INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recent reports from our laboratory and from other groups have documented that lung, as are other organs, is susceptible to ischemia/reperfusion (I/R) injury. Thus, in a model of ventilated isolated rat lungs, I/R led to increased oxidant generation in the pulmonary tissue and to oxidative pulmonary injury (1). Oxidative injury to the pulmonary tissue induced by I/R has been supposed to contribute to complications of lung transplantation (6, 7).

In the present study we specifically addressed injury to the pulmonary endothelium in lung I/R. Endothelial cells themselves are capable of generating oxidants (8, 9). Our recent findings indicate that lung ischemia induces depolarization of the endothelial plasma membrane in the pulmonary vasculature, which in turn leads to generation of oxidants (10). Therefore, endothelium may play a role in the initiation of pulmonary oxidative injury induced by I/R, and endothelial cells could be either a source or a target of oxidants. Oxidative injury to endothelium has been found to set the stage for secondary pulmonary injury by leukocytes (5, 11).

Previously, endothelial cell injury induced by lung I/R was specifically addressed by Grosso and coworkers, who demonstrated that lung I/R leads to reduction of endothelium-specific antigen, Factor VIII, in lung tissue (12). In their study, however, isolated lungs were not ventilated during the ischemic period, and the endothelial injury could have represented a manifestation of pulmonary injury caused by tissue anoxia. Anoxia leads to tissue accumulation of xanthine (due to degradation of adenosine triphosphate [ATP]) and xanthine dehydrogenase (due to conversion of xanthine dehydrogenase), which in turn leads to oxidant generation after restoration of the oxygen supply (13). However, oxygen tension in the lung tissue is normally maintained by ventilation rather than by perfusion. Thus, the pulmonary ATP level during I/R in perfused rat lungs continuously ventilated with air did not differ significantly from that in control perfused lungs (4). Moreover, oxidants and products of tissue oxidation accumulate in the lung during the ischemic period before the restoration of blood flow (4). Therefore, mechanism(s) of initiation and pathway(s) for the development of oxidative injury induced by I/R in ventilated lung (a model used in our laboratory) seem to differ from those in anoxic lung.

In the present study, we utilized isolated rat lungs continuously ventilated with 20% oxygen for 2 h, and compared normal perfusion with an I/R protocol. In order to specifically characterize endothelial injury, we assessed angiotensin-converting enzyme (ACE) activity in the perfusate and in the lung tissue, as well as ACE immunoreactivity in the lung. Our results show that normoxic lung I/R leads to endothelial injury manifested by: (1) an increased ACE level in the perfusate; (2) a reduction of ACE content in the lung tissue; and (3) an increase in pulmonary vascular permeability.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents

Iodogen was obtained from Pierce (Rockford, IL), Na¹²⁵I from Amersham (Arlington Heights, IL), fatty acid-free bovine serum albumin (BSA) from Boehringer-Mannheim (Indianapolis, IN), Triton X-114 and mouse IgG from Sigma (St. Louis, MO), and O-phthalaldehyde and Z-Phe-His-Leu from Serva (Heidelberg, Germany). mAb 9B9, a mouse monoclonal anti-ACE antibody (IgG₁ class), has been characterized previously (14).

Iodination of Anti-ACE Monoclonal Antibody and Nonspecific IgG

mAb 9B9 and control mouse IgG were labeled with ¹²⁵I using Iodogen-coated tubes according to the manufacturer's recommendation. Iodogen-coated tubes were prepared by evaporation of a chloroform solution of Iodogen with nitrogen gas. Briefly, 100 µg of mouse monoclonal antibody (mAb) 9B9 or IgG were incubated with 100 µCi of Na¹²⁵I in 200 µl of 50 mM borate-buffered saline, pH 8.1, for 10 min. Excess isotope was removed through Sephadex G-25 gel filtration. Specific radioactivity for both preparations was 0.3 to 1.0 × 10⁶ cpm/µg. More than 98% of the radioactivity was TCA-precipitable.

The Isolated Lung Preparation

Male Sprague-Dawley rats weighing 170 to 200 g were anesthetized with sodium pentobarbital, 50 mg/kg intraperitoneally, and prepared for isolated lung perfusion with a recirculating perfusate as previously described (4). The trachea was cannulated and the lungs were ventilated with a humidified gas mixture (Airco Inc., Philadelphia, PA) containing 5% CO₂ and 95% air. Ventilation was done with an SAR-830 rodent ventilator (CWE Inc., PA) at 60 cycles/min, a tidal volume (VT) of 2 ml, and 2 cm H₂O end-expiratory pressure. The thorax was then opened and a cannula was placed in the main pulmonary artery through the transected heart. The lungs were isolated from the thorax and transferred to the water-jacketed perfusion chamber, which was maintained at 37° C. There was no interruption of ventilation during this transfer process, and interruption of lung perfusion was for < 5 s. Perfusion through the pulmonary artery was maintained by a peristaltic pump at a constant flow rate of 10 ml/min. The perfusate (45 ml) was Krebs-Ringer buffer (KRB) (pH 7.4) containing 10 mM glucose and 3% fatty acid-free BSA KRB-BSA solution. The perfusate was filtered through a 0.4 µm filter prior to perfusion, in order to eliminate particulates. In a separate series of experiments, hydrogen peroxide was added to the perfusate (final H₂O₂ concentration: 5 mM), and lungs were perfused for 1 h under the described conditions in order to assess the effect of an extracellular oxidant on ACE activity in the perfusate. Intratracheal and pulmonary arterial pressures were continuously recorded throughout the experiment with pressure transducers (PM 131TC and P23DC, Statham Instruments, Oxnard, CA) and direct-writing oscillographs (Gould, Cleveland, OH) with two-pen AC recorders (Primeline, Sun Valley, CA).

Perfusion of Radiolabeled Proteins

Control isolated lungs were perfused for 2 h under standard conditions after an initial 5-min equilibration period. In the I/R group, lungs were perfused for an initial 5-min equilibration period, after which perfusion was discontinued while ventilation continued. After an ischemic period of 1 h, reperfusion was restarted at the same flow rate and continued for another 1 h. Aliquots of perfusate (0.5 ml) were collected after every 15 min of perfusion and frozen in liquid nitrogen for further study of ACE activity. Aliquots were not collected from I/R lungs during the ischemia period, since there was no circulation of the perfusate through the lung. One microgram of radiolabeled nonspecific mouse IgG or mouse mAb 9B9 cross-reacting with rat ACE was added to the perfusate after the first hour of perfusion in the control group, or at the start of reperfusion in the I/R group. After 1 h of perfusion with perfusate containing radiolabeled protein, the lungs were perfused with a single-pass perfusion for 5 min with perfusate that did not contain radiolabeled protein, in order to eliminate nonbound radioactivity. In a previous study, we documented that this procedure provides effective elimination of nonbound radiolabeled material and reduces radioactivity of the lung tissue to the background level (15). At the end of experiment, the lungs were removed from the chamber, rinsed with saline, and blotted with filter paper, and the lung wet weight was determined. The left lobe was removed, its wet weight was determined, and its radioactivity was measured in a gammacounter and expressed as a percentage of perfused radioactivity per lung. Lungs minus left lobes were frozen in liquid nitrogen.

Biochemical Examination of Tissue Homogenates

Lungs and perfusates were stored at 70° C before assay of ACE activity. ACE activity in the lungs and perfusates was measured through the rate of generation of His-Leu formed from the ACE substrate Z-Phe-His-Leu, using a fluorometric assay (16). Frozen lung was thawed and homogenized in a Polytron and a Potter-Elvehjem homogenizer in Tris-HCl buffer, pH 8.3, at 4° C, at a ratio of 1:10 (wt/ vol). Ten microliters of the lung homogenate or perfusate were added to 200 µl of 50 mM Tris-HCl, 0.15 M NaCl, pH 7.4, buffer containing 0.5 mM substrate. The sample of lung homogenate was diluted 1/100 (vol/vol) in the same buffer because this dilution gave optimal fluorescence while maintaining the initial reaction-rate conditions (16). Samples of lung homogenate and perfusate were incubated at 37° C for 60 min and 120 min, respectively, after which the reaction was terminated by the addition of 1.5 ml of 0.28 N NaOH. O-phthalaldehyde (1 mg in 100 µl methanol) was added for 10 min before stopping this reaction with 200 µl 2 N HCl. His-Leu was measured with a fluorescence spectrophotometer at an excitation wavelength of 363 nm and an emission wavelength of 500 nm. Results were calculated as milliunits (mU) of ACE activity. Protein concentration was determined with a protein assay kit (Bio-Rad, Hercules, CA), using IgG as a standard.

Separation of the perfusate into aqueous and detergent phases was performed according to the procedure for separating integral and surface-membrane proteins developed by Bordier (17). Briefly, aliquots of the perfusates were prepared in 200 µl of 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 1% Triton X-114 at 0° C. The clear sample was overlaid on 300 µl of a 6% (wt/vol) sucrose cushion in 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 0.06% Triton X-114; incubated for 5 min at 30° C for condensation; and centrifuged for 5 min at 300 × g in a swinging-bucket rotor. Supernatant was collected and fresh, ice-cold Triton X-114 was added to the supernatant to a final concentration of 0.5%. After dissolution of the mixture at 0° C the sample was again overlaid on the sucrose cushion used previously, incubated for 5 min at 30° C, and centrifuged on the previous detergent phase. At the end of the separation procedure, ACE activity was assessed in the aqueous (upper layer) and the detergent phases as described earlier.

Statistical Analysis

Data are expressed as mean ± SE. Statistical comparisons of the I/R and control group were done through analysis of variance (ANOVA) for two groups, with repeated measurements in the control and I/R groups. Statistically significant differences in all cases were defined at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Effect of I/R on ACE Activity in the Perfusate and Lung Tissue of Ventilated Perfused Lung

Figure 1 shows the kinetics of appearance of ACE activity in aliquots of perfusate obtained from the control and I/R groups. Normal perfusion of isolated rat lungs with standard KRB- BSA solution caused a gradual time-dependent increase in ACE activity in the perfusate, from the initial level of 15 ± 1 mU to 108 ± 8 mU after 1 h of perfusion, and to 145 ± 14 mU after 2 h of perfusion (all results of ACE activity determination in the perfusates from all groups are presented as ACE mU in total perfusate, mean ± SE). In the I/R group, ACE activity in the perfusate at the end of the 1 h ischemic period was 56 ± 2 mU. This value was less than the ACE activity after 1 h of normal perfusion, reflecting decreased ACE exchange between the lung tissue and perfusate in the absence of circulation, but was significantly higher than the zero-time value. Importantly, within 15 min after the start of reperfusion, levels of ACE activity in the perfusate obtained from the I/R group were detectably higher than those in the control group (Figure 1). After 1 h of ischemia followed by 1 h of reperfusion, ACE activity in the perfusate was 252 ± 1 mU, versus 145 ± 14 mU in perfusate of lungs continuously perfused during the 2-h period. Therefore, at this time point, I/R led to a 174% increase in ACE activity in the perfusate as compared with control (the difference between the I/R and control groups was statistically significant at p < 0.05).

View larger version (19K): [in this window] [in a new window]   Figure 1.   Kinetics of the appearance of ACE activity in perfusate during control perfusion and I/R. ACE activity was measured in the perfusate after control perfusion (closed circles) and after I/R (closed squares). Zero time indicates the end of 5-min equilibration period. Samples were obtained at indicated times after the start of the experiment. ACE activity was not measured in the perfusate of the I/R group during the 60-min ischemia period, since perfusate was not circulating through the lung tissue. The data are presented as mU of ACE activity in total perfusate. (45 ml), mean ± SE for n = 6.

Increased ACE activity in the perfusate of the I/R group could reflect either release of ACE from lung tissue or desquamation of ACE-containing endothelial cells and/or plasma-membrane vesicles from the pulmonary vasculature. To discriminate between these two possibilities, samples of perfusate were filtered through a 0.2-µm filter. Table 1 shows that filtration did not reduce significantly ACE activity in the perfusates from either the control or I/R groups, indicating that the circulating ACE activity was not particulate. In contrast, filtration led to a marked reduction of ACE activity in the perfusate obtained from lungs perfused with 5 mM H₂O₂. Therefore, in contrast to the effect of H₂O₂, I/R does not induce disruption or desquamation of endothelial cells to the pulmonary perfusate.

The physiologic mechanism for ACE shedding from the plasma membrane requires specific proteolytic cleavage of the extracellular domain of ACE by an extracellular protease, or secretase, associated with the plasma membrane (18). This process leads to appearance of the hydrophilic form of ACE in the extracellular medium. The complete ACE molecule contains a hydrophobic transmembrane domain that anchors ACE in the plasma membrane (19). Both complete ACE and membrane-associated ACE are hydrophobic. We assessed the distribution of ACE activity between the aqueous and detergent phases of the perfusates. Figure 2 shows that in both control and I/R perfusates, about 80% of ACE activity was associated with the aqueous phase. In contrast, only 30% of ACE activity was associated with the aqueous phase in the perfusate obtained from lungs perfused with 5 mM H₂O₂. Therefore, the time-dependent increase in ACE activity in the perfusates of the control and I/R groups seems to result from shedding of ACE, whereas the increase in ACE in the perfusate after perfusion with H₂O₂ seems to result from cellular desquamation and/ or release of the complete ACE molecule from endothelium.

View larger version (51K): [in this window] [in a new window]   Figure 2.   Distribution of perfusate ACE activity between aqueous and detergent phases. Perfusates were obtained after 2 h of control perfusion (control), 1 h of ischemia followed by 1 h of reperfusion (I/R), or 1 h of perfusion with 5 mM H2O2 (H2O2). Phase separation of the perfusates was done by centrifugation in preconditioned Triton X-114, as described in METHODS. ACE activity in the aqueous phase (closed bars) or detergent phase (open bars) is shown as a percent of total ACE activity found in both phases. The data are shown as mean ± SE for n = 3.

ACE activity determined in the lung tissue homogenates obtained after control perfusion and after I/R was 0.38 ± 0.07 versus 0.29 ± 0.05 U/mg of protein. Total ACE activity in the lung tissue (Figure 3A) was 29.7 ± 2.3 U/lung for the control group, versus 22.7 ± 1.7 U/lung for the I/R group (n = 5, p < 0.05). Therefore, I/R induced a marked reduction of ACE activity in the lung-tissue homogenate.

View larger version (28K): [in this window] [in a new window]   Figure 3.   ACE activity in the perfusate and lung homogenate. Isolated ventilated rat lungs were perfused for 2 h (open bars) or subjected to 1 h of ischemia followed by 1 h of reperfusion (closed bars). At the end of perfusion, ACE activity in lung homogenates (A) and perfusates (B) was measured as described in METHODS. The data are presented as units of ACE activity in the lung homogenate (A) and as units of ACE activity in total perfusate (B) (mean ± SE for n = 5; *p < 0.05).

Effect of I/R on Interaction of Pulmonary ACE with Radiolabeled Anti-ACE mAb 9B9

We also studied the availability of ACE localized on the luminal surface of the pulmonary vascular endothelium for binding of circulating monoclonal anti-ACE antibody. Radiolabeled anti-ACE mAb 9B9 was added to the perfusate after 1 h of control perfusion or after the start of reperfusion. Figure 4 shows that 1 h of perfusion with radiolabeled mAb 9B9 after 1 h of control perfusion led to accumulation of 32.1 ± 1.7% of the perfused radioactivity (n = 4). One hour of perfusion with the same preparation of mAb 9B9 after 1 h of ischemia led to accumulation of 24.8 ± 0.9% of the perfused radioactivity (n = 3, p < 0.05). Therefore, I/R leads to reduction of the pulmonary uptake of circulating radiolabeled anti-ACE mAb 9B9 to 77% of the control level. In a previous study we demonstrated that addition of nonlabeled anti-ACE mAb 9B9 to the perfusate inhibits pulmonary uptake of radiolabeled mAb 9B9, indicating the specificity of interaction of radiolabeled anti-ACE mAb 9B9 with ACE localized on the luminal surface of the pulmonary vascular endothelium (15).

View larger version (31K): [in this window] [in a new window]   Figure 4.   Pulmonary uptake of radiolabeled IgG and radiolabeled mAb 9B9 in isolated perfused rat lungs. Control perfusion: open bars, I/R: closed bars. The data are presented as percentages of perfused radioactivity in the lung tissue, mean ± SE for n = 4; *p < 0.01.

Pulmonary uptake of nonspecific mouse [¹²⁵I]IgG in the ventilated perfused rat lungs was less than 1% of perfused radioactivity in both the control and I/R groups (Figure 4). In contrast to its effect on the pulmonary uptake of anti-ACE mAb 9B9, I/R led to a significant increase in the pulmonary uptake of circulating nonspecific IgG (see the subsequent discussion). This result provides additional evidence for the specificity of the I/R-induced reduction of pulmonary uptake of anti-ACE mAb.

Effect of I/R on Vascular Permeability

Pulmonary uptake of radiolabeled albumin (20) or transferrin (2) is a useful parameter for quantitative characterization of pulmonary vascular permeability. Although radiolabeled nonspecific IgG is a larger protein than albumin, we have previously shown that its pulmonary uptake is quantitatively similar to that of radiolabeled BSA in both normal animals and in animals with acute lung injury (ALI) (21). Therefore, pulmonary accumulation of radiolabeled nonspecific IgG added to the perfusate can be used to characterize the barrier function of the pulmonary vasculature. In the present study, pulmonary uptake of [¹²⁵I]IgG was 0.67 ± 0.04% of perfused radioactivity in the I/R group, versus 0.17 ± 0.02% in the control group (n = 4, p < 0.001). Therefore, I/R induced a statistically significant increase in the pulmonary uptake of [¹²⁵I]IgG to 384% of the control level. This result, suggesting enhanced vascular permeability in the lungs treated with I/R, is supported by the measurements of lung weight gain during the experiment. After completion of the experiment (2 h ventilation), there was a significantly higher ratio of lung wet weight to body weight in the I/R group (0.78 ± 0.08, n = 4) as compared with the control value (0.57 ± 0.02, n = 6; p < 0.05).

Pulmonary artery pressure in the control lungs was constant during the perfusion. Lung ischemia caused an increase in the perfusion pressure measured in the pulmonary artery at the start of reperfusion with respect to the preischemic equilibration level; the elevated pressure gradually returned to baseline over 30 min of reperfusion. These results coincide with previously published physiologic parameters monitored in control and I/R lungs in our laboratory (1, 3, 4). Peak tracheal pressure at constant VT was similar in control and I/R lungs, and there were no indications of overt pulmonary edema in either the experimental or control groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we specifically investigated injury of the pulmonary endothelium induced by normoxic I/R. We evaluated ACE as a cell-specific marker of endothelial injury. ACE is preferentially localized on the luminal surface of endothelial cell membranes (19). Several groups have attempted to use ACE as an indicator of pulmonary vascular injury. Pulmonary injury in animal models has been reported to lead to an increase of ACE activity in the plasma and a reduction of activity in lung tissue (22), albeit clinical utilization of ACE as a marker of endothelial injury is not that straightforward (23). In some cases, plasma ACE activity increases during pulmonary and vascular diseases, and thus correlates with two other markers of endothelial injury: Factor VIII and thrombomodulin (23, 24). In other cases, ACE activity in plasma is rather reduced, whereas the Factor VIII level is increased (25). Nevertheless, several groups have reported that in the perfused lungs, vascular injury causes a reduction of ACE activity in the lung tissue (26) and elevation of ACE activity in the perfusate (27).

Our present results show that I/R induces significant elevation of ACE activity in the perfusate of isolated rat lungs, as well as reduction of ACE activity in the lung tissue. Endothelial cells normally shed ACE from the plasma membrane via extracellular cleavage of ACE through the activity of a specific protease associated with the cellular plasma membrane (18). This shedding explains the gradual increase in ACE activity in the perfusate during control perfusion. I/R induced a marked increase in ACE activity in the perfusate as compared with control perfusion. Filtration of the perfusate did not significantly reduce ACE activity in the perfusate obtained from the control or I/R groups. About 80% of ACE activity in the perfusates obtained from either the control or I/R groups was associated with the aqueous phase. These data imply that accelerated ACE shedding from endothelium represents the most likely mechanism for increased ACE activity in the perfusate in normoxic pulmonary I/R. Whether this mechanism operates via upregulation of ACE-specific membrane protease remains to be elucidated.

I/R induces a reduction of ACE activity in lung tissue, thus indicating the source of ACE activity in the perfusate. Endothelial cells, however, are not the only ACE-containing cell population in lung tissue; for example, pulmonary macrophages also contain ACE (28). In order to specifically assess endothelium-associated ACE, we utilized a new approach based on the specific interaction of anti-ACE mAb 9B9 with pulmonary ACE localized on the plasma membrane of the pulmonary endothelium. Radiolabeled anti-ACE mAb 9B9 accumulates in the lung of various animal species because of its specific interaction with pulmonary endothelial ACE (14). Pulmonary accumulation mAb 9B9 decreases in animals treated with agents inducing injury to the pulmonary endothelium (21, 29), and reduction of the pulmonary uptake of mAb 9B9 therefore, seems to be a cell-specific marker of endothelial injury in the lung.

The results presented here clearly demonstrate that I/R leads to reduction of the pulmonary uptake of perfused radiolabeled mAb 9B9. This reduction is specific, since pulmonary uptake of control, nonimmune IgG in a parallel experiment was not only not decreased, but was actually enhanced 2-fold with I/R. Enhanced uptake of IgG could be explained either by an increase in pulmonary vascular permeability or/and an increase in the perfusing surface area in the lung tissue. The increase in wet lung weight in the I/R group favors the first possibility. A previous report by Taylor and coworkers also documented that I/R leads to an increase in pulmonary vascular permeability (30). Reduction of the pulmonary uptake of mAb 9B9 correlates with the reduction of ACE activity in lung tissue and elevation of ACE activity in the perfusate.

This reduction could be explained either by a decrease in the surface density of ACE in the pulmonary vasculature and/ or a decrease in perfusing surface area in lung tissue. For example, Jackson and associates have documented that both factors may contribute to a reduction in the metabolism of radiolabeled ACE substrate in the lungs of rats exposed to hyperoxia (31). Our experiments, however, were performed in isolated perfused lungs, which allows the exclusion of systemic effects (e.g., alterations of cardiac output) affecting lung perfusion. In addition, pulmonary artery pressure increased after the start of the reperfusion and returned to the normal control level during reperfusion. Since pulmonary artery pressure correlates with alterations in flow through the pulmonary vasculature, this result suggests normalization of the perfusion flow during reperfusion. These considerations support the conclusion that the observed reduction in the pulmonary uptake of mAb 9B9 was due mainly to a decrease in surface density of ACE in the pulmonary vasculature, whereas alterations in the perfusing area may have made a minor contribution.

As reported by several laboratories, lung I/R leads to the accumulation of oxidants in lung tissue (1). A previous study with perfused rat lungs (31), as well as our data presented in Table 1, documented that exogenous oxidants induce an increase in the ACE level in the perfusate. One may speculate that endogenous oxidants produced in the lung tissue during I/R enhances ACE shedding from endothelium. ACE shedding through the activity of membrane-associated specific protease is upregulated by phorbol myristate acetate (PMA) (32). PMA induces a wide spectrum of functional alterations in endothelial cells, including generation of oxidants (8). Vice versa, sublethal doses of oxidants induce functional alterations in endothelial cells similar to those induced by PMA (33). Therefore, and again speculatively, oxidants generated in the lung tissue during I/R could upregulate an ACE-specific membrane protease via PMA-like mechanism(s). However, the source of active oxygen metabolites generated in the lung tissue in normoxic I/R as well as mechanism(s) for I/R-induced upregulation of ACE shedding, remain to be identified.