- Email: [email protected]
Hyperoxic lung injury in Fischer-344 and Sprague-Dawley rats in vivo
Free Radical Biology & Medicine, Vol. 14, pp. 531-539, 1993 Printed in the USA. All fights reserved.
0891-5849/93 $6.00 + .00 Copyright © 1993 Pergamon Press Ltd.
Original Contribution HYPEROXIC
LUNG
INJURY
IN FISCHER-344 RATS IN VIVO
A N D SPRAGUE-DAWLEY
JOEL D . STENZEL, STEPHEN E. WELTY, A R T H U R E. BENZICK, E. O ' B R I A N SMITH, CHARLES V. SMITH, and THOMAS N . HANSEN Department of Pediatrics, Baylor College of Medicine, Houston, TX 77030, USA (Received 23 November 1992; Accepted 11 December 1992) Abstract--Supplemental oxygen remains an important therapy for pulmonary insufficiency, despite the potential adverse effects of hyperoxic exposures. Recently, He et al. reported that hyperoxic ventilation more readily damaged isolated perfused lungs from Fischer-344 rats than from Sprague-Dawley rats (Am. J. Physiol. 259:L451), which correlates with the previously reported strain differences in hepatic responses to diquat-induced oxidant stress in vivo (J. Pharmacol. Exp. Ther. 235:172). We therefore examined the differences in hyperoxic lung injury in Fischer-344 and Sprague-Dawley rats in vivo. Adult male rats were exposed to > 95% 02 and were sacrificed after 24, 48, or 60 h. Control animals were maintained in room air. Dramatically greater increases in pleural effusions and bronchoalveolar lavage protein concentrations in response to hyperoxia were observed in the Fischer-344 rats than in the Sprague-Dawley rats Co < .05 at both 48 and 60 h for both measurements). Additionally, the glutathione concentrations in alveolar lining fluid decreased from 800 #M to 115 zM in Fischer-344 rats after 60 h of> 95% 02, but did not change in Sprague-Dawley rats. We conclude that the greater susceptibility of Fischer-344 than of Sprague-Dawley rats to hyperoxic lung injury in vitro reported previously also is observed in vivo and that this strain difference offers unique opportunities to study mechanisms of hyperoxic lung injury. Keywords--Oxygen toxicity, Glutathione, Fischer-344 rats, Pleural effusions, Bronchoalveolar lavage fluid
tion to oxidant-induced organ injury in Fischer-344 rats when compared with Sprague-Dawley rats. 6,7 Smith et al., 6 using diquat to generate intracellular oxidants in the liver, found that Fischer-344 rats developed liver necrosis when treated with diquat while similarly treated Sprague-Dawley rats did not. He et al. 7 compared the effects of oxidant stress in isolated perfused lungs from the same two strains and found that lungs from Fischer-344 rats developed injury and edema when exposed to oxidant stresses, including exposure to hyperoxia, while isolated perfused lungs from Sprague-Dawley rats treated identically did not show similar signs of injury. If the predilection to hyperoxia-induced lung injury in Fischer-344 rats compared with Sprague-Dawley rats would be expressed in vivo in the responses of these strains to hyperoxia, it would provide a useful paradigm for investigation of the biological and molecular mechanisms of oxygen-toxic lung injury in vivo. Therefore, we chose to investigate the susceptibilities to lung injury of Fischer-344 rats versus SpragueDawley rats in response to continuous exposure to greater than 95% oxygen. As indicators of lung injury,
INTRODUCTION
Despite significant improvements in biotechnology, and in many circumstances in the application of these improvements to clinical medicine, supplemental oxygen administration remains an important therapy for patients with respiratory system disease. As with almost any therapeutic intervention, the unwanted side effects of supplemental oxygen exposure must be weighed against the potential benefits of the therapy. The changes that take place in oxygen toxic lung injury have been investigated in a number of well-designed and executed studies, 1-5 but many fundamental questions remain unanswered and additional investigations providing new perspectives are needed. Previously, investigators studying mechanisms of oxidant-induced injury found an interesting predilecSupported by Baylor Child Health Research Center New Project Development Award 5 P30 HD27823, USDA/ARS Children's Nutritional Research Center. Address correspondence to: Stephen E. Welty, MD, Dept. of Pediatrics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. 531
J . D . STENZEL ~ aL
532
we measured pleural effusion weights and bronchoalveolar lavage protein concentrations as a function of time of exposure to hyperoxia. As possible indirect indicators of lung injury, we measured plasma and lavage lactate dehydrogenase (LDH) activities in each strain as a function of time exposed to hyperoxia. In addition, we measured concentrations of total glutathione (GSH + GSSG) and GSSG in plasma and in bronchoalveolar lavage fluid along with simultaneously measured urea nitrogen concentrations so that we could calculate alveolar lining fluid concentrations of GSH and GSSG by the method of Rennard et al. s Previously, investigators have estimated that GSH concentrations in alveolar lining fluid in humans were more than 100-fold greater than in plasma and speculated that these high GSH levels may be critical in the protection against environmental oxidants, 9 while others have found that bronchoalveolar lavage GSH concentrations are increased in rats exposed to 80% oxygenJ ° In our studies, Fischer-344 rats did prove to be more susceptible to hyperoxic lung injury in vivo than were Sprague-Dawley rats, as evidenced by increased pleural effusion weights and bronchoalveolar lavage protein concentrations after 48-60 h of hyperoxic exposure seen in Fischer-344 rats. Additionally, we found that hyperoxic exposure caused alveolar lining fluid concentrations of GSH to decrease significantly in Fischer-344 rats, whereas we did not observe a similar decrease in GSH levels in hyperoxia-exposed Sprague-Dawley rats. Despite the obvious oxidant stresses of hyperoxic exposure and the marked differences in physiological responses, we did not observe a significant increase in GSSG concentrations in any of the fluids examined. MATERIALS AND M E T H O D S
Animal protocol The study was done using either Sprague-Dawley or Fischer-344 adult male rats weighing 200 to 265 g obtained from Harlan Industries (Houston, TX). The rats were housed in a sealed chamber with a 12:12 hour light-dark cycle with free access to food and water throughout the experimental period. The FiO2 was monitored twice daily and maintained at greater than 0.95 by the administration of pure oxygen at 5 L per minute. Calcium carbonate (Sodasorb, Dewey & Almy Chemical Division, Grace & Co., Lexington, MA) was placed in the chamber to absorb carbon dioxide. Throughout the study period, both strains were exposed to identical concentrations of oxygen simultaneously in the same sealed chamber with a partition separating the two strains.
Measurements Equal numbers of Fischer-344 and Sprague-Dawley rats were sacrificed after 24, 48, and 60 h of exposure to > 95% oxygen. At each time point air-breathing control animals from each strain were sacrificed and the samples were processed in parallel with samples obtained from the hyperoxia-exposed rats. Data from the control animals are presented as time -- 0 (for 0 h in > 95% 02). The rats were anesthetized with 200 mg/kg ofintraperitoneal pentobarbital. Next, the trachea was surgically exposed and cannulated using an 18-gauge 1¼-inch Angiocath (Deseret Medical, Inc., Becton Dickinson & Co., Sandy, UT). The thoracic cavity was then opened using a midline thoracotomy. Pleural effusion weights were measured by absorbing the pleural fluid with a preweighed dry gauze pad and reweighing the pad immediately. Next, intracardiac blood samples were obtained for measurement of plasma GSSG, total glutathione (GSH + GSSG), LDH activity, hemoglobin, and blood urea nitrogen (BUN). Heparinized syringes were used for the collection of plasma for the measurement of LDH activity, hemoglobin, and BUN. Following the collection of the blood and plasma samples, lung tissue from the right lung was obtained and frozen at -20°C and subsequently used to measure extravascular lung water by a modification of the wet-to-dry determination ll described by Pearce et alJ 2 After removing the right lung, the left lung was lavaged. Lavage fluid was obtained using 5 cc of 4°C saline through the previously inserted tracheostomy tube. Lavage fluid was aliquoted for measurement of LDH activity, protein, urea nitrogen, GSSG, as well as total glutathione determinations. If the amount of lavage fluid recovered was less than 1.6 cc (volume oflavage per animal required for the above measurements), then the lavage procedure was repeated using an additional 5 cc of 4°C saline. Recovery volumes of lavage fluid were recorded and typically fell between 50% and 80%. Low recoveries, hence second lavages, were required in only a few of the animals having clinically evident respiratory distress following prolonged hyperoxic exposure. Plasma and lavage GSSG and total glutathione concentrations were determined by the method of Adams et alj3 with modifications. The samples for determination of plasma total glutathione were immediately added to equal volumes of 4°C 0.05 M potassium phosphate buffer containing 17.5 mM ethylenediaminetetraacetic acid (EDTA), 0.05 M serine, and 0.05 M borate, pH 7.4 (Buffer I). The samples were mixed immediately by gentle inversion and centrifuged for 3 min. The supernatants (plasma + buffer)
Hyperoxiclung injury in rats were removed and frozen at - 7 0 ° C for later analysis. The samples for determination of plasma GSSG were immediately added to 500 #L of Buffer I plus 15 #L of 1 M N-ethylmaleimide (NEM). The samples were similarly mixed and centrifuged and the supernatants removed and stored at - 7 0 ° C for later analysis. For the measurements of lavage GSSG and total glutathione, 500 #L of lavage fluid was added to 100 #L of 4°C Buffer I and then centrifuged in a microfuge for 3 min. Supernatants were removed, 10 #L of 1 M NEM was added to the lavage GSSG sample, and both GSSG and total glutathione lavage samples were frozen at -70°C. To measure plasma total glutathione, a 100-#L aliquot of the plasma/Buffer I supernatant was added to a cuvette containing 800 #L of 0.1 M phosphate, 5 mM Na2EDTA, pH 7.4 (Buffer II), 50 #L of 10 mM 5,5'-dithio-bis(2-nitrobenzoic acid) (DTNB), and 0.7 U of GSSG reductase (Sigma Type III, Sigma Chemical Co., St. Louis, MO). This mixture was incubated for 1 min at room temperature. To initiate the reaction, 240 nmol of reduced nicotinamide adenine dinucleotide phosphate (NADPH) in Buffer II was added, making the final reaction volume 1.1 ml. The change in absorbance at 412 nm was continuously monitored for 5 min. Lavage total glutathione measurements were made using 100 #L of buffer-lavage supernatant with the same reaction mixture. Plasma GSSG determinations were made by taking a 250-tzL aliquot of the plasma-buffer-NEM solution and passing it at 1 drop/s through a C 18 Sep-Pack cartridge (Waters Associates, Framingham, MA) followed by 800 ~tL of Buffer II. An 800-#L aliquot of the combined eluate was added to a cuvette along with 50 #L of 10 mM DTNB and 0.7 U ofglutathione reductase. The assay then proceeded as for the measurement of total glutathione as noted above. Lavage GSSG was measured as above, substituting lavagebuffer-NEM solution for the plasma solution. Sample GSSG and total glutathione concentrations were calculated from standard curves generated with each set of measurements. Plasma and lavage GSH concentrations were calculated by subtracting the GSSG concentrations from the total glutathione (GSH + GSSG) concentrations for each sample. To interpret and compare the levels of GSH and GSSG in lavage, we estimated the amount of actual alveolar lining fluid recovered in the lavage fluid for each rat by comparing plasma urea nitrogen with lavage urea nitrogen concentrations as described by Rennard et al., 8 using the assumption that the concentrations of freely diffusable urea nitrogen in alveolar lining fluid and plasma are the same. From the quantity of alveolar lining fluid recovered, we then calculated the concentrations
533
of GSSG and total glutathione per milliliter of alveolar lining fluid for each rat. For the measurement of LDH activities, heparinized blood samples were centrifuged, the plasma removed, immediately placed on ice, and analyzed according to the method of Bergmeyer and Bernt.14 Lavage sample supernatants for measurement of LDH were analyzed similarly. Plasma and lavage samples for the measurement of BUN were frozen at - 2 0 ° C and later analyzed using kits purchased from Sigma (#535 or #66-UV). Blood hemoglobin samples were frozen at - 2 0 ° C and later analyzed using BoehringerMannheim Diagnostics Cyanmethemoglobin Reagent (Boehringer-Mannheim Biochemicals, Indianapolis, IN). Lavage protein samples were stored at - 2 0 ° C until analysis by the method of Lowry et al. 15
Statistics All data are expressed as mean _ standard error of the mean. Two-way analysis of variance (ANOVA) was used to assess the effects of strain, time of hyperoxic exposure, and the interaction between these two factors. If the interaction was significant, strains were compared at each time point and times of hyperoxic exposure were compared within strain using the least significant difference (LSD) multiple comparison procedure. Data that appeared different but were not indicated as different by the initial two-way ANOVA (i,e., no significant interaction detected) were evaluated using the Bonferroni multiple comparison procedure. Once interactions were ruled out, the overall effects of strain and time of hyperoxic exposure were interpreted, with various time means being compared using the LSD procedureJ 6 P < .05 was used to determine statistical significance. RESULTS
General appearance Neither Fischer-344 nor Sprague-Dawley rats displayed evidence of illness or respiratory distress after 48 h of hyperoxic exposure. Fischer-344 rats developed evidence of respiratory distress after 60 h of hyperoxic exposure, manifested primarily by tachypnea, while none of the similarly exposed Sprague-Dawley rats developed similar signs of respiratory distress. There were no deaths in hyperoxia-exposed rats.
Indices of lung injury Hyperoxic exposure increased the bronchoalveolar lavage protein concentrations significantly in Fischer344 rats at 48 and 60 h (fourfold increase at 48 h and
534
J.D. STENZELet al.
MG/ML
+ . 1.2
•
F344
[]
S.D.
0.8
0.4
"
0.0 0
48
24
60
54% and 85% at 48 and 60 h respectively, while in similarly treated Sprague-Dawley rats alveolar lining fluid GSH concentrations did not decrease (Fig. 4). Mean alveolar lining fluid GSSG concentrations were 66% lower at 60 h in the Fischer-344 rats, but this effect was not statistically significant, while hyperoxic exposure had less effect on GSSG concentrations in the alveolar lining fluids of the Sprague-Dawley rats (Fig. 5). Furthermore, concentrations of GSSG in the alveolar lining fluid were not different between the two strains. At 48 and 60 h ofhyperoxic exposure, plasma GSH concentrations were higher in Fischer-344 rats (although not reaching statistical significance) than in similarly treated Sprague-Dawley rats (Fig. 6). Hyperoxic exposure did not affect plasma GSH concentrations in the Sprague-Dawley rats. Compared with airbreathing controls, plasma concentrations of GSSG decreased in Sprague-Dawley rats at 48 and 60 h of hyperoxic exposure and were significantly lower than in similarly treated Fischer-344 rats (Fig. 7).
TIME (h)
Fig. 1. Bronchoalveolar lavage protein. Concentrations of bronchoalveolar lavage protein (BAL) were measured in hyperoxia-exposed Fischer-344 (F344) rats, hyperoxia-exposed Sprague-Dawley (SD) rats, and in their respective air-breathing controls (0 h). Compared to BAL protein concentrations in SD rats, F344 BAL protein concentrations rise significantly with increasing duration of hyperoxic exposure (n = 5 to 8 of each strain per time point). *p < .005 at 60 h F344 vs. SD and p < .002 at 48 h F344 vs. SD. tp < .001 for F344 60 h and F344 48 h vs. F344 0 h.
fivefold at 60 h), but had no effect on bronchoalveolar lavage protein concentrations in Sprague-Dawley rats (Fig. 1). While exposure to hyperoxia increased pleural effusion weights in both strains of rats, the increase was greater in the Fischer-344 rats (fivefold at 48 h and 15-fold at 60 h) than in the Sprague-Dawley rats (threefold increase at 60 h) (Fig. 2). Extravascular lung water increased in both Fischer-344 rats and Sprague-Dawley rats at 48 and 60 h of exposure to hyperoxia (Fig. 3). Within each strain, exposure to hyperoxia did not affect LDH activities in bronchoaveolar lavage (Table 1) or plasma (Table 2) measured over time. However, plasma LDH activities in the Fischer-344 rats at 60 h were approximately twice the respective activities observed in the Sprague-Dawley rats at the same time point.
GSH and GSSG concentrations in alveolar lining fluid and plasma Hyperoxic exposure decreased GSH concentrations in the alveolar lining fluid of Fischer-344 rats by
DISCUSSION
The primary finding of this investigation is that Fischer-344 rats are more sensitive to hyperoxic lung GM
8.0
= []
F344 S.D.
4.0
++
,4,- 'k 2.0
0.0
0
24
48
80
TIME (h)
Fig. 2. Pleural effusions. Pleural effusion weights were measured in hyperoxia-exposed F344 rats, hyperoxia-exposed SD rats, and their respective air-breathing controls (0 h). Pleural effusion weights ot hyperoxia-exposed F344 rats were significantly higher than pleural effusion weights in identically exposed SD rats (n = 7 to 8 of each strain per time point). *p < .001 at 60 h F344 vs. SD and p < .05 at 48 h F344 vs. SD. tp < .001 F344 60 h vs. F344 0 h and p < .01 F344 48 h vs. F344 0 h. *p < .02 SD 60 h vs. SD 0 h.
535
Hyperoxic lung injury in rats Table 2. Plasma LDH Activity (IU/L)
GM H20/GM DRY BLOODLESS LUNG 9
• F344 [] S.D. +4-
4--I-
Time exposed to > 95% O2
0h
24 h
48 h
60 h
Fischer-344 Sprague-Dawley
244 _+ 40 187 _+ 19
208 _+ 33 155 +_ 17
292 _+ 44 238 +_ 48
324 _+ 52 149 +_ 12
Plasma LDH activities were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Mean plasma LDH activities in hyperoxia-exposed F344 rats trend upward, but this effect is not statistically significant. In Sprague-Dawley rats, plasma LDH activities are unchanged by hyperoxic exposure (n = 7 to 8 of each strain per time point).
ability pulmonary edema. This pulmonary edema has been determined as an increase in lung lymph flow or lymph-to-plasma total protein ratios, ]7-20increases in extravascular lung water, or increases in broncholalveolar lavage protein concentrations.17-2s In the present study, we found that hyperoxia-exposed Fischer-
uM 0
24
48
12oo 1
60
[] F344 [] S.D.
TIME(h) Fig. 3. Extravascular lung water. Extravascular lung water was measured using the wet-to-dry method in hyperoxia-exposed F344 and SD rats and in air-breathing controls (0 h) of both strains. Extravascular lung water was significantly higher in hyperoxia-exposed animals of both strains compared with their respective air-breathing controls. However, no statistical difference was detected between F344 and SD rats during hyperoxic exposure (n = 7 to 8 of each strain per time point). *p < .001 F344 48 and 60 h vs. F344 0 h. *p < .001 SD 48 and 60 h vs. SD 0 h.
injury in vivo than are Sprague-Dawley rats. The increased susceptibility of Fischer-344 rats was seen in the increased pleural effusion weights and bronchoalveolar lavage protein concentrations after 48 and 60 h of continuous hyperoxic exposure. Previous studies have demonstrated that lung injury secondary to hyperoxic exposure is manifested primarily as a perme-
800
400
0
Table 1. Lavage LDH Activity (IU/L)
24
48
60
TiME (h)
Time exposed to > 95% 02
0h
24 h
48 h
60 h
Fischer-344 Sprague-Dawley
170 +_41 127 _+ 12
201 + 25 141 + 15
163 _ 21 123 _ 23
186 _+ 43 99 + 15
Lavage LDH activities were measured in hyperoxia-exposed F344 rats, hyperoxia-exposed SD rats, and in their respective airbreathing controls (0 h). No statistical differences were detected between strains relative to the duration of hyperoxic exposure (n = 6 to 8 of each strain per time point).
Fig. 4. Alveolar lining fluid glutathione. GSH was measured in BAL fluid and corrected (by the urea method of determining volume of alveolar lining fluid [ALl:'] recovered in the BAL fluid) to give GSH concentrations in ALF of hyperoxia-exposed F344 rats, hyperoxiaexposed SD rats, and their respective air-breathing controls (0 h). GSH concentrations in ALF of F344 rats following 60 h of hyperoxic exposure are significantly lower than in identically exposed SD rats and also lower than in F344 air-breathing controls (n = 4 to 8 of each strain per time point). *p < .05 F344 vs. SD at 60 h. *p < .002 F344 60 h vs. F344 0 h.
536
J.D. STENZEL el al.
uM 45
•
F344
[]
S.D.
30
15
0
24
48
80
TIME(h)
to increase in Fischer-344 rats in response to hyperoxic exposure, but this increase was not quite statistically significant. Previous investigators have shown that measurement of specific LDH isoforms enhances the sensitivity of plasma LDH measurements for determination of lung injury. 2~ In the context of this investigation comparing plasma and lavage LDH activities in Fischer-344 rats versus Sprague-Dawley rats, the relative increased activities of lavage LDH and plasma LDH at 60 h in Fischer-344 rats suggest a predilection to oxygen-toxic lung injury in Fischer344 rats when compared to Sprague-Dawley rats, but the effects of hyperoxic exposure on the manifestations of permeability pulmonary edema are far more dramatic. Estimating alveolar lining fluid concentrations of GSH and GSSG by the urea method as a function of time exposed to hyperoxia, to our knowledge, has not been reported. Others have determined baseline alveolar lining fluid levels of GSH and GSSG in humans (approximately 800 pM in smokers and 400 pM
uM
30 1
Fig. 5. Alveolar lining fluid giutathione disulfide. GSSG concentrations were measured in BAL fluid and corrected (by the urea method of determining volume of ALF recovered in the BAL fluid) to give GSSG concentrations in ALF of hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). No statistical differences were detected between strains relative to hyperoxic exposure (n = 3 to 8 of each strain per time point).
i
I
•
F344
[]
S.D.
20
344 rats developed dramatic evidence of pulmonary edema, which was reflected in the early (48 h) and sustained (60 h) increases in bronchoalveolar lavage protein concentrations and pleural effusion weights. Sprague-Dawley rats were not as affected by identical hyperoxic exposures. After 60 h of hyperoxic exposure, the pleural effusion weights in Fischer-344 rats were 3.68 g greater than the same weights in SpragueDawley rats. The differences between strains in pleural effusion weights and in bronchoalveolar lavage protein concentrations (Fischer-344 > SpragueDawley) quantitatively overshadow the rather modest apparent (not statistically significant) difference in extravascular lung water (Sprague-Dawley > Fischer344), but the qualitative discordance of the parameters is somewhat puzzling at this time. Measurement of LDH activity has also been used previously as an indicator of cell and/or organ injury, including activities in plasma and lavage as measurements of lung injury in oxygen-toxic mice in v i v o . 2t In the present study, plasma LDH activities do appear
10
0
24
48
80
TIME(h) Fig. 6. Plasma glutathione. Concentrations of plasma GSH were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Plasma GSH concentrations increased in F344 rats following 48 and 60 h of hyperoxic exposure although the apparent increase was not statistically significant. In SD rats, plasma GSH concentrations were unchanged by hyperoxic exposure (n = 6 to 8 of each strain per time point).
Hyperoxic lung injury in rats
uM
3
0
24
48
•
F344
[]
S.D.
60
TIME(h) Fig. 7. Plasma glutathione disulfide. Plasma GSSG concentrations were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Plasma GSSG concentrations in SD rats are significantly lower than in F344 rats following 48 and 60 h of hyperoxic exposure. Additionally, plasma GSSG concentrations in SD rats following 48 h and 60 h of hyperoxic exposure are significantly lower than in air-breathing controls (n = 6 to 8 of each strain per time point). *p < .001 F344 vs. SD at 48 and 60 h. )p < .005 SD 48 and 60 h vs. SD 0 h.
in nonsmokers), finding the alveolar lining fluid levels to be much higher than simultaneously measured plasma levels? The concentrations reported by these investigators are relatively consistent with the alveolar lining fluid and plasma levels we observed in rats. Sutherland et al. 26 reported measurements of [GSH + GSSG] in lavage fluids of rats with corrections derived from distributions of intravenously administered 3Hpolyethylene glycol (PEG) (mol wt = 4000), and ~4Cdextran (mol wt = 70,000). These investigators estimated mean concentrations of (GSH + GSSG) in alveolar lining fluid of 1.95 mM in air-breathing controls and 0.77 mM in rats after 30 h of hyperoxic exposure. Unfortunately, no variance estimates or statistical analyses of the data were reported. In addition, Sutherland et al.26 mention a difference in the distribution into the lavaged spaces between dextran and PEG, which further complicates interpretation of their report. The explanation for the greater concentrations of
537
GSH in the alveolar lining fluid than in plasma is unclear, but there are a number of cell types that inhabit the lower respiratory tract that are thought to export GSH. 27'2s This idea of active export of GSH from cells (most of which have 1-2 mM GSH concentrations) is supported by the 0.8 mM concentration of GSH we estimate in the alveolar lining fluid, which more closely resembles an intracellular fluid or the active secretion of GSH to the alveolar lining fluid. In addition, the activities of the enzyme responsible for degradation and uptake of GSH by parenchymal cells, namely gamma-glutamyltranspeptidase, 29-3~are lower in lungs than in other organs, 32 which suggests that GSH is not cleared as rapidly from the alveolar lining fluid. The decrease in alveolar lining fluid GSH concentrations in Fischer-344 rats exposed to hyperoxia we observed also differs somewhat from the results of Jenkinson et al.) ° who observed an increase in GSH concentrations in bronchoalveolar lavage fluid of Sprague-Dawley rats as a function of time exposed to 80% oxygen. Eighty percent oxygen is not lethal, but induces adaptive responses. The increase observed by Jenkinson et alJ ° may reflect an adaptive response or it may represent an effect of cell injury that is more subtle than are the effects of hyperoxia normally monitored to estimate the extent of oxygen-induced injury. There are a number of possible explanations for the apparent decrease in alveolar lining fluid GSH in Fischer-344 rats. The low GSH concentrations we estimate in alveolar lining fluid may represent actual decreases. Such decreases may arise from injury to the cells responsible for maintaining alveolar lining fluid GSH concentrations. In addition, damage to the alveolar-capillary barrier, which would allow alveolar lining fluid to mix with plasma, could contribute to the apparent decrease in GSH concentration in alveolar lining fluid. This latter mechanism is supported by our finding of an increase in the plasma GSH concentrations seen in Fischer-344 rats as early as 48 h of hyperoxic exposure. Further support for this mechanism comes from Matalon et al., a3who found that the alveolar epithelium became more permeable to solute as a function of time exposed to 100% oxygen in rabbits, even before there was evidence of lung injury in the form of pulmonary edema. Another potential reason for the decrease in alveolar lining fluid GSH concentrations involves the estimation of alveolar lining fluid volume. The estimation of alveolar lining fluid volume by using lavage urea nitrogen as a marker may become less valid as the time exposed to hyperoxia increases. Rennard et al., 8 when describing the technique, state that the al-
538
J.D. STENZELel al.
veolar lining fluid volume calculated using lavage urea nitrogen as a marker may be an overestimation of the actual volume of alveolar lining fluid recovered because of the diffusion to the lavage fluid of urea nitrogen from sources other than the alveolar lining fluid. In this event, the measured lavage urea nitrogen would overestimate the actual concentration of urea nitrogen in vivo, which, according to the equation for estimating alveolar lining fluid recovered [alveolar lining fluid volume = (lavage volume × lavage urea nitrogen)/plasma urea nitrogen], would then increase the estimated alveolar lining fluid volume recovered, thus artifactually lowering the calculated concentrations of GSH and GSSG in the alveolar lining fluid. We conclude that Fischer-344 rats have a greater susceptibility to oxygen-toxic lung injury in vivo than do Sprague-Dawley rats, as evidenced by the increases in pleural effusions weights and in lavage protein concentrations. This strain difference affords a unique model allowing for further investigation of hyperoxic lung injury in vivo. Although the importance of GSHdependent systems in antioxidant defense mechanisms is well established by numerous previous reports, the present data do not reveal a marked depletion of GSH or increase in production of GSSG that would readily explain the strain differences. The critical roles played by GSH in maintaining critical physiological functions therefore would appear to be more subtle than are the effects reflected by the parameters measured in our study. Hopefully, the strain difference in oxidant susceptibility observed in vivo by He et al. 7 and in vivo in this study will prove helpful in identifying these subtle, but pathologically critical, effects. Although the present studies are consistent with the greater sensitivity of Fischer-344 rats than of Sprague-Dawley rats to hepatic oxidant stresses in vivo 6 and to pulmonary oxidant stresses in vitro, 7 the present report does not explain the reasons for the strain difference. Recently, Gupta et al. 34 have reported that the strain susceptibility to hepatic necrosis mediated by diquat-induced redox stresses coincides with alterations in intrahepatic iron compartmentation in the Fischer-344 rats, but not in the SpragueDawley rats. The present working hypothesis derived from the studies with diquat is that tissues in vivo can withstand exposure to substantial amounts of oxidants such as hydrogen peroxide, but relatively small amounts of chemically reactive iron (or copper) chelates in the presence of a flux of hydrogen peroxide are damaging. Whether the loss of homeostatic control of intracellular iron compartmentation in hyperoxia-exposed lungs occurs to a greater extent in Fischer-344
rats than in Sprague-Dawley rats in not known at this time, but we are investigating this possibility.
REFERENCES 1. Clark, J. M.; Lambertson, C. J. Pulmonary oxygen toxicity. Pharmacol. Rev. 23:37-133; 1971. 2. Fanburg, B. L.; Deneke, S. M. Normobaric oxygen toxicity of the lung. N. Engl. Z Med. 303:76-86; 1980. 3. Crapo, J. D.; Barry, B. E.; Foscue, H. A.; Shelburne, J. Structural and biochemical changes in rats lungs occurring during exposures to lethal and adaptive doses of oxygen. Am. Rev. Respir. Dis. 122:123-143; 1980. 4. Hansen, T. N.; Gest, A. L. Oxygen toxicity and other ventilatory complications of treatment of infants with persistent pulmonary hypertension. Clin. Perinatology 11:653-672; 1984. 5. Frank, L. Developmental aspects of experimental pulmonary oxygen toxicity. J. Free Radic. Biol. Med. 11:463-494; 1991. 6. Smith, C. V.; Hughes, H.; Lauterburg, B. H.; Mitchell, J. R. Oxidant stress and hepatic necrosis in rats treated with diquat. J. Pharmacol. Exp. Ther. 235:172-177; 1985. 7. He, L.; Shang, S.; Ortiz de Montellano, P.; Burke, T. J.; Voelkel, N. F. Lung injury in Fischer but not Sprague-Dawley rats after short-term hyperoxia. Am. J. Physiol. 259:L451-L458; 1990. 8. Rennard, S. I.; Basset, G.; Lecossier, D.; O'Donnell, K. M.; Pinkston, P.; Martin, P. G.; Crystal, R. G. Estimation of volume of epithelial lining fluid recovered by lavage using urea as marker of dilution. J. Appl. Physiol. 60:532-538; 1986. 9. Cantin, A. M.; North, S. L.; Hubbard, R. C.; Crystal, R. G. Normal alveolar epithelial lining fluid contains high levels of glutathione. J. Appl. Physiol. 63:152-157; 1987. 10. Jenkinson, S. G.; Black, R. D.; Lawrence, R. A. Glutathione concentrations in rat lung bronchoalveolar lavage fluid: Effects of hyperoxia. J. Lab. Clin. Med. 112:345-351; 1988. 11. Erdmann, A. J.; Vaughan, T. R., III; Brigham, K. L.; Woolverton, W. C.; Staub, N. C. Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ. Res. 37:271-284; 1975. 12. Pearce, M. L.; Yamashita, J.; Beazell, J. Measurement of pulmonary edema. Circ. Res. 16:482-488; 1965. 13. Adams, J. D.; Lauterberg, L. B. H.; Mitchell, J. R. Plasma glutathione and glutathione disulfide in the rat: Regulation and response to oxidative stress. J. Pharmacol. Exp. Ther. 227:749754; 1983. 14. Bergmeyer, H. U.; Bernt, E. Methods of enzymatic analysis. New York: Academic Press; 1974. 15. Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. Protein measurement with the Folin reagent. J. Biol. Chem. 132:265-275; 1951. 16. Millikan, G. E.; Johnson, D. E. Analysis of messy data. New York: Van Nostrand Reinhold Co.; 1984. 17. Hansen, T. N.; Hazinski, T. A.; Bland, R. D. Vitamin E does not prevent oxygen-induced lung injury in newborn lambs. Pediatr. Res. 16:583-587; 1982. 18. Newman, J. H.; Loyd, J. E.; English, D. K.; Ogletree, M. L.; Fulderson, W. J.; Brigham, K. L. Effects of 100% oxygen on lung vascular function in awake sheep. J. Appl. Physiol. 54:1379-1386; 1983. 19. Hazinski, T. A.; Kennedy, K. A.; France, M. L.; Hansen, T. N. Pulmonary 02 toxicity in lambs: Physiological and biochemical effects of endotoxin infusion. J. Appl. Physiol. 65:15791585; 1988. 20. Fukushima, M.; King, L. S.; Kang, K.; Banerjee, M.; Newman, J. H. Lung mechanics and airway reactivity in sheep during development of oxygen toxicity. J. Appl. Physiol. 69:1779-178; 1990. 21. Smith, L. J. Hyperoxic lung injury: Biochemical, cellular, and
Hyperoxic lung injury in rats
22. 23. 24.
25. 26.
27.
morphologic characterization in the mouse. J. Lab. Clin. Med. 106:269-278; 1985. Smith, L. J.; Friedman, H.; Anderson, J. Hyperoxic lung injury in mice: Effect of neutrophil depletion and food deprivation. J. Lab. Clin. Med. 111:449-458; 1988. Shasby, D. M.; Fox, R. B.; Harada, R. N.; Repine, J. E. Reduction of the edema of acute hyperoxic lung injury by granulocyte depletion. J. AppL PhysioL 52:1237-1244; 1982. De Los Santos, R.; Seidenfeld, J. J.; Anzueto, A.; Collins, J. F.; Coalson, J. J.; Johanson, W. G.; Peters, J. I. One hundred percent oxygen lung injury in adult baboons. Am. Rev. Respir. Dis. 136:657-661; 1987. Holm, B. A.; Notter, R. H.; Siegle, J.; Matalon, S. Pulmonary physiological and surfactant changes during injury and recovery from hyperoxia. J. AppL Physiol. 59:1402-1409; 1985. Sutherland, M. W.; Glass, M.; Nelson, J.; Lyen, Y.; Forman, H. J. Oxygen toxicity: Loss of lung macrophage function without metabolite depletion. J. Free Radic. BioL Med. 1:209-214; 1985. Bannai, S.; Tsukeda, H. The export ofglutathione from human diploid cells in culture. J. Biol. Chem. 254:3444-3450; 1979.
539
28. Dethmers, J. K.; Meister, A. Glutathione export by human lymphoid cells: Depletion of glutathione by inhibition of its synthesis decreases export and increases sensitivity to irradiation. Proc. NatL Acad. Sci. USA 78:7492-7496; 1981. 29. Meister, A. New aspects ofglutathione biochemistry and transport: Selective alteration ofglutathione metabolism. FASEB J. 43:3031-3042; 1984. 30. Meister, A.; Anderson, M. E. Glutathione. Ann. Rev. Biochem. 52:711-760; 1983. 31. Meister, A. Glutathione metabolism and its selective modification. J. BioL Chem. 263:17205-17208; 1988. 32. Albert, Z.; Orlowska, J.; Orlowskki, M.; Szewczuk, A. Histochemical and biochemical investigations of gamma-giutamyl transpeptidase in the tissues of man and laboratory animals. Acta. Histochem. Suppl. 18:$78-$89; 1964. 33. Matalon, S.; Egan, E. A. Effects of 100% O2 breathing on permeability of alveolar epithelium to solute. J. AppL PhysioL 50:859-863; 198 I. 34. Gupta, S.; Chang, A.; Rogers, L. K.; Smith, C. V. Diquat-induced subcellular redistribution of iron in rats. FASEB J. 6:1056; 1992 (abstract).
0891-5849/93 $6.00 + .00 Copyright © 1993 Pergamon Press Ltd.
Original Contribution HYPEROXIC
LUNG
INJURY
IN FISCHER-344 RATS IN VIVO
A N D SPRAGUE-DAWLEY
JOEL D . STENZEL, STEPHEN E. WELTY, A R T H U R E. BENZICK, E. O ' B R I A N SMITH, CHARLES V. SMITH, and THOMAS N . HANSEN Department of Pediatrics, Baylor College of Medicine, Houston, TX 77030, USA (Received 23 November 1992; Accepted 11 December 1992) Abstract--Supplemental oxygen remains an important therapy for pulmonary insufficiency, despite the potential adverse effects of hyperoxic exposures. Recently, He et al. reported that hyperoxic ventilation more readily damaged isolated perfused lungs from Fischer-344 rats than from Sprague-Dawley rats (Am. J. Physiol. 259:L451), which correlates with the previously reported strain differences in hepatic responses to diquat-induced oxidant stress in vivo (J. Pharmacol. Exp. Ther. 235:172). We therefore examined the differences in hyperoxic lung injury in Fischer-344 and Sprague-Dawley rats in vivo. Adult male rats were exposed to > 95% 02 and were sacrificed after 24, 48, or 60 h. Control animals were maintained in room air. Dramatically greater increases in pleural effusions and bronchoalveolar lavage protein concentrations in response to hyperoxia were observed in the Fischer-344 rats than in the Sprague-Dawley rats Co < .05 at both 48 and 60 h for both measurements). Additionally, the glutathione concentrations in alveolar lining fluid decreased from 800 #M to 115 zM in Fischer-344 rats after 60 h of> 95% 02, but did not change in Sprague-Dawley rats. We conclude that the greater susceptibility of Fischer-344 than of Sprague-Dawley rats to hyperoxic lung injury in vitro reported previously also is observed in vivo and that this strain difference offers unique opportunities to study mechanisms of hyperoxic lung injury. Keywords--Oxygen toxicity, Glutathione, Fischer-344 rats, Pleural effusions, Bronchoalveolar lavage fluid
tion to oxidant-induced organ injury in Fischer-344 rats when compared with Sprague-Dawley rats. 6,7 Smith et al., 6 using diquat to generate intracellular oxidants in the liver, found that Fischer-344 rats developed liver necrosis when treated with diquat while similarly treated Sprague-Dawley rats did not. He et al. 7 compared the effects of oxidant stress in isolated perfused lungs from the same two strains and found that lungs from Fischer-344 rats developed injury and edema when exposed to oxidant stresses, including exposure to hyperoxia, while isolated perfused lungs from Sprague-Dawley rats treated identically did not show similar signs of injury. If the predilection to hyperoxia-induced lung injury in Fischer-344 rats compared with Sprague-Dawley rats would be expressed in vivo in the responses of these strains to hyperoxia, it would provide a useful paradigm for investigation of the biological and molecular mechanisms of oxygen-toxic lung injury in vivo. Therefore, we chose to investigate the susceptibilities to lung injury of Fischer-344 rats versus SpragueDawley rats in response to continuous exposure to greater than 95% oxygen. As indicators of lung injury,
INTRODUCTION
Despite significant improvements in biotechnology, and in many circumstances in the application of these improvements to clinical medicine, supplemental oxygen administration remains an important therapy for patients with respiratory system disease. As with almost any therapeutic intervention, the unwanted side effects of supplemental oxygen exposure must be weighed against the potential benefits of the therapy. The changes that take place in oxygen toxic lung injury have been investigated in a number of well-designed and executed studies, 1-5 but many fundamental questions remain unanswered and additional investigations providing new perspectives are needed. Previously, investigators studying mechanisms of oxidant-induced injury found an interesting predilecSupported by Baylor Child Health Research Center New Project Development Award 5 P30 HD27823, USDA/ARS Children's Nutritional Research Center. Address correspondence to: Stephen E. Welty, MD, Dept. of Pediatrics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. 531
J . D . STENZEL ~ aL
532
we measured pleural effusion weights and bronchoalveolar lavage protein concentrations as a function of time of exposure to hyperoxia. As possible indirect indicators of lung injury, we measured plasma and lavage lactate dehydrogenase (LDH) activities in each strain as a function of time exposed to hyperoxia. In addition, we measured concentrations of total glutathione (GSH + GSSG) and GSSG in plasma and in bronchoalveolar lavage fluid along with simultaneously measured urea nitrogen concentrations so that we could calculate alveolar lining fluid concentrations of GSH and GSSG by the method of Rennard et al. s Previously, investigators have estimated that GSH concentrations in alveolar lining fluid in humans were more than 100-fold greater than in plasma and speculated that these high GSH levels may be critical in the protection against environmental oxidants, 9 while others have found that bronchoalveolar lavage GSH concentrations are increased in rats exposed to 80% oxygenJ ° In our studies, Fischer-344 rats did prove to be more susceptible to hyperoxic lung injury in vivo than were Sprague-Dawley rats, as evidenced by increased pleural effusion weights and bronchoalveolar lavage protein concentrations after 48-60 h of hyperoxic exposure seen in Fischer-344 rats. Additionally, we found that hyperoxic exposure caused alveolar lining fluid concentrations of GSH to decrease significantly in Fischer-344 rats, whereas we did not observe a similar decrease in GSH levels in hyperoxia-exposed Sprague-Dawley rats. Despite the obvious oxidant stresses of hyperoxic exposure and the marked differences in physiological responses, we did not observe a significant increase in GSSG concentrations in any of the fluids examined. MATERIALS AND M E T H O D S
Animal protocol The study was done using either Sprague-Dawley or Fischer-344 adult male rats weighing 200 to 265 g obtained from Harlan Industries (Houston, TX). The rats were housed in a sealed chamber with a 12:12 hour light-dark cycle with free access to food and water throughout the experimental period. The FiO2 was monitored twice daily and maintained at greater than 0.95 by the administration of pure oxygen at 5 L per minute. Calcium carbonate (Sodasorb, Dewey & Almy Chemical Division, Grace & Co., Lexington, MA) was placed in the chamber to absorb carbon dioxide. Throughout the study period, both strains were exposed to identical concentrations of oxygen simultaneously in the same sealed chamber with a partition separating the two strains.
Measurements Equal numbers of Fischer-344 and Sprague-Dawley rats were sacrificed after 24, 48, and 60 h of exposure to > 95% oxygen. At each time point air-breathing control animals from each strain were sacrificed and the samples were processed in parallel with samples obtained from the hyperoxia-exposed rats. Data from the control animals are presented as time -- 0 (for 0 h in > 95% 02). The rats were anesthetized with 200 mg/kg ofintraperitoneal pentobarbital. Next, the trachea was surgically exposed and cannulated using an 18-gauge 1¼-inch Angiocath (Deseret Medical, Inc., Becton Dickinson & Co., Sandy, UT). The thoracic cavity was then opened using a midline thoracotomy. Pleural effusion weights were measured by absorbing the pleural fluid with a preweighed dry gauze pad and reweighing the pad immediately. Next, intracardiac blood samples were obtained for measurement of plasma GSSG, total glutathione (GSH + GSSG), LDH activity, hemoglobin, and blood urea nitrogen (BUN). Heparinized syringes were used for the collection of plasma for the measurement of LDH activity, hemoglobin, and BUN. Following the collection of the blood and plasma samples, lung tissue from the right lung was obtained and frozen at -20°C and subsequently used to measure extravascular lung water by a modification of the wet-to-dry determination ll described by Pearce et alJ 2 After removing the right lung, the left lung was lavaged. Lavage fluid was obtained using 5 cc of 4°C saline through the previously inserted tracheostomy tube. Lavage fluid was aliquoted for measurement of LDH activity, protein, urea nitrogen, GSSG, as well as total glutathione determinations. If the amount of lavage fluid recovered was less than 1.6 cc (volume oflavage per animal required for the above measurements), then the lavage procedure was repeated using an additional 5 cc of 4°C saline. Recovery volumes of lavage fluid were recorded and typically fell between 50% and 80%. Low recoveries, hence second lavages, were required in only a few of the animals having clinically evident respiratory distress following prolonged hyperoxic exposure. Plasma and lavage GSSG and total glutathione concentrations were determined by the method of Adams et alj3 with modifications. The samples for determination of plasma total glutathione were immediately added to equal volumes of 4°C 0.05 M potassium phosphate buffer containing 17.5 mM ethylenediaminetetraacetic acid (EDTA), 0.05 M serine, and 0.05 M borate, pH 7.4 (Buffer I). The samples were mixed immediately by gentle inversion and centrifuged for 3 min. The supernatants (plasma + buffer)
Hyperoxiclung injury in rats were removed and frozen at - 7 0 ° C for later analysis. The samples for determination of plasma GSSG were immediately added to 500 #L of Buffer I plus 15 #L of 1 M N-ethylmaleimide (NEM). The samples were similarly mixed and centrifuged and the supernatants removed and stored at - 7 0 ° C for later analysis. For the measurements of lavage GSSG and total glutathione, 500 #L of lavage fluid was added to 100 #L of 4°C Buffer I and then centrifuged in a microfuge for 3 min. Supernatants were removed, 10 #L of 1 M NEM was added to the lavage GSSG sample, and both GSSG and total glutathione lavage samples were frozen at -70°C. To measure plasma total glutathione, a 100-#L aliquot of the plasma/Buffer I supernatant was added to a cuvette containing 800 #L of 0.1 M phosphate, 5 mM Na2EDTA, pH 7.4 (Buffer II), 50 #L of 10 mM 5,5'-dithio-bis(2-nitrobenzoic acid) (DTNB), and 0.7 U of GSSG reductase (Sigma Type III, Sigma Chemical Co., St. Louis, MO). This mixture was incubated for 1 min at room temperature. To initiate the reaction, 240 nmol of reduced nicotinamide adenine dinucleotide phosphate (NADPH) in Buffer II was added, making the final reaction volume 1.1 ml. The change in absorbance at 412 nm was continuously monitored for 5 min. Lavage total glutathione measurements were made using 100 #L of buffer-lavage supernatant with the same reaction mixture. Plasma GSSG determinations were made by taking a 250-tzL aliquot of the plasma-buffer-NEM solution and passing it at 1 drop/s through a C 18 Sep-Pack cartridge (Waters Associates, Framingham, MA) followed by 800 ~tL of Buffer II. An 800-#L aliquot of the combined eluate was added to a cuvette along with 50 #L of 10 mM DTNB and 0.7 U ofglutathione reductase. The assay then proceeded as for the measurement of total glutathione as noted above. Lavage GSSG was measured as above, substituting lavagebuffer-NEM solution for the plasma solution. Sample GSSG and total glutathione concentrations were calculated from standard curves generated with each set of measurements. Plasma and lavage GSH concentrations were calculated by subtracting the GSSG concentrations from the total glutathione (GSH + GSSG) concentrations for each sample. To interpret and compare the levels of GSH and GSSG in lavage, we estimated the amount of actual alveolar lining fluid recovered in the lavage fluid for each rat by comparing plasma urea nitrogen with lavage urea nitrogen concentrations as described by Rennard et al., 8 using the assumption that the concentrations of freely diffusable urea nitrogen in alveolar lining fluid and plasma are the same. From the quantity of alveolar lining fluid recovered, we then calculated the concentrations
533
of GSSG and total glutathione per milliliter of alveolar lining fluid for each rat. For the measurement of LDH activities, heparinized blood samples were centrifuged, the plasma removed, immediately placed on ice, and analyzed according to the method of Bergmeyer and Bernt.14 Lavage sample supernatants for measurement of LDH were analyzed similarly. Plasma and lavage samples for the measurement of BUN were frozen at - 2 0 ° C and later analyzed using kits purchased from Sigma (#535 or #66-UV). Blood hemoglobin samples were frozen at - 2 0 ° C and later analyzed using BoehringerMannheim Diagnostics Cyanmethemoglobin Reagent (Boehringer-Mannheim Biochemicals, Indianapolis, IN). Lavage protein samples were stored at - 2 0 ° C until analysis by the method of Lowry et al. 15
Statistics All data are expressed as mean _ standard error of the mean. Two-way analysis of variance (ANOVA) was used to assess the effects of strain, time of hyperoxic exposure, and the interaction between these two factors. If the interaction was significant, strains were compared at each time point and times of hyperoxic exposure were compared within strain using the least significant difference (LSD) multiple comparison procedure. Data that appeared different but were not indicated as different by the initial two-way ANOVA (i,e., no significant interaction detected) were evaluated using the Bonferroni multiple comparison procedure. Once interactions were ruled out, the overall effects of strain and time of hyperoxic exposure were interpreted, with various time means being compared using the LSD procedureJ 6 P < .05 was used to determine statistical significance. RESULTS
General appearance Neither Fischer-344 nor Sprague-Dawley rats displayed evidence of illness or respiratory distress after 48 h of hyperoxic exposure. Fischer-344 rats developed evidence of respiratory distress after 60 h of hyperoxic exposure, manifested primarily by tachypnea, while none of the similarly exposed Sprague-Dawley rats developed similar signs of respiratory distress. There were no deaths in hyperoxia-exposed rats.
Indices of lung injury Hyperoxic exposure increased the bronchoalveolar lavage protein concentrations significantly in Fischer344 rats at 48 and 60 h (fourfold increase at 48 h and
534
J.D. STENZELet al.
MG/ML
+ . 1.2
•
F344
[]
S.D.
0.8
0.4
"
0.0 0
48
24
60
54% and 85% at 48 and 60 h respectively, while in similarly treated Sprague-Dawley rats alveolar lining fluid GSH concentrations did not decrease (Fig. 4). Mean alveolar lining fluid GSSG concentrations were 66% lower at 60 h in the Fischer-344 rats, but this effect was not statistically significant, while hyperoxic exposure had less effect on GSSG concentrations in the alveolar lining fluids of the Sprague-Dawley rats (Fig. 5). Furthermore, concentrations of GSSG in the alveolar lining fluid were not different between the two strains. At 48 and 60 h ofhyperoxic exposure, plasma GSH concentrations were higher in Fischer-344 rats (although not reaching statistical significance) than in similarly treated Sprague-Dawley rats (Fig. 6). Hyperoxic exposure did not affect plasma GSH concentrations in the Sprague-Dawley rats. Compared with airbreathing controls, plasma concentrations of GSSG decreased in Sprague-Dawley rats at 48 and 60 h of hyperoxic exposure and were significantly lower than in similarly treated Fischer-344 rats (Fig. 7).
TIME (h)
Fig. 1. Bronchoalveolar lavage protein. Concentrations of bronchoalveolar lavage protein (BAL) were measured in hyperoxia-exposed Fischer-344 (F344) rats, hyperoxia-exposed Sprague-Dawley (SD) rats, and in their respective air-breathing controls (0 h). Compared to BAL protein concentrations in SD rats, F344 BAL protein concentrations rise significantly with increasing duration of hyperoxic exposure (n = 5 to 8 of each strain per time point). *p < .005 at 60 h F344 vs. SD and p < .002 at 48 h F344 vs. SD. tp < .001 for F344 60 h and F344 48 h vs. F344 0 h.
fivefold at 60 h), but had no effect on bronchoalveolar lavage protein concentrations in Sprague-Dawley rats (Fig. 1). While exposure to hyperoxia increased pleural effusion weights in both strains of rats, the increase was greater in the Fischer-344 rats (fivefold at 48 h and 15-fold at 60 h) than in the Sprague-Dawley rats (threefold increase at 60 h) (Fig. 2). Extravascular lung water increased in both Fischer-344 rats and Sprague-Dawley rats at 48 and 60 h of exposure to hyperoxia (Fig. 3). Within each strain, exposure to hyperoxia did not affect LDH activities in bronchoaveolar lavage (Table 1) or plasma (Table 2) measured over time. However, plasma LDH activities in the Fischer-344 rats at 60 h were approximately twice the respective activities observed in the Sprague-Dawley rats at the same time point.
GSH and GSSG concentrations in alveolar lining fluid and plasma Hyperoxic exposure decreased GSH concentrations in the alveolar lining fluid of Fischer-344 rats by
DISCUSSION
The primary finding of this investigation is that Fischer-344 rats are more sensitive to hyperoxic lung GM
8.0
= []
F344 S.D.
4.0
++
,4,- 'k 2.0
0.0
0
24
48
80
TIME (h)
Fig. 2. Pleural effusions. Pleural effusion weights were measured in hyperoxia-exposed F344 rats, hyperoxia-exposed SD rats, and their respective air-breathing controls (0 h). Pleural effusion weights ot hyperoxia-exposed F344 rats were significantly higher than pleural effusion weights in identically exposed SD rats (n = 7 to 8 of each strain per time point). *p < .001 at 60 h F344 vs. SD and p < .05 at 48 h F344 vs. SD. tp < .001 F344 60 h vs. F344 0 h and p < .01 F344 48 h vs. F344 0 h. *p < .02 SD 60 h vs. SD 0 h.
535
Hyperoxic lung injury in rats Table 2. Plasma LDH Activity (IU/L)
GM H20/GM DRY BLOODLESS LUNG 9
• F344 [] S.D. +4-
4--I-
Time exposed to > 95% O2
0h
24 h
48 h
60 h
Fischer-344 Sprague-Dawley
244 _+ 40 187 _+ 19
208 _+ 33 155 +_ 17
292 _+ 44 238 +_ 48
324 _+ 52 149 +_ 12
Plasma LDH activities were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Mean plasma LDH activities in hyperoxia-exposed F344 rats trend upward, but this effect is not statistically significant. In Sprague-Dawley rats, plasma LDH activities are unchanged by hyperoxic exposure (n = 7 to 8 of each strain per time point).
ability pulmonary edema. This pulmonary edema has been determined as an increase in lung lymph flow or lymph-to-plasma total protein ratios, ]7-20increases in extravascular lung water, or increases in broncholalveolar lavage protein concentrations.17-2s In the present study, we found that hyperoxia-exposed Fischer-
uM 0
24
48
12oo 1
60
[] F344 [] S.D.
TIME(h) Fig. 3. Extravascular lung water. Extravascular lung water was measured using the wet-to-dry method in hyperoxia-exposed F344 and SD rats and in air-breathing controls (0 h) of both strains. Extravascular lung water was significantly higher in hyperoxia-exposed animals of both strains compared with their respective air-breathing controls. However, no statistical difference was detected between F344 and SD rats during hyperoxic exposure (n = 7 to 8 of each strain per time point). *p < .001 F344 48 and 60 h vs. F344 0 h. *p < .001 SD 48 and 60 h vs. SD 0 h.
injury in vivo than are Sprague-Dawley rats. The increased susceptibility of Fischer-344 rats was seen in the increased pleural effusion weights and bronchoalveolar lavage protein concentrations after 48 and 60 h of continuous hyperoxic exposure. Previous studies have demonstrated that lung injury secondary to hyperoxic exposure is manifested primarily as a perme-
800
400
0
Table 1. Lavage LDH Activity (IU/L)
24
48
60
TiME (h)
Time exposed to > 95% 02
0h
24 h
48 h
60 h
Fischer-344 Sprague-Dawley
170 +_41 127 _+ 12
201 + 25 141 + 15
163 _ 21 123 _ 23
186 _+ 43 99 + 15
Lavage LDH activities were measured in hyperoxia-exposed F344 rats, hyperoxia-exposed SD rats, and in their respective airbreathing controls (0 h). No statistical differences were detected between strains relative to the duration of hyperoxic exposure (n = 6 to 8 of each strain per time point).
Fig. 4. Alveolar lining fluid glutathione. GSH was measured in BAL fluid and corrected (by the urea method of determining volume of alveolar lining fluid [ALl:'] recovered in the BAL fluid) to give GSH concentrations in ALF of hyperoxia-exposed F344 rats, hyperoxiaexposed SD rats, and their respective air-breathing controls (0 h). GSH concentrations in ALF of F344 rats following 60 h of hyperoxic exposure are significantly lower than in identically exposed SD rats and also lower than in F344 air-breathing controls (n = 4 to 8 of each strain per time point). *p < .05 F344 vs. SD at 60 h. *p < .002 F344 60 h vs. F344 0 h.
536
J.D. STENZEL el al.
uM 45
•
F344
[]
S.D.
30
15
0
24
48
80
TIME(h)
to increase in Fischer-344 rats in response to hyperoxic exposure, but this increase was not quite statistically significant. Previous investigators have shown that measurement of specific LDH isoforms enhances the sensitivity of plasma LDH measurements for determination of lung injury. 2~ In the context of this investigation comparing plasma and lavage LDH activities in Fischer-344 rats versus Sprague-Dawley rats, the relative increased activities of lavage LDH and plasma LDH at 60 h in Fischer-344 rats suggest a predilection to oxygen-toxic lung injury in Fischer344 rats when compared to Sprague-Dawley rats, but the effects of hyperoxic exposure on the manifestations of permeability pulmonary edema are far more dramatic. Estimating alveolar lining fluid concentrations of GSH and GSSG by the urea method as a function of time exposed to hyperoxia, to our knowledge, has not been reported. Others have determined baseline alveolar lining fluid levels of GSH and GSSG in humans (approximately 800 pM in smokers and 400 pM
uM
30 1
Fig. 5. Alveolar lining fluid giutathione disulfide. GSSG concentrations were measured in BAL fluid and corrected (by the urea method of determining volume of ALF recovered in the BAL fluid) to give GSSG concentrations in ALF of hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). No statistical differences were detected between strains relative to hyperoxic exposure (n = 3 to 8 of each strain per time point).
i
I
•
F344
[]
S.D.
20
344 rats developed dramatic evidence of pulmonary edema, which was reflected in the early (48 h) and sustained (60 h) increases in bronchoalveolar lavage protein concentrations and pleural effusion weights. Sprague-Dawley rats were not as affected by identical hyperoxic exposures. After 60 h of hyperoxic exposure, the pleural effusion weights in Fischer-344 rats were 3.68 g greater than the same weights in SpragueDawley rats. The differences between strains in pleural effusion weights and in bronchoalveolar lavage protein concentrations (Fischer-344 > SpragueDawley) quantitatively overshadow the rather modest apparent (not statistically significant) difference in extravascular lung water (Sprague-Dawley > Fischer344), but the qualitative discordance of the parameters is somewhat puzzling at this time. Measurement of LDH activity has also been used previously as an indicator of cell and/or organ injury, including activities in plasma and lavage as measurements of lung injury in oxygen-toxic mice in v i v o . 2t In the present study, plasma LDH activities do appear
10
0
24
48
80
TIME(h) Fig. 6. Plasma glutathione. Concentrations of plasma GSH were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Plasma GSH concentrations increased in F344 rats following 48 and 60 h of hyperoxic exposure although the apparent increase was not statistically significant. In SD rats, plasma GSH concentrations were unchanged by hyperoxic exposure (n = 6 to 8 of each strain per time point).
Hyperoxic lung injury in rats
uM
3
0
24
48
•
F344
[]
S.D.
60
TIME(h) Fig. 7. Plasma glutathione disulfide. Plasma GSSG concentrations were measured in hyperoxia-exposed F344 and SD rats, and in their respective air-breathing controls (0 h). Plasma GSSG concentrations in SD rats are significantly lower than in F344 rats following 48 and 60 h of hyperoxic exposure. Additionally, plasma GSSG concentrations in SD rats following 48 h and 60 h of hyperoxic exposure are significantly lower than in air-breathing controls (n = 6 to 8 of each strain per time point). *p < .001 F344 vs. SD at 48 and 60 h. )p < .005 SD 48 and 60 h vs. SD 0 h.
in nonsmokers), finding the alveolar lining fluid levels to be much higher than simultaneously measured plasma levels? The concentrations reported by these investigators are relatively consistent with the alveolar lining fluid and plasma levels we observed in rats. Sutherland et al. 26 reported measurements of [GSH + GSSG] in lavage fluids of rats with corrections derived from distributions of intravenously administered 3Hpolyethylene glycol (PEG) (mol wt = 4000), and ~4Cdextran (mol wt = 70,000). These investigators estimated mean concentrations of (GSH + GSSG) in alveolar lining fluid of 1.95 mM in air-breathing controls and 0.77 mM in rats after 30 h of hyperoxic exposure. Unfortunately, no variance estimates or statistical analyses of the data were reported. In addition, Sutherland et al.26 mention a difference in the distribution into the lavaged spaces between dextran and PEG, which further complicates interpretation of their report. The explanation for the greater concentrations of
537
GSH in the alveolar lining fluid than in plasma is unclear, but there are a number of cell types that inhabit the lower respiratory tract that are thought to export GSH. 27'2s This idea of active export of GSH from cells (most of which have 1-2 mM GSH concentrations) is supported by the 0.8 mM concentration of GSH we estimate in the alveolar lining fluid, which more closely resembles an intracellular fluid or the active secretion of GSH to the alveolar lining fluid. In addition, the activities of the enzyme responsible for degradation and uptake of GSH by parenchymal cells, namely gamma-glutamyltranspeptidase, 29-3~are lower in lungs than in other organs, 32 which suggests that GSH is not cleared as rapidly from the alveolar lining fluid. The decrease in alveolar lining fluid GSH concentrations in Fischer-344 rats exposed to hyperoxia we observed also differs somewhat from the results of Jenkinson et al.) ° who observed an increase in GSH concentrations in bronchoalveolar lavage fluid of Sprague-Dawley rats as a function of time exposed to 80% oxygen. Eighty percent oxygen is not lethal, but induces adaptive responses. The increase observed by Jenkinson et alJ ° may reflect an adaptive response or it may represent an effect of cell injury that is more subtle than are the effects of hyperoxia normally monitored to estimate the extent of oxygen-induced injury. There are a number of possible explanations for the apparent decrease in alveolar lining fluid GSH in Fischer-344 rats. The low GSH concentrations we estimate in alveolar lining fluid may represent actual decreases. Such decreases may arise from injury to the cells responsible for maintaining alveolar lining fluid GSH concentrations. In addition, damage to the alveolar-capillary barrier, which would allow alveolar lining fluid to mix with plasma, could contribute to the apparent decrease in GSH concentration in alveolar lining fluid. This latter mechanism is supported by our finding of an increase in the plasma GSH concentrations seen in Fischer-344 rats as early as 48 h of hyperoxic exposure. Further support for this mechanism comes from Matalon et al., a3who found that the alveolar epithelium became more permeable to solute as a function of time exposed to 100% oxygen in rabbits, even before there was evidence of lung injury in the form of pulmonary edema. Another potential reason for the decrease in alveolar lining fluid GSH concentrations involves the estimation of alveolar lining fluid volume. The estimation of alveolar lining fluid volume by using lavage urea nitrogen as a marker may become less valid as the time exposed to hyperoxia increases. Rennard et al., 8 when describing the technique, state that the al-
538
J.D. STENZELel al.
veolar lining fluid volume calculated using lavage urea nitrogen as a marker may be an overestimation of the actual volume of alveolar lining fluid recovered because of the diffusion to the lavage fluid of urea nitrogen from sources other than the alveolar lining fluid. In this event, the measured lavage urea nitrogen would overestimate the actual concentration of urea nitrogen in vivo, which, according to the equation for estimating alveolar lining fluid recovered [alveolar lining fluid volume = (lavage volume × lavage urea nitrogen)/plasma urea nitrogen], would then increase the estimated alveolar lining fluid volume recovered, thus artifactually lowering the calculated concentrations of GSH and GSSG in the alveolar lining fluid. We conclude that Fischer-344 rats have a greater susceptibility to oxygen-toxic lung injury in vivo than do Sprague-Dawley rats, as evidenced by the increases in pleural effusions weights and in lavage protein concentrations. This strain difference affords a unique model allowing for further investigation of hyperoxic lung injury in vivo. Although the importance of GSHdependent systems in antioxidant defense mechanisms is well established by numerous previous reports, the present data do not reveal a marked depletion of GSH or increase in production of GSSG that would readily explain the strain differences. The critical roles played by GSH in maintaining critical physiological functions therefore would appear to be more subtle than are the effects reflected by the parameters measured in our study. Hopefully, the strain difference in oxidant susceptibility observed in vivo by He et al. 7 and in vivo in this study will prove helpful in identifying these subtle, but pathologically critical, effects. Although the present studies are consistent with the greater sensitivity of Fischer-344 rats than of Sprague-Dawley rats to hepatic oxidant stresses in vivo 6 and to pulmonary oxidant stresses in vitro, 7 the present report does not explain the reasons for the strain difference. Recently, Gupta et al. 34 have reported that the strain susceptibility to hepatic necrosis mediated by diquat-induced redox stresses coincides with alterations in intrahepatic iron compartmentation in the Fischer-344 rats, but not in the SpragueDawley rats. The present working hypothesis derived from the studies with diquat is that tissues in vivo can withstand exposure to substantial amounts of oxidants such as hydrogen peroxide, but relatively small amounts of chemically reactive iron (or copper) chelates in the presence of a flux of hydrogen peroxide are damaging. Whether the loss of homeostatic control of intracellular iron compartmentation in hyperoxia-exposed lungs occurs to a greater extent in Fischer-344
rats than in Sprague-Dawley rats in not known at this time, but we are investigating this possibility.
REFERENCES 1. Clark, J. M.; Lambertson, C. J. Pulmonary oxygen toxicity. Pharmacol. Rev. 23:37-133; 1971. 2. Fanburg, B. L.; Deneke, S. M. Normobaric oxygen toxicity of the lung. N. Engl. Z Med. 303:76-86; 1980. 3. Crapo, J. D.; Barry, B. E.; Foscue, H. A.; Shelburne, J. Structural and biochemical changes in rats lungs occurring during exposures to lethal and adaptive doses of oxygen. Am. Rev. Respir. Dis. 122:123-143; 1980. 4. Hansen, T. N.; Gest, A. L. Oxygen toxicity and other ventilatory complications of treatment of infants with persistent pulmonary hypertension. Clin. Perinatology 11:653-672; 1984. 5. Frank, L. Developmental aspects of experimental pulmonary oxygen toxicity. J. Free Radic. Biol. Med. 11:463-494; 1991. 6. Smith, C. V.; Hughes, H.; Lauterburg, B. H.; Mitchell, J. R. Oxidant stress and hepatic necrosis in rats treated with diquat. J. Pharmacol. Exp. Ther. 235:172-177; 1985. 7. He, L.; Shang, S.; Ortiz de Montellano, P.; Burke, T. J.; Voelkel, N. F. Lung injury in Fischer but not Sprague-Dawley rats after short-term hyperoxia. Am. J. Physiol. 259:L451-L458; 1990. 8. Rennard, S. I.; Basset, G.; Lecossier, D.; O'Donnell, K. M.; Pinkston, P.; Martin, P. G.; Crystal, R. G. Estimation of volume of epithelial lining fluid recovered by lavage using urea as marker of dilution. J. Appl. Physiol. 60:532-538; 1986. 9. Cantin, A. M.; North, S. L.; Hubbard, R. C.; Crystal, R. G. Normal alveolar epithelial lining fluid contains high levels of glutathione. J. Appl. Physiol. 63:152-157; 1987. 10. Jenkinson, S. G.; Black, R. D.; Lawrence, R. A. Glutathione concentrations in rat lung bronchoalveolar lavage fluid: Effects of hyperoxia. J. Lab. Clin. Med. 112:345-351; 1988. 11. Erdmann, A. J.; Vaughan, T. R., III; Brigham, K. L.; Woolverton, W. C.; Staub, N. C. Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ. Res. 37:271-284; 1975. 12. Pearce, M. L.; Yamashita, J.; Beazell, J. Measurement of pulmonary edema. Circ. Res. 16:482-488; 1965. 13. Adams, J. D.; Lauterberg, L. B. H.; Mitchell, J. R. Plasma glutathione and glutathione disulfide in the rat: Regulation and response to oxidative stress. J. Pharmacol. Exp. Ther. 227:749754; 1983. 14. Bergmeyer, H. U.; Bernt, E. Methods of enzymatic analysis. New York: Academic Press; 1974. 15. Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. Protein measurement with the Folin reagent. J. Biol. Chem. 132:265-275; 1951. 16. Millikan, G. E.; Johnson, D. E. Analysis of messy data. New York: Van Nostrand Reinhold Co.; 1984. 17. Hansen, T. N.; Hazinski, T. A.; Bland, R. D. Vitamin E does not prevent oxygen-induced lung injury in newborn lambs. Pediatr. Res. 16:583-587; 1982. 18. Newman, J. H.; Loyd, J. E.; English, D. K.; Ogletree, M. L.; Fulderson, W. J.; Brigham, K. L. Effects of 100% oxygen on lung vascular function in awake sheep. J. Appl. Physiol. 54:1379-1386; 1983. 19. Hazinski, T. A.; Kennedy, K. A.; France, M. L.; Hansen, T. N. Pulmonary 02 toxicity in lambs: Physiological and biochemical effects of endotoxin infusion. J. Appl. Physiol. 65:15791585; 1988. 20. Fukushima, M.; King, L. S.; Kang, K.; Banerjee, M.; Newman, J. H. Lung mechanics and airway reactivity in sheep during development of oxygen toxicity. J. Appl. Physiol. 69:1779-178; 1990. 21. Smith, L. J. Hyperoxic lung injury: Biochemical, cellular, and
Hyperoxic lung injury in rats
22. 23. 24.
25. 26.
27.
morphologic characterization in the mouse. J. Lab. Clin. Med. 106:269-278; 1985. Smith, L. J.; Friedman, H.; Anderson, J. Hyperoxic lung injury in mice: Effect of neutrophil depletion and food deprivation. J. Lab. Clin. Med. 111:449-458; 1988. Shasby, D. M.; Fox, R. B.; Harada, R. N.; Repine, J. E. Reduction of the edema of acute hyperoxic lung injury by granulocyte depletion. J. AppL PhysioL 52:1237-1244; 1982. De Los Santos, R.; Seidenfeld, J. J.; Anzueto, A.; Collins, J. F.; Coalson, J. J.; Johanson, W. G.; Peters, J. I. One hundred percent oxygen lung injury in adult baboons. Am. Rev. Respir. Dis. 136:657-661; 1987. Holm, B. A.; Notter, R. H.; Siegle, J.; Matalon, S. Pulmonary physiological and surfactant changes during injury and recovery from hyperoxia. J. AppL Physiol. 59:1402-1409; 1985. Sutherland, M. W.; Glass, M.; Nelson, J.; Lyen, Y.; Forman, H. J. Oxygen toxicity: Loss of lung macrophage function without metabolite depletion. J. Free Radic. BioL Med. 1:209-214; 1985. Bannai, S.; Tsukeda, H. The export ofglutathione from human diploid cells in culture. J. Biol. Chem. 254:3444-3450; 1979.
539
28. Dethmers, J. K.; Meister, A. Glutathione export by human lymphoid cells: Depletion of glutathione by inhibition of its synthesis decreases export and increases sensitivity to irradiation. Proc. NatL Acad. Sci. USA 78:7492-7496; 1981. 29. Meister, A. New aspects ofglutathione biochemistry and transport: Selective alteration ofglutathione metabolism. FASEB J. 43:3031-3042; 1984. 30. Meister, A.; Anderson, M. E. Glutathione. Ann. Rev. Biochem. 52:711-760; 1983. 31. Meister, A. Glutathione metabolism and its selective modification. J. BioL Chem. 263:17205-17208; 1988. 32. Albert, Z.; Orlowska, J.; Orlowskki, M.; Szewczuk, A. Histochemical and biochemical investigations of gamma-giutamyl transpeptidase in the tissues of man and laboratory animals. Acta. Histochem. Suppl. 18:$78-$89; 1964. 33. Matalon, S.; Egan, E. A. Effects of 100% O2 breathing on permeability of alveolar epithelium to solute. J. AppL PhysioL 50:859-863; 198 I. 34. Gupta, S.; Chang, A.; Rogers, L. K.; Smith, C. V. Diquat-induced subcellular redistribution of iron in rats. FASEB J. 6:1056; 1992 (abstract).