Published ahead of print on September 11, 2003, doi:10.1164/rccm.200305-684OC
© 2003 American Thoracic Society Nicotinamide Adenine Dinucleotide (Phosphate) Reduced:Quinone Oxidoreductase and Glutathione S-Transferase M1 Polymorphisms and Childhood AsthmaDivision of Intramural Research, National Institute of Environmental Health Sciences, National Institutes of Health, Department of Health and Human Services, Research Triangle Park, North Carolina; National Institute of Public Health, Cuernavaca; and Hospital Infantil de Mexico Frederico Gomez, Mexico City, Mexico Correspondence and requests for reprints should be addressed to Gloria David, Ph.D., National Institute of Environmental Health Sciences, PO Box 12233, MD D2-01, Research Triangle Park, NC 27709. E-mail: davidbe1{at}niehs.nih.gov
Nicotinamide adenine dinucleotide (phosphate) reduced:quinone oxidoreductase (NQO1) and glutathione S-transferase (GST) M1 are phase II enzymes important in response to oxidative stress, such as occurs during exposure to ozone. We examined the relationship between functionally significant polymorphisms in NQO1 (Pro187Ser) and GSTM1 (homozygous deletion) and asthma risk in children with high lifetime exposure to ozone. We enrolled children with asthma from the allergy referral clinic at a public pediatric hospital in Mexico City, together with their parents. We assayed for the Pro187Ser polymorphism in NQO1 using a polymerase chain reaction–restriction fragment length polymorphism assay and for the presence of GSTM1 by polymerase chain reaction among 218 case–parent triads. We did not find strong evidence of an association between NQO1 genotype alone and asthma risk. However, among subjects with homozygous deletion of GSTM1, carriers of a serine allele were at significantly reduced risk of asthma compared with Pro/Pro homozygotes (relative risk = 0.4; 95% confidence interval, 0.2–0.8). The p value for difference in relative risk for NQO1 by GSTM1 genotype = 0.013. These data are consistent with a protective effect of the NQO1 Ser allele in this population of GSTM1-null children with high ozone exposure.
Key Words: case–parent triad • oxidative stress • environmental tobacco smoke • GSTM1 • NQO1 Ozone is a potent oxidant known to induce a variety of respiratory effects, including symptoms of cough and inspiratory pain, increased airway reactivity, reduction in lung function, and bronchoalveolar inflammation (1, 2). Exposure to ozone has been clearly related to asthma severity (1). Recent evidence supports a role of chronic exposure to ambient ozone in development of asthma in adults (3) and children (4). Response to ozone varies between individuals. The pattern of variability in response to ozone, with high interindividual variability combined with a low within-subject variability, is consistent with genetic susceptibility (5, 6). Mudway and Kelly have estimated that only 10 to 20% of healthy subjects are susceptible to ozone-induced decrements in lung function and airway reactivity (2). Studies in inbred mice have identified regions linked to pulmonary response to ozone (7, 8). The responsible genes in humans have not been identified.
Oxidative stress mechanisms are involved in response to ozone. Ozone imposes an oxidative burden on the lung by directly oxidizing biomolecules, thereby generating reactive oxygen species (ROS), and by activating inflammatory cells, such as macrophages, neutrophils, and eosinophils, which release ROS (2). Oxidative stress mechanisms may also be important in asthma. Markers of oxidative stress, including F2-isoprostanes, ethane, and 3-nitrotyrosine, are elevated in patients with asthma compared with control subjects (9–11). Antioxidants and antioxidant enzymes are present in the lung and provide protection against oxidative stress. Levels of the antioxidants Nicotinamide adenine dinucleotide (phosphate) reduced:quinone oxidoreductase (NQO1) and glutathione S-transferase (GST) M1 are phase II detoxifying enzymes induced in response to oxidative stress, such as occurs during ozone exposure (15, 16). NQO1 catalyzes the two-electron reduction of quinones to hydroquinones, thus bypassing the potentially toxic semiquinone radical intermediate (17). NQO1 can also function as an antioxidant enzyme by reducing ubiquinone (coenzyme Q10) and vitamin E quinone to their antioxidant forms (18, 19). In addition to its detoxifying role, NQO1 can also catalyze the bioactivation of some quinones to more reactive hydroquinones that, in turn, autooxidize to produce ROS or undergo rearrangement to generate alkylating species (17). For example, NQO1 activates dinitropyrenes in diesel exhaust and nitro compounds, such as 4-nitroquinoline-1-oxide, in cooked foods (20, 21). GSTs catalyze the conjugation of a wide variety of electrophiles, including reactive products of lipid, protein, and DNA oxidation, to the thiol of reduced glutathione producing less reactive water-soluble compounds that are more readily excreted (22). Functional polymorphisms have been identified for both NQO1 and GSTM1. Traver and coworkers identified a common coding polymorphism in NQO1: a C to T substitution resulting in an amino acid change Pro187Ser (23). Subjects homozygous for the variant Ser allele have no detectable NQO1 enzyme activity and no detectable or only trace levels of NQO1 protein due to increased degradation of the variant protein (24–26). NQO1 protein and activity levels are significantly lower in heterozygous compared with homozygous wild-type individuals (25, 27). A common homozygous deletion polymorphism of the GSTM1 gene abolishes enzyme activity (28). Bergamaschi and colleagues recently studied ozone-induced acute effects on respiratory function in 24 healthy nonsmokers performing 2-hour bike rides in summer ambient ozone and found that lung function of subjects with the NQO1 Pro/Pro genotype and null for GSTM1 was diminished at ozone concentrations below the U.S. standard (29). In a recent chamber study, biomarkers of oxidative stress and inflammation were increased only among individuals with a combination of NQO1 Pro/Pro and GSTM1-null genotypes (30). These findings suggest a protective effect for the NQO1 Ser allele in GSTM1-null subjects. To date, there are no published studies of the NQO1 genotype in relation to asthma risk, either alone or in combination with GSTM1. We used the case–parent triad design to examine the association between the NQO1 Pro187Ser polymorphism and asthma risk in children from Mexico City, an area with the highest ozone levels in North America. According to data from the Mexico City Monitoring Network for 1998 to 2000, the 1-hour daily maximum ozone level ranged from 12 to 309 ppb, with a mean ± SD of 102 ± 47 ppb (13). We also examined whether the interaction between NQO1 and GSTM1 polymorphisms observed with pulmonary responses to ozone (29) is present in relation to asthma risk in this highly ozone exposed population. Some of the results of this study have been previously reported in the form of an oral presentation/abstract (31).
Study Design and Subject Enrollment Using the case–parent triad design (32, 33), nuclear families consisting of a case and both parents were recruited. We enrolled cases of asthma at the allergy clinic of a large public hospital in central Mexico City (Hospital Infantil Frederico Gomez) between October 1998 and November 2001. Children were aged 4 to 17 years, and asthma was diagnosed in the children by a pediatric allergist. Clinical work-up included skin-prick and pulmonary function testing. Mothers were interviewed to obtain questionnaire data on risk factors for asthma, and clinical data were abstracted from the medical records. Children and parents provided a blood sample as a source of DNA. The protocol was reviewed and approved by the Institutional Review Boards of the Mexican National Institute of Public Health and the U.S. National Institute of Environmental Health Sciences. Written informed consent for the children's participation was obtained from parents.
Genotyping of NQO1 Pro187Ser and GSTM1 The GSTM1 gene was detected by polymerase chain reaction as described by Bell and coworkers (34). This assay distinguishes homozygous null genotypes from those with one or two copies of the gene present. ß-globin was coamplified as a positive control. Among the 245 complete triads enrolled, we were able to assign genotypes for both NQO1 Pro187Ser and GSTM1 for 239 triads. Twenty-one triads were excluded due to probable nonpaternity on the basis of analysis of nine tetranucleotide short-tandem repeats (AmpFLSTR Profiler Plus; Applied Biosystems, Foster City, CA), leaving a total of 218 triads for analysis.
Statistical Analysis We used a log-linear–based method to test for asymmetric distribution of a particular variant allele among affected offspring and their biologic parents (32). Under the null hypothesis that asthma risk is equal among offspring marker genotypes, the distribution of offspring genotypes within parental genotype combinations is expected to follow Mendelian proportions. Under an alternative hypothesis, in which asthma risk varies among offspring genotypes, the distribution of offspring genotypes within parental genotypes would be expected to differ from Mendelian proportions. This can happen because nuclear families are sampled conditional on an affected offspring. The method uses genotypes of cases and their parents stratified into the 15 possible types of triads. By expressing Mendelian proportions within parental mating types in terms of genotype relative risks (RRs), a log-linear model allows direct estimation of point estimates and 95% confidence intervals (CIs) for genotype RRs as well as valid likelihood ratio tests of the null hypothesis that genotype RRs equal one. Likelihood ratio tests are valid (i.e., do not exhibit excess false-positive [type I] errors) in the presence of cryptic population structure because they are based on fitting a model for Mendelian proportions stratified by parental mating type. For NQO1, we estimated 95% CIs for genotype RRs and performed likelihood ratio tests of the null hypothesis that genotype RRs equal one. We tested for genotype interaction between NQO1 and GSTM1 by treating the GSTM1 genotype of the child (deletion homozygote vs. others) as a binary stratification factor and performed a one or two degree of freedom likelihood ratio test for interaction between NQO1 and GSTM1. This approach produces a valid test because NQO1 and GSTM1 are unlinked and, as a consequence, are conditionally independent (35). One and two degrees of freedom likelihood ratio tests for interaction were performed, depending on whether NQO1 genotypes bearing one or two copies of the serine allele were combined or not. Data were analyzed using SAS version 8.2 (SAS Institute, Cary, NC) and STATA version 7.0 (Stata Corporation, College Station, TX).
Demographic and clinical characteristics of the children are presented in Table 1 . The mean ± SD age of the children was 9.0 ± 2.5 years (range 4–17 years). Boys (58%) outnumbered girls (42%). Skin test data were available for 204 children with 82% testing positive (wheel diameter > 15 mm) for at least one antigen. Only 6% of the children were exposed in utero, but 51% were currently exposed to environmental tobacco smoke by a parent in the home. All families were of low socioeconomic status, being seen at the public hospital covering the uninsured. There was no significant difference in the characteristics by NQO1, GSTM1, or combined genotypes of the children (data not shown); therefore, they are not potential confounders.
For NQO1 Pro187Ser, the frequency of the variant allele among the parents was 0.430 (95% CI, 0.397–0.463). These genotype frequencies were in Hardy–Weinberg equilibrium (p = 0.21). The frequency of the GSTM1-null genotype among the children was 0.394 (95% CI, 0.347–0.442). The frequencies of both the NQO1 variant allele and GSTM1-null genotype are similar to previous reports in U.S. populations of Mexican origin (36, 37). Table 2 shows the frequency distributions of the 15 possible mating types of triads for the Pro187Ser polymorphism. The triad genotype (mother–father–child) is based on the number of copies of the variant allele (Ser). Subjects carrying at least one variant Ser allele for NQO1 exhibited slightly, but not statistically significantly, decreased risk of asthma (RR = 0.8; 95% CI, 0.5–1.1) relative to subjects with the Pro/Pro genotype (Table 3) . We observed no gene dosage effect.
We examined the RRs for the NQO1 Pro187Ser polymorphism in relation to asthma risk within strata of the child's genotype for GSTM1 (null or positive). The frequency distributions of the 15 possible triad types stratified by GSTM1 are given in Table 2. Among the GSTM1-null, subjects carrying at least one Ser allele for NQO1 were at statistically significant decreased risk of asthma (RR = 0.4; 95% CI, 0.2–0.8) relative to those with Pro/Pro genotype (Table 3). This reduction in risk was not observed in children who were GSTM1 positive. The p value for the difference in RR by GSTM1 genotype was 0.013, consistent with a statistically significant interaction between GSTM1 and NQO1 genotypes. It has been suggested that the role of the GSTM1 deletion polymorphism in relation to asthma risk in children may vary according to in utero exposure to maternal smoking (38). Maternal smoking during pregnancy is uncommon in our population (6%) compared with the California population (16%) previously studied. Thus, with only 13 children exposed, we could not further stratify the NQO1–GSTM1 association by in utero exposure. However, we did examine the association stratified by exposure to a smoking parent in the home. The protective effect for the NQO1 Ser allele in GSTM1-null children was limited to those with nonsmoking parents (RR = 0.1; 95% CI, 0.0–0.5) (Table 4) . The p value for the difference in RR by parental smoking status was 0.05, consistent with a statistically significant interaction between NQO1–GSTM1 combined genotypes and exposure to environmental tobacco smoke in the home.
Subjects carrying an inactive NQO1 allele exhibited only a slight, nonstatistically significant, decreased risk of asthma. However, a significant reduction in asthma risk was observed among individuals carrying at least one Ser allele for NQO1 and also having homozygous deletion of GSTM1. The finding of lower risk of asthma for subjects with an inactive NQO1 allele who also have deficiency of GSTM1, in this highly ozone exposed population, is consistent with previous data on acute respiratory effects in ozone-exposed subjects (29). Bergamaschi and colleagues reported that subjects with the combined NQO1 Pro/Pro and GSTM1-null genotypes are more susceptible to adverse effects of ambient ozone. Their studies imply decreased susceptibility to ozone among subjects carrying at least one Ser allele and who are GSTM1-null. They reported this susceptibility both with pulmonary function decrements after exercise in ambient ozone and increases in biomarkers of oxidative stress after short-term controlled exposures (29, 30). A protective effect of NQO1 Ser alleles in GSTM1-null children in relation to asthma risk, in this highly ozone exposed population, is biologically plausible. The decreased asthma risk observed in children with the variant NQO1 Ser allele, which is expected to have reduced activity, is consistent with the role of NQO1 in metabolic activation. Although often detoxifying, NQO1 can catalyze the reduction of some quinones to hydroquinones, which are more reactive and which can autooxidize to generate ROS (17). For example, some nitroaromatic compounds and heterocyclic amines present in diesel exhaust are activated by NQO1 (21, 39, 40). Diesel exhaust, consisting of a complex mixture of particulate matter, has been associated with asthma exacerbation (41–43). The effect of diesel exhaust on respiratory function may be partially mediated by ROS (44, 45). Free radical scavengers can attenuate diesel exhaust particle–induced expression of proinflammatory mediators by human bronchial epithelial cells (46). A recent study by Squadrito and coworkers suggests that generation of ROS by inhaled airborne particulate matter in the lung is sustained by redox cycling of semiquinones (47). In vitro studies in human airway epithelial cells and in vivo studies in rats have shown that treatment with diesel exhaust particles induces expression of NQO1 (48, 49). According to data from the Mexico City Monitoring Network for 1998 to 2000, the 24-hour average particulates with mass median diameter less than 10 µm ranged from 9.92 to 249.06 µg/m3, with a mean ± SD of 56.68 ± 27.36 µg/m3, which exceeds the U.S. standard of 50 µg/m3. We postulate that ROS generated during NQO1-mediated activation of quinones present in particulates with mass median diameter less than 10 µm may interact with ozone in increasing oxidative stress in the lung and may subsequently lead to an increased risk of asthma in susceptible individuals (Figure 1) . An interactive effect between particulates with mass median diameter less than 10 µm and ozone levels has been observed in Mexico City in relation to respiratory mortality (50). In subjects who are GSTM1 positive, quinones could be conjugated to reduced glutathione (51–53) and excreted and thus unavailable for activation by NQO1. GSTs also can conjugate reactive products of lipid, protein, and DNA oxidation (52, 53). Individuals without the added protection of GSTM1 could therefore be at increased risk of oxidative stress. This may explain the decreased risk of asthma in those who are both GSTM1-null and who carry the variant inactive Ser allele for NQO1.
Although it is possible that the GSTM1 genotype on its own, without interaction with the NQO1 genotype, could influence asthma risk, studies are inconsistent (38, 54). We could not determine the association between asthma risk and effect of the GSTM1 genotype alone because the polymerase chain reaction assay does not distinguish subjects with one versus two nondeleted alleles, and detection of heterozygosity is required for analyses of triads data. We have also observed in a sample of study subjects enrolled in a trial of antioxidants that children with GSTM1-null genotype were more sensitive to ozone-induced decrements in pulmonary function (55). Environmental tobacco smoke contains quinones and polyaromatic hydrocarbons that may be activated by NQO1. We observed a significant protective effect for the NQO1 Ser allele among GSTM1-null children of nonsmoking parents only. Children of smoking parents were not protected. Particle levels in smoking homes will be largely driven by this exposure, not outdoor air pollution sources (56). It is conceivable that, at very high levels of exposure, genetic variation in the ability to activate quinones could be largely irrelevant. NQO1 and GSTM1 are only two of the many genes involved in the metabolism of quinones and in response to oxidative stress. Two-electron reduction by NQO1 directly competes with one-electron reductases, such as phase I cytochrome P450s. One-electron reduction of quinones generates unstable semiquinones that undergo redox cycling in the presence of molecular oxygen to form highly ROS (Figure 1). Baulig and coworkers have observed that CYP1A1 is induced in human airway epithelial cells exposed to diesel exhaust particles (49). GSTP1 and GSTT1 can also conjugate quinones and ROS, thereby reducing oxidative stress, and may be associated with asthma (54, 57). Genetic variants affecting asthma risk can be found at any point on the pathway from ozone/diesel exhaust/environmental tobacco smoke exposure to increased pulmonary inflammation and decreased lung function. Asthma is indeed a complex disease, and susceptibility likely involves multiple gene–gene and gene–environment interactions. In this study of children with a high lifetime exposure to ozone, subjects carrying at least one NQO1 Ser allele and homozygous for the GSTM1 deletion were at a decreased risk of asthma. These data are consistent with previous reports that GSTM1 and NQO1 genotypes interact in relation to adverse consequences of acute ozone exposure (29, 30). Additional genetic epidemiologic studies, perhaps using alternative designs (i.e., population-based case–control) are needed to confirm this observation and to extend it to other ethnic groups.
G.L.D. has no declared conflict of interest; I.R. has no declared conflict of interest; J.J.S-M. has no declared conflict of interest; W.J.C. has no declared conflict of interest; M.R-A. has no declared conflict of interest; B.E.d.R-N. has no declared conflict of interest; N.I.R-R. has no declared conflict of interest; R.W.M. has no declared conflict of interest; J.M.M. has no declared conflict of interest; S.J.L. has no declared conflict of interest. The authors acknowledge the efforts of Bioserve Biotechnologies (Laurel, MD) for GSTM1 genotyping; Rolv T. Lie for providing the STATA analysis program; Grace Chiu, Ph.D. (Westat, Research Triangle Park, NC) for data management and analysis; and Mark Bruno, Susan Baker, and Shirley Richards (CODA, Research Triangle Park, NC) for specimen handling. The authors thank Irma Lara and Dulce Ramirez for their participation in the fieldwork, and Drs. Lesley Butler and Darryl Zeldin for a critical review of the manuscript.
Supported by National Council of Science and Technology (CONACYT) grant 26206-M, the National Center for Environmental Health at Centers for Disease Control, and National Institute of Environmental Health Sciences, Department of Intramural Research Z01 ES49019, National Institutes of Health, Department of Health and Human Services. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form May 22, 2003; accepted in final form September 9, 2003
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
-tocopherol (vitamin E) and ascorbic acid (vitamin C) are decreased in the lung lining fluid of patients with asthma compared with control subjects (
2 test. 
