Specific pathogen-free male Hartley guinea pigs (Sankyo Labo Service, Tokyo, Japan), 4 weeks of age at the beginning of exposure, weighing 270–300 g, were used (n = 159). All of the experimental protocols were approved by the institutional Animal Experiment Committee and conformed to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.

Oxygen exposure was conducted by placing animals in a semi-sealed vinyl isolation chamber (volume 450 L), throughout the experimental period. Sterile food and water were provided regularly through a double outlet, and the chamber was kept clean with daily care. Through a 0.2 micrometer filter, 40% oxygen or room air was delivered at a flow rate of 15 L/min. Oxygen and carbon dioxide concentrations in the chamber were measured with a gas analyzer (Type 1312, Instrumentation Laboratory, Lexington, MA). The oxygen concentration was maintained and consistently confirmed to be 21 ± 1% (mean ± SD), 40 ± 2% or 90 ± 3%, and the carbon dioxide concentration was always below 0.5% at the outlet of the chamber. Humidity was maintained at 55–60%.

Total animal numbers for the examination of BAL/Serum ALB and TP, lung W/D or collagen metabolism were n = 10 in the pre-exposure group; n = 8 (2 weeks), 10 (4 weeks), 6 (8 weeks) and 6 (16 weeks) in the 40% oxygen exposure groups; n = 5 (2 weeks), 8 (4 weeks), 6 (8 weeks) and 6 (16 weeks) in the 21% oxygen exposure groups; n = 5 (24 hours), 5 (48 hours), 5 (72 hours), 10 (96 hours) and 5 (120 hours) in the 90% oxygen exposure groups. For examination of the clearance of inhaled Tc-DTPA: n = 5 in the pre-exposure group; n = 5 in each (2, 4, 8 or 16 weeks) 40% oxygen exposure group; n = 5 in each (2, 4, 8 or 16 weeks) 21% oxygen exposure group; n = 5 (24 hours), 6 (48 hours) and 8 (72 hours) in the 90% oxygen exposure groups.

Sixty millicuries of ^{99 m}TcO₄^- were eluted from a ^{99 m}Mb-^{99 m}Tc generator and bound to DTPA (Daiichi Radioisotope Labs, Tokyo). Eight milliliters of Tc-DTPA saline solution were aerosolized using an ultrasonic nebulizer (Model 65, Devilbiss Sunrise Medical, Carlsbad, CA). The aerosol was kept in the bag for 5 min to allow the large droplets to settle to the bottom 14. The animals were placed supine, and Tc-DTPA aerosol was delivered via the cannulated inspiratory circuit under spontaneous breathing. Radioactivity was measured each minute for 10 minutes in the anterior view using a gamma camera (GCA 401-5, Toshiba, Tokyo). A logarithmic plot of activity vs time was obtained from a region of interest. We corrected for physical decay and the background was subtracted. Monoexponential equations were fitted to the curves by the least squares method and the clearance rate was expressed as kep.

Deep surgical anesthesia was achieved with 100 mg/kg of ketamine (Ketalar; Parke Davis & Co., Detroit, MI) and 5 mg/kg of xylazine (Rompun; Cutter Laboratories, Shawnee, KS) injected intraperitonealy. The animals were then sacrificed by exsanguination. The chest was opened, and the lung was taken out of the thorax. A catheter was inserted into the right bronchus. The right lung tissue was embedded in paraffin after fixation with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 24 hours at 4°C. Three sections (each 3 μm thick) were made from the upper, middle and lower lobes at an equal distance from the top to the bottom of the right lung and stained with aniline blue or hematoxylin-eosin.

The area of aniline blue-stained fibrous tissue, which mainly consists of collagen fibers, was measured with a color image analyzing system (SP500, Olympus, Tokyo). An area of fibrous tissue around a bronchus stained with aniline blue (Fig. 1A) was selected with the color image analyzing system (Fig. 1B). The stained portion and the extracted portion were matched. The selected area was measured and expressed as a percentage of the total area (Fig. 1C) 15. The fibrous tissue area was measured using ten different fields of three different sections corresponding to the upper, middle and lower lobes of the right lung. These areas were selected at random from peripheral parts of the lung. If there was a bronchus, a large vessel or pleura inside the area, another area was selected because collagen fibers were abundant around these tissues. The fibrous tissue area to total area was expressed as the mean ± SD. Histological changes in hematoxylin-eosin stained lung sections with 0, 2, 4, 8 and 16 week-exposure periods in the 40% oxygen groups and pre, 72, 96 and 120 hour-exposure periods in the 90% oxygen groups were also examined.

The lung W/D in the left caudal lobe was measured to quantify the pulmonary edema induced by 40% or 90% oxygen exposure. After the wet weights of the lung tissue samples had been measured, the tissues were completely dried in a vacuum oven (DP22; Yamato Scientific, Tokyo) at 95°C and at 0 cmH₂O for 48 hours 16.

Bronchoalveolar infiltrating cells were assessed by BAL of the left cranial lobe 16. The lobe was weighed and its bronchus was catheterized, then lavaged twice with 5 ml of saline and 1 ml of air. BAL fluid was collected, placed on slides, then fixed and stained with a modified Wright's staining technique (Diff-Quik; American Scientific Products, McGraw Park, IL). BAL cell differentials were based on the means of five separate counts of 100 cells on the high power field of light microscopic slides. BAL fluid samples were centrifuged at 400 × g at 4°C for 10 min to obtain a cell pellet. Each cell pellet was suspended in 1 ml of saline, and a cell count was obtained using a modified hematocytometer method (Unopet Microcollection System; Becton Dickinson, Rutherford, NJ). BAL cell counts were expressed as counts per 1 mg lung tissue. The albumin concentrations in plasma and BAL fluid were determined using the bromcresol green dye binding method (Sigma Chemical Co., St. Louis, MO) 17. Total protein was determined using the Biuret method. The BAL fluid-to-serum (BAL/Serum) albumin (ALB) and total protein (TP) ratios were used as parameters of alveolar-capillary permeability 18 and lung injury 19, respectively.

Measurement of total collagen contents in the peripheral portions of the lung was based on the estimation of hydroxyproline contents. Saline solution (1.5 ml) was added to an approximately 30 mg tissue sample which was collected from the peripheral part of the left lower lobe. Then, the sample was homogenized on ice and used for the hydroxyproline content measurement, and to determine types I and III collagenolytic enzyme activities. Hydroxyproline contents were measured by the spectrophotometrical method 2021, which allows quantification of the amount of hydroxyproline in a small sample.

Pure type III collagen was prepared according to the method of Glanville 22, and type III collagen labeling was performed as previously described 23.

In order to measure type I and type III collagenolytic activities in lung tissues, we utilized bacterial collagenase to obtain a standard curve 23. Purified collagenase from Clostridium histolyticum (EC 3.4.24.3, bacterial collagenase type II, Sigma Chemical, St. Louis, MO) was dissolved in Tris-HCl buffer to obtain a 1 mg/ml concentration of the standard solution, which was then diluted to various concentrations (10^-1 to 10^-5 mg/ml) to provide the standard curves of types I and III collagenolytic enzyme activities. One unit of collagenolytic enzyme activity was defined as the weight of the bacterial collagenase having the same collagenolytic activity (μg bacterial collagenase/g protein) as the lung tissue sample. This standardization with bacterial collagenase made it possible to compare types I and III collagenolytic activities 23.

Protein contents of the samples were determined by the Lowry method 24. Hydroxyproline content was expressed as μg per 1 mg protein, and types I and III collagenolytic activities were expressed as μg bacterial collagenase per 1 g protein.

The results were presented as means ± standard deviation (SD). Statistical analyses were conducted using one-way analysis of variance (ANOVA) followed by Tukey post-hoc test to detect differences among groups. Differences between two groups were assessed using the unpaired t-test or Welch's test. A p-value < 0.05 was considered statistically significant.

The average count of BAL cells was higher in the 40% than in the 21% oxygen groups, and this increase was statistically significant after 2 weeks of exposure (Fig. 2). BAL cell counts in the 40% oxygen 2 week-, 4 week- and 8 week-exposure groups were significantly increased as compared to those in the pre-exposure group. After 72 hour exposure to 90% oxygen, BAL cell counts were increased as compared to the pre-exposure level. The percentages of the BAL neutrophils were not different between the 40% and the 21% oxygen groups at any time examined (Table 1). The percentages of the BAL macrophages and lymphocytes were not different between the 40% and the 21% oxygen groups either (Table 1). On the other hand, after 24 and 48 hour exposure to 90% oxygen, the percentages of the BAL eosinophils were significantly increased as compared to those in the pre-exposure condition (Table 2). After 72 hour exposure to 90% oxygen, the percentage of the BAL neutrophils was significantly increased as compared to the pre-exposure level (Table 2).

The BAL/Serum Alb ratio in the 40% oxygen groups did not differ from that in the 21% exposure groups for any of the durations examined (Fig. 2). The Alb ratio in the 40% or 21% oxygen 16 week-exposure group was significantly increased as compared to those in the pre-exposure group. This increase, however, did not differ between the 40% and the 21% oxygen groups. After 96 hour exposure to 90% oxygen, the Alb ratio was significantly increased as compared to the pre-exposure level.

The BAL/Serum TP ratio in the 40% oxygen groups did not differ from that in the 21% exposure groups with 2, 4 and 8 week exposures (Fig. 2). After 16 weeks of exposure, the BAL/Serum TP ratio in the 40% oxygen group was higher than that in the 21% exposure group. The BAL/Serum TP ratio in the 40% oxygen 8 and 16 week-exposure groups and the 21% oxygen 4 and 16 week-exposure groups were significantly increased as compared to those in the pre-exposure group. After 96 hours of exposure to 90% oxygen, the increased TP ratio was not statistically significant as compared to the pre-exposure level.

The kep of the 40% oxygen 2 week-exposure group was higher than that of the 21% oxygen 2 week-exposure group (Fig. 3). The kep of the 40% oxygen 2, 4, 8 and 16 week-exposure groups and the 21% oxygen 8 week-exposure group were significantly increased as compared to that of the pre-exposure group. This increase did not differ between the 40% and the 21% oxygen groups. The kep of the 90% oxygen 24 hour-exposure group was higher than that of the pre-exposure group, and that of the 48 hour-exposure group was higher than that of the 24 hour group.

No significant differences were observed between the 40% and 21% oxygen groups at any of the durations examined (Fig. 4). The lung W/D of the 40% or 21% oxygen 16 week-exposure group was significantly decreased as compared to that of the pre-exposure group. Lung W/D of the 120 hour-exposure group was increased as compared to that of the pre-exposure group.

Type I collagenolytic enzyme activity in the 40% oxygen 2 week-exposure group was significantly higher than that in the 21% oxygen 2 week-exposure group (Fig. 5). This increase was, however, not significantly different as compared to that in the pre-exposure group. Type I collagenolytic activity in the 40% oxygen 4 week-exposure group was also higher than that in the 21% oxygen 4 week-exposure group, although not significantly higher than that before exposure. Type I collagenolytic activity in the 40% oxygen 8 week-exposure group was lower than that in the 40% oxygen 2 week-exposure group. After 8 or 16 weeks of exposure, there were no differences between the 40% and 21% exposure groups or between the 40% and pre-exposure groups. There were no significant differences among the 90% oxygen exposure groups.

Type III collagenolytic enzyme activity in the 40% oxygen 2 week-exposure group was significantly higher than that in the 21% oxygen 2 week-exposure group (Fig. 6). This increase was, however, not significantly different as compared to that in the pre-exposure group. After 4, 8 or 16 weeks of exposure, there were no differences in type III collagenolytic enzyme activity between the 40% and 21% exposure groups, or between the 40% and pre-exposure groups. Type III collagenolytic activities in the 40% oxygen 8 week-exposure group was lower than that in the 40% oxygen 2 week-exposure group. There were no significant differences among the 90% oxygen exposure groups.

The hydroxyproline content with 2 week exposure to 40% oxygen was significantly increased as compared to that with 2 week exposure to 21% oxygen (Fig. 7). This increase was not significantly different as compared to that in the pre-exposure group. After 4, 8 or 16 weeks of exposure, there were no differences in hydroxyproline content between the 40% and 21% exposure groups, or between the 40% and pre-exposure groups. There were no significant differences among the 90% oxygen exposure groups.

Sections of peripheral parts of the lungs stained with aniline blue and the estimated area of fibrous tissue with 2 week exposure to 21% or 40% oxygen are shown in Fig. 1. The percentage of fibrous tissue with 2 week exposure to 40% oxygen was significantly increased as compared to that with 2 week exposure to 21% oxygen (Fig. 8). There were no differences between the 40% and 21% exposure groups at 4, 8 or 16 weeks. There were no significant differences among the 90% oxygen exposure groups. Fibrous tissue levels in 2 and 16 week-21% oxygen groups were not lower as compared to that in pre-exposure

Although the image analyzing system revealed that the fibrous tissue in the aniline blue stained section was increased with 2 weeks exposure to 40% oxygen, neither discernible injury nor architectural disorganization was observed in these groups (Fig. 9). In the 90% oxygen groups, inflammatory cell infiltration was recognized after 72 hours of exposure, and alveolar septal destruction was apparent by 96 hours (Fig. 10). Diffuse alveolar damage was apparent after 120 hours of exposure.