviation=14 years). In comparison with controls, cases were more likely to be European-American and to be less physically active (Table 1). Furthermore, cases had much stronger family histories of asthma than controls: over 60% of cases reported that one or more of their first-degree relatives had asthma, while only 25% of controls reported such a family history (P<0.001).

Table 1 Characteristics of study population
Full size table

Cases were more likely than controls to carry one or two copies of the TNFα-308*2 allele: 30% of the cases, vs 22% of the controls, had one or more copies of TNFα-308*2 (P=0.03) (Table 2). Logistic regression analyses indicated that having one or two copies of the TNFα-308*2 allele appeared to increase the risk of asthma (Table 2). The magnitude of this association was increased when restricting the cases to those with acute asthma (ie, one or more attacks per month) (adjusted OR=1.86; 95% CI=1.03–3.34; P=0.04). Further restricting the subjects to those with a family history of asthma (ie, ⩾first-degree relative), and those of European-American ancestry increased the association even more: adjusted OR=3.16 (95% CI=1.04–9.66; P=0.04).

Table 2 Case-control comparisons of TNF-308 and LTα-NcoI genotypes
Full size table

A substantially weaker relation was observed for LTα-NcoI: 57% of the cases vs 53% of the controls, had one or more copies of LTα-NcoI*1 (P=0.26), and possessing one or two copies of the LTα-NcoI*1 allele was marginally associated with an increased risk of asthma (adjusted OR=1.41; 95% CI=0.95–2.10; P=0.09) (Table 2).

An analysis of TNFα-308 and LTα-NcoI haplotypes gave results similar to those observed for TNFα-308 alone. In particular, comparing the TNFα-308–LTα-NcoI haplotype *2–*1 vs the haplotype *1–*2 gave an OR=1.54 (95% CI=1.05–2.24; P=0.02). Furthermore, genotype analyses that compared subjects with at least one copy of TNFα-308*2 or one copy of LTα-NcoI*1 to those without any of these alleles gave even further attenuated results (adjusted OR=1.30; 95% CI=0.88–1.92; P=0.20).

Discussion

We observed a statistically significant positive association between carrying one or two copies of the TNFα-308*2 allele and asthma risk. The magnitude of this association was increased when looking at subjects with more acute asthma, a family history, and European-American ethnicity. The weak relation between LTα-NcoI*1 and asthma may simply reflect the strong linkage disequilibrium between this allele and TNFα-308*2 (P<0.001 here), which are only 2781 bp apart.21 This would also help explain why the haplotype analysis did not strengthen the association beyond that observed for TNFα-308*2 alone.

The finding for TNFα-308*2 is biologically plausible because TNFα is a potent pro-inflammatory cytokine involved in the airway's inflammation and atopic response in asthma,2,3,7,8 and TNFα-308*2 has been shown to be associated with higher levels of TNFα production.22 Furthermore, previous genetic epidemiologic studies support our results.15,16,17,18 Other genetic epidemiologic studies, however, have been equivocal with regard to the relation between TNFα-308*2 and asthma.19,24,25,26 These conflicting results may reflect differing: study designs; sample sizes; environmental interactions; or even molecular/statistical analyses. Nonetheless, the results observed here help to further establish that TNFα-308 (or a nearby gene) is causally related to asthma. (The TNFα-308 alleles could simply be a marker for one or more functional polymorphisms in the MHC that affect asthma susceptibility.) Furthermore, this relation may be strongest among individuals with acute asthma and a family history of this phenotype.

The present study was more ethnically diverse than previous reports, and this diversity differed between cases and controls. Nevertheless, we addressed this issue in our analyses by first adjusting for ethnicity, and then restricting subjects to those with European-American ancestry. The observation that family history is a strong risk factor for disease may in part reflect the potential over-reporting of family history among cases and under-reporting among controls. In an attempt to make the self-report of family history as accurate as possible, we asked study subjects whether anyone in their immediate family currently or previously had asthma, or had ever been prescribed an asthma medicine.

In conclusion, our observations suggest that TNFα-308*2 – or a neighbouring polymorphism – is involved in the pathogenesis of asthma, and this relation is strengthened among cases with acute asthma, a family history, and European-American ancestry. Additional genetic epidemiologic studies that focus on individuals from these sub-groups, and directly measure TNFα levels in conjunction with TNF genotypes will help to further clarify the potentially important role of TNF polymorphism in asthma pathogenesis.