The relationship of glutathione-S-transferases copy number variation and indoor air pollution to symptoms and markers of respiratory disease

Authors


  • Authorship and contributorship: Lars-Georg Hersoug contributed to data collection, study design, analysed data, and wrote the paper; Charlotte Brasch-Andersen developed assays and supervised genotyping; Lise Lotte Nystrup Husemoen supervised data analyses; Torben Sigsgaard contributed data analyses; Allan Linneberg designed and planned the study, supervised data analyses. All authors contributed to the discussion of results, critically revising the paper, and have approved the final version of the paper.

  • Ethics: All participants gave written informed consent; the study protocol was approved by the local ethics committee of Copenhagen County and the danish Data Protection Agency.

  • Conflicts of interest: The authors have declared that they have no conflicts of interest.

Lars-Georg Hersoug, MSc, Research Centre for Prevention and Health, Glostrup, Nordre Ringvej 57, Building 84/85, DK-2600 Glostrup, Denmark. Tel: +45 38633281, Fax: +45 38633977, email: lagehe01@glo.regionh.dk

Abstract

Introduction:  Exposure to particulate matter (PM) may induce inflammation and oxidative stress in the airways. Carriers of null polymorphisms of glutathione S-transferases (GSTs), which detoxify reactive oxygen species, may be particularly susceptible to the effects of PM.

Objectives:  To investigate whether deletions of GSTM1 and GSTT1 modify the potential effects of exposure to indoor sources of PM on symptoms and objective markers of respiratory disease.

Methods:  We conducted a population-based, cross-sectional study of 3471 persons aged 18–69 years. Information about exposure to indoor sources of PM and respiratory symptoms was obtained by a self-administered questionnaire. In addition, measurements of lung function (spirometry) and fractional exhaled nitric oxide were performed. Copy number variation of GSTM1 and GSTT1 was determined by polymerase chain reaction-based assays.

Results:  We found that none of the symptoms and objective markers of respiratory disease were significantly associated with the GST null polymorphisms. An increasing number of positive alleles of the GSTM1 polymorphism tended to be associated lower prevalence of wheeze, cough, and high forced expiratory volume in 1 s (FEV1), but these trends were not statistically significant. Furthermore, we did not observe any statistically significant interactions between GST copy number variation and exposure to indoor sources of PM in relation to respiratory symptoms and markers.

Conclusions:  In this adult population, GST copy number variations were not significantly associated with respiratory outcomes and did not modify the effects of self-reported exposure to indoor sources of PM on respiratory outcomes.

Please cite this paper as: Hersoug L-G, Brasch-Andersen C, Husemoen LLN, Sigsgaard T and Linneberg A. The relationship of glutathione-S-transferases copy number variation and indoor air pollution to symptoms and markers of respiratory disease. Clin Respir J 2012; 6: 175–185.

Abbreviations
CI

confidence interval

ETS

environmental tobacco smoke

FeNO

fractional exhaled nitric oxide

FEV1

forced expiratory volume in 1 s

GST

glutathione S-transferase

NO

nitric oxide

OR

odds ratio

PCR

polymerase chain reaction

PM

particulate matter

ROS

reactive oxygen species

Introduction

Complex interactions between genes and environment are implicated in the pathogenesis of asthma and allergic diseases (1–6). Exposure to environmental pollutants such as particles, tobacco smoke, and oxidant gasses may trigger production of reactive oxygen species (ROS) in inflammatory cells and bronchial epithelial cells (7), which in turn may facilitate the development of asthma (8). Exposure to particulate matter (PM) generated in the indoor environment may be particularly relevant to respiratory disease pathogenesis in developed countries, where people spend much of their time indoors. For example, studies have indicated that fine particles generate more ROS than larger particles (9, 10). In Health 2006, a population-based, cross-sectional study of adults in Denmark, we found that self-reported exposure to indoor sources of environmental tobacco smoke (ETS) was associated with a significantly increased prevalence of wheeze, chronic cough and decreased lung function [forced expiratory volume in 1 s (FEV1)% predicted](11). However, we did not find any effects of exposure to the use of woodstoves or candles in the home.

Enzymatic detoxification of ROS and other products generated by inflammatory cells may ameliorate the deleterious effects of environmental exposures (2, 12). Enzymes in the glutathione S-transferases (GSTs) family detoxify ROS and xenobiotics, which are present in smoke from combustion sources (7). There exist well-known deletions (null polymorphisms) of two GST genes GSTM1 and GSTT1 genes. The genotype frequencies of these genes vary in different populations. In Caucasians the frequency of two copies of the null genotype has been reported to be 53% for GSTM1 and 20% for GSTT1 (13). Results of an earlier Danish study suggested that positive alleles of the GSTM1 and GSTT1 genes might have a protective effect on development of atopic asthma (14). Conversely, deficiencies of the GSTM1 and GSTT1 genes have been shown to increase the effect of passive smoking on the risk of childhood asthma and wheezing (15, 16). We hypothesised that individuals homozygous for GSTT1 or GSTM1 null mutations might be more susceptible to the effects of exposure to PM than individuals without those mutations and that positive GSTM1 and GSTT1 alleles might protect against ROS-induced adverse effects in a dose-dependent manner.

We used a polymerase chain reaction (PCR)-based assay to determine the copy number of GSTT1 and GSTM1 alleles for all participants in the Health 2006 study (14). We then performed logistic regression analysis to investigate possible effects of copy number of positive GSTM1 or GSTT1 alleles on symptoms and markers of respiratory disease. Additionally we aimed to investigate whether null alleles of GST genes modify the observed effects of exposure to ETS and whether effects of exposure to woodstoves/candles would be detectable in carriers of the GST null polymorphisms.

Materials and Methods

Study population

The Health 2006 study is a population-based, cross-sectional study investigating the prevalence and risk factors of chronic diseases such as asthma, allergies, cardiovascular disease and diabetes (17). The study took place between June 2006 and June 2008. A random sample of the general population aged 18–69 years living in the South-Western part of Copenhagen was drawn from the Civil Registration System. A total of 7931 persons were invited to a health examination and 3471 (43.8%) participated. Participation was significantly associated with female gender and increasing age (P < 0.001) (17). The health examination included a self-administered questionnaire and blood drawing. The study including the present analyses was approved by the ethics committee of Copenhagen County (KA-20060011) and the Danish Data Protection Agency. Informed written consent was been obtained from all participants.

Genotyping of copy number variation of GSTM1 and GSTT1

Two multiplex real-time PCR-based assays were used for the detection of GSTM1 and GSTT1 genomic DNA copy number. These assays are based on amplification and quantification of GSTM1 or GSTT1 sequence in relation to a reference gene, albumin, in a multiplex PCR using StepOnePlus™ real-time PCR system (Life Technologies, Carlsbad, CA, USA). The method has been slightly modified from previously described by Brasch-Andersen et al. (14): PCR was performed in a volume of 5 µL using 10 ng of DNA. Failed samples were re-run once. Furthermore the PCR was performed in a fast mode using TaqMan® Fast Universal PCR Master Mix (Applied Biosystems) thus reducing both costs and reaction time. A pilot study genotyping 93 samples from the cohorte in triplicate was performed to investigate the quality of the DNA. For GSTT1 all 93 samples gave the same genotype. For GSTM1 one sample repeatedly resulted in a delta Ct value in the interval between the cluster of samples with one copy and the cluster of samples with two copies. This sample was not genotyped. The remaining cohorte was genotyped once for each GST-gene. If a sample did not fall within a cluster it was not assigned a genotype. Results were analysed using StepOne software v 2.1. (Life Technologies, Carlsbad, CA, USA). Gene copy numbers were assigned using the Delta Ct method as previously described (14) using a fixed threshold (0.1) for all genes. The genotyping was performed blinded. A total 32 (0.9%) and 52 (1.5%) samples could not be assigned to GSTM1 or GSTT1 genotypes, respectively, because of poor DNA quality.

Definition of respiratory symptoms and asthma

Information on asthma and lower respiratory symptoms was obtained from a self-administrated questionnaire. Asthma was defined as a positive answer to the question ‘Have a doctor ever told you that you have asthma?’ Hay fever was defined as a positive answer to the question ‘Has a doctor ever told you that you have hay fever?’

Wheezing was defined as a confirmative answer to the question: ‘Have you had wheezing or whistling in your chest at any time during the last 12 months?’‘Wheezing without a cold’ was defined as wheezing in combination with a confirmative answer to the question: ‘If yes, have you had this wheezing or whistling when you did not have a cold?’ Chronic cough was defined as a confirmative answer to both the questions: ‘Have you had coughing most days or nights?’ and ‘Do you cough this way for a minimum of 3 months a year and for at least 2 successive years?’

Allergic rhinitis symptoms was defined as a positive answer to at least one of the following three questions: (i) Have you within the last 12 months had an itchy or stuffy nose or sneezing when near grass, trees or flowers'; or (ii) ‘Have you within the last 12 months had an itchy or stuffy nose or sneezing when near furred animals as horse, dog, cat, rabbit, guinea pig or hamster'; or (iii) ‘Have you within the last 12 months had an itchy or stuffy nose or sneezing when cleaning rooms or making beds or when in bed?’

Definition of atopy

Serum samples were analysed for serum-specific immunoglobulin E (IgE) against the four most clinical important inhalant allergens in Denmark, i.e. birch, grass, cat and a house dust mite (Dermatophagoides pteronyssinus), by using the ADVIA Centaur® Specific IgE assay (Bayer HealthCare Diagnostics Divition, Tarrytown, NY, USA) (18). Atopy was defined as at least one positive test (≥0.35 KU/L) for specific IgE.

Assessment of lung function

Spirometry was performed according to international standards (19) using the SpiroUSB from MicroMedical Limited, Rochester, Kent, UK. Every morning the two spirometers were checked with a 3-L syringe. Every 6 months the spirometers were checked with a decompression flow simulator (20). Predicted FEV1 was calculated from measured standing height, age and sex (21).

Fractional exhaled nitric oxide (FeNO) measurement

FeNO measurements were performed using the NIOX MINO (Aerocrine AB, Stockholm, Sweden). Participants were tested in a standing upright position without a nose clip. Through a mouthpiece, participants inhaled to total lung capacity and exhaled at a constant pressure guided by visual and auditory cues to stabilise flow rate. A dynamic flow restrictor yielded a constant flow rate of 50 mL/s, in accordance with recommendations of the American Thoracic Society and European Respiratory Society guidelines for FeNO measurement (22). The device measured NO concentrations between 5 and 300 ppb. Measurements below 5 ppb were automatically set to 0.

Definition of exposure to indoor sources of particles

Exposure to ETS was assessed by the question ‘How many hours per day do you usually stay in rooms exposed to tobacco smoke?’ with the following answer categories: ‘More than 5 h’, ‘1–5 h’, ‘1/2–1 h’, ‘Never’.

Exposure to woodstove smoke was assessed by the questions:

  • (i) ‘Is it possible to supplement the heating in your residence with a fireplace/woodstove?’
  • (ii) If yes, ‘How often are you using the woodstove/fireplace during the winter season?’ with the following answer categories: ‘Daily (7 days per week)’, ‘2–6 days per week’, ‘1 day per week or less’, ‘Never or very seldom’.

Exposure to use of candles was assessed by the question: ‘How often are you lighting candles during the evening in winter season?’ with the following answer categories: ‘Daily (7 days per week)’, ‘2–6 days per week’, ‘1 day per week or less’, ‘Never or very seldom’.

Statistical analyses

Statistics were performed using the statistical program SAS, version 9.2 (SAS Institute Inc. Cary, NC, USA). All P values reported are two-tailed and statistical significance was defined as P < 0.05.

Relative risks of symptoms associated with exposure to sources of PM were estimated as odds ratios (with 95% confidence intervals) by logistic regression models. The FeNO and FEV1% predicted were analysed by linear regression. Occasional smokers were excluded from the analyses (n = 113). Smoking status was scored as: never smoker, previous smoker, current smoker <5 g tobacco, current smoker 5 to <10 g tobacco, current smoker 10 to <15 g tobacco, current smoker 15 to <20 g tobacco and current smoker >20 g tobacco. The following potential confounders were considered: gender; age group (18–28 years, ≥28–38 years, ≥38–48 years, ≥48–58 years and >58–69 years); season of examination (December–February, March–May, June–August, September–November). Before linear regression analyses FeNO values = 0 was set to 3 ppb and the distribution normalised by the natural logarithm (ln). Inverse transformations of model estimates were performed to provide proportional differences in FeNO. Linear trends (dose–response relationships) across ordered categories were tested by scoring the categories and modeling the variables as continuous variables in the models. P values for linear regression were log-likelihood ratio tests, and P values for logistic regression were Wald χ2 tests. Effect modifications by GST genotypes (GSTT1 or GSTM1) were examined by including the relevant interaction terms: GST*smoking status, GST*ETS, GST*use of woodstoves, and GST*use of candles. Hardy–Weinberg test were performed by using the Hardy–Weinberg Calculator from the ‘Online Encyclopedia for Genetic Epidemiology studies’(23).

Results

Characteristics of the study population are presented in Table 1. Distributions of both GSTM1 and GSTT1 genotypes were in Hardy–Weinberg equilibrium (P = 0.179 and P = 0.920, respectively). Of 3364 subjects with analysable samples, 1770 subjects (52.6%) were homozygous for the null GSTM1 allele, 1322 subjects (39.3%) were heterozygous, with one positive GSTM1 allele and a null allele, and 272 subjects (8.1%) had two or more positive alleles. Six subjects had more than two positive GSTM1 alleles (0.002%). For the GSTT1 analysis, data were available for 3344 samples, with the genotype frequencies as follows: homozygous null genotype 493 subjects (14.7%); heterozygous genotype, 1578 subjects (47.2%); and homozygous positive genotype, 1273 subjects (38.1%). None of the participants had more than two copies of the GSTT1 positive alleles.

Table 1. Characteristics of the Health 2006 study population
 n/total (%)
  • *

    The glutathione S-transferase (GST) genotype distributions of GSTM1 and GSTT1 were both in Hardy–Weinberg equilibrium (P = 0.179 and P = 0.920, respectively).

Sex 
 Men1553/3471 (44.7%)
 Women1918/3471 (55.3%)
Age in years 
 18–28269/3471 (7.8%)
 ≥28–38444/3471 (12.8%)
 ≥38–48872/3471 (25.1%)
 ≥48–58854/3471 (24.6%)
 ≥58–691032/3471 (29.7%)
GSTM1 genotype* 
 GSTM1 null1770/3364 (52.6%)
 GSTM1 one copy1322/3364 (39.3%)
 GSTM1 two or more copies272/3364 (8.1%)
GSTT1 genotype* 
 GSTT1 null493/3344 (14.7%)
 GSTT1 one copy1578/3344 (47.2%)
 GSTT1 two or more copies1273/3344 (38.1%)
Smoking status 
 Current773/3437 (22.5%)
 Occasionally113/3437 (3.3%)
 Previously1116/3437 (32.5%)
 Never1435/3437 (41.8%)
Type of housing 
 Apartment963/3424 (28.1%)
 House1629/3424 (47.6%)
 Row house782/3424 (22.8%)
 Other50/3424 (1.5%)
Size of residence; square meters (m2) 
 Less than 70 m2364/3438 (10.6%)
 70–110 m21380/3438 (40.1%)
 More than 110 m21694/3438 (49.3%)
Persons living in the residence 
 1 person512/3447 (14.9%)
 2 persons1592/3447 (46.2%)
 3 persons524/3447 (15.2%)
 4 persons617/3447 (17.9%)
 >4 persons202/3447 (5.9%)
Exposure to environmental tobacco smoke (hours per day)
 Never2216/3434 (64.5%)
 ½–1422/3434 (12.3%)
 1–5391/3434 (11.4%)
 ≥5405/3434 (11.8%)
Use of woodstove during wintertime 
 Daily (7 days per week)357/3416 (10,5%)
 2–6 days per week304/3416 (8.9%)
 1 day per week or less.148/3416 (4.3%)
 Never or very seldom164/3416 (4.8%)
 No woodstove2443/3416 (71.5%)
Use of candles during wintertime 
 Daily (7 days per week)1150/3421 (33.6%)
 2–6 days per week1400/3421 (40.9%)
 1 day per week or less.578/3421 (16.9%)
 Never or very seldom293/3421 (8.6%)

Table 2 presents the prevalence of diagnoses and symptoms of respiratory diseases, lung function, and FeNO in relation to GSTM1 and GSTT1 genotype copy number variation. GST copy number variation was not significantly associated with any of the respiratory outcomes. However, increased GSTM1 copy number was weakly associated with decreased risk of wheeze, wheeze without a cold, chronic cough, and mean FEV1% predicted, though these trends were not statistically significant. In addition, when GST genotypes were dichotomised (homozygous null genotype vs other genotypes), as has been employed in most previous studies, there were no statistically significant associations of GST genotypes with respiratory outcomes (Supporting Information Table S1).

Table 2. Prevalence of diagnoses, symptoms, and objective markers of respiratory diseases in relation to GSTM1 and GSTT1 copy number variation (number of positive alleles)
CharacteristicsGSTM1 null % (n/total n)GSTM1 1 copy % (n/total n)GSTM1 2 copies % (n/total n)P values*GSTT1 null % (n/total n)GSTT1 1 copy % (n/total n)GSTT1 2 copies % (n/total n)P values*
  • *

    P value Wald χ2 test/P for trend performed by logistic regression with genotype coded as ordered categorical variable.

  • Atopy was defined as a positive test for serum-specific immunoglobulin E against at least one of birch, grass, cat or D. pteronyssinus.

  • P value Kruskal–Wallis test. P for trend performed by linear regression with the continuous variable as dependent variable and genotype (coded as ordered categorical variable) as independent variable.

  • FEV1, forced expiratory volume in 1 s; FeNO, fractional exhaled nitrogen oxide; GST, glutathione S-transferase; OR, odds ratio; CI, confidence interval.

Atopy22.2 (392/1769)25.0 (330/1321)24.3 (66/272)0.177/0.11324.1 (119/493)23.2 (366/1577)23.6 (300/1272)0.909/0.907
Wheeze23.4 (409/1746)21.2 (279/1314)19.7 (53/269)0.203/0.07624.0 (117/488)21.7 (338/1560)22.1 (279/1261)0.562/0.564
Wheeze without a cold16.1 (279/1738)14.6 (191/1309)13.8 (37/269)0.414/0.19014.6 (71/486)15.3 (237/1554)15.5 (194/1256)0.909/0.688
Asthma (doctor-diagnosed)10.7 (186/1737)11.2 (146/1303)10.0 (27/269)0.823/0.98111.1 (54/486)10.2 (158/1551)11.4 (143/1253)0.565/0.611
Hay fever (doctor-diagnosed)17.8 (308/1734)18.5 (240/1297)18.2 (49/270)0.871/0.68318.5 (90/586)17.2 (266/1544)19.3 (241/1252)0.379/0.444
Chronic cough16.0 (279/1748)14.6 (191/1310)13.8 (37/269)0.448/0.21014.3 (70/489)15.8 (245/1556)15.1 (190/1262)0.718/0.891
Allergic rhinitis symptoms54.2 (949/1751)55.3 (727/1314)49.6 (135/272)0.229/0.25029.8 (146/490)31.6 (494/1566)33.9 (428/1263)0.197/0.073
Allergic rhinitis symptoms and atopy15.8 (276/1751)18.5 (243/1314)17.7 (48/272)0.131/0.09415.9 (78/490)16.8 (263/1565)17.8 (225/1262)0.592/0.308
FEV1 < 80% of predicted9.5 (168/1762)11.6 (152/1313)10.3 (28/272)0.185/0.1879.8 (48/491)10.0 (157/1567)10.9 (138/1269)0.693/0.418
FeNO > 20 ppm31.4 (505/1611)29.6 (357/1205)30.8 (76/247)0.618/0.48828.1 (129/459)30.4 (436/1434)31.8 (368/1158)0.345/0.151
Mean FEV1% predicted97.998.099.50.231/0.21498.298.497.80.568/0.466
Mean FeNO in ppm14.314.614.50.847/0.86514.614.214.60.612/0.897

We also examined whether there was a combined effect of having both GSTM1 null and GSTT1 null compared with the other genotypes, but this combined GST null genotype was not significantly associated with any of the respiratory outcomes (data not shown).

Table 3 shows the association of self-reported exposure to indoor sources of PM with ‘wheeze without a cold’ stratified by GSTM1 and GSTT1 genotype. Overall, smoking and exposure to ETS tended to be associated with increased prevalence of ‘wheeze without a cold’ independent of GST genotype. However, none of the tested interactions between indoor exposures and GST genotype were found to be statistically significant (all P values > 0.2) (Table 3). Similarly, none of the tested interactions between indoor exposures and GST genotypes were found to be statistically significant in relation to lung function (all P values > 0.1) (Table 4). However, we observed a significant effect of smoking and ETS exposure on lung function independent of GST genotypes. In line with these results, we did not find any statistically significant interactions between exposure to indoor sources of PM and GST genotypes in relation to the prevalence of atopy, wheeze, self-reported doctor-diagnosed asthma, self-reported doctor-diagnosed hay fever, chronic cough, and FeNO (data not shown).

Table 3. The association of indoor sources of particulate matter with ‘wheeze without a cold’ stratified by GSTM1 and GSTT1 genotype
Indoor exposuresGSTM1 with null copiesGSTM1 with 1 or 2 copiesP value for interactionGSTT1 with null copiesGSTT1 with 1 or 2 copiesP value for interaction
OR (CI 95%)*OR (CI 95%)*OR (CI 95%)*OR (CI 95%)*
  • *

    Odds ratios (95% confidence intervals) obtained by logistic regression analyses adjusted for gender, age, season of examination and variables shown in the table.

  • P value Wald χ2 test/P value for trend test.

  • ETS, environmental tobacco smoke; OR, odds ratio; CI, confidence interval.

Smoking status
Never1.00 Reference1.00 Reference 1.00 Reference1.00 Reference 
Previous1.53 (1.07; 2.17)1.40 (0.96; 2.04) 1.41 (0.72; 2.77)1.49 (1.12; 1.98) 
Current 1–10 g2.01 (1.01; 4.01)1.16 (0.51; 2.62) 0.97 (0.20; 4.75)1.79 (1.03; 3.10) 
Current 10–15 g2.02 (1.10; 3.72)1.64 (0.78; 3.46) 1.15 (0.32; 4.18)2.06 (1.24; 3.41) 
Current 15–20 g3.10 (1.71; 5.65)1.79 (0.86; 3.74) 3.07 (0.91; 10.37)2.52 (1.53; 4.14) 
Current >20 g2.91 (1.73; 4.89)4.99 (2.87; 8.70) 2.61 (0.87; 7.83)4.04 (2.70; 6.06) 
 P < 0.001/<0.001P < 0.001/<0.0010.316P = 0.388/0.055P < 0.001/<0.0010.365
Exposure to ETS (hours per day)
Never1.00 Reference1.00 Reference 1.00 Reference1.00 Reference 
½–11.34 (0.86; 2.06)1.09 (0.67; 1.78) 1.53 (0.70; 3.36)1.19 (0.83; 1.71) 
1–51.19 (0.74; 1.90)1.17 (0.70; 1.95) 0.80 (0.25; 2.50)1.26 (0.87; 1.82) 
≥52.31 (1.47; 3.61)1.65 (0.98; 2.78) 1.36 (0.51; 3.59)2.21 (1.54; 3.19) 
 P = 0.002/<0.001P = 0.309/0.0860.821P = 0.602/0.586P < 0.001/<0.0010.321
Use of woodstove during wintertime (days per week)
No woodstove1.00 Reference1.00 Reference 1.00 Reference1.00 Reference 
Rare users0.97 (0.49; 1.93)1.25 (0.67; 2.36) 0.62 (0.08; 3.01)1.14 (0.70; 1.86) 
1 day0.91 (0.43; 1.91)1.07 (0.50; 2.28) 1.52 (0.44; 5.18)0.89 (0.49; 1.61) 
2–6 days0.81 (0.48; 1.37)1.51 (0.93; 2.45) 1.51 (0.65; 3.51)1.08 (0.72; 1.61) 
7 days1.04 (0.65; 1.66)0.62 (0.35; 1.10) 0.34 (0.08; 1.50)0.88 (0.61; 1.29) 
 P = 0.946/0.755P = 0.153/0.6800.245P = 0.417/0.726P = 0.895/0.7000.475
Use of candles during wintertime (days per week)
Never users1.00 Reference1.00 Reference 1.00 Reference1.00 Reference 
1 day1.04 (0.65; 1.66)0.96 (0.51; 1.80) 0.84 (0.27; 2.64)0.77 (0.49; 1.20) 
2–6 days1.34 (0.86; 2.06)0.84 (0.48; 1.49) 0.77 (0.28; 2.14)0.83 (0.56; 1.24) 
7 days1.19 (0.74; 1.90)1.02 (0.57; 1.83) 1.02 (0.37; 2.84)0.78 (0.52; 1.18) 
 P = 0.457/0.500P = 0.746/0.8400.604P = 0.855/0.807P = 0.651/0.4460.793
Table 4. Association of indoor sources of particulate matter with lung function (FEV1% predicted) stratified by GSTM1 and GSTT1 genotype
Indoor exposuresGSTM1 with null copiesGSTM1 with 1 or 2 copiesP value for interactionGSTT1 with null copiesGSTT1 with 1 and 2 copiesP value for interaction
ß estimate*ß estimate*ß estimate*ß estimate*
  • *

    β estimate (95% confidence intervals) obtained by linear regression analyses adjusted for gender, age, season of examination and variables shown in the table.

  • P value Wald χ2 test/P value for trend test.

  • ETS, environmental tobacco smoke.

Smoking status
Never0.00 Reference0.00 Reference 0.00 Reference0.00 Reference 
Previous−2.76 (−4.38; −1.13)−2.05 (−3.83; −0.28) −3.65 (−6.80; −0.50)−2.52 (−3.84; −1.20) 
Current 1–10 g−4.44 (−8.20; −0.68)−6.03 (−9.88; −2.17) −8.03 (−15.24; −0.83)−5.42 (−8.39; −2.46) 
Current 10–15 g−3.98 (−7.37; −0.57)−5.04 (−8.99; −1.10) −3.65 (−9.94; 2.64)−4.72 (−7.58; −1.87) 
Current 15–20 g−4.86 (−8.49; −1.22)−5.72 (−9.88; −1.55) −6.26 (−13.86; 1.35)−5.39 (−8.34; −2.44) 
Current >20 g−8.31 (−11.34; −5.29)−12.95 (−16.31; −9.59) −6.33 (−12.43; −0.22)−10.98 (−13.43; −8.54) 
 P < 0.001/<0.001P < 0.001/<0.0010.281P = 0.061/0.010P < 0.001/<0.0010.871
Exposure to ETS (hours per day)
Never0.00 Reference0.00 Reference 0.00 Reference0.00 Reference 
½–11.50 (−0.73; 3.74)0.02 (−2.42; 2.46) 0.46 (−3.59; 4.51)0.63 (−1.21; 2.47) 
1–5−1.88 (−4.39; −0.61)−0.93 (−3.63; 1.78) −4.27 (−9.80; 1.25)−0.96 (−2.92; 1.01) 
≥5−4.12 (−6.77; −1.47)−2.12 (−5.10; 0.86) −6.65 (−11.89; −1.42)−2.64 (−4.82; −0.47) 
 P = 0.003/0.006P = 0.545/0.1860.567P = 0.057/0.016P = 0.064/0.0350.740
Use of woodstove during wintertime (days per week)
No woodstove0.00 Reference0.00 Reference 0.00 Reference0.00 Reference 
Rare users0.88 (−2.54; 4.30)1.07 (−2.28; 4.43) −1.66 (−9.33; 6.00)1.25 (−1.28; 3.78) 
1 day−0.80 (−4.39; 2.79)5.42 (1.68; 9.16) 3.41 (−3.27; 10.10)2.38 (−0.44; 5.19) 
2–6 days−0.31 (−2.80; 2.18)2.18 (−0.51; 4.87) −0.96 (−5.61; 3.70)1.28 (−0.76; 3.33) 
7 days0.72 (−1.64; 3.08)0.97 (−1.53; 3.46) 2.83 (−2.02; 7.68)0.50 (−1.36; 2.36) 
 P = 0.920/0.281P = 0.039/0.0650.145P = 0.597/0.420P = 0.336/0.1910.761
Use of candles during wintertime (days per week)
Never users0.00 Reference0.00 Reference 0.00 Reference0.00 Reference 
1 day1.56 (−1.41; 4.54)0.31 (−2.83; 3.46) 1.96 (−3.91; 7.82)1.04 (−1.32; 3.40) 
2–6 days1.47 (−1.22; 4.17)2.91 (0.07; 5.74) 3.97 (−1.29; 9.24)1.99 (−0.15; 4.13) 
7 days2.14 (−0.61; 4.89)2.14 (−0.78; 5.05) 1.19 (−4.17; 6.55)2.47 (0.27; 4.66) 
 P = 0.489/0.173P = 0.052/0.0600.362P = 0.249/0.919P = 0.101/0.0150.367

Discussion

In this general adult population, we could not confirm that GSTT1 and GSTM1 null polymorphisms were associated with increased prevalence of symptoms, diagnoses or objective markers of respiratory disease. Furthermore, persons who were homozygous for the GST null polymorphisms did not appear to be more susceptible to the potential health effects of exposure to indoor sources of PM.

We have previously reported data from our study population showing that self-reported exposure to ETS was associated with increased prevalence of wheeze and ‘wheeze without a cold’ and decreased lung function (FEV1% predicted) (11). In contrast, exposure to the use of woodstoves and candles was not associated with any of the respiratory outcomes (11). In the present analyses, we hypothesised that the potential effects of exposure to indoor PM would be strongest in persons homozygous for the GST null polymorphisms. However, we did not find any statistically significant interactions between GST genotype and indoor exposure to PM (or own smoking) in relation to any of the respiratory outcomes. In this study we have only assessed the exposure variables of PM by questionnaires. Direct measurements of PM exposures would have been preferable, but is difficult in large epidemiological studies. There are great differences on toxicity and concentrations among the PM sources. Cigarette smoke, ETS and wood smoke contain transition metals and chemicals that induce ROS in the airways. Although, wood stoves are potential sources of PM it also improves ventilation of the homes as the air that escape through the chimney is replaced by outside air. PM from burning candles generate ultrafine particles but with low content of transition metals. It is well known that smoking cause respiratory symptoms and from our own study on this cohort we have seen respiratory symptoms of ETS, but not from use of woodstove and candles (11). We found in the present study that persons homozygous for the GST null polymorphisms did not appear to be more susceptible to the effects of PM sources even among smokers and persons exposed to ETS. In contrast, a few other studies have observed significant interactions between in utero exposure to maternal smoking or ETS and the GSTM1 and GSTT1 null alleles with respect to risk of asthma and decreased lung function during childhood (16, 24–26). Thus, it may be hypothesised that the effects of GST copy number variation may mainly play a role in early life. This effect on early life development is suggested because smoking during pregnancy affects the growth of the fetus and the maturation of the fetal lung, resulting in an impairment of lung function in the newborn (27). It seems that GSTM1 and GSTT1 play a protective role on lung function growth because studies have suggested that pulmonary function only was affected in children with a GSTT1 deficiency and exposed to ETS (28). Another study has shown that also GSTM1 deficient children appear to have impaired lung function growth (25).

We also investigated whether GST genotypes in themselves (GST nulls, GSTs with only one positive allele and GSTs with two positive alleles) were associated with respiratory outcomes. This was possible as the used PCR assay could quantify the number of GST alleles (14) and not just determine presence or absence of alleles as most prior studies involving GSTM1 and GSTT1. However, none of the respiratory symptoms and objective markers of respiratory diseases were significantly associated with GST genotypes. This result that none of the polymorphisms, even the GST nulls, were associated with respiratory symptoms and objective markers of respiratory diseases appears to be somewhat inconsistent with the results of earlier studies. A few small case-control studies have suggested that the prevalence of GSTM1 and GSTT1 null polymorphisms is increased in patients with asthma with concurrent atopy (29, 30). On the other hand, our results are in line with a recent meta-analysis (31). In that meta-analysis, asthma risk was associated with the homozygous null genotype of both GSTT1 and GSTM1, but these associations disappeared when the smallest studies were excluded from the meta-analysis suggesting that publication bias might have biased the results of the meta-analysis. The authors concluded that there was limited evidence to support an association of GSTM1 and GSTT1 null genotypes with asthma phenotypes. They also analysed studies including data on copy number variation. These results showed no dose–response effect of the positive allele of the GSTM1 gene, but a small protective effect on asthma of the GSTT1 positive allele could not be excluded. In our study, we found no statistically significant dose–response effect of having one or two functional alleles of either GSTM1 or GSTT1 compared with those homozygous for the null alleles. Nevertheless, the increasing number of the positive allele of the GSTM1 tends to be associated with lower risk of some asthma phenotypes, but these trends were not statistically significant.

In another meta-analysis four studies were identified where the combined effect of GSTM1 null and GSTT1 null on asthma were analysed. That analysis showed a significantly increased risk of asthma among persons homozygous for the null allele of both GSTM1 and GSTT1 (32). In the present study, we could not confirm this.

In our study we could not replicate previously reported associations between GST variants and asthma phenotypes. Failure of replication of results from genetic association studies is a well-known and a common problem. There may be several reasons for this. First of all, some associations may arise by chance (type 1 errors) and since positive findings may more likely be published than negative findings, the literature may not accurately reflect the overall balance of evidence. Furthermore, complex genetic diseases, such as asthma, may result from effects of multiple genes and their interactions with multiple environmental factors, each of which may have differential impact in different populations (33, 34).

In our study population, the prevalence of the homozygous null genotype of the GSTM1 and GSTT1 polymorphisms were 52.6% and 14.7%, respectively. This corresponds well with what has been reported in another Danish cohort where the prevalence of the homozygous null polymorphisms genotype for GSTM1 was 55.3% (24). In other Caucasian populations the prevalence of homozygous null polymorphisms of GSTM1 and GSTT1 genes has been reported to be 53% and 20%, respectively (13).

There are several strengths of the present study. Firstly, this was a population-based study in a genetically relatively homogenous population. Secondly, we used several objective markers of respiratory disease. Thirdly, we used quantitative PCR to genotype GSTM1 and GSTT1 for zero, one or two copies in a large cohort providing the opportunity to study gene-dose–response effects on respiratory outcomes. In most of the previous studies the GSTM1 or GSTT1 were genotyped only as presence or absence of the positive allele, and therefore these studies were unable to investigate a potential gene-dose-responce effect. Several limitations should be considered as well. Firstly, information about exposure to indoor sources of PM was obtained by questionnaires and there were no objective measurements of PM exposure. Secondly, the generalisability of the study may have been decreased by the relatively low participation rate. Nevertheless, it seems less plausible that genotype status has influenced participation. Finally, the cross-sectional study design does not allow us to make strong conclusions about the temporal relationship between exposure and disease. A longitudinal follow-up of the present cohort is scheduled for 2011–2012.

In conclusion, we found no evidence that GST null polymorphisms modify the potential effects of self-reported exposure to indoor sources of PM on respiratory health. Furthermore, we did not find a significant gene-dose-response effect of the positive alleles of GSTT1 and GSTM1 on respiratory outcomes.

Acknowledgements

We thank the staff at Research Centre for Prevention and Health for skilfully performed data collection and Pernille Jordan at Odense University Hospital for genotyping of the GST genotypes. The project is financially supported by The Velux Foundation; The Danish Medical Research Council, Danish Agency for Science, Technology and Innovation; The Aase and Ejner Danielsens Foundation; ALK-Abelló A/S, Hørsholm, Denmark; and Research Centre for Prevention and Health, the Capital Region of Denmark.

Ancillary