SEARCH

SEARCH BY CITATION

Keywords:

  • ADAM33;
  • asthma;
  • bronchial hyperresponsiveness;
  • in utero cigarette smoke exposure;
  • lung function

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Introduction: ADAM33 is the first identified asthma gene by positional cloning, especially asthma combined with bronchial hyperresponsiveness (BHR). Moreover, ADAM33 is associated with early-life lung function and decline of forced expiratory volume in 1 s (FEV1) in the general population. In utero and postnatal cigarette smoke exposure (CSE) are associated with reduced lung function, and development of BHR and asthma. We hypothesized that this may occur via interaction with ADAM33.

Aim:  To replicate the role of ADAM33 in childhood lung function and development of BHR and asthma. Furthermore, we investigated gene–environment interaction of ADAM33 with in utero and postnatal CSE in the Dutch PIAMA cohort.

Methods:  Six ADAM33 single-nucleotide polymorphisms (SNPs) were genotyped. Rint was measured at age 4 and 8 years, FEV1 and BHR at age 8 years; asthma was based on questionnaire data at age 8.

Results:  In the total cohort, the rs511898 A, rs528557 C, and rs2280090 A alleles increased the risk to develop asthma (+BHR). There existed interaction between in utero but not postnatal CSE and the rs528557 and rs3918396 SNPs with respect to development of BHR, the rs3918396 SNP with Rint at age 8 and the rs528557 SNP with FEV1% predicted.

Conclusions:  We confirm associations between ADAM33 and the development of asthma (+BHR). This is the first study suggesting that interaction of in utero CSE with ADAM33 results in reduced lung function and the development of BHR, which needs further confirmation.

Impaired lung function and bronchial hyperresponsiveness (BHR) in early childhood are well-known risk factors for the development of asthma later in life (1–3). The origins of impaired lung function and BHR have not been elucidated so far, but it is plausible that biologic processes that underlie in utero and early-life lung development are important contributing factors. It is clear that asthma and BHR have strong genetic and environmental components. These environmental factors often start in pregnancy and early infancy, and include a.o. pre and postnatal cigarette smoke exposure (CSE) (4, 5).

ADAM33 is the first asthma gene identified by positional cloning (6). This gene was in particular found by combining a doctor’s diagnosis of asthma with the presence of BHR. Interestingly, ADAM33 SNPs have also been associated with impaired early-life lung function at age 3–5 years, and with forced expiratory volume in 1 s (FEV1) decline in both adult asthma patients and the general population (7–10).

Multiple ADAM33 protein isoforms exist in human embryonic lung (11), as is the case in the lungs of adult asthma patients as well (11, 12). Not only airway smooth muscle cells but also primitive mesenchymal cells of the fetal lung, forming a cuff at the end of the growing lung bud, stain with ADAM33 (11). Thus, ADAM33 may well play a role in orchestrating branching morphogenesis and polymorphic variations in the ADAM33 gene may influence the subsequent susceptibility of the lung to asthma.

Several epidemiologic studies have provided evidence that CSE in utero and early life is associated with reduced lung function and constitutes a risk factor for the development of asthma and BHR (13, 14). It is thus conceivable that in utero and early-life CSE interact with genes that are important in asthma development, thereby affecting in utero and early-life lung development, airway remodeling, and subsequent development of early-life lung function changes and asthma.

We studied the role of ADAM33 in early-life lung function, BHR and asthma in a Dutch birth cohort (prevention and incidence of asthma and mite allergy; PIAMA) that allowed us to prospectively follow environmental exposure, symptoms and lung function, including BHR measurements up to 8 years of age. We analyzed whether CSE in utero and in early life, in combination with ADAM33 genotypes influences early-life lung function, BHR, and asthma.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Study design

Prevention and incidence of asthma and mite allergy is a birth cohort study that started in 1996. At baseline, 4146 children were included, 1327 high-risk children (with a positive maternal history for atopy) and 2819 children of nonallergic mothers. The children were recruited during the first trimester of pregnancy. Parents completed annual questionnaires derived from the International Study of Asthma and Allergies in Childhood (ISAAC) (15) about allergic symptoms in the child, information about general health, indoor environment like CSE, socio-economic characteristics, lifestyle, breastfeeding, and demographic factors. All 1327 high-risk children and a random sample of 663 low-risk children were selected for medical examinations. Resistance interrupter technique (Rint; Micro Medical, Rochester, Kent, UK) measurements were obtained at age 4 years, and Rint as well as spirometry and BHR to methacholine bromide were measured at age 8 years. The Rint technique measures interrupter resistance, which is similar to airway resistance. It is a noninvasive method that requires only tidal breathing. Therefore, it can be easily performed in preschool children (16–18). A further description of the study protocol has been published previously (19). All parents provided written informed consent.

DNA

Genomic DNA was extracted from buccal swabs or blood (20) at age 4 years (n = 1037). Genetic studies were approved by local medical ethics committees of participating institutes. Six single-nucleotide polymorphisms (SNPs), i.e. rs511898, rs612709, rs3918396, rs528557, rs2280090 and rs2787094 were analyzed. These SNPs were chosen as they were associated with asthma and lung function in the Dutch population (8, 9, 21). Children with a non-Caucasian ethnicity were excluded (n = 51).

For all SNPs, primers and probes were ordered from Applied Biosystems TaqMan® SNP Genotyping Assays (Nieuwekerk aan de IJssel, The Netherlands), using the Assay-by-Design service for which we provided sequences, as the rs528557 SNP previously failed (8) we used primers and probes with the Primer Express package and obtained primers from Biolegio (Malden, The Netherlands) and probes from Applied Biosystems. Sequences of all primers and probes had been shown previously (8). Reactions were performed in 5 μl volumes of 10 ng DNA, 1× Taqman Universal Mastermix (Applied Biosystems), 200 nM probe, and 900 nM primer. Cycling conditions on the ABI PRISM® 7900HT Sequence Detection System (Applied Biosystems) were 10 min at 95ºC followed by 40 cycles of 15 s at 95ºC and 1 min at 60ºC. End-point fluorescence was measured immediately after cycling. Alleles were assigned using sds 2.1 software (Applied Biosystems). Genotyping call rates ranged from 89% to 94.5%, comparable with the study by Schedel et al. (22).

Phenotypes

Rint measurements were performed as described previously (23). Rint measurements at age 4 and 8 years and FEV1% predicted at age 8 years were used as lung-function variables. Bronchial hyperresponsiveness was defined as a fall of ≥20% in FEV1 after inhalation of a maximum of 0.62 mg methacholine bromide. We used two definitions of asthma: (i) asthma at age 8, based on at least one episode of wheeze and/or dyspnea in the last year and/or the use of inhaled steroids; and (ii) asthma in combination with BHR.

In utero CSE was defined as a confirmative answer to the question ‘do you smoke cigarettes?’ or ‘have you been smoking, and when did you quit smoking?’ asked to the mother during the third trimester of pregnancy. Cigarette smoke exposure in the first year of life was defined as a confirmative answer to the question ‘does anyone smoke, at least once a week, in the home?’ asked to the parents when the child was 1 year of age. Lung function, BHR, asthma, and asthma with BHR were analyzed in the total group and in children with in utero CSE and CSE at age 1 year separately.

Statistical analysis

Lung-function parameters were normally distributed and analyzed using anova and linear regression. Bronchial hyperresponsiveness, asthma, and asthma with BHR were analyzed using chi-squared tests and logistic regression. An interaction term of genotype and CSE (in utero and in the first year of life) was included into linear and logistic regression models. All analyzes were corrected for gender. All genotypes were analyzed in the best fitting, dominant model, i.e. by comparing children with one or two copies of the minor alleles with those carrying no copy. Analyzes of ADAM33 in combination with CSE analyzes were corrected for gender, atopy, or asthma of the mother, atopy, or asthma of the father, siblings, breastfeeding, pet, and in utero CSE or CSE at age 1 year. Calculations were performed using spss 14.0 statistical software. Benjamini and Hochberg’s (24) false discovery rate was used to correct for multiple testing.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

DNA and Rint values at age 4 years were available in 768 children, DNA and spirometry at age 8 years were available in 746 children. As inhalation medication taken before Rint measurement could influence the outcome, we excluded children who took inhalation medication <12 h prior to Rint measurements (n = 25 at 4 years and n = 7 at 8 years). General characteristics of the study population are shown in Table 1.

Table 1.   Characteristics
CharacteristicsGenotyped children
  1. BHR, bronchial hyperresponsiveness.

  2. *Within children with complete methacholine provocation measurements.

Number1037
Boys (%)532 (51.2)
Family history (%)
 Atopy mother690 (66.5)
 Atopy father327 (31.6)
Environmental exposures (%)
 In utero smoke exposure154 (15.0)
 Smoke exposure in first year230 (23.2)
Asthma153 (16.5)
Lung function
 Number779
 FEV1% predicted mean (SD)107.0 (11.5)
Methacholine provocation
 Number692
 BHR (%)313 (45.2)
 Asthma with BHR* (%)72 (10.2)

Genotyping

All SNPs were in Hardy–Weinberg equilibrium (P > 0.05). There existed strong linkage disequilibrium (LD) between the SNPs rs511898, rs612709, rs3918396, rs528557, and rs2280090, (D’>0.8), (Fig. 1). Allele frequencies are described in Table 2.

image

Figure 1.  Linkage disequilibrium (LD) structure of ADAM33 SNPs in the PIAMA cohort (r2 below and D’ above the diagonal).

Download figure to PowerPoint

Table 2. ADAM33 SNPs and their minor allele frequency in PIAMA cohort
rs numberAlleles*SNP†LocationMAF
  1. MAF, minor allele frequency.

  2. *First allele, wild type.

  3. †Single-nucleotide polymorphism (SNP) numbering as shown in previous publications (6, 8, 9, 21).

rs511898G/AF+1Catalytic domain0.35
rs612709C/TQ-1Cysteine-rich domain0.13
rs3918396G/AS_1Trans membrane domain0.09
rs528557G/CS_2Trans membrane domain0.28
rs2280090G/AT_2Cytoplasmic domain0.14
rs2787094C/GV_43′UTR0.25

Association analysis

Lung function.  No association was found between SNPs in ADAM33 and FEV1 and Rint in the total cohort (data not shown).

BHR, asthma, and asthma combined with BHR.  No association was found between SNPs in ADAM33 and BHR in the total cohort. Children carrying the rs511898 A allele or rs528557 C allele were more likely to develop asthma compared with children with the rs511898 GG or rs528557 GG genotypes. Children carrying the rs528557 C allele or rs2280090 A allele were more likely to develop asthma in combination with BHR compared with the rs528557 GG or rs2280090 GG genotypes (Table 3).

Table 3.   Association of ADAM33 SNPs with asthma and asthma combined with BHR in the PIAMA cohort
  1. BHR, bronchial hyperresponsiveness.

  2. The values in bold are significant at P ≤ 0.05, in a dominant model. Odds ratios and (95% CI), wild type is reference group.

  3. *Numbers genotyped.

SNPAllelen*BHRAsthmaAsthma with BHR
rs511898GG397111
GA3971.1 (0.8–1.5)1.4 (1.0–2.1)1.5 (0.8–2.8)
AA1221.2 (0.8–2.0)1.7 (1.0–2.9)2.0 (0.9–4.3)
rs612709CC684111
CT2160.8 (0.5–1.2)1.2 (0.8–1.9)1.1 (0.6–2.1)
TT110.2 (0.0–1.9)1.9 (0.5–7.4)1.2 (0.1–12.2)
rs3918396GG745111
GA1531.0 (0.6–1.5)1.4 (0.9–2.3)1.2 (0.6–2.4)
AA90.4 (0.0–3.9)0.7 (0.8–5.3)1.3 (0.1–13.1)
rs528557GG458111
GC3301.1 (0.8–1.6)1.4 (0.9–2.0)1.8 (1.0–3.2)
CC730.8 (0.4–1.5)1.9 (1.0–3.6)1.8 (0.7–4.6)
rs2280090GG659111
GA2191.2 (0.8–1.7)1.5 (1.0–2.3)1.7 (0.9–3.2)
AA172.6 (0.8–8.9)1.0 (0.2–4.4)3.8 (0.6–23.9)
rs2787094CC520111
CG3340.9 (0.7–1.3)1.0 (0.7–1.5)1.0 (0.6–1.8)
GG591.1 (0.8–1.5)0.4 (0.1–1.0)0.9 (0.3–2.8)

Interaction of ADAM33 genotypes and in utero smoke exposure. FEV1: In children whose mother smoked during pregnancy, the rs528557 C allele was associated with lower FEV1% predicted compared with children with the GG genotype. In children whose mother did not smoke during pregnancy, the rs528557 C allele was associated with higher FEV1% predicted levels compared with children with the GG genotype (interaction term P = 0.006, Fig. 2A). A similar trend was present for rs3918396 [smoke exposed children mean FEV1% predicted 104.0% (A allele) vs 105.9% (GG genotype) and children not exposed to smoke in utero mean FEV1% 108.1% (A allele) vs 106.9% (GG genotype)].

image

Figure 2.  (A) Interaction of ADAM33 SNPs and in utero CSE in level of FEV1 at 8 years of age. CSE+, smoke exposure during pregnancy; CSE−, no smoke exposure during pregnancy. *P ≤ 0.05; **P ≤ 0.01. (B) Interaction of ADAM33 SNPs and in utero CSE in the development of BHR at 8 years of age. CSE+, smoke exposure during pregnancy; CSE−, no smoke exposure during pregnancy. *P ≤ 0.05; **P ≤ 0.01.

Download figure to PowerPoint

Rint: A similar type of interaction as seen with FEV1 was present for Rint at age 8 years and rs3918396 SNP, showing increased Rint values in children with the rs3918396 A allele in children whose mother smoked during pregnancy (interaction term P = 0.02). A similar trend was present for rs528557. No interaction was present for Rint at age 4 years and in utero CSE.

BHR: We found a significant interaction between in utero CSE and the ADAM33 SNPs rs528557 and rs3918396 with respect to the development of BHR. Within children whose mother smoked during pregnancy, the rs528557 C or rs3918396 A allele increased the risk for BHR development compared with rs528557 GG or rs3918396 GG [rs528557 odds ratio (OR) 4.0; 95% confidence interval (CI): 1.5–10.4, P-value interaction term = 0.004; rs3918396 OR 4.0; 95% CI: 1.1–14.2, P-value interaction term = 0.02]. This genotype effect was not found in children whose mothers did not smoke during pregnancy. Conversely, in children with the rs528557 GG or rs3918396 GG genotype in utero CSE decreased the risk of BHR, compared with all the children who were not exposed to cigarette smoke (Fig. 2B). The same trend was present for the rs511898 SNP (data not shown).

The groups became too small to analyze an interaction between in utero CSE and ADAM33 genotypes with respect to asthma development.

Interaction of ADAM33 genotypes and smoke exposure in the first year of life.  No significant interaction was found between CSE in the first year of life and the ADAM33 genotypes with respect to the development of BHR and level of lung function at the age of 4 and 8 years (data not shown).

Correction for multiple testing.  None of the above described analyzes reached significance after correction for multiple testing.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Our results confirm the previously reported association of ADAM33 SNPs with asthma as well as asthma combined with BHR (6). We extend previous findings in that we provide data suggesting that gene–environment interaction of ADAM33 genotypes and in utero CSE exist with respect to the level of lung function, and the development of BHR at age 8 years, an effect not observed with CSE in the first year of life.

ADAM33 is expressed in embryonic lung tissue on mesenchymal cells (11), which are responsible for branching, a lung developmental phase that continues until ±17 weeks of gestation (12). The respiratory airways and future respiratory units develop during the third trimester. This is followed by an alveolar phase of lung development, which proceeds into the first 2–3 years of postnatal life, during which the number of alveoli increases (25). ADAM33 may thus play a role especially in prenatal lung development, for instance by orchestrating branching morphogenesis. In this way, it can affect the development of lung function, BHR, and asthma. Furthermore, soluble ADAM33 in bronchial alveolar lavage fluid is increased in asthma patients, correlates with disease severity, and reduced lung function (26). A recent in vitro and in vivo study showed that soluble ADAM33 causes endothelial cell differentiation and affects angiogenesis in embryonic lung tissue (27). Angiogenesis is considered an important aspect of tissue inflammation and remodeling. How would in utero CSE influence ADAM33-shedding? This may well run by increased levels of transforming growth factor-β (TGF-β) by smoke exposure (28), as Puxeddu et al. showed that TGF-β2 enhances ADAM33-shedding from cells expressing ADAM33 (27).

A recent report on another prospective birth cohort study, the Manchester Asthma and Allergy Study (MAAS) cohort, showed at the age of 5 years that four SNPs in ADAM33 were associated with reduced FEV1 (rs511898, rs3918395, rs2280091, and rs2280090; P < 0.04) (7). Our study extends the latter observations. It appears that smoking by mothers during pregnancy is an important contributor to the effects of ADAM33 on lung function, as the rs528557 SNP was associated with reduced FEV1 in children with in utero CSE, with an opposite effect on FEV1 in children not exposed. Interestingly, the percentage of children with in utero CSE in the MAAS cohort is much higher than in the PIAMA cohort (respectively 22.3% and 15%) (29). It is therefore plausible that the association of ADAM33 with reduced lung function in the MAAS cohort is predominantly present in the children whose mother smoked during pregnancy. In addition to the interaction of in utero CSE with ADAM33 SNPs on lung function, we found interaction on the development of BHR as well. In utero CSE is known to influence lung function development and to increase the risk to develop asthma and BHR (13, 30–32). Our data suggests that this association may at least in part be caused by an interaction with ADAM33.

Of interest and requiring further functional study is the observation that the rs528557 GG and rs3918396 GG genotype have a protective effect on BHR when in utero smoke exposure occurred. It is yet unclear, and has to be determined whether this implies an effect on smooth muscle proliferation, epithelial integrity, and/or airway branching.

The interaction of ADAM33 with in utero CSE and CSE in childhood was also analyzed by Schedel et al. (22) in children from the ISAAC II and Multicentre Study of Allergy (MAS) cohorts. They did not find evidence for interaction of ADAM33 SNPs with asthma, BHR, or lung function. The difference in outcome between their and our study may be caused by, up to now unidentified environmental factors, or by methodological differences between the studies (e.g. methacholine provocation vs histamine and cold air provocation).

We replicate the initial association of ADAM33 genotypes with asthma and asthma with BHR as reported by Van Eerdewegh et al. (6), thereby confirming the importance of ADAM33 in especially asthma with BHR. Not all studies have confirmed association of ADAM33 SNPs and asthma (33, 34), and the studies that do replicate the results often find association with different SNPs or different alleles. There are several explanations for this discrepancy. First, there is variability in the definition of asthma used in the different studies. We defined asthma at age 8 based on at least one episode of wheeze or dyspnea in the last year and/or the use of inhaled steroids. Furthermore, we analyzed asthma in combination with BHR, as linkage became stronger combining asthma and BHR in the initial study by Van Eerdewegh et al. (6). Some studies with negative results analyzed asthma solely based on a doctor’s diagnosis (34, 35). Secondly, a meta-analysis showed that ADAM33 has only a moderate effect on the development of asthma (OR: 1.4 for SNPs known to date) (36). Therefore, as previously suggested (22, 36), it is plausible that even with large datasets the chance of finding associations with such a moderate risk is limited. Thirdly, it may be possible that discrepant results between the studies are caused by differences in environmental factors and population heterogeneity. We show in our study that the effects of ADAM33 polymorphisms on lung function, and BHR in the total population were only present in combination with CSE, suggesting that ADAM33 is a gene that can be influenced by environmental factors. Finally, it is also possible that the associated SNPs are not the causative ones, but in LD with the causative SNPs within ADAM33 or in genes located near ADAM33. This might also explain why different SNPs and alleles are associated in the studies. This phenomenon is also known as loose genotypic replication (37).

This study has some pitfalls, which should be acknowledged. First, our study was based on the selection of SNPs that were previously associated with lung function and asthma. We found strong LD between several genotyped SNPs, so it is plausible that the associations we observed with different SNPs and similar phenotypes is caused by one single SNP that is in strong LD with the other SNPs. Therefore, a full description of ADAM33 in terms of polymorphic variations and LD structure is needed. Secondly, our results did not remain significant after correction for multiple testing. However, as this is a hypothesis testing study and (part of) the results are a replication of previously published data we feel that correction for multiple testing will underestimate the results. Notwithstanding this, we do recognize that our results might be attributable to chance and that other studies have to replicate especially our new findings. Furthermore, as mothers who are smoking during pregnancy are likely to smoke after pregnancy it is difficult to study the independent effect of CSE. However, several studies have shown a strong effect of especially in utero smoke exposure on lung function and asthma development (30–32). In our study, the majority of the children who were exposed in the first year of life (59%) were not exposed during pregnancy. Cigarette smoke exposure in the first year of life only did not show any interaction with ADAM33.

We conclude that ADAM33 genotypes increase the risk to develop asthma and asthma in combination with BHR. Furthermore, we have provided data that need confirmation, but suggest gene–environment interaction of ADAM33 genotypes and in utero CSE with the level of early-life lung function and development of BHR. In combination with data in the literature, our study highlights that it is important to investigate cigarette smoking during pregnancy as a risk factor for early-life lung function and BHR development, in interaction with genetic factors like ADAM33.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

N.E. Reijmerink is a MD-medical research trainee and funded by The Netherlands Organisation for Health Research and Development, number 920-03-379. The genetic study was financially supported by ZonMW grant number 912-03-031 and the Spinoza premium. The PIAMA study is supported by the Netherlands Asthma Foundation, the ZonMW Netherlands Organisation for Health Research and Development, and the Ministry of the Environment.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References