Effect of Fasting on Hyperoxic Lung Injury in Mice The Role of Glutathione1-3
LEWIS J. SMITH, JAMES ANDERSON, MIR SHAMSUDDIN, and WEI HSUEH
Introduction
Exposure to high concentrations of oxygen produces acute lung injury in several animal species, including mice (1,2). We recently found that when mice are simultaneously fasted and exposed to 100070 oxygen, they develop lung damage sooner than when they are allowed free access to food (3). The mechanism responsible for this phenomenon is unclear. One possibility is a decrease in lung antioxidant levels. Frank and coworkers (4) reported decreases in glutathione peroxidase (GP) and glucose-e-phosphate dehydrogenase (G6PD) in intrauterine growth-retarded rat pups, but no change in catalase or superoxide dismutase (SOD) (4). In contrast, Deneke and colleagues (5) did not identify any changes in lung levels of G6PD, glutathione reductase, glutathione peroxidase, SOD, or glutathione in rats fed a proteindeficient diet. Other studies from these same two groups, however, have reported no change in GP and G6PD in undernourished rat pups (6) and decreases in G6PD and glutathione in protein-· deficient rats (7). Despite these somewhat contradictory data, there is reason to suspect that reduced levels of glutathione, a tripeptide (gamma-glutamyl-cysteinylglycine) that acts as an intracellular antioxidant by supplying reducing equivalents in the. form of electrons to reduce peroxides (8), . and enzymes of the glutathione redox cycle playa role in the increased susceptibility of fasted mice to hyperoxic lung damage (9). First, fasting (starvation) decreases hepatic glutathione levels in mice and increases susceptibility to oxidant damage (10). One might suspect that fasting also decreases lung glutathione levels. This could result from decreased availability of sulfur-containing amino acids (e.g., methionine and cysteine), which are used in glutathione synthesis and obtained mainly from dietary sources (11, 12), or from increased extrahepatic
SUMMARY Fasted mice exposed to 100% oxygen have more lung damage and die sooner than do fad mice. The mechanism responsible for this phenomenon has not been Identified. We performed the following experlmenta to test the hypothesis that reduced glutathione content In lung tluue of fasted mice contributes to the Increased susceptibility to hyperoxlc lung damage. Flret, alr-exposed mice were fasted for as long as 3 days. They had little change In lung levels of superoxIde dlsmutese (SOD) or cetalase, but they had a 41% decrease In glutathione by Dey 3 (p < 0.001). Second, fad mice and fasted mice were exposed to 100% oxygen for as long as 4 days. Both groups had nearly Identical values of lung SOD and cetalase, but the fasted mice had lower levels of glutathione (p < 0.001).Third, fad mice received the glutathione synthesis Inhibitor buthlonlne suifoxlmlne (BSO; 20 mM) In their drinking water for 2 wk and were then exposed to either air or 100% oxygen. Alr-exposed mice receiving SSO for 14 days had no change In lung SOD content, a 43% Increase In catalase (p < 0.001), and a 41% decrease In glutathione (p < 0.01). Oxygen-exposed, BSG-treated mice had no change In SOD and an Increase In catalase, but lower glutathione levela, more desths, and Increased lung damage on Day 3 (SAL protein: 1.72 ± 0.21 versus 0.94 ± 0.08 mg/ml; p < 0.01)than did diluent-treated mice. Fourth, fasted mice were given llposomes containing glutathione Intratracheally. When exposed to air, they had smaller decreases In lung glutathione levels than did fasted mice receiving either nothing or llposomes containing PBS. When exposed to 100% oxygen, they had higher lung glutathione levels on Days 1, 2, and 3, fewer deaths, and lower SAL protein values on Day 3 than did the untreated fasted mice. Only partial protection was achieved after glutathione administration because the llposomes localized to the left lung. These resulta support our hypothesis that the Increased susceptibility of fasted mice to hyperoxlc lung damage may be due to a reduced lung content of glutathione. AM REV RESPIR DIS 1990: 141:141-149
metabolism of glutathione (13).Second, to 100%oxygenand the effect of glutathifasting may interfere with converting the one supplementation in fasted mice simioxidized form of glutathione (GSSG) to larly exposed. Our findings suggest that its active form (GSH) by depleting glu- glutathione depletion plays an important cose stores and thereby slowing the hex- role in the increased susceptibility of fastose monophosphate (HMP) shunt and ed mice to hyperoxic lung damage. reducing the NADPH levels that are Methods needed for the conversionprocess (14-16). Third, some studies suggest that lung Experimental Animals and Lung Injury glutathione levels correlate with suscep- Male Balb-c mice (Charles River, Portage, tibility to hyperoxia (17-19), although MI), 8 to 10 wk of age and weighing 20 to other studies have questioned this rela- 25 g, were housed 10 mice to a cage, placed tionship (7, 20). The present study was designed to test the hypothesis that fasting decreases lung glutathione levels in mice and that this (Received in original form April 3, 1989 and in decrease contributes to the increased sus- revised form June 5, 1989) ceptibility of fasted mice to hyperoxic 1 From the Departments of Medicine and Pathollung damage. We measured lung levels ogy, Northwestern University, VA Lakeside Mediof glutathione, as wellas two antioxidant cal Center, and Children's Memorial Hospital, enzymes, catalase and SOD, during fast- Chicago, Illinois. 2 Supported by the VA Research Service. ing, oxygen exposure, and the combina3 Correspondence and requests for reprints should tion of the two. We also examined the be addressed to Lewis J. Smith, M.D., Pulmonary effect of chemically induced glutathione Section, Northwestern University Medical Center, depletion in normally fed mice exposed Wesley456, 250 East Superior, Chicago, IL 60611. 141
SMITH, ANDERSON, SHAMSUDDIN, AND HSUEH
142
in polyethylene "glove bags" (American Scientific Products, McGaw Park, IL), and exposed to either 100% oxygen (98 ± 1070) or air for I, 2, 3, or 4 days. Flow rates were set to maintain the carbon dioxide concentration below 0.5070. Oxygen and carbon dioxide concentrations were monitored with Beckman OM-ll and LB-2 analyzers (Beckman Instruments, Fullerton, CA). The temperature was maintained between 23.9° and 26.6° C, and the relative humidity between 40 and 50% . Mice were fed either a normal diet during the exposure (fed) or had all food removed at the start of the exposure (fasted). Water wasprovided to all mice. In some experiments, fed mice received 20 mM DL-buthionine-SRsulfoximine (BSO) (Sigma Chemical, St. Louis, MO), a selective inhibitor of gammaglutamylcysteine synthetase and therefore of glutathione synthesis (21), in their drinking water for the 14 days before and during the exposure. In other experiments, fasted mice received liposome-encapsulated glutathione intratracheally at the start of the exposure. At the end 0 f the exposure, mice were anesthetized with pentobarbital, 50 mg/kg, given intraperitoneally. The lungs were perfused via the right ventricle at 20 cm H 2 0 with cold phosphate-buffered saline (PBS) at pH 7.4 until they became white. Then they were removed from the chest, trimmed of extraneous tissue includ ing trachea and main bronchi, weighed, and homogenized (see below). In the experiments in which bronchoalveolar lavage was also performed, the lungs were perfused, the trachea was cannulated with a no. 18,1.25inch plastic intravenous catheter filled with PBS, and the lungs were lavaged four times with separate l-ml aliquots of cold PBS. The BAL effluent, > 90% of the instilled volume, was immediately placed on ice, and centrifuged at 4° C for 10 min at 500 x g. The lungs were then processed as described above in preparation for homogenization. Except for wet lung weight, which was higher in the lavaged lungs, the lavage procedure had no effect on the lung antioxidant measurements or on the lung protein concentration. An aliquot of the BAL supernatant was stored at 4° C and analyzed within 1 wk for protein concentration by the method of Lowry and coworkers (22) using bovine serum albumin as the standard. The BAL protein concentration was used as a measure of lung injury (2).
Lung Homogenization Lungs were homogenized for 60 s in cold, hypotonic buffer (5 roM potassium phosphate at pH 7.4) using a weight (mg) to volume (ml) ratio of25:1.The homogenate wascentrifuged at 500 x g for 10 min, the pellet was discarded, and the supernatant was centrifuged at 20,000 x g for 10min. Aliquots ofthe supernatant were saved for protein, SOD, catalase, and glutathione assays.
Antioxidant Assays Protein was quantitated by the method of Lowry and coworkers (22) as noted above.
SOD was measured by the method of McCord and Fridovich (23) in which cytochrome c reduction is measured spectrophotometrical1y at 550 nm . Catalase was quantitated spectrophotometrically at 240 nm as the decrease in H 202 over time. Total glutathione was measured by the method of Griffith (24) in which the rate of reduction of 5-5'-dinitrobis{-2-nitrobenzoic acid) (DfNB) is measured spectrophotometrically at 412 nm.
Liposome Preparation Lecithin, cholesterol, and either dicetyl phosphate (negative liposomes) or stearylamine (positive liposomes), obtained from Avanti Polar Lipids (Pelham, AL), were dissolved in chloroform at a molar ratio of 63:9:18 and placed in a round-bottom flask (25). The chloroform was evaporated in a rotary evaporator leaving a thin film. The film was dislodged with buffer solution (PBS or glutathione in PBS, 25 mg/ml) by gentle shaking. The resulting solution was sonicated for 2 min and then passed through a 30-gauge needle to disperse aggregates. This method produces mostly small vesicles approximately 50 to 100 nm in size (26). The liposome preparation was incubated at room temperature for 1h and then instilled intratracheally as described below. To determine glutathione incorporation into the liposome carrier, liposomes were prepared as above, washed twice with buffer, and centrifuged at 170,000 x g for 2 h after each wash. The supernatant was decanted and saved. The pellet was suspended in 0.2% Triton X-IOO to disrupt the liposomes, and the resulting solution was also saved. Glutathione concentration was determined as described above. We found that less than 10% of the glutathione was incorporated into liposomes. Thus, the administered dose of liposomeassociated glutathione was closer to 2 mg/ml (0.25 mg/0.12 ml) rather than to the initial 25 mg/ml solution. In preliminary experiments, we tested the toxicity of intratracheal administration of the positive and negative liposomes. For a control group, mice received the same volume (0.12 ml) of PBS intratracheally. Five mice in each group were killed 2 days after the intratracheal instillations, their lungs were lavaged as described above, and the fluid was analyzed for protein concentration. The negatively charged liposomes caused an increase in protein concentration (0.30 ± 0.04 versus 0.12 ± 0.01 mg/ml; p
LEWIS J. SMITH, JAMES ANDERSON, MIR SHAMSUDDIN, and WEI HSUEH
Introduction
Exposure to high concentrations of oxygen produces acute lung injury in several animal species, including mice (1,2). We recently found that when mice are simultaneously fasted and exposed to 100070 oxygen, they develop lung damage sooner than when they are allowed free access to food (3). The mechanism responsible for this phenomenon is unclear. One possibility is a decrease in lung antioxidant levels. Frank and coworkers (4) reported decreases in glutathione peroxidase (GP) and glucose-e-phosphate dehydrogenase (G6PD) in intrauterine growth-retarded rat pups, but no change in catalase or superoxide dismutase (SOD) (4). In contrast, Deneke and colleagues (5) did not identify any changes in lung levels of G6PD, glutathione reductase, glutathione peroxidase, SOD, or glutathione in rats fed a proteindeficient diet. Other studies from these same two groups, however, have reported no change in GP and G6PD in undernourished rat pups (6) and decreases in G6PD and glutathione in protein-· deficient rats (7). Despite these somewhat contradictory data, there is reason to suspect that reduced levels of glutathione, a tripeptide (gamma-glutamyl-cysteinylglycine) that acts as an intracellular antioxidant by supplying reducing equivalents in the. form of electrons to reduce peroxides (8), . and enzymes of the glutathione redox cycle playa role in the increased susceptibility of fasted mice to hyperoxic lung damage (9). First, fasting (starvation) decreases hepatic glutathione levels in mice and increases susceptibility to oxidant damage (10). One might suspect that fasting also decreases lung glutathione levels. This could result from decreased availability of sulfur-containing amino acids (e.g., methionine and cysteine), which are used in glutathione synthesis and obtained mainly from dietary sources (11, 12), or from increased extrahepatic
SUMMARY Fasted mice exposed to 100% oxygen have more lung damage and die sooner than do fad mice. The mechanism responsible for this phenomenon has not been Identified. We performed the following experlmenta to test the hypothesis that reduced glutathione content In lung tluue of fasted mice contributes to the Increased susceptibility to hyperoxlc lung damage. Flret, alr-exposed mice were fasted for as long as 3 days. They had little change In lung levels of superoxIde dlsmutese (SOD) or cetalase, but they had a 41% decrease In glutathione by Dey 3 (p < 0.001). Second, fad mice and fasted mice were exposed to 100% oxygen for as long as 4 days. Both groups had nearly Identical values of lung SOD and cetalase, but the fasted mice had lower levels of glutathione (p < 0.001).Third, fad mice received the glutathione synthesis Inhibitor buthlonlne suifoxlmlne (BSO; 20 mM) In their drinking water for 2 wk and were then exposed to either air or 100% oxygen. Alr-exposed mice receiving SSO for 14 days had no change In lung SOD content, a 43% Increase In catalase (p < 0.001), and a 41% decrease In glutathione (p < 0.01). Oxygen-exposed, BSG-treated mice had no change In SOD and an Increase In catalase, but lower glutathione levela, more desths, and Increased lung damage on Day 3 (SAL protein: 1.72 ± 0.21 versus 0.94 ± 0.08 mg/ml; p < 0.01)than did diluent-treated mice. Fourth, fasted mice were given llposomes containing glutathione Intratracheally. When exposed to air, they had smaller decreases In lung glutathione levels than did fasted mice receiving either nothing or llposomes containing PBS. When exposed to 100% oxygen, they had higher lung glutathione levels on Days 1, 2, and 3, fewer deaths, and lower SAL protein values on Day 3 than did the untreated fasted mice. Only partial protection was achieved after glutathione administration because the llposomes localized to the left lung. These resulta support our hypothesis that the Increased susceptibility of fasted mice to hyperoxlc lung damage may be due to a reduced lung content of glutathione. AM REV RESPIR DIS 1990: 141:141-149
metabolism of glutathione (13).Second, to 100%oxygenand the effect of glutathifasting may interfere with converting the one supplementation in fasted mice simioxidized form of glutathione (GSSG) to larly exposed. Our findings suggest that its active form (GSH) by depleting glu- glutathione depletion plays an important cose stores and thereby slowing the hex- role in the increased susceptibility of fastose monophosphate (HMP) shunt and ed mice to hyperoxic lung damage. reducing the NADPH levels that are Methods needed for the conversionprocess (14-16). Third, some studies suggest that lung Experimental Animals and Lung Injury glutathione levels correlate with suscep- Male Balb-c mice (Charles River, Portage, tibility to hyperoxia (17-19), although MI), 8 to 10 wk of age and weighing 20 to other studies have questioned this rela- 25 g, were housed 10 mice to a cage, placed tionship (7, 20). The present study was designed to test the hypothesis that fasting decreases lung glutathione levels in mice and that this (Received in original form April 3, 1989 and in decrease contributes to the increased sus- revised form June 5, 1989) ceptibility of fasted mice to hyperoxic 1 From the Departments of Medicine and Pathollung damage. We measured lung levels ogy, Northwestern University, VA Lakeside Mediof glutathione, as wellas two antioxidant cal Center, and Children's Memorial Hospital, enzymes, catalase and SOD, during fast- Chicago, Illinois. 2 Supported by the VA Research Service. ing, oxygen exposure, and the combina3 Correspondence and requests for reprints should tion of the two. We also examined the be addressed to Lewis J. Smith, M.D., Pulmonary effect of chemically induced glutathione Section, Northwestern University Medical Center, depletion in normally fed mice exposed Wesley456, 250 East Superior, Chicago, IL 60611. 141
SMITH, ANDERSON, SHAMSUDDIN, AND HSUEH
142
in polyethylene "glove bags" (American Scientific Products, McGaw Park, IL), and exposed to either 100% oxygen (98 ± 1070) or air for I, 2, 3, or 4 days. Flow rates were set to maintain the carbon dioxide concentration below 0.5070. Oxygen and carbon dioxide concentrations were monitored with Beckman OM-ll and LB-2 analyzers (Beckman Instruments, Fullerton, CA). The temperature was maintained between 23.9° and 26.6° C, and the relative humidity between 40 and 50% . Mice were fed either a normal diet during the exposure (fed) or had all food removed at the start of the exposure (fasted). Water wasprovided to all mice. In some experiments, fed mice received 20 mM DL-buthionine-SRsulfoximine (BSO) (Sigma Chemical, St. Louis, MO), a selective inhibitor of gammaglutamylcysteine synthetase and therefore of glutathione synthesis (21), in their drinking water for the 14 days before and during the exposure. In other experiments, fasted mice received liposome-encapsulated glutathione intratracheally at the start of the exposure. At the end 0 f the exposure, mice were anesthetized with pentobarbital, 50 mg/kg, given intraperitoneally. The lungs were perfused via the right ventricle at 20 cm H 2 0 with cold phosphate-buffered saline (PBS) at pH 7.4 until they became white. Then they were removed from the chest, trimmed of extraneous tissue includ ing trachea and main bronchi, weighed, and homogenized (see below). In the experiments in which bronchoalveolar lavage was also performed, the lungs were perfused, the trachea was cannulated with a no. 18,1.25inch plastic intravenous catheter filled with PBS, and the lungs were lavaged four times with separate l-ml aliquots of cold PBS. The BAL effluent, > 90% of the instilled volume, was immediately placed on ice, and centrifuged at 4° C for 10 min at 500 x g. The lungs were then processed as described above in preparation for homogenization. Except for wet lung weight, which was higher in the lavaged lungs, the lavage procedure had no effect on the lung antioxidant measurements or on the lung protein concentration. An aliquot of the BAL supernatant was stored at 4° C and analyzed within 1 wk for protein concentration by the method of Lowry and coworkers (22) using bovine serum albumin as the standard. The BAL protein concentration was used as a measure of lung injury (2).
Lung Homogenization Lungs were homogenized for 60 s in cold, hypotonic buffer (5 roM potassium phosphate at pH 7.4) using a weight (mg) to volume (ml) ratio of25:1.The homogenate wascentrifuged at 500 x g for 10 min, the pellet was discarded, and the supernatant was centrifuged at 20,000 x g for 10min. Aliquots ofthe supernatant were saved for protein, SOD, catalase, and glutathione assays.
Antioxidant Assays Protein was quantitated by the method of Lowry and coworkers (22) as noted above.
SOD was measured by the method of McCord and Fridovich (23) in which cytochrome c reduction is measured spectrophotometrical1y at 550 nm . Catalase was quantitated spectrophotometrically at 240 nm as the decrease in H 202 over time. Total glutathione was measured by the method of Griffith (24) in which the rate of reduction of 5-5'-dinitrobis{-2-nitrobenzoic acid) (DfNB) is measured spectrophotometrically at 412 nm.
Liposome Preparation Lecithin, cholesterol, and either dicetyl phosphate (negative liposomes) or stearylamine (positive liposomes), obtained from Avanti Polar Lipids (Pelham, AL), were dissolved in chloroform at a molar ratio of 63:9:18 and placed in a round-bottom flask (25). The chloroform was evaporated in a rotary evaporator leaving a thin film. The film was dislodged with buffer solution (PBS or glutathione in PBS, 25 mg/ml) by gentle shaking. The resulting solution was sonicated for 2 min and then passed through a 30-gauge needle to disperse aggregates. This method produces mostly small vesicles approximately 50 to 100 nm in size (26). The liposome preparation was incubated at room temperature for 1h and then instilled intratracheally as described below. To determine glutathione incorporation into the liposome carrier, liposomes were prepared as above, washed twice with buffer, and centrifuged at 170,000 x g for 2 h after each wash. The supernatant was decanted and saved. The pellet was suspended in 0.2% Triton X-IOO to disrupt the liposomes, and the resulting solution was also saved. Glutathione concentration was determined as described above. We found that less than 10% of the glutathione was incorporated into liposomes. Thus, the administered dose of liposomeassociated glutathione was closer to 2 mg/ml (0.25 mg/0.12 ml) rather than to the initial 25 mg/ml solution. In preliminary experiments, we tested the toxicity of intratracheal administration of the positive and negative liposomes. For a control group, mice received the same volume (0.12 ml) of PBS intratracheally. Five mice in each group were killed 2 days after the intratracheal instillations, their lungs were lavaged as described above, and the fluid was analyzed for protein concentration. The negatively charged liposomes caused an increase in protein concentration (0.30 ± 0.04 versus 0.12 ± 0.01 mg/ml; p
