INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ozone, a strong oxidant, is a toxic air pollutant associated with respiratory symptoms, decrements in lung function, airway injury and inflammation, limited athletic performance, exacerbations of asthma, and increased mortality (1). The biochemical processes that initiate airway effects have not been elucidated, and there are no identified determinants or markers of susceptibility to ozone-related health effects.

Ozone likely reacts completely within the epithelial lining fluid (ELF), and interacts with apical cell membranes only where ELF is markedly attenuated (2). The protean effects of ozone exposure may thus be mediated in large part by reaction products of ozone. For example, various lipid ozonolysis products initiate signal transduction (3), and are chemotactic for polymorphonuclear leukocytes (4), alter alveolar macrophage function (5), suppress T-lymphocyte mitogenesis (4), and activate eicosanoid metabolism in airway epithelial cells (6). Products of the Criegee ozonation process include aldehydes (7):

which are sufficiently stable to be isolated and quantitated in ELF. Pryor and colleagues (8) have proposed that aldehydes may serve as biomarkers of ozone reactivity with ELF lipids. Moreover, the relative yield of various aldehyde species should be predictable from the known fatty acid content of surfactant or membrane lipids.

Aldehydes are identifiable in the bronchoalveolar lavage (BAL) fluid of animals after ozone exposure. Pryor and colleagues (8) found increases in hexanal, nonanal, and heptanal in BAL fluid from rats following 30- to 120-min exposures to ozone at levels as low as 0.5 parts per million (ppm). Concentrations increased with the addition of CO₂ to the air, indicating that the increased ventilatory rate was accompanied by increased reaction of ozone with ELF lipids.

We hypothesized that exposure of healthy humans to ozone, at concentrations found in ambient air, causes both ozonation and peroxidation of lipids in ELF or epithelial cell membranes. Ozonation would yield nonanal (C9) from oleic acid and hexanal (C6) from any n-6 unsaturated fatty acid. In addition, C6 can be produced from the ozone-initiated autoxidation of any n-6 polyunsaturated fatty acid. Both C9 and C6 would therefore be expected to increase in BAL fluid following ozone exposure. We further hypothesized that cigarette smoking alters the recovery of aldehydes, both at baseline and following ozone exposure, because of the high burden of oxidants in cigarette smoke.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study utilized BAL fluid collected from subjects in a previous investigation of the relationship between lung function responsiveness, and airway inflammation following ozone exposure. Details of the exposure facilities, methodology, and findings from that study have been reported (9, 10).

Subjects

Informed consent was obtained from all subjects, and the study was approved by the Research Subjects Review Board of the University of Rochester. Subjects (age 18-40 yr) were free of cardiorespiratory disease, denied symptoms of respiratory infection within the 3 wk preceding exposure, completed the exercise protocol, and had normal spirometry without exercise-induced bronchoconstriction. Nonsmokers had never used tobacco regularly, with no tobacco use in the 3 yr preceding the study. Smokers were currently smoking at least one pack per day, with at least 3 pack-years of smoking history.

Subjects were selected for this study based on lung function responsiveness to a 4-h exposure to 0.22 ppm ozone, with intermittent exercise (9). For the purposes of this study, ozone "responders" and "nonresponders" were selected based on decrement (> 15%) or lack of decrement (< 5%) in FEV₁, respectively. Because of the low rate of ozone responsiveness among the smokers (9), smokers were considered as a single group. Thus three groups were studied: nonsmoker nonresponders (n = 7), nonsmoker responders (n = 8), and smokers (n = 12, including one ozone responder).

Study Design

Each subject underwent a total of three exposures in an environmental chamber (two ozone and one air) and three BAL procedures, with each exposure-BAL sequence separate by at least 3 wk. Bronchoalveolar lavage was performed approximately 30 min after one of the ozone exposures (referred to subsequently as "ozone early") and 18 h after the other ozone exposure ("ozone late"). For the air exposures, BAL was randomized to either early or late.

All ozone exposures were 0.22 ppm ozone for 4 h, with exercise 20 of each 30 min, sufficient to achieve a minute ventilation (E) of approximately 25 L/min/m² body surface area. The order of the exposures was randomized and double-blinded. Smokers were not permitted to smoke during exposure or prior to BAL, but were not advised to abstain from smoking prior to exposure. Spirometry was performed before and after each exposure, and E was measured at rest and during exercise using inductive plethysmography (9).

Bronchoalveolar lavage was performed using fiberoptic bronchoscopy in both the lingula and the right middle lobe, using sterile saline stored in the original plastic containers, as previously described (10). Total and differential cell counts were performed, lavage fluids were centrifuged to remove cells, and the supernatant fluids stored at 80° C until assayed. Fluids used for analysis of aldehydes were from BAL of the right middle lobe in all subjects.

Measurement of Proteins in BAL Fluid

Concentrations of total protein, albumin, and immunoglobulin M (IgM) were determined simultaneously on all samples from each subject to provide indices of changes in epithelial permeability. Total protein was determined using the method of Lowry and coworkers (11), with crystalline bovine serum albumin as the standard. Albumin was measured using a modified antibody-capture ELISA, as described previously (12). IgM was measured using a sandwich ELISA with sensitivity in the range of 5 to 200 ng/ml. Immunoassays were validated for BAL fluid using serial dilutions and "add back" of purified antigen to confirm accurate recovery.

Aldehyde Analyses

The chemicals utilized in these assays were purchased from Sigma Chemical Co. (St. Louis, MO). The aldehydes were analyzed as oximes of pentafluorobenzylhydroxylamine (PFBHA) by gas chromatography (GC) using electron capture detection (ECD), a modification of the method described by Glaze and colleagues (13, 14).

Briefly, 2 ml of a solution (1-20 µg/L) containing the aldehydes hexanal or nonanal, or 5 ml of BAL fluid was allowed to react with 0.5 ml of a PFBHA solution (1.0 mg/ml) for 2 h. Then three drops of 18N H₂SO₄ were added and the oximes were extracted, with 1 ml of hexane containing decafluorobiphenyl (50 µg/L) as the internal standard. The hexane layer was then washed with 5 ml of 0.1 N H₂SO₄ and dried over anhydrous sodium sulfate. A Hewlett-Packard GC model 5890 series II with a ⁶³Ni electron capture detector and an autosampler (Hewlett-Packard 7361A; Palo Alto, CA) connected to a cool on-column injector with electronic pressure control was used for the analysis. An HP-5 25-m × 0.2-mm × 0.33-µm column with a 5-m × 0.53-mm retention gap was used for the separation. Helium (0.9 ml/min) was used as a carrier, and argon/methane was used as a makeup gas. The chromatographic conditions were as follows: detector temperature, 280° C; temperature programming, 50° C for 1 min; temperature ramp, 5° C/min; final temperature, 220° C. Two microliters of sample were injected. The area of the chromatograph peak was divided by the area of the internal standard peak (decafluorobiphenyl) and expressed as nmol/L of BAL fluid, based on standard calibration curves. Identity of the aldehydes was confirmed by mass spectrometry and by "spiking" of BAL fluid samples with authentic aldehyde standards.

Data Handling and Statistical Methods

The primary analyses for the data from BAL were based on a two-way mixed model or repeated measures analysis of variance (ANOVA). For a small number of end points, analysis of covariance (ANCOVA) was performed. A residual analysis was included and outliers were removed for these analyses. A level of 5% was required for statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The three subject groups did not differ with regard to age, baseline FEV₁, or E during exposure. Decrements in FEV₁ and FVC following ozone exposure were similar to the initial classification ozone exposure (9). For ozone responders, FEV₁ returned to near baseline at 18 h after exposure.

Airway Inflammation and Injury

All three subject groups showed evidence for airway inflammation in response to ozone exposure, with increases in polymorphonuclear leukocytes (PMN) and lymphocytes recovered by BAL after ozone exposure in comparison with air exposure (10). The cellular influx was greater 18 h after exposure than immediately after exposure in all groups, and there was no significant difference between groups in the intensity or time course of the response.

As shown in Figure 1, total protein, albumin, and IgM increased in response to ozone exposure in all subject groups, reaching maximal levels 18 h after exposure. Albumin showed the greatest increase and IgM the least, consistent with a permeability effect. Increases in total protein showed no differences between subject groups, but a highly significant effect of ozone exposure (p < 0.0001). The increase in albumin following ozone exposure was delayed for smokers compared with nonsmokers (Figure 1B). Ozone exposure did not alter the recovery of BAL fluid.

View larger version (37K): [in this window] [in a new window]   Figure 1.   Concentration of total protein (A), albumin (B), and IgM (C ) in BAL fluid. Total protein, albumin, and IgM increased after ozone exposure (p < 0.003 for all). The increase in albumin was delayed in smokers relative to nonsmokers (p = 0.033). Open bars: air exposure; cross-hatched bars: ozone early; solid bars: ozone late. Data are expressed as means ± SE.

Aldehydes

Both C6 and C9 were detectable in all samples measured (Figure 2). Ozone exposure resulted in a significant early increase in C9 in all groups (p = 0.0001), with no significant difference among groups. The increase in C6 was not statistically significant (p = 0.16). Both C9 and C6 levels returned to baseline by 18 h after exposure. BAL fluid from smokers contained less C6 than nonsmokers after both air and ozone exposure (p = 0.049). The levels of C6 and C9 after air exposure were similar when BAL was done either "early" or "late," indicating that timing of BAL did not influence the findings.

View larger version (34K): [in this window] [in a new window]   Figure 2.   Recovery of nonanal (C9, top panel ) and hexanal (C6, bottom panel ) in BAL fluid. Open bars: air exposure; cross-hatched bars: ozone early; solid bars: ozone late. Data are expressed as means ± SE.

ANCOVA revealed no significant relationship between aldehyde levels and group, confirming the absence of a relationship between aldehyde levels and changes in pulmonary function following ozone exposure. There was also no relationship with changes in PMN, total protein or albumin, estimated ozone dose, or subject age or sex. The increases in C9 correlated with the increases in C6 early after ozone exposure (r = 0.55, p = 0.034).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

These studies demonstrate that aldehydes are detectable in BAL fluid from smoking and nonsmoking humans, and that the recovery of C9 increases following 4-h exposures to 0.22 ppm ozone with exercise. Levels of both C9 and C6 returned to baseline by 18 h after exposure. These findings suggest that exposure to ozone at environmentally relevant levels, with exercise, results in production of lipid ozonation products in the human lung.

Recovery of C6 in BAL fluid was less in smokers than in nonsmokers (Figure 2). Explanations for this may include alterations in lipid composition of ELF of smokers, increased levels of antioxidants in ELF, or increased epithelial permeability to plasma proteins that form adducts with aldehydes in BAL fluid (5). However, the pattern of response to ozone for smokers was similar to that for nonsmokers for both C6 and C9, in spite of the significant daily exposure to oxidants (15) and the thicker mucous layer covering the airways associated with smoking (16). This surprising finding suggests that airways already exposed to a substantial oxidant burden remain susceptible to ozone-induced lipid oxidation.

These studies suggest that increases in C6 and C9 are unrelated to ozone lung function responsiveness, indicating that reactive aldehydes do not play a significant role in the airway irritant receptor response to ozone. We also found no relationship between C6 or C9 levels and airway inflammation, increases in total protein or albumin, subject gender, or age. Aldehyde levels did not correlate with estimated ozone dose, although the use of only one exposure concentration in this study does not provide sufficient data for evaluation of the concentration- response relationship.

The current data show both similarities and differences in experiments in rats exposed to ozone. Pryor and coworkers (8) found that, in rats, both C6 and C9 increased in BAL fluid following ozone exposure, with levels dependent on ozone concentration, duration, and inhalation of CO₂. In the human studies, baseline and post-exposure concentrations of C6 and C9 were lower than in the rat studies; however, this may be related to differences in dilution of ELF by the lavage procedure. Using published data on the estimated volume of ELF in rats and humans (17), we estimate that the C9 concentration in ELF at baseline was approximately 1.5 µM for rats and 6.7 µM for nonsmoking humans. Concentrations of C9 increased approximately eightfold in the rat after 90 min of exposure to 0.5 ppm with 5% CO₂, and in nonsmoking humans increased twofold following exposure to 0.22 ppm ozone for 4 h, with intermittent exercise. Levels returned to baseline by 18 h after exposure in both the human and rat studies. Thus, baseline and post-ozone concentrations of C9 in ELF appear to be of the same order of magnitude in rats and humans.

It is unclear whether the smaller increase in C6 in humans compared with rats represents a species difference or the effect of a lower exposure concentration. It may depend on the fact that C6, unlike C9, arises from both Criegee ozonation and from ozone-initiated lipid peroxidation. The relative importance of these processes may differ among species.

In conclusion, these studies suggest that exposure to ozone with exercise, at concentrations relevant to urban outdoor air, results in ozonation of lipids in ELF, with generation of C9. This effect occurs independent of smoking status or decrements in lung function following exposure. Furthermore, in this study, C9 levels did not correlate with indices of airway inflammation or injury. Further studies are needed to evaluate the utility of C9 as a marker or dosimeter of ozone exposure and to determine the role of reactive aldehydes in the airway effects of ozone.