Permeability characteristics perfused rat lungs
of isolated
JANINA CZARTOLOMNA, NORBERT F. VOELKEL, AND SHIH-WEN CHANG Cardiovascular Pulmonary Research Laboratory and Webb- Waring Lung Institute, University of Colorado Health Sciences Center, Denver, Colorado 80262 CZARTOLOMNA,JANINA,NORBERT F. VOELKEL,AND SHIHCHANG. Permeability characteristics of isolated perfused rat Lungs. J. Appl. Physiol. 70(4): 1854-1860, 1991.-We examined the factors that influence the permeability characteristics of isolated perfused rat lungs and compared the ex vivo permeability-surface area product (PS) with that obtained in vivo. In lungs perfused for 20 min with homologous blood or a physiological salt solution (PSS) containing 4 g/100 ml albumin, mean PS values, obtained by the single-sample method of Kern et al. [Am. J. Physiol. 245 (Heart Circ. Physiol. 14): H229-H236, 19831, were 9.9 t 0.6 (SE) and 6.8 * 0.3 cm3 min-’ g wet 10m2, respectively. These values were similar to lung PS lung-’ obtained in intact rats (7.7 t 0.4 cm”. min-’ g wet lung-’ 10M2). In perfused lungs, PS values were influenced by the perfusate albumin concentration, the length of perfusion time, and the degree of vascular recruitment. Twenty minutes after lung isolation, PS was 126% higher in lungs perfused with albumin-free PSS containing Ficoll than in lungs perfused with albumin-PSS. Moreover, PS in Ficoll-PSS-perfused lungs increased even higher after 2 h of perfusion, and this time-dependent increase in PS was attenuated by addition of 0.1 g/100 ml albumin to the perfusate. Two hours of ex vivo ventilation with hypoxic (0 or 3% 0,) or hyperoxic (95% 0,) gas mixture did not affect PS values in perfused lungs. However, PS was elevated in lungs perfused ex vivo with protamine, which causes endothelial cell injury, or in lungs from rats exposed in vivo to human recombinant tumor necrosis factor. We conclude that lung isolation and perfusion per se do not markedly alter pulmonary vascular permeability and that PS measured in perfused lungs can be a sensitive index of vascular injury. WEN
l
l
l
l
l
permeability-surface area product; protamine sulfate; tumor necrosis factor; lung injury; hypoxia; hyperoxia; isolated rat lung
such as isolated perfused lungs and endothelial cell monolayers are used extensively in studies examining the pulmonary circulatory response to physiological and injurious stimuli. While much useful information has been gained from these studies, the in vivo relevance of observations made in these ex vivo settings has been questioned. In part, this skepticism has been due to the perception that during lung removal or in the process of endothelial cell isolation and culture, variables may be introduced that impair the physiological characteristics of the preparations. Indeed, Albelda et al. (1) have recently pointed out that cultured endothelial cell monolayers have baseline permeability values that are 10-100 times greater than those measured for endothelium in situ and that this finding is attributable to the
EX VIVO PREPARATIONS
1854
0161-7567/91
$1.50 Copyright
presence of small gaps between cultured endothelial cells. In regard to the question whether isolated perfused lungs are per se leaky, conflicting reports exist in the literature. Spontaneous edema has been observed in isolated perfused lungs from rabbits, dogs, and sheep (9,16, 18, 27), and this may depend in part on the perfusate characteristics (16, 27) or activation of pulmonary eicosanoids (18). However, Allison et al. (2) have noted the stability of isolated canine left lower lobe during 4 h of artificial perfusion, and we have not observed spontaneous edema in isolated rat lungs perfused with blood or a variety of physiological salt solutions (PSS) (5,14,20). In dogs, Morriss et al. (15) reported a threefold higher capillary filtration coefficient (K,,) in isolated compared with intact lungs. On the other hand, Wagensteen et al. (29) and Kern et al. (12) reported comparable values for permeability-surface area product (PS) in isolated and intact rabbit lungs. Thus whether lung isolation and pump perfusion invariably result in microvascular damage and increased vascular permeability is unclear. The purpose of this study is to identify factors that affect PS measurements in the isolated perfused rat lung preparation and to compare lung PS in intact rats with that obtained in ex vivo perfused lungs. We measured lung PS using the “single-sample” method described by Kern and Malik (12) and Kern et al. (13) and investigated the effect of perfusate albumin concentration, perfusion time, vascular recruitment, and 0, tension. In addition, we measured PS in lungs injured ex vivo with protamine sulfate or in vivo with recombinant human tumor necrosis factor (TNF). Finally, we compared the PS values in lungs perfused with homologous blood or PSS containing albumin with those in intact rats. METHODS
Male Sprague-Dawley rats weighing 275-345 g were purchased from Harlan (Indianapolis, IN) and given free access to food and water. Rats were anesthetized with pentobarbital sodium (60 mg/kg body wt ip) before the experiments. IJnless otherwise stated, all reagents were purchased from Sigma Chemical (St. Louis, MO). Isolated lungs. The isolated lung preparation was performed according to the procedures previously described by McMurtry et al. (14). Briefly, after cannulation of the trachea, main pulmonary artery, and left ventricle, lungs were removed from the chest and suspended in a humid
0 1991 the American Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (128.205.114.091) on December 1, 2018. Copyright © 1991 the American Physiological Society. All rights reserved.
PERMEABILITY
OF
PERFUSED
chamber. The lungs were ventilated with a humid mixture of 95% air-5% CO, at 55 breaths/min with a peak inspiratory pressure of 7 cmH,O and a positive end-expiratory pressure of 2.5 cmH,O. Lungs were perfused at constant flow (0.03 ml min-l g body wt-l) with either homologous blood (removed from 3 ether-anesthetized donor rats) or PSS by using a Holter peristaltic pump. The maximal time period between the interruption of pulmonary blood flow and the reestablishment of artificial perfusion was 5 min, during which time the lungs were continually ventilated. The composition of PSS was (in mM) 119 NaCl, 4.7 KCl, 1.17 MgSO,, 22.6 NaHCO,, 1.18 KH,PO,, 3.2 CaCl,, and 5.5 glucose. Either bovine serum albumin or Ficoll (Sigma Chemical, 4 g/100 ml of PSS) was used to maintain adequate oncotic pressure. When Ficoll was used, the PSS was modified to contain 19.0 mM NaHCO, and 1.6mM CaCl,. Mean pulmonary arterial pressure (Ppa) was measured with a Statham transducer and continuously recorded on a Soltec recorder. In all lungs perfused with PSS, the first 50 ml of perfusate effluent from the left ventricle were discarded. The temperature of the recirculating perfusate was maintained at 38OC. Lung weight was continuously measured and recorded in some experiments using a Statham Gl-15-300 force displacement transducer. Measurement of PS in perfused lungs. Lung PS was measured using a modification of the single sample technique described by Kern and Malik (12) and Kern et al. (13). To ensure low free iodine content, 1251-labeled albumin was separated from free 12!jI daily using a Sephadex minicolumn as described by Tuszynski et al. (26). In general, the percentage of 12!jI bound to albumin exceeded 99.9%. At 20 min, or in longer experiments at 60 and 120 min of perfusion, -1 &i of 1251-labeled bovine albumin (ICN Radiochemicals, Irvine, CA) was added to the perfusate reservoir. Exactly 3 min later, the intravascular space was perfused free of radioactivity for an additional 3 min by using a fresh nonradioactive solution that was not recirculated. Blood-perfused lungs were washed with PSS containing 4 g/100 ml albumin. In preliminary studies, we determined that >97-98% of the intravascular radioactivity is removed by this washout protocol. Perfusate samples (1 ml) before and after vascular washout were collected. Lungs were then dissected free, lightly blotted, and weighed. Radioactivity of perfusate samples and lungs was measured in a Packard auto gamma scintillation spectrometer (Downers Grove, IL). Lungs were left in a transit oven (Blue M Electric) at 60°C until constant weight (-3 wk) and the wet weight-to-dry weight ratios were calculated. Lung PS was calculated using the equation l
l
PS = 1251-activity in lung tissue/ (1251-activity
in 1.0 g of perfusate
X
3 min)
Effect of perfusate composition. To investigate whether perfusate composition affects lung PS, 24 perfused lungs were divided into four groups (n = 6) and perfused with one of the four solutions: 1) homologous rat blood, 2) PSS containing 4 g /lOO ml albumin, 3) PSS containing 4 g/100 ml Ficoll, and 4) PSS containing 0.1 g albumin +
RAT
LUNGS
1855
3.9 g/100 ml Ficoll. Lungs were equilibrated for 20 min before the PS measurements. Effect of floul and venous pressure. To assess the influence of increased flow and venous pressure on PS, the venous outflow pressure (Ppv) was elevated to 7.5 cmH,O in six lungs perfused with PSS-albumin. In another group of seven lungs, the flow was doubled (to 0.06 ml min-l g body wt-‘) and outflow pressure was elevated to 7.5 cmH,O. These adjustments were made 10 min before adding the 1251-albumin to the reservoir. Effect of albumin concentration and perfusion time. To determine the effect of perfusate albumin concentration on macromolecular permeability during prolonged perfusion, two groups of lungs (n = 6 each) were perfused for 1 h with PSS containing either 4 g/l00 ml albumin or 4 g/100 ml Ficoll, and three groups of lungs (n = 6 each) were perfused for 2 h with PSS containing 4 g/l00 ml albumin, 4 g/100 ml Ficoll, or 0.1 g/100 ml albumin with 3.9 g/100 ml Ficoll. Lung PS was measured and compared with the values obtained in lungs perfused for only 20 min. Effect of 0, tension. To assess the influence of either low or high 0, tensions on PS in isolated lungs, four groups of lungs (n = 5 each) were perfused with PSS 4 g/100 ml albumin and ventilated for 2 h with gas mixtures containing 0,3,21, or 95% O,, 5% CO,, and balance N, before PS measurement. To ensure full vascular recruitment, we elevated Ppv to +5 cmH,O for 10 min before measurement of PS. Effect of injury. To assess the effect of cationic protein-induced endothelial cell injury on lung PS, 18 lungs were perfused with PSS-albumin and divided into three groups (n = 6 each). After 10 min of equilibration, protamine sulfate (500 pg/ml), saline (vehicle control, 0.3 ml), or protamine + heparin (20 U/ml) was added to the perfusate reservoir, and lung PS was determined 15 min later. Heparin was used as a polyanionic compound to neutralize the charge-mediated effect of protamine (4). To see whether the PS measurement in perfused lungs can be used in the assessment of lung vascular permeability in rats exposed in vivo to injurious agents, we measured lung PS in 16 rats injected intravenously with either purified recombinant tumor necrosis factor (TNF, 1 mg/kg, n = 6), saline vehicle (n = 6), or TNF + monoclonal antibody against human TNF (anti-TNF, 10 mg/kg injected iv 1 h before TNF injection, n = 4). TNF and anti-TNF were kindly provided by Dr. George Kuo of Chiron, Emeryville, CA. The TNF preparation was >98% pure, contained
of isolated
JANINA CZARTOLOMNA, NORBERT F. VOELKEL, AND SHIH-WEN CHANG Cardiovascular Pulmonary Research Laboratory and Webb- Waring Lung Institute, University of Colorado Health Sciences Center, Denver, Colorado 80262 CZARTOLOMNA,JANINA,NORBERT F. VOELKEL,AND SHIHCHANG. Permeability characteristics of isolated perfused rat Lungs. J. Appl. Physiol. 70(4): 1854-1860, 1991.-We examined the factors that influence the permeability characteristics of isolated perfused rat lungs and compared the ex vivo permeability-surface area product (PS) with that obtained in vivo. In lungs perfused for 20 min with homologous blood or a physiological salt solution (PSS) containing 4 g/100 ml albumin, mean PS values, obtained by the single-sample method of Kern et al. [Am. J. Physiol. 245 (Heart Circ. Physiol. 14): H229-H236, 19831, were 9.9 t 0.6 (SE) and 6.8 * 0.3 cm3 min-’ g wet 10m2, respectively. These values were similar to lung PS lung-’ obtained in intact rats (7.7 t 0.4 cm”. min-’ g wet lung-’ 10M2). In perfused lungs, PS values were influenced by the perfusate albumin concentration, the length of perfusion time, and the degree of vascular recruitment. Twenty minutes after lung isolation, PS was 126% higher in lungs perfused with albumin-free PSS containing Ficoll than in lungs perfused with albumin-PSS. Moreover, PS in Ficoll-PSS-perfused lungs increased even higher after 2 h of perfusion, and this time-dependent increase in PS was attenuated by addition of 0.1 g/100 ml albumin to the perfusate. Two hours of ex vivo ventilation with hypoxic (0 or 3% 0,) or hyperoxic (95% 0,) gas mixture did not affect PS values in perfused lungs. However, PS was elevated in lungs perfused ex vivo with protamine, which causes endothelial cell injury, or in lungs from rats exposed in vivo to human recombinant tumor necrosis factor. We conclude that lung isolation and perfusion per se do not markedly alter pulmonary vascular permeability and that PS measured in perfused lungs can be a sensitive index of vascular injury. WEN
l
l
l
l
l
permeability-surface area product; protamine sulfate; tumor necrosis factor; lung injury; hypoxia; hyperoxia; isolated rat lung
such as isolated perfused lungs and endothelial cell monolayers are used extensively in studies examining the pulmonary circulatory response to physiological and injurious stimuli. While much useful information has been gained from these studies, the in vivo relevance of observations made in these ex vivo settings has been questioned. In part, this skepticism has been due to the perception that during lung removal or in the process of endothelial cell isolation and culture, variables may be introduced that impair the physiological characteristics of the preparations. Indeed, Albelda et al. (1) have recently pointed out that cultured endothelial cell monolayers have baseline permeability values that are 10-100 times greater than those measured for endothelium in situ and that this finding is attributable to the
EX VIVO PREPARATIONS
1854
0161-7567/91
$1.50 Copyright
presence of small gaps between cultured endothelial cells. In regard to the question whether isolated perfused lungs are per se leaky, conflicting reports exist in the literature. Spontaneous edema has been observed in isolated perfused lungs from rabbits, dogs, and sheep (9,16, 18, 27), and this may depend in part on the perfusate characteristics (16, 27) or activation of pulmonary eicosanoids (18). However, Allison et al. (2) have noted the stability of isolated canine left lower lobe during 4 h of artificial perfusion, and we have not observed spontaneous edema in isolated rat lungs perfused with blood or a variety of physiological salt solutions (PSS) (5,14,20). In dogs, Morriss et al. (15) reported a threefold higher capillary filtration coefficient (K,,) in isolated compared with intact lungs. On the other hand, Wagensteen et al. (29) and Kern et al. (12) reported comparable values for permeability-surface area product (PS) in isolated and intact rabbit lungs. Thus whether lung isolation and pump perfusion invariably result in microvascular damage and increased vascular permeability is unclear. The purpose of this study is to identify factors that affect PS measurements in the isolated perfused rat lung preparation and to compare lung PS in intact rats with that obtained in ex vivo perfused lungs. We measured lung PS using the “single-sample” method described by Kern and Malik (12) and Kern et al. (13) and investigated the effect of perfusate albumin concentration, perfusion time, vascular recruitment, and 0, tension. In addition, we measured PS in lungs injured ex vivo with protamine sulfate or in vivo with recombinant human tumor necrosis factor (TNF). Finally, we compared the PS values in lungs perfused with homologous blood or PSS containing albumin with those in intact rats. METHODS
Male Sprague-Dawley rats weighing 275-345 g were purchased from Harlan (Indianapolis, IN) and given free access to food and water. Rats were anesthetized with pentobarbital sodium (60 mg/kg body wt ip) before the experiments. IJnless otherwise stated, all reagents were purchased from Sigma Chemical (St. Louis, MO). Isolated lungs. The isolated lung preparation was performed according to the procedures previously described by McMurtry et al. (14). Briefly, after cannulation of the trachea, main pulmonary artery, and left ventricle, lungs were removed from the chest and suspended in a humid
0 1991 the American Physiological
Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (128.205.114.091) on December 1, 2018. Copyright © 1991 the American Physiological Society. All rights reserved.
PERMEABILITY
OF
PERFUSED
chamber. The lungs were ventilated with a humid mixture of 95% air-5% CO, at 55 breaths/min with a peak inspiratory pressure of 7 cmH,O and a positive end-expiratory pressure of 2.5 cmH,O. Lungs were perfused at constant flow (0.03 ml min-l g body wt-l) with either homologous blood (removed from 3 ether-anesthetized donor rats) or PSS by using a Holter peristaltic pump. The maximal time period between the interruption of pulmonary blood flow and the reestablishment of artificial perfusion was 5 min, during which time the lungs were continually ventilated. The composition of PSS was (in mM) 119 NaCl, 4.7 KCl, 1.17 MgSO,, 22.6 NaHCO,, 1.18 KH,PO,, 3.2 CaCl,, and 5.5 glucose. Either bovine serum albumin or Ficoll (Sigma Chemical, 4 g/100 ml of PSS) was used to maintain adequate oncotic pressure. When Ficoll was used, the PSS was modified to contain 19.0 mM NaHCO, and 1.6mM CaCl,. Mean pulmonary arterial pressure (Ppa) was measured with a Statham transducer and continuously recorded on a Soltec recorder. In all lungs perfused with PSS, the first 50 ml of perfusate effluent from the left ventricle were discarded. The temperature of the recirculating perfusate was maintained at 38OC. Lung weight was continuously measured and recorded in some experiments using a Statham Gl-15-300 force displacement transducer. Measurement of PS in perfused lungs. Lung PS was measured using a modification of the single sample technique described by Kern and Malik (12) and Kern et al. (13). To ensure low free iodine content, 1251-labeled albumin was separated from free 12!jI daily using a Sephadex minicolumn as described by Tuszynski et al. (26). In general, the percentage of 12!jI bound to albumin exceeded 99.9%. At 20 min, or in longer experiments at 60 and 120 min of perfusion, -1 &i of 1251-labeled bovine albumin (ICN Radiochemicals, Irvine, CA) was added to the perfusate reservoir. Exactly 3 min later, the intravascular space was perfused free of radioactivity for an additional 3 min by using a fresh nonradioactive solution that was not recirculated. Blood-perfused lungs were washed with PSS containing 4 g/100 ml albumin. In preliminary studies, we determined that >97-98% of the intravascular radioactivity is removed by this washout protocol. Perfusate samples (1 ml) before and after vascular washout were collected. Lungs were then dissected free, lightly blotted, and weighed. Radioactivity of perfusate samples and lungs was measured in a Packard auto gamma scintillation spectrometer (Downers Grove, IL). Lungs were left in a transit oven (Blue M Electric) at 60°C until constant weight (-3 wk) and the wet weight-to-dry weight ratios were calculated. Lung PS was calculated using the equation l
l
PS = 1251-activity in lung tissue/ (1251-activity
in 1.0 g of perfusate
X
3 min)
Effect of perfusate composition. To investigate whether perfusate composition affects lung PS, 24 perfused lungs were divided into four groups (n = 6) and perfused with one of the four solutions: 1) homologous rat blood, 2) PSS containing 4 g /lOO ml albumin, 3) PSS containing 4 g/100 ml Ficoll, and 4) PSS containing 0.1 g albumin +
RAT
LUNGS
1855
3.9 g/100 ml Ficoll. Lungs were equilibrated for 20 min before the PS measurements. Effect of floul and venous pressure. To assess the influence of increased flow and venous pressure on PS, the venous outflow pressure (Ppv) was elevated to 7.5 cmH,O in six lungs perfused with PSS-albumin. In another group of seven lungs, the flow was doubled (to 0.06 ml min-l g body wt-‘) and outflow pressure was elevated to 7.5 cmH,O. These adjustments were made 10 min before adding the 1251-albumin to the reservoir. Effect of albumin concentration and perfusion time. To determine the effect of perfusate albumin concentration on macromolecular permeability during prolonged perfusion, two groups of lungs (n = 6 each) were perfused for 1 h with PSS containing either 4 g/l00 ml albumin or 4 g/100 ml Ficoll, and three groups of lungs (n = 6 each) were perfused for 2 h with PSS containing 4 g/l00 ml albumin, 4 g/100 ml Ficoll, or 0.1 g/100 ml albumin with 3.9 g/100 ml Ficoll. Lung PS was measured and compared with the values obtained in lungs perfused for only 20 min. Effect of 0, tension. To assess the influence of either low or high 0, tensions on PS in isolated lungs, four groups of lungs (n = 5 each) were perfused with PSS 4 g/100 ml albumin and ventilated for 2 h with gas mixtures containing 0,3,21, or 95% O,, 5% CO,, and balance N, before PS measurement. To ensure full vascular recruitment, we elevated Ppv to +5 cmH,O for 10 min before measurement of PS. Effect of injury. To assess the effect of cationic protein-induced endothelial cell injury on lung PS, 18 lungs were perfused with PSS-albumin and divided into three groups (n = 6 each). After 10 min of equilibration, protamine sulfate (500 pg/ml), saline (vehicle control, 0.3 ml), or protamine + heparin (20 U/ml) was added to the perfusate reservoir, and lung PS was determined 15 min later. Heparin was used as a polyanionic compound to neutralize the charge-mediated effect of protamine (4). To see whether the PS measurement in perfused lungs can be used in the assessment of lung vascular permeability in rats exposed in vivo to injurious agents, we measured lung PS in 16 rats injected intravenously with either purified recombinant tumor necrosis factor (TNF, 1 mg/kg, n = 6), saline vehicle (n = 6), or TNF + monoclonal antibody against human TNF (anti-TNF, 10 mg/kg injected iv 1 h before TNF injection, n = 4). TNF and anti-TNF were kindly provided by Dr. George Kuo of Chiron, Emeryville, CA. The TNF preparation was >98% pure, contained
