INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PULMONARY MICROVASCULAR INJURY is a characteristic feature of reperfusion of lungs after a period of ischemia (1); however, the mechanisms mediating reperfusion-induced lung microvascular injury remain unclear. Sequestration of polymorphonuclear leukocytes (PMN) subsequent to the expression of vascular endothelial adhesion molecules and the activation of PMN in the pulmonary vascular bed may contribute to the pathogenesis of injury (14). However, their presence may not be obligatory, because vascular injury occurred in lung preparations perfused with a cell-free buffer and subjected to unventilated ischemia followed by reperfusion and reinstatement of ventilation (3, 16, 19). It has also been suggested that humoral factors such as arachidonic acid metabolites, reactive oxygen species, nitric oxide metabolites, and cytokines, released by the lung and blood cells during ischemia and reperfusion, are involved in the mechanism of vascular endothelial injury; however, reactive oxygen species are generated primarily during ventilated ischemia in proportion to the oxygen concentration of the inspired gas (see Ref. 6). Because the role of these mediators remains uncertain, we evaluated, in the present study, the possibility that reperfusion-induced lung vascular injury could occur under experimental conditions that minimized the contribution of humoral factors and activated PMN. We addressed the hypothesis that the absence of ventilation during ischemia is an essential causative factor mediating the development of vascular injury during reperfusion.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The experimental protocol was approved by the Animal Care Committee of University of Illinois at Chicago. Sprague-Dawley rats weighing between 300 and 350 g were anesthetized with 3% halothane in 20% O₂ (balance N₂) using a vaporizer, which delivered the gas into a bell jar for induction and then into a specially designed face mask (flow rate 2 l/min) for maintenance. The trachea was cannulated, and lungs were ventilated with a small animal ventilator at 60 cycles/min, 3 ml tidal volume, and 2 cmH₂O end-expiratory pressure with continued administration of the anesthetic gas mixture. A thoracotomy was performed, and heparin (1,000 U/ml) was injected into the right ventricle. The pulmonary artery was cannulated and perfused with modified Krebs-Henseleit solution (composition in mM: 118 NaCl, 4.7 KCl, 1.0 CaCl₂, 0.5 MgCl₂, 4.45 HEPES Na⁺, 5.55 HEPES, 11 glucose, and 0.025 EDTA) supplemented with 5 g/100 ml BSA (Sigma Chemical, St. Louis, MO) using a peristaltic pump (Gilson, Minipuls 2) at a rate of 0.03 ml · g body wt¹ · min¹. A cannula was placed in the left atrium to drain the venous effluent. The length of the cannula was kept constant at 4 cm in all the experiments. The heart and lungs were then excised en bloc and suspended from a calibrated beam balance that continuously monitored the wet weight gain of lung. The initial portion of the venous effluent was discarded until the effluent was clear of blood cells. Subsequently, 50 ml of the venous effluent were recirculated. The temperature of the recirculating fluid was maintained at 37°C. A bubble trap and a pressure transducer [to measure the pulmonary artery pressure (P_pa)] were placed in the circuit. With the use of counterweights, the lung wet weight was nulled, and subsequent changes in the lung wet weight and the P_pa were continuously monitored. Data were digitized (µDAS 8 PGA board, Keithley Metrabyte), logged, and displayed on the video monitor of a computer (IBM PS2 model 50Z) with the use of commercially available software (Labtech Notebook Pro for Windows). Baseline recording of lung weight and P_pa were obtained for 15 min, and subsequently, the different experimental protocols were implemented.

Experimental protocols. The experiment lasted for 2 h, which included a 15-min period of equilibration followed by 75 min of ischemia and 30 min of reperfusion (Fig. 1). At the end of reperfusion, the vascular permeability to ¹²⁵I-labeled albumin (¹²⁵I-albumin) was measured. Six groups of lung preparations were studied, including a control group and five experimental groups. The control lung preparations were continuously ventilated with 20% O₂ and perfused without interruption for 120 min (n = 4). The experimental groups were ventilated with 20% O₂ (balance N₂) before ischemia and during the reperfusion phase. These groups differed only on the basis of the ventilatory gas mixtures used during the phase of ischemia: group I, lungs ventilated with 20% O₂ (n = 4); group II, lungs ventilated with 100% N₂ (n = 4); group III, lungs remaining collapsed and unventilated (n = 4); group IV, lungs similar to group III but instilled with 4 ml/kg bovine surfactant containing 25 mg/ml of phospholipid (SURVANTA, Ross Abbott Laboratories, Columbus, OH) before ischemia (n = 4); and group V, lungs similar to group III but instilled with 4 ml/kg saline before ischemia (n = 4). Two additional lung preparations were treated identically to group III lung preparations, except that the postischemic rise in P_pa (30 mmHg) normally seen in group III preparations was not permitted to exceed 7.5 mmHg above the preischemia baseline by means of clamp-circuit limiting P_pa to a preset maximum level.

View larger version (7K): [in this window] [in a new window]   Fig. 1.   Time line for experiments. Lung preparations initially underwent a 15-min period of equilibration during which they were ventilated with 20% O2 (balance N2) (a). Preparations were subsequently subjected to an ischemic (no flow) period of 75 min, during which ventilation was varied depending on experimental group (See METHODS) (b). During 30-min period of reperfusion, all lung preparations were ventilated with 20% O2 (balance N2) (c). This was followed by a brief exposure to 125I-labeled albumin (125I-albumin) for measurement of albumin permeability-surface area (PS) product (d). In two groups of lung preparations either surfactant or saline vehicle was instilled intratracheally at time 0, as described in METHODS.

Measurement of vessel wall ¹²⁵I-albumin permeability-surface area product. BSA was labeled with Na ¹²⁵I (New England Nuclear, Boston, MA) using the standard chloramine T method (2). Free ¹²⁵I label was separated from ¹²⁵I-albumin using a Sephadex G column. A ¹²⁵I-albumin-labeled Krebs solution was prepared by adding 80,000 counts/ml of ¹²⁵I-albumin plus a small quantity of Evans blue dye (4 mg/g albumin) as a vascular marker to 50 ml of Krebs solution; any preparation exhibiting an obviously nonuniform distribution of dye was discarded. At the end of 30 min of reperfusion, ¹²⁵I-albumin-labeled Krebs solution was perfused through the lungs for a period of 3 min followed by a washout period of 6 min with Krebs solution containing no label. The lungs were detached from the perfusion apparatus and cut into 12 samples. The samples were weighed and placed in separate containers. A 0.5-ml sample of the ¹²⁵I-labeled Krebs solution was also placed in one of the containers. The number of counts due to labeled albumin in each of the containers was measured by a gamma counter (Minaxi  Auto Gamma 5000 Series Gamma Counter; Packard Instrument, Downers Grove, IL). The ¹²⁵I-albumin permeability-surface area product (PS) was then calculated according to the formula (10): A /(C × t), where A represents measured tissue counts per gram of lung tissue, t is the duration of exposure (in minutes), and C is the concentration of labeled albumin in the perfusing liquid (counts/ml).

Statistical analysis. All values are expressed as means ± SE. Statistical comparisons between group means were made by Student's t-test. The significance level was set at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilation during ischemia prevents vascular injury after reperfusion. During a 120-min perfusion, control lung preparations did not gain weight but rather showed a slight weight loss, probably secondary to evaporative loss from the surface (Table 1). The ¹²⁵I-albumin PS value calculated at the end of 120 min of perfusion was 4.9 ± 0.3 × 10³ ml · min¹ · g¹ (Table 1; control group). The lung preparations in groups I and II also did not develop edema, and the albumin PS values were similar to those of control lungs (Table 1). The unventilated lung preparations (group III) showed significant (P < 0.05) vascular injury as evidenced by marked increases in wet weight of the lung (1.45 ± 0.35 g) and ¹²⁵I-albumin PS (17.7 ± 2.3 × 10³) compared with controls (Table 1).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   A: representative recording of increase in lung wet weight in an injured preparation (group III). During ischemia wet weight decreased by 0.6 g, and on reperfusion the value increased by 0.8 g relative to baseline weight (1.5 g), which was electronically nulled at time 0. B: representative recording of changes in lung wet weight in an uninjured preparation (i.e., group II lung ventilated with 100% N2 during no-flow period). During ischemia lung wet weight was decreased by 1.3 g. At end of 30 min of reperfusion, lung preparation showed no evidence of injury (i.e., lung wet weight was stable at 0.5 g below baseline value).

Effects of intratracheal surfactant administration on reperfusion-induced lung vascular injury. Pretreatment with 4 ml/kg of surfactant before ischemia (group IV) prevented the reperfusion injury as evidenced by the lack of significant changes in the lung wet weight and albumin PS compared with control (Table 1). Instillation of 4 ml/kg of saline (group V) caused a significant (P < 0.05) increase in both lung wet weight (4.00 ± 0.24) and albumin PS (72.2 ± 4.4 × 10³) compared with group IV (Table 1).

Changes in P_pa. The baseline P_pa at the end of the equilibration period (see Fig. 1) was 5-8 mmHg in all groups. In all experimental groups, during the 30-min reperfusion period, there was a marked but transient increase in P_pa that declined toward the normal baseline value. Figure 3 shows a typical recording for a group III lung preparation. For purposes of analyzing the data, we defined peak P as the difference between the peak pressure attained on reperfusion and the baseline P_pa, and we defined 30-min P as the difference between the P_pa at the end of the 30-min reperfusion period and the baseline P_pa (see Fig. 3). Peak P significantly increased (P < 0.05) over the values for the protected lungs (group I) only in lungs of groups III and V (Fig. 4), which also showed evidence of reperfusion injury (Table 1). There was no significant difference in the 30-min P among any of the experimental groups (see Fig. 4). To determine whether peak P was related to reperfusion-induced lung injury, we evaluated the correlations between the magnitude of peak P and either lung wet weight gain or albumin PS values at the end of the reperfusion period. These correlations were absent for lungs from groups I, II, and IV in which there was also no evidence of lung injury (Fig. 5, A and B). In contrast, the peak P and wet lung weight gain values (Fig. 5A; r = 0.78, P < 0.05) and peak P and albumin PS values (Fig. 5B; r = 0.61, P < 0.05) were significantly correlated in groups III and V (Fig. 5, A and B).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Representative recording of pulmonary artery pressure (Ppa) in an injured preparation (group III). From a baseline pressure of 9 mmHg, peak pressure rose to 24 mmHg (i.e., peak P = 15 mmHg) at onset of reperfusion, and then pressure gradually declined to 8 mmHg at end of 30 min of reperfusion (i.e., 30-min P = 1 mmHg). Peak P, difference between peak pressure attained on reperfusion and baseline Ppa; 30-min P, difference between Ppa at end of reperfusion period and baseline Ppa.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Changes in Ppa after reperfusion in all experimental groups. All values are represented as means ± SE. Baseline Ppa (before ischemia) varied between 5 and 8 mmHg in all groups. Note peak P was augmented significantly in unprotected lung preparations (groups III and V). Group I, 20% O2 ventilation; group II, N2 ventilation; group III, no ventilation; group IV, no ventilation and surfactant treatment; and group V, no ventilation and saline vehicle. * Significance (P < 0.05) from group I.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Correlation plots of peak P with either lung wet weight gain (A) or albumin PS (B). Square symbols in both graphs represent protected lung preparations (groups I, II, and IV); crosses denote injured preparations (groups III and V). Regression lines are shown. In comparison with protected preparations, injured preparations showed higher wet weight gain or albumin PS for a given peak P value observed during reperfusion. See text for additional detail.

3

1

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The lung preparations in group II, which were in a state of anoxia (i.e., ventilated with 100% N₂) during the cessation of perfusion, did not show evidence of vascular injury on reperfusion and ventilation with the oxygen in gas mixture. Reperfusion injury was evident only in lungs in groups III and V (both of which were unventilated) in agreement with other reports (8, 16). The results are consistent with the hypothesis that the state of lung inflation rather than anoxia followed by reintroduction of oxygen is a critical determinant of reperfusion-induced lung vascular injury.

Continuous recordings of P_pa showed a transient but marked rise immediately at the onset of reperfusion. The magnitude of the peak rise in P_pa in groups III and V lungs, which showed evidence of vascular injury, was significantly greater than that of P_pa values in other groups without vascular injury. Normoxic, perfused rabbit lungs exposed to similar high transmural pressures demonstrated injury of both capillary endothelium and alveolar epithelium (22). Vascular injury was evident even though the vessel wall was exposed to the high pressure for 1-5 min (12). The rapid onset of injury may be the result of stress failure because the capillaries were subjected to high wall stress during the episode of increased hydrostatic pressure (22). Fu et al. (7), using rabbit lungs exposed to high capillary transmural pressure, demonstrated that P_pa rise above a critical value (>32.5 cmH₂O) is required to produce capillary wall stress failure. However, the present results cannot be explained on this basis. In lung preparations showing evidence of reperfusion vascular injury (i.e., groups III and V), peak P during reperfusion was correlated with lung wet weight gain, consistent with greater liquid permeability in the injured lung preparations. In contrast, in lung preparations protected by either mechanical ventilation (groups I and II) or surfactant instillation (group IV) during the no-flow period, no such correlation was observed even though the rise in P_pa was similar (Fig. 5). Moreover, in an experiment in which peak P during reperfusion was regulated to a rise of only 7.5 mmHg, the increases in lung wet weight and ¹²⁵I-albumin PS persisted during the reperfusion period. Thus these findings indicate that the increase of in hydrostatic pressure occurring during the reperfusion period cannot explain the development of lung vascular injury.

We observed that reperfusion injury occurred only when the lungs were both ischemic and collapsed (as in group III). Lung inflation or instillation of surfactant (in groups I, II, and IV) prevented the injury. Two forces are influenced by the state of lung inflation (7, 15) and thus may be involved in the mechanism of vascular injury. Both forces act at the alveolar-capillary wall: 1) surface tension of the alveolar lining layer (15) and 2) alveolar distending pressure (7). Both of these factors are discussed below.

With respect to alveolar surface tension, this force primarily depends on the amount of surfactant lining the alveolar epithelium. It is known that alveolar distention stimulates surfactant release. In studies on freshly excised dog lungs, Faridy (5) showed that surface tension and lecithin content of bronchoalveolar fluid were increased when the lobe was inflated with air, and the increases were directly related to the inflating pressure, indicating that lung distention enhanced the release of surfactant. In excised rat lungs, Hilderbren et al. (9) showed that air inflation to total lung capacity is a major physiological stimulus to the release of lung surfactant into the alveolar space. In studying the effects of atelectasis on surfactant production in rabbit lungs, Levine and Johnson (13) showed a variable but significant loss of surface activity with atelectasis. Moreover, surfactant synthesis is also known to be impaired during ischemia. Veldhuizen et al. (21) showed that phosphatidylglycerol and surfactant-associated protein A levels were decreased after 12 h of lung graft storage and 6 h of reperfusion in dogs. In a study on rat lung graft transplantation after warm ischemia of 60-120 min, Erasmus et al. (4) showed that the concentration of phosphatidylcholine decreased in association with severity of lung injury. Klepetko et al. (11) documented a steady decrease in dipalmitoylphosphatidylcholine concentrations of lungs stored at 4°C and showed that this correlated with ischemic damage to type II pneumocytes. Hence, it can be concluded that combination of atelectasis and ischemic damage of type II alveolar epithelial cells during the phase of ischemia can result in decreased surfactant content in alveoli, thereby leading to an increase in surface tension and alveolar collapse. Moreover, alveolar collapse itself can influence the integrity of pulmonary capillaries (15) because at high transmural pressures, the capillaries bulge into the alveolar space, and the alveolar lining layer acts to support the bulging capillary wall by compressing the endothelial and epithelial cell layers (22). Thus lung microvessel injury may, in part, be attributed to stress failure of pulmonary microvessels secondary to the loss of surfactant and the resultant alveolar collapse. Our finding that surfactant instillation prevented the reperfusion-induced lung vascular injury in the collapsed lung supports this mechanism of lung vascular injury. In this regard, a possible explanation for the higher labeled-albumin PS product in group V (saline instilled) lungs compared with unventilated lungs of group III (with no instillation) could be dilution or washout of surfactant by saline (17).

An elevated alveolar distending pressure during inflation of previously collapsed alveolar units can also lead to lung microvascular injury (7). This can be appreciated by consideration of the Young-Laplace equation, P = 2T/R, where P represents the distending pressure needed to expand the alveoli, T, the alveolar surface tension, and R, the radius of curvature. The pressure required to inflate the alveoli is greatly increased if the surface tension rises. Hence, decreased surfactant production (due to both collapse of alveoli during ischemia and ischemic damage of type II alveolar epithelial cells) will require a significantly greater distending pressure to expand alveoli during the reperfusion phase. If the alveolar distending pressure is directly transmitted to the alveolar wall and leads to increased longitudinal tension in the capillary (22), this could be a mechanism of vascular injury in the reperfused lung. Therefore, the combined effects of increased alveolar surface tension and increased distending pressure transmitted to open alveoli could explain our findings that only unventilated and ischemic lung preparations demonstrated microvascular injury on reperfusion.

There is a possibility that the increase of ¹²⁵I-albumin PS observed in injured lung preparations may have occurred as a result of changes in the lung vascular surface area. Because our experiments were made at constant flow (see METHODS) and the perfusion pressure (except for a brief period immediately after reperfusion) remained near control levels during the 30-min reperfusion phase, it is unlikely that the vascular surface area was greatly increased by ischemia-reperfusion. Also, we observed a threefold increase in albumin PS (measured in the final 3 min of reperfusion period). Such a large increase in albumin PS cannot be accounted for by a vascular surface area shift. Finally, Fig. 5 shows the relationship between peak P and albumin PS. The slope of the line for lungs developing edema was significantly greater than for the protected lungs. The markedly greater increase in albumin PS product for a given change in pressure in the lungs developing edema suggests increased vascular permeability in these lungs. Moreover, in the experiments in which the rise in P_pa was limited to 7.5 mmHg, we still observed an increase in albumin PS after reperfusion. The albumin PS increased markedly in these lungs even though vascular surface area increase (due to vascular distension and recruitment) would be expected to be less than in the lungs that experienced a marked increase in reperfusion P_pa.

It is also unlikely that vascular volume expansion contributes significantly to the measured increases of wet weight in the injured lungs. Vascular volume in the rat lung is ~0.1 ml/g lung wet wt when determined at a steady-state perfusion pressure of 10 cmH₂O (Vogel and Malik, unpublished observations) and consequently could not account for weight gains of several grams observed in injured lungs. Thus a doubling of vascular volume would account for only ~3% of the observed lung wet weight gain. It is also unlikely that vascular distension during reperfusion explains the weight gain in injured lungs (groups III and V). The peak rise in pulmonary arterial pressure was transient, and therefore any resulting increase in lung wet weight attributable to distension of microvessels would be short lived.

The observed protective effect of surfactant could conceivably be a result of its potential anti-inflammatory effects (23), because several groups have demonstrated that there is a population of resident neutrophils even in the buffer-perfused lung preparation (18). However, the anti-inflammatory effect of surfactant required several hours to develop (20) and consequently is not expected to be marked in the time frame of our experiments. Moreover, if inflammation is important in the development of vascular injury in this model, then we would have expected to see signs of barrier breakdown (lung wet weight gain and an increased albumin PS product) in group II lung preparations (ventilated with 100% N₂). Because this was not the case, the present results suggest that the reperfusion injury cannot be explained on the basis of a resident population of neutrophils.

In conclusion, the present findings suggest that alveolar collapse during the period of lung ischemia is a key determinant of reperfusion-induced lung vascular injury. The rise in pulmonary vascular hydrostatic pressure and increased vascular surface area cannot account for the marked increases in the vessel wall albumin PS product observed during reperfusion. The study raises the possibility that cellular and humoral factors may not be dominant, although they may contribute to the mechanism of the response. Alveolar collapse secondary to the loss of surfactant during ischemia may be an important factor in the mechanism of reperfusion injury, because ventilation of the lungs during ischemia or instillation of surfactant prevented the response. Mechanical factors resulting from the loss of alveolar tethering forces during alveolar collapse may injure the pulmonary capillaries on reperfusion.