TRACHEAL AND BRONCHOALVEOLAR PERMEABILITY CHANGES IN RATS INHALING OXIDANT ATMOSPHERES DURING REST OR EXERCISE

Permeability of tracheal and bronchoalveolar airways of rats was measured and used to examine the effects of inhaled oxidant-containing atmospheres. The atmospheres studied were (a) ozone (O 3 ) at 0.6 ppm (1.2 mg/m 3 ) or 0.8 ppm (1.6 mg/m 3 ); (b) nitrogen dioxide (NO 2 )at 6 ppm (11.3 mg/m 3 ) or 12 ppm (22.6 mg/m 3 ); (c) O 3 + NO 2 at 0.6 ppm (1.2 mg/m 3 ) and 2.5 ppm (4.7 mg/m 3 ), respectively; and (d) a 7-component particle and gas mixture (complex atmosphere) representing urban air pollution in a photochemical environment. The rats were exposed for 2 h. The effects of exercise during exposure were evaluated by exposing additional groups in an enclosed treadmill. Exposure of resting rats to 0.8 ppm O 3 increased tracheal permeability to DTPA and bronchoalveolar permeability to diethylenetriamine pentaacetate (DTPA) and bovine serum albumin (BSA) at 1h after the exposure. Bronchoalveolar, but not tracheal, permeability remained elevated at 24 h after the exposure. Exercise during exposure to O 3 increased permeabil- ity to both tracers in the tracheal and the brochoalveolar zones, and prolonged the duration of increased permeability in the tracheal zone from 1 h to 24 h, and in the bronchoalveolar zone from 24 h to 48 h. Permeability in the tracheal and brochoalveolar zones of rats exposed at rest to 6 or 12 ppm NO 2 did not differ from controls. However, rats exposed during exercise


INTRODUCTION
Exposure of resting animals to single oxidant pollutants is known to produce pulmonary injury detectable by morphologic, biochemical, physiologic, or immunologic criteria. The nature and extent of injury depend on the pollutant concentration and the duration of the exposure. Altered epithelial permeability has been utilized to analyze pulmonary injury after exposure of laboratory animals or humans to gaseous pollutants. Increased mucosal permeability was observed in humans smoking cigarettes (Jones et a!., 1980;Mason et al., 1983) and animals exposed at rest to O 3 (Bhalla et al., 1986), NO 2 (Selgrade et al., 1981;Ranga et al., 1980), and cigarette smoke (Boucher et al., 1980;Hulbert et al., 1981). Mustafa et al. (1984), noting the coexistence of NO 2 and O 3 in the atmosphere, compared the biochemical effects of these gases when administered alone or in combination. Their studies indicated that concentrations of either NO 2 or O 3 that produced nonsignificant biochemical changes in mouse lungs produced changes that were more than additive when the gases were combined. Watanabe et al. (1980) found that when NO 2 and O 3 were administered in sequence, the biochemical changes produced in mouse lungs were additive. Such additive or synergistic effects of exposure to NO 2 + O 3 are, however, not always observed. Ehrlich and Findlay (1979) reported additive effects of O 3 + NO 2 on pulmonary host resistance to inhaled bacteria, but Goldstein et al. (1974) suggested that response to O 3 and NO 2 is equivalent to that expected from each gas. Histologic changes produced by O 3 + NO 2 in rats were also similar to the changes produced by O 3 alone (Freeman et al., 1974).
Studies of the toxicologic effects of atmospheres that result from adding sulfur emissions from coal and oil burning to an atmosphere already dominated by photochemical air pollutants have been reported from this laboratory. Phalen et al. (1979Phalen et al. ( ,1980 described the effects of O 3 , sulfuric acid, and ammonium sulfate aerosols on clearance of labeled insoluble particles from rat lungs following single 4-h exposures to single or combined gases or aerosols at high or low relative humidity. Ozone slowed the early particle clearance and stimulated clearance during the later phase. High humidity usually amplified these effects of ozone, as well as many of the other atmospheres studied. Combinations of ozone and sulfate aerosols resulted in effects that were similar to those of ozone alone, sulfuric acid being a probable exception. One consequence of studying the health effects of atmospheric mixtures is that the components will often interact to form new, possibly toxic, reaction products. In the case of mixtures containing NO 2 , O 3 , and nonacidic aerosols, it is possible to form acidic particles and acidic vapors that were not in the original mixture but that might significantly alter the biological response upon exposure. For example, in recent stud-ies of atmospheric products formed by O 3 plus NO 2 and a combination of O 3 , NO 2 , SO 2 , (NH 4 ) 2 SO 4 , Fe 2 (SO 4 ) 3/ and MnSO 4 , Kleinman et al. (1985a, b) found the formation of HNO 3 vapor and of acid aerosol. Histologic sections of rat lungs were examined quantitatively at 18 and 48 h after exposure to score the percent of the alveolar zone of the section that was occupied by focal inflammatory lesions . The pathogenic effect of O 3 was stable over time, but the effect of the two atmospheres that formed HNO 3 was greater than for O 3 alone and was progressive with time, indicating potentiation of the O 3 effect by an acidic atmosphere. Last and Cross (1978) and Last et al. (1983Last et al. ( , 1984 reported that (NH 4 ) 2 SO 4 and H 2 SO 4 aerosols potentiated the effect of O 3 on the inflammatory lesions of rat lung and induced increased collagen synthesis, with the conclusion that the more acidic aerosol was the major contributor. The present study is therefore relevant to the evolving description of the effects of O 3 , and O 3 plus acidic atmospheres, on the rat respiratory tract.
In humans, exercise has been shown to alter pulmonary function (Silverman et al., 1976;DeLucia and Adams, 1977). Exercise increased minute ventilation and led to deeper penetration and a different distribution of inhaled toxic agents. In a recent study aimed at analyzing histopathologic changes in rats exposed to O 3 during exercise, Mautz et al. (1985a) found enhancement of lung lesions by exercise. O 3 concentrations that did not produce detectable histologic changes when the rats were exposed at rest induced lung lesions in exercising rats. In a series of similar studies, Last et al. (1985) reported that acid aerosols potentiated the effect of O 3 upon blood-to-lung transfer of serum proteins, and of radiolabeled serum albumin in rats exposed at rest.
Although the adverse effects of exposure to individual pollutant gases under resting conditions have been widely investigated (Plopper et al., 1973;Bils and Christie, 1980;Hu et al., 1982;Bhalla et al., 1986), the literature on the effects of combined gases inhaled during exercise is rather limited. From evidence available at the time this work was undertaken, exercise was known to modify the response produced by exposure to individual gases at rest. It is now clear that the response to inhaled pollutants depends on the level of ventilation as well as pollutant composition and concentration (Silverman et al., 1976;Mautz et al., 1985a;Last et al., 1985). Reaction products, including acids, can also be expected to modify responses measured in the laboratory.
This report represents an effort to study permeability changes in animals exposed under conditions that approach the complex atmospheres to which humans are exposed at rest or during periods of physical activity. The study analyzes permeability changes following exposure of rats to O 3 , NO 2 , O 3 + NO 2 or O 3 + NO 2 + SO 2 + ammonium sulfate + ferric and manganese salts that catalyze oxidation of S(IV) to S(VI) in the presence of an insoluble aerosol of colloidal Fe 2 O 3 .

Animals
Male barrier-reared Sprague Dawley rats (Hilltop Laboratory, Scottdale, Pa.) that weighed 225-274 g (about 47-52 d of age) were delivered in filtered containers, maintained on a standard diet, and held in a laminar air barrier caging system for at least 1 wk prior to pollutant exposure. Ten percent of each batch was autopsied upon arrival to assure their freedom from lung disease.

Generation and Characterization of Atmospheres
Exposure Chambers and Air Purification Exposures were in stainlesssteel chambers  or an enclosed treadmill . Each system was supplied with purified air from a highpressure air-purification system. Outside air was first filtered and then compressed to about 100 psig using a liquid ring compressor (Nash Engineering Company, Norwalk, Conn.). The high-pressure air subsequently passed through a fixed Purafil bed (KMnO 4 on alumina; H. E. Burroughs, Inc.) in order to remove gaseous contaminants such as O 3 , SO 2 , NO 2 , and hydrocarbons. Carbon monoxide was removed as it passed through a catalyst bed; a heatless dryer located just upstream of the catalyst bed prevented catalyst inactivation by moisture. Following filtration and thermal equilibration to laboratory temperature, the high-pressure purified air was throttled down to ambient pressure and humidified to 85 ± 2% relative humidity (RH). Delivery of purified air to the control chamber followed a final HEPA filtration. In the pollutant exposure chamber, test gases and/or aerosols were injected into the airstream following the HEPA filtration, and these diluted atmospheric pollutants then entered an expanding cone-shaped diffuser-mixer (Walters et al., 1982) before reaching the exposure chamber (Figs. 1 and 2). Atmospheric "aging" time was about 2.5 min from initial pollutant mixing to the time of inhalation.
Generation Techniques Ozone was generated by passing medicalgrade oxygen through an electrical ozone generator (Sander ozonizer, type III, Osterberg, West Germany). NO 2 and SO 2 were obtained from cylinders of research-quality compressed gas, metered into purified air using calibrated rotameters with needle valves and then injected into the chamber air-supply stream through separate ports located downstream of the O 3 injection port. Ammonium sulfate aerosols (about 0.25 /im mass median aerodynamic diameter) containing catalytic metal ions were generated by nebulizing an aqueous solution of ammonium sulfate, ferric sulfate, and manganese(ll) sulfate using a Collison nebulizer (Ho et al., 1980). Aerosols were passed through an 85 Kr discharger to reduce static charges on particles to the Boltzmann equilibrium level. Ferric oxide particles were generated by nebulization of dilute colloidal suspensions of Fe 2 O 3 , yielding spherical aggregate particles. These particles had a mass AIR INLET AIR OUTLET FIGURE 1. Nose-only exposure system. Air pollutants are introduced into the Walters' chamber air flow before the expanding cone-shaped diffuser-mixer. Rats in plastic tubes are exposed noseonly to chamber atmosphere by way of ports in the chamber wall. Arrows show direction of air flow. median aerodynamic diameter of 0.3 /*m and geometric standard deviation of about 2. The aerosols were diluted with warm, dry air, discharged by passage through an 85 Kr discharger and introduced into the chamber air.
Characterization Techniques The following methods have been described in greater detail by Kleinman et al. (1984). All atmospheric gas measurements were made at the rat breathing zone using Teflon tubing sampling lines. Nitrogen dioxide and nitric oxide were monitored using a chemiluminescent detector (Monitor Labs 8840, San Diego, Calif.). Sulfur dioxide was monitored using a Teco series 43 pulsed fluorescence detector (Thermal Electron Corp., Hopkinton, Mass.). Ozone was measured by ultraviolet (UV) absorption (Dasibi model 1003-AH, Dasibi Environmental Corp., Glendale, Calif.). During exposure, the aerosol stability was monitored using "real-time" detectors: a Climet 208 optical particle detector (Redlands, Calif.) for aerosols with diameters from 0.3 to 10 fim and a TSI model 3030 (St. Paul, Minn.) electrical aerosol size analyzer for particles with diameters ranging from 0.05 to 0.5 ^m. Cascade impactor samples (Sierra eight-stage impactor, Anderson Samplers, Atlanta, Ga.) were collected to provide size-classified particles for gravimetric and chemical analyses. Aerosols were collected on precleaned quartz fiber or Teflon AIR INLET AIR OUTLET FIGURE 2. Exercise exposure system in which air pollutants are drawn from ports in the chamber walls and through a modified ten-channel Quinton small-animal treadmill shown in the foreground. Each channel is separated from the others by stainless-steel-lined partitions. Arrows show direction of air flow. filters for gravimetric determinations and for chemical analyses of atmospheric components and reaction products. Nylon backup filters were used to collect nitric acid vapor from the reaction of NO 2 and O 3 . Ionexchange chromatography was used to analyze sulfate, sulfite, nitrate, and nitrite ions in extracts from both particulate and vapor-phase filters. Sample aliquots of extracts were analyzed for manganese, "soluble iron" (ferric sulfate), and "insoluble iron" (ferric oxide) using a differential extraction technique. Iron and manganese were determined by atomic absorption spectrophotometry and H + by potentiometry.

Exposure of Resting or Exercising Rats
Rats were exposed at rest or during exercise for 2 h to (a) air, (b) O 3 (0.6 or 0.8 ppm), (c) NO 2 (6 or 12 ppm), (d) O 3 (0.6 ppm) + NO 2 (2.5 ppm), (e) O 3 (0.6 ppm) + NO 2 (2.5 ppm) + SO 2 (5 ppm) + Fe 2 O 3 (241 /xg/m) + (NH 4 ) 2 SO 4 (308-364 nglm% Fe 2 (SO 4 ) 3 (411-571 j*g/m 3 ), and MnSO 4 (7-9 /ng/m 3 ). In each experiment, at least 10 rats were exposed, either during rest or while exercising to each atmosphere. Only 9 of the 10 exposed rats were used for permeability measurements. The extra rats served as a backup in case of technical problems with any of the scheduled rats. Three of the exposed rats were tested for tracheal or bronchoalveolar permeability immediately after the exposure. Three of the remaining rats were tested at 24 h and 3 at 48 h after the exposure to determine the duration of effect of exposure on permeability. Each experiment was repeated at least once for testing permeability immediately after the exposure and 24 h later. Since permeability at 48 h after the exposure was not different from controls in representative experiments, the repeats were not performed at this time point. The data for 48 h after the exposure are stated in the results but not included in the tables.
In experiments in which all exposures were at rest, rats were placed in cylindrical plastic tubes fitted with aluminum end pieces at the head end. The end piece bears a cone-shaped interior cavity tapered toward the forward end so that only the tip of the rat's nose protrudes into the atmosphere. The exterior of the end piece bears a cylindrical fitting by which it is inserted and sealed into one of a set of 26 exposure ports placed in the wall of the chamber (Fig. 1) as described by Mannix et al. (1982). Each port was supplied by a channel on the interior wall of the chamber; the channel directed the atmosphere downward past the nose of each rat, to prevent rebreathing of that atmosphere by the exposed rat and to prevent sharing of that sample of atmosphere by another rat.
In experiments designed to test the effects of exercise during exposure, all rats were trained on a 10-channel treadmill (Fig. 2) (Quinton model 42-15, Quinton Instruments, Seattle, Wash.) for 2 d prior to the experimental exposure; they ran at 15 m/min at 20% grade for 1 h on the first day and for 2 h on the second day. All rats underwent identical training and were then randomly assigned to exposure during rest or exercise. Rats exposed at rest were held in the restraining plastic tubes for nose-only exposure and placed in the exposure treadmill in order to maintain the exposure positions as close as possible to those for exercising rats to assure that the breathing zone of each resting rat was located at the same point in relation to entry of O 3 in the treadmill as the breathing zone of rats exposed while running. Concentrations of test gases were measured continuously during a 2-h exposure. The exposure treadmill  has been used to determine the mean metabolic rate of 10 rats at rest or during exercise; the exposure regimen produced about a twofold increase in oxygen consumption in exercising rats as compared to resting rats.

Permeability Measurement and Data Analysis
Methods for measuring transport of tracers across the tracheal and bronchoalveolar mucosa have been reported (Bhalla et al., 1986). The methods are reviewed briefly here. Animals were anesthetized intraperitoneally with sodium pentobarbital (Nembutal sodium, Abbott Laboratories, North Chicago, III.), 5 mg per 100 g body weight. A polyethylene tube (PE-90) was placed in the trachea, and a polyethylene catheter (PE-10, Clay Adams, Parsippany, N.J.) was placed in the femoral artery. The radiolabeled tracer inoculum contained [ 99m Tc]diethylenetriamine pen-taacetate (""Tc-DTPA, molecular weight 492) and 125 l-labeled bovine serum albumin ( 125 I-BSA, molecular weight 69,000).
For tracheal instillations, 0.1 ml of the inoculum was manually delivered into the trachea using a PE-10 tube attached to 1-ml syringe; inoculum was introduced through the PE-90 tracheostomy tube while the rat was held supine in a nearly horizontal position. The tracer solution was delivered into the middle zone of the trachea about 0.5 cm beyond the tracheostomy tube over a 5-min period.
For bronchoalveolar instillation, rats were held on an incline. Inocula were delivered directly to a main stem bronchus using a longer PE-10 tube inserted through the tracheostomy tube. When India ink was used as a tracer inoculum, the dye distributed in the major airways following instillation into the tracheas of supine rats, but the dye could be localized in the alveolar region upon instillation into a mainstem bronchus of rats held on an incline (Bhalla et al., 1986). After delivery of radiolabeled tracers but prior to blood sampling, heparin solution (20 units in 0.2 ml saline) was injected into rats through femoral-artery catheters prior to blood sampling. Blood samples of 0.10 ml were drawn 6, 7, 8, 9, and 10 min after the start of instillation. Blood samples were transferred to glass microtubes for immediate counting of " m Tc in a gamma counter (1275 Minigamma, LKB Instruments, Gaithersburg, Md.). Same blood samples were then counted for 125 l radioactivity.
Isotope counts in 0.10-ml blood samples at the 6, 7, 8, 9, and 10-min time points for the two tracers were expressed as the percent of the inoculum transferred from the site of application to the entire blood volume of each rat. Blood volume was calculated as 6.4% of the body weight (Altman and Dittmer, 1974). The label in blood samples was analyzed by obtaining the regression curve extrapolated to the intercept at the mid-time point, T if for the five samples. Differences in the fractions of each labeled molecule transferred to the blood at time T, in both the control and the O 3 -exposed groups were analyzed by analysis of variance, using a significance level of 0.05. Data from different experiments were pooled and permeability values in rats exposed to pollutant gases or their combinations were compared between the groups as well as with clean-air controls. To delineate the effect of exercise on permeability, comparisons were made between groups exposed to same pollutant while at rest or during exercise.

Atmospheres
Analytic results for 6 atmospheres of O 3 + NO 2 and for two 7component atmospheres that include O 3 and NO 2 (Table 1) reveal that the dominant acidic product in these atmospheres was HNO 3 at concen-trations of from 1179 to 2558 /*g/m 3 (0.46-1.02 ppm). The amount of HNO 3 formed was consistent with calculated products of the reaction of O 3 and NO 2 at the concentrations employed in this study (M. T. Kleinman and B. Finlayson-Pitts, personal communications). Some formation of acid sulfates (most likely ammonium bisulfate, NH 4 HSO 4 ) was also observed. Extracts from aerosols collected on the quartz-fiber or Teflon filters were analyzed for hydrogen ion content, shown in Table 1 in terms of microequivalents of hydrogen ion per cubic meter of air. The particles contained very little NO 3 ion; thus particle acidity values were regarded as arising from SO 2 conversion to SO 4 . Some hydrogen ion is also released by hydrolysis of water by Fe 3+ during storage of the extract prior to analysis; the values in Table 1 were empirically corrected by subtracting the amount of hydrogen ion formed by hydrolysis due to iron in the sample. The residual acidity, regarded as arising from NH 4 HSO 4 , is shown as acidic aerosol in Table 1.

Rats Exposed to O 3 at Rest or during Exercise
Exercise during shamexposure to clean air did not alter tracheal mucosal permeability to DTPA or BSA as compared to permeability in rats exposed to clean air while at rest (Table 2). In control rats exposed to clean air, either at rest or during exercise, the fraction of the instilled quantity of DTPA transferred from the tracheal lumen to the blood was greater than the fraction of BSA (Table 2). In exposed rats, 1 h after a 2-h exposure to O 3 at rest, both DTPA and BSA fractions appearing in blood were higher than in control rats (Table 2). In a previous series of experiments (Bhalla et al., 1986), both DTPA and BSA were transferred in significantly greater fractions in O 3exposed, as compared to control, resting rats, but in Table 2 the increase in transfer of DTPA only is significantly greater in rats exposed at rest to 0.8 ppm O 3 for 2 h.
Tracheal mucosal permeability to DTPA was significantly greater in animals examined 1 h after a 2-h exposure to ozone at rest and was not significantly elevated at 24 and 48 h after the exposure, while in rats exposed during exercise, tracheal permeability at 1 h was increased to both DTPA and BSA and remained elevated at 24 h after the exposure (Table 2). Although transfer of DTPA and BSA at 1 h and 24 h postexposure in rats exposed during exercise was greater than at the corresponding time points in rats exposed at rest, only DTPA transfer at 24 h postexposure was significantly different between the two groups (p < 0.05).
Rats Exposed to NO 2 at Rest or during Exercise Tracheal permeability in rats exposed to 6 or 12 ppm NO 2 while at rest, or to 12 ppm during exercise, has been studied twice. In rats exposed at rest to 12 ppm NO 2 , the fractions of DTPA and BSA in blood were not consistently increased either immediately after, or at 24 h after, the exposure (Table 3). In rats exposed during exercise, there was a significant increase in transfer of DTPA and BSA 1 h after the exposure, but only DTPA transfer was significantly elevated at 24 h. Only DTPA transfer at 1 h after exposure to NO 2 was significantly greater (p < 0.05) in rats exposed during exercise than at the same time point in rats exposed at rest.

Rats Exposed to O 3 at Rest or during Exercise
Ozone increased the transfer of DTPA and BSA to blood at 1 h after exposure. The percent transferred was greater for DTPA than for BSA, and the permeability induced by exposure to O 3 persisted for DTPA at 24 h after the exposure ( Table 4).
As in the case of trachea-to-blood transfer of molecules, exercise during exposure to clean air did not produce any consistent, significant change in permeability from the bronchoalveolar zone to the blood. Exercise during exposure to O 3 resulted in increased transmural transfer of both DTPA and BSA (Table 4). Bronchoalveolar permeability in rats ex-  Data as mean ± standard deviation. Significant at p < 0.05 when compared with clean-air exercise. Significant at p < 0.05 when compared with corresponding time points after exposure to NO 2 at rest. 1.14 ± 0.31 b 0.60 ± 0.12 3.27 ± 0.78 b ' c 2.59 ± 0.37 b ' c 1.08 ± 0.28* BSA 0.70 ± 0.19 0.70 ± 0.14 0.71 ± 0.09 0.79 ± 0.14 1.46 ± 0.45 6 0.84 ± 0.15 0.81 ± 0.21 2.15 ± 0.62 6 ' c 1.50 ± 0.17 6 ' c 1.19 ± 0.20 fa ' c (n -16) (n -6) (n -6) (n -6) (n -7) (n -8) (n -5) (n -7) (n -6) (n -5) a Data as mean ± standard deviation. Significant at p < 0.05 when compared with clean-air rest. Significant at p < 0.05 when compared with corresponding time points after exposure to O 3 at rest. posed during exercise to O 3 was greatest at 1 h after the exposure. Permeability to DTPA remained high at 24 h, while permeability to BSA was still significantly elevated at 48 h after the exposure. Permeability to DTPA and BSA in rats exposed during exercise was significantly greater than in rats exposed at rest (p < 0.05).
Rats Exposed to NO 2 at Rest or during Exercise Rats exposed while at rest or during exercise to 6 ppm or at rest to 12 ppm NO 2 did not differ from controls, but in rats exposed at exercise to 12 ppm, the bronchoalveolar zone showed significantly increased permeability to DTPA and BSA at 1 h after a 2-h exposure; the increased permeability was absent at 24 and 48 h (Table 5). Permeability to DTPA and BSA at 1 h after exposure of exercising rats to 12 ppm NO 2 was significantly greater than 1 h after exposure of resting rats (p< 0.05).
Rats Exposed to O 3 + NO 2 at Rest or during Exercise Bronchoalveolar permeability to DTPA in rats exposed at rest to 0.6 ppm O 3 alone was elevated only at 1 h after exposure, while in rats exposed at rest to 0.6 ppm O 3 + 2.5 ppm NO 2 for 2 h, permeability for DTPA was significantly greater both at 1 h and 24 h after the exposure ( Table 6). The increase, however, was not significantly greater than in rats exposed to O 3 (p < 0.05). In rats exposed during exercise to 0.6 ppm O 3 alone, the significant increase in permeability to DTPA that persisted at 24 h after the exposure was similar to that already described for 0.8 ppm O 3 (Table 4). The increase in permeability in rats exposed during exercise to 0.6 ppm O 3 + 2.5 ppm NO 2 was significantly greater than the increase induced by these gases in rats exposed at rest both at 1 h and at 24 h after the exposure (p < 0.05). In rats exposed to O 3 + NO 2 during exercise, permeability was significantly greater at 1 h and 24 h after the exposure than at the corresponding time points in rats exposed during exercise to O 3 alone (p < 0.05).
Rats Exposed to Complex Atmosphere at Rest Bronchoalveolar permeability to DTPA in rats exposed for 2 h to a complex atmosphere containing O 3 (0.6 ppm) + NO 2 (2.5 ppm) + SO 2 (5 ppm) + Fe 2 O 3 (0.3 mg/m 3 ) + ammonium, ferric, and manganese sulfates (1.0 mg/m 3 ) increased by over threefold immediately after the exposure when compared with clean-airexposed controls ( Table 6). As after exposure to O 3 + NO 2 , increased bronchoalveolar permeability after exposure to the complex atmosphere persisted at 24 h after the exposure. The increase at 1 h after the exposure was greater than that after exposure of resting rats to O 3 , or O 3 + NO 2 . The difference between the effects of complex atmosphere and O 3 was, however, not significant (p = 0.07), but the difference between the effects of complex atmosphere and O 3 + NO 2 was significant (p < 0.05). The increase was significantly greater at 24 h after the exposure to complex atmosphere than at the corresponding time point in rats exposed to O 3 alone at rest (p < 0.05).  0.73 ± 0.21 b ' c 0.46 ± 0.11 (n -6) (n -6) a Data as mean ± standard deviation.
Significant at p < 0.05 when compared with clean-air exercise. Significant at p < 0.05 when compared with corresponding time points after exposure to 12 ppm NO 2 at rest.  Data as mean ± standard deviation. fa Significant at p < 0.05 when compared with clean-air rest.
""Significant at p < 0.05 when compared with corresponding time points after exposure to NO 2 + O 3 at rest.
Significant at p < 0.05 when compared with corresponding time points after exposure to O 3 during exercise. Significant at p < 0.05 when compared with corresponding time points after exposure to O 3 at rest.

DISCUSSION
We have previously reported (Bhalla et al., 1986) that the simultaneous application of DTPA and BSA for detecting airway permeability provides a reliable means for studying epithelial injuries localized to various airway regions, i.e., nose, trachea, and bronchoalveoli. No interaction between DTPA and BSA was evident, and the two tracers appeared to be transferred independently across the airway mucosa. A 2-h exposure of resting rats to 0.8 ppm O 3 resulted in an approximately twofold increase in tracheal and bronchoalveolar permeability to DTPA; permeability to BSA also increased, but to a lesser extent. The rates of transfer from lumen to blood and from blood to either a storage or excretion route were fit to a three-component model: The transfer coefficients fc, and k 2 were estimated by an iterative fitting technique. In experiments where concentration in blood was monitored during the first 6 min, concentration rose rapidly in the first 2 min, reached a plateau when the rate of transfer into blood equaled the rate of transfer out of blood, and in some cases began to decline when the rate of transfer out of blood exceeded the rate of transfer in. The increased tracheal permeability was observed for at least 4.5 h after the exposure but not for 24 h after the exposure. The increased bronchoalveolar permeability persisted at 24 h after the exposure, suggesting a more persistent effect of O 3 in this region than in the trachea. We have now extended these studies to evaluate permeability changes in the tracheal and bronchoalveolar regions of rats exposed under conditions that approach more realistic environmental exposures, i.e., exposure during periods of moderate exercise activity and to atmospheres composed of multiple pollutants. Due to the lack of information on the permeability changes produced by exercise or combined gases and aerosols, we compare our results with reported studies evaluating pulmonary-function changes following exposure during exercise, and with morphologic injuries or biochemical alterations produced by gas or gas-aerosol combinations.
In humans, pulmonary-function changes have been studied after exposure of exercising healthy subjects to O 3 concentrations that do not produce effects during resting exposures. Exercise during exposure to O 3 caused substantial decrements in pulmonary-function measures (Silverman et al., 1976;DeLucia and Adams, 1977;Adams and Schelegle, 1983). Exercise appeared to be a major enhancer of toxicity. Greater total dosages of O 3 inhaled during exercise and the delivery of higher concentra-tions of the pollutant to distal alveolar surfaces are likely to be the causes of increased toxicity during exercise.
The exercise protocol used in our experiments has been shown to cause a twofold increase in oxygen consumption and was postulated to lead to a twofold increase in minute ventilation . Exposure to O 3 under these conditions increased the proportion of the alveolar zone of lung involved in focal lesions (as measured histopathologically) by about fourfold as compared to the percent of the lung parenchyma bearing lesions after exposure of rats at rest. This degree of alveolitis was greater than predicted on the basis of twofold increase in minute ventilation. A more widespread distribution of inhaled toxic gases in the lung was expected to increase alveolar lesion formation in proportion to increased minute ventilation upon exposure during exercise; hence, the reason for the disproportionately greater lesion formation is unknown .
Increased bronchoalveolar permeability in our studies is consistent with increased histologic changes following O 3 exposure in the alveolar zone Bils and Christie, 1980;Evans, 1984). Delivery of greater cumulative O 3 doses to the bronchoalveolar region during exercise would explain both increased permeability and percent of alveolar tissue bearing lesions. Although the greater flow rate reduces the fractional absorption in the upper respiratory tract (DeLucia and Adams, 1977), there is an increase in total inhaled O 3 . Exposures to higher pollutant concentrations for short periods may produce a greater effect than exposure to lower concentrations for extended periods (Silverman et al., 1976;Hazucha et al., 1973). It is likely that higher total dosage of O 3 per minute were delivered to the trachea and upper airways during exercise and compensated for shorter contact periods in the upper respiratory tract resulting from increased flow rates.
Following exposure of resting rats to O 3 , we have noted a more prolonged permeability response in the bronchoalveolar region than in the trachea (Bhalla et al., 1986). A similar response after exposure to O 3 + NO 2 suggests that the gas combination tends to enhance the O 3 effect but does not necessarily alter the target sites for O 3 . Since NO 2 by itself does not alter permeability or produce histologically detectable lesions even at much higher concentrations, i.e., 12 ppm exposure at rest, an NO 2 effect as such is not likely to be an influencing factor in the response elicited by the gas combination. Freeman et al. (1974) studied the alveolar lesions produced by a combination of O 3 and NO 2 and concluded that the injury following exposure to this gas combination reflected largely the effect of O 3 . In our study, the O 3 injury was probably potentiated by nitric acid vapor formed at 85% relative humidity, rather than by NO 2 .
The rationale for and the experimental system for producing experimental atmospheres of O 3 combined with gases that form acid or with acid aerosols have been described by Kleinman et al. (1984). Kleinman et