Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The mechanism(s) for ischemia/reperfusion (I/R) injury of tissues has generated considerable interest in recent years (1, 2). The paradoxical nature of this phenomenon consists in increased damage to tissues after restoration of blood flow following a nonlethal ischemic period. The injury has been studied in the intestine, kidneys, heart, brain, and other organs; and it has been shown that oxidation of tissue components occurs during reperfusion (1, 3). As one possible mechanism for these effects, generation of oxidants has been attributed to conversion of xanthine dehydrogenase to xanthine oxidase and to accumulation of hypoxanthine (a substrate for xanthine oxidase) as a result of anoxia during ischemia (4). Reintroduction of oxygen during reperfusion provides the other substrate for xanthine oxidase leading to the formation of superoxide anion and hydrogen peroxide. Therefore, the operation of the xanthine dehydrogenase pathway requires that tissues experience a cycle of anoxia/reoxygenation with I/R.

Anoxia/reoxygenation does occur during transient arterial obstruction in the systemic circulation. However, it is important to recognize that lung ischemia as a result of pulmonary artery occlusion does not result in tissue anoxia, if ventilation continues during ischemia. Actually, local lung tissue PO₂ may increase in ischemia because O₂ is no longer removed by pulmonary capillary blood. Therefore, a cycle of I/R in lung does not necessarily mean a cycle of anoxia/reoxygenation. Nevertheless, oxidant generation and oxidative damage to tissue components have been shown to occur during both in vivo and ex vivo lung I/R (5). Moreover, air- or O₂-ventilated isolated lungs showed the initiation of oxidative injury during the ischemic period before the onset of reperfusion (5).

It is clear that mechanistic models which depend on anoxia and reoxygenation cannot explain the occurrence of tissue oxidation in lung ischemia. The lack of change in adenosine triphosphate (ATP) during lung ischemia in a ventilated lung model provides support for this conclusion (5). To cast additional doubt on the relevance of the anoxic model for the lung, it was shown that the degree of tissue oxidation in lung ischemia was directly proportional to ventilation gas PO₂, and that anoxia was protective (5, 6). Further, the mechanism could not be ascribed to any of the likely biochemical events that would result from decreased substrate delivery or decreased metabolite removal secondary to ischemia (5). Therefore, the conclusion was that primary biochemical alterations commonly associated with ischemia in other organs are unlikely to be the primary initiators of lung oxidative injury. The question remains: what is the trigger for the initiation of oxygen radical production in lung ischemia? One possibility is that another factor that accompanies ischemianamely, the cessation of the mechanical component of flowcould be responsible for initiation of subsequent biochemical events.

Blood flow evokes physical forces, such as pressure, shear stress, and stretch, that act on the vessel wall (8). The endothelium transforms these forces (mechanotransduction) into electrical and biochemical signals (9). Although the precise mechanism(s) of mechanotransduction have not been elucidated, stretch and shear stress-responsive ion channels may be involved. During ischemia, channel-mediated transduction could involve ionic shifts resulting in changes of membrane potential as an initiating event for oxidant generation. Our previous studies with cultured endothelial cells and the isolated rat lung have demonstrated oxyradical production and lipid peroxidation with K⁺-induced membrane depolarization (10, 11).

A key element of our hypothesis is that cessation of flow leads to pulmonary endothelial cell membrane depolarization. Ischemia in other organs has been associated with depolarization of the cell membrane which has generally been ascribed to depressed energy state (12). As an alternate possibility, ischemia-induced depolarization could be mediated by channels or regulators that are responsive to mechanical factors. An example of a shear stress-sensitive mechanism is the endothelial inwardly rectifier potassium channel (I_K,S) which, in the flow-activated open state, maintains membrane polarization (13). In this study we evaluated whether endothelial depolarization occurs with ischemia in the isolated perfused lung under conditions that maintain tissue oxygenation and ATP.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Isolated Lung Preparation

Male Sprague-Dawley rats, CRL:CD(SD)BR (Charles River Breeding Laboratories, Kingston, NY) weighing 150-170 g, were anesthetized with sodium pentobarbital, 50 mg/kg, intraperitoneally. A tracheostomy was performed and ventilation with a respirator was started at 1 Hz, 2 ml tidal volume, and 2 cm H₂O positive end-expiratory pressure. The lungs were cleared of blood by perfusion through the pulmonary artery, removed from the thorax, and transferred to the isolated lung ventilation-perfusion system as previously described (5). Transfer from the thorax involved suspension of perfusion for < 5 s, but no interruption of ventilation. Perfusion into the pulmonary artery was maintained by a peristaltic pump generally at a flow rate of 7 ml/min; the effluent dripped freely from the transected left atrium and was collected and recirculated. The total volume of recirculating perfusate was 50 ml. Intratracheal and pulmonary arterial pressures were continuously recorded throughout the experiment with pressure transducers (PM 131TC and P23DC; Statham Instruments, Oxnard, CA), direct writing oscillographs (Gould, Inc., Cleveland, OH), and two pen AC recorders (Primeline, Sun Valley, CA). A digital pressure monitor (BMP 832; CWE Inc., Ardmore, PA) was calibrated to provide direct readout of pulmonary artery mean pressure in cm H₂O. The ventilation gas (Airco, Philadelphia, PA) used for these experiments was 5% CO₂ in O₂ or 5% CO₂ in air. Perfusate was Krebs-Ringer bicarbonate buffer (KRB; pH 7.4) containing 10 mM glucose and either 3% (wt/vol) dextran (clinical grade, MW 70-90 kD; Sigma, St. Louis, MO) or fatty acid-free bovine serum albumin (BSA; Boehringer-Mannheim Biochemicals, Indianapolis, IN).

After 20-30 min of equilibration-perfusion to allow uptake of fluorescent dye, global ischemia was produced by turning off the perfusion pump while ventilation continued. After a period of ischemia, perfusion was restarted with the same pre-ischemic settings for flow. Control perfusions were carried out with continuous perfusion and ventilation as above.

Lung Surface Fluorometry

A surface fluorometer was used to monitor fluorescence from the lung surface as previously described (10, 14). Briefly, the principle involves perfusing the isolated ventilated rat lung with a fluorescent dye in the dark; illuminating the lung surface with narrow-bandwidth (10 nm) excitation light delivered via a fiberoptic guide; collecting the emitted light from the lung surface with a fiberoptic probe and amplifying it in a photomultiplier tube after passing through a narrow bandpass (10 nm) emission filter; and recording the signal with a data acquisition computer. To monitor membrane potential changes in the intact lung, we used a fluorescent dye, bis-oxonol, that is sensitive to cell membrane potential and increases in its intensity of fluorescence with membrane depolarization as previously described (11). Lungs were perfused with 20 nM solution of bis-1,3 dibutylbarbituric acid trimethane oxonol (Molecular Probes, Eugene, OR). This concentration yielded the optimum signal-to-noise ratio. Fluorescence was excited at 490 nm and collected at 520 nm. Additional studies were carried out with another membrane potential-sensitive probe, 5,5',6,6'-tetrachloro-1,1',3,3'-tetraethylbenzimidazolyl-carbocyanine iodide (JC-1; Molecular Probes), that decreases in fluorescence with depolarization (15). JC-1 was administered at 200 nM in the perfusate and data was collected with 490 nm excitation and 610 nm emission. With this dye, fluorescence was collected during a 5-10-s excitation period at approximately 1-min intervals to minimize photobleaching which would lead to an artifactual decrease in fluorescence.

Since the tissue fluorescence field that is sampled by our light-collection system theoretically could increase due to decreased capillary fluid volume with ischemia, we used a membrane potential-insensitive, membrane-localizing dye, 1-(4-trimethylammonium phenyl)-6-phenyl-1,3,5-hexatriene p-toluene sulfonate (TMA-DPH; Molecular Probes), to evaluate this potential artifact. The dye was administered at 1 µM in perfusate. Fluorescence was excited at 350 nm and collected at 460 nm.

Bis-oxonol Fluorescence Response to Ischemia against the Background of K⁺-channel Inhibition, High K⁺-perfusion, and Air versus Oxygen Ventilation

To inhibit K⁺-channels, the lungs were pretreated with 1 mM BaCl₂ or 10 mM tetraethylammonium (TEA), administered in the perfusate for 30 min. For high K⁺-perfusion, we used KRB solution containing 24 mM K⁺ instead of the usual 5.9 mM. This concentration of K⁺ was chosen because lungs could be continuously perfused for at least 2 h without the development of overt edema. To evaluate the effect of tissue oxygenation, the ventilation gas was switched from 95% O₂ + 5% CO₂ to air + 5% CO₂ and vice versa in the same lung, and the response to sequential episodes of ischemia was recorded.

Intravital Subpleural Microvascular Endothelial Cell Microscopy

The surface fluorometric assessment of gross subpleural capillary endothelial membrane potential was taken to the individual cellular level by the use of an intravital technique to observe and image in situ capillary endothelial cells in the isolated rat lung in real time using an epifluorescence microscope. This method has been applied previously to study Ca²⁺ waves in rat lungs (16). The isolated blood-free rat lung was placed horizontally on the 48 × 60 mm coverglass window in a specially designed Plexiglas chamber and connected to ventilation and perfusion ports of the chamber. The chamber was placed on the stage of an epifluorescence microscope fitted with a ×100 objective (Nikon Diaphot TMD; Nikon Inc., Melville, NY) and equipped with an optical filter changer (Lambda 10-2; Sutter Instrument Co., Novato, CA). Lungs were ventilated and perfused as described above with the exception that the heart was not transected and a cannula was placed in the left atrial appendage to collect the effluent and return it to the reservoir via a pump. Since the transmitted pulsation of the sino-atrial node caused significant lung movement artifact even in the absence of ventilation, we injected a local anesthetic (0.05 mg xylazine) subepicardially into the posterior wall of the right atrium.

The fluorescence system utilized a mercury lamp fiberoptic light source, narrow bandpass filters (FITC-485/ 10; Rhod-560/10), and a triple-band dichroic mirror (D/F/ R-BS&M; Chroma Technology Corp., Brattleboro, VT) for excitation of the lung surface. The emitted light passed through a narrow triple-bandpass filter (470/40, 535/40, and 630 /60 peak transmission/half bandwidth, in nm). The objective was positioned with the maximum microvascular diameter at the focal plane. Images from the microscope were acquired at 500 ms exposure time with a computer-controlled, cooled CCD camera (MicroMAX; Princeton Instruments, Inc., Trenton, NJ) using graphics control software (Metamorph Imaging System; Universal Imaging Corp., West Chester, PA).

After an equilibration period of 30 min with the isolated lung to allow bis-oxonol (20 nM) uptake, the intravascular dye was removed by perfusion with dye-free medium for 5 min to reduce background fluorescence. Ventilation was stopped briefly (< 10 s) to permit collection of fluorescence images. Subpleural vessels were identified by observing real-time passage of bis-oxonol-labeled red blood cells added to the perfusate. To further assess the localization of the bis-oxonol signal, some lungs were double-labeled with 20 nM bis-oxonol administered in the perfusate and 100 µl of 25 µM Nile Red (Molecular Probes) instilled intratracheally. Fluorescence images from the same area were taken sequentially using appropriate filter combinations for each dye, and were overlaid for co-localization after assigning separate pseudocolor to each image of a pair with the graphics software program.

Venous-outflow Pressure

To test whether increasing the outflow pressure during ischemia would affect bis-oxonol fluorescence, the left atrium was catheterized as for intravital microscopy described above and the tip of the catheter was elevated during ischemia to generate intravascular pressure equivalent to that of control perfusion. Pressure is expressed relative to zero pressure at the level of the left atrium.

Lung ATP Assay

Lung ATP content was measured as previously described (5). Briefly, a portion of the lung homogenate was extracted with cold ethanolic perchloric acid and assayed enzymatically with hexokinase and glucose-6-phosphate dehydrogenase.

Statistical Analysis

Results are expressed as mean ± SE for each condition unless otherwise stated. Significance of parametric differences among groups was evaluated with one-way ANOVA and Bonferroni's test for multiple comparisons using SigmaStat (Jandel Scientific, San Rafael, CA). Differences were considered statistically significant at P < 0.05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

At the wavelengths used to evaluate bis-oxonol, lungs show intrinsic autofluorescence (Figure 1). With infusion of the dye bis-oxonol and continuous monitoring, fluorescence from the lung surface increased and after about 30 min attained a steady level at about four times the baseline signal (Figure 1). The level of bis-oxonol fluorescence that was achieved after equilibration did not change significantly during the subsequent 90 min of continuous perfusion, indicating stability of membrane potential for this period as detected in our system (Figure 1).

View larger version (9K): [in this window] [in a new window]   Figure 1.   Equilibration of bis-oxonol with the isolated perfused lung and stability of the signal as revealed by surface fluorometry during a 2-h control perfusion. The initial part of the tracing indicates lung intrinsic fluorescence. Bis-oxonol administration in the perfusate at 20 nM was initiated as indicated by the arrow and continued for approximately 2 h. Fluorescence was collected as described in MATERIALS AND METHODS. Lungs were continuously ventilated throughout the experiment with 5% CO2 in O2. The increased fluorescence signal with bis-oxonol reached a plateau value in 20-30 min and was then stable for the remainder of the perfusion period.

We evaluated the effect of ischemia on lung-surface fluorescence. Lung intrinsic fluorescence prior to bis-oxonol addition did not change with a cycle of I/R (initial segment, Figure 2). When evaluated after the period of dye-equilibration, there was a significant increase in bis-oxonol fluorescence with ischemia. This result is compatible with membrane depolarization (Figure 2). The onset of membrane potential change was observed within 1-2 min and reached a plateau value after about 10 min of ischemia. Using another potential-sensitive probe (JC-1), we found that ischemia led to a significant decrease in fluorescence (Figure 3). This different direction of response with membrane depolarization is expected for these two dyes (11, 15). Reperfusion rapidly reversed the fluorescence back to its pre-ischemic level, indicating repolarization (Figures 2 and 3). We also observed transient hyperpolarization upon reperfusion with the JC-1 probe (Figure 3). The fluorescence changes could be reproduced with subsequent cycles of I/R in the same lung (for at least four cycles) and mean values for the first and second cycles were not significantly different (Table 1). In contrast with the membrane potential-sensitive dyes, lung-surface fluorescence with the membrane potential-insensitive fluorophore TMA-DPH was unchanged with ischemia (Figure 4).

View larger version (13K): [in this window] [in a new window]   Figure 2.   Effect of ischemia (I) and reperfusion (R) on bis-oxonol fluorescence continuously recorded from the lung surface in the presence and absence of intravascular dye. Before the administration of bis-oxonol (Oxonol), a cycle of I/R had no effect on lung intrinsic fluorescence. After dye equilibration (indicated by dotted line), ischemia led to a significant increase in fluorescence, compatible with membrane depolarization. Fluorescence returned to the baseline level with reperfusion, indicating membrane repolarization. Perfusion with dye-free dextran-KRB had no effect on lung-associated dye, as indicated by steadiness of fluorescence. Tracing is representative of 12 experiments with separate perfused-lung preparations.

View larger version (40K): [in this window] [in a new window]   Figure 3.   Effect of ischemia (I) and reperfusion (R) on JC-1 fluorescence measured intermittently from the lung surface for 5-10 s at approximately 1-min intervals. The fluorescence level is indicated by the filled circles. After dye equilibration to obtain the control level of fluorescence, ischemia led to a significant decrease in fluorescence, compatible with membrane depolarization. Fluorescence increased above the baseline with reperfusion, indicating membrane hyperpolarization. Tracing is representative of three experiments with separate perfused-lung preparations.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Effect of ischemia and reperfusion on lung surface fluorescence with TMA-DPH, a membrane-localizing, potential-insensitive dye. The dye was administered at 1 µM in the perfusate. All other conditions were similar to those described for bis-oxonol experiments in Figure 2. The tracing is representative of two experiments.

To evaluate whether the increase in bis-oxonol fluorescence during ischemia was due to decreased intravascular pressure rather than removal of shear stress, we raised the venous outflow pressure during ischemia (Figure 5). During the initial ischemic episode (Figure 5A), the left atrial catheter was maintained at zero pressure relative to the left atrium. Intravascular pressure was raised during the second episode of ischemia by elevating the tip of the left atrial catheter. This manipulation did not reverse the bis-oxonol response. A calibration curve for bis-oxonol fluorescence versus intravascular pressure generated by raising venous outflow pressure during ischemia showed a decrease in fluorescence by only about 5% of control for an 8-cm H₂O pressure elevation (Figure 5B; left-to-right arrow). By contrast, there was a 35% fluorescence increase for the standard lung ischemic model (intravascular pressure change on discontinuing perfusion was 8 cm H₂O; Figure 5B, right-to-left arrow).

View larger version (14K): [in this window] [in a new window]   Figure 5.   Effect of increased capillary pressure during ischemia on lung surface bis-oxonol fluorescence. (A) Typical tracing showing lung fluorescence and pulmonary artery pressure. During the first ischemic episode, the left atrial catheter that was placed for these experiments was maintained at the level of the left atrium (zero pressure). Intravascular pressure was raised by approximately 8 cm H2O during the second episode of ischemia by elevating the tip of the left atrial catheter. (B) Bis- oxonol fluorescence calibration curve for intravascular pressure generated by raising venous outflow pressure during ischemia. Bis-oxonol lung surface fluorescence is expressed as percent of control (perfusion at a rate of 7 ml/min with an intravascular pressure of 8 cm H2O). With ischemia and 0 cm H2O pressure, fluorescence increased up to 135% of control (right-to-left arrow). With stepwise elevation of intravascular pressure during ischemia, the decrease in fluorescence was only about 5% of control for an 8 cm H2O pressure elevation (left-to-right arrow). Data represent mean values for two experiments.

When intravascular free bis-oxonol was removed by perfusion with dye-free medium containing dextran, neither the baseline fluorescence nor the magnitude of change with I/R was altered, indicating a high affinity association of the dye with the lung tissue (Figure 6). However, the level of lung-surface fluorescence progressively decreased during perfusion with BSA-containing medium (Figure 6), indicating removal of dye from the tissue. Addition of BSA to perfusate resulted in a markedly decreased magnitude of the response to I/R, providing further evidence for removal of lung-associated bis-oxonol (Figure 6). Perfusing again with dextran stabilized the fluorescence response to I/R (Figure 6). This result indicated ready accessibility of the lung-associated bis-oxonol dye to the perfusate albumin and suggested its endothelial localization.

View larger version (12K): [in this window] [in a new window]   Figure 6.   Lung-associated dye dislodgment by the high-affinity ligand, BSA. With dextran in perfusate, bis-oxonol fluorescence exhibited a typical response to ischemia (I) and reperfusion (R). Substitution of perfusate dextran with BSA, which has high affinity for the dye, led to progressive decline in fluorescence, indicating dye dislodgment, and the magnitude of response to I/R was significantly lower. Reversal to perfusion with dextran halted the decline. Conditions and legends are as in Figure 2. Results are representative of four experiments.

Further evidence for endothelial localization of bis- oxonol was obtained by direct evaluation of fluorescence from in situ subpleural alveolar microvascular endothelial cells by epifluorescence microscopy of the isolated rat lung surface (Figures 7A and 7B). The capillary and postcapillary venule margins, along with endothelial cells, could be clearly identified in bis-oxonol-labeled lungs. Basal bis- oxonol fluorescence was associated mainly with endothelium, and endothelial cell-associated fluorescence increased with ischemia. Double-labeling experiments provided further evidence that the bis-oxonol fluorescence signal was associated primarily with the alveolar microvasculature, since it was clearly separated from the Nile Red signal (Figure 7C), which labeled the epithelium (type II alveolar epithelial cells). We did not evaluate fluorescence from central lung structures because these would be unlikely to contribute to the changes in fluorescence detected by the pleural surface light guide.

View larger version (112K): [in this window] [in a new window]   Figure 7.   Bis-oxonol fluorescence from in situ capillary endothelial cells as revealed by epifluorescence microscopy of the isolated rat-lung surface. The capillaries were identified by passage of bis-oxonol-labeled red blood cells during control perfusion (not shown). The lungs were ventilated during ischemia and perfusion; ventilation was stopped for 5-10 s to permit imaging. In panels A and B, the endothelial cells outlining a branched microvessel are visible as brighter areas of bis-oxonol fluorescence (EC); the alveolar space is the darker area outside the microvessel. Panel A: control perfusion for 30 min after dye equilibration; Panel B: ischemia for 15 min; Panel C: simultaneous labeling with intratracheal Nile Red and bis-oxonol. Fluorescence overlay imaging shows that the bis-oxonol signal is confined to the vasculature (CAP), while Nile Red predominantly labels alveolar type II epithelial cells (T2). The bar-length in panel C represents 20 µm; a bar of the same length in panels A and B would represent 10 µm, since they were magnified ×2 with the software to reveal detail. Results represent six experiments with bis-oxonol and three double-labeling experiments.

Bis-oxonol fluorescence exhibited a proportional decrease during stepwise increase in flow rate above control values, compatible with endothelial hyperpolarization with increasing shear stress (Figure 8A). Likewise, there was a proportional increase in fluorescence with stepwise decrement in flow rate from control to 1.8 ml/min (Figure 8A). This indicates an inverse linear relationship between bis-oxonol fluorescence and flow in the range of 26 to 1.8 ml/ min (Figure 8B). However, the increase in fluorescence with complete cessation of flow showed a markedly greater effect. The change in fluorescence on going from a perfusion rate of 1.8 ml/min to zero was 9-fold higher than the increase in fluorescence for a flow-rate change from 7 to 1.8 ml/min (Figure 8B). This relatively great increase in fluorescence at zero flow rate suggests that the total absence of shear has the greatest effect on membrane potential. As can be seen from Figure 5, the effects of flow were not mediated by associated changes in the intravascular pressure.

View larger version (13K): [in this window] [in a new window]   Figure 8.   Effect of stepwise changes in flow rate on bis-oxonol fluorescence from lung surface. Control perfusion flow rate was 7 ml/min. (A) Typical tracings for fluorescence and pulmonary artery perfusion pressure. The numbers in italics on the pulmonary artery perfusion pressure tracing indicate flow rates. Fluorescence is expressed as arbitrary units. (B) Summary of the effect of perfusate flow rate on bis-oxonol fluorescence from lung surface. Fluorescence is expressed as a percentage of control. Results are mean ± SE for n = 4.

The role of K⁺-channels in mediating membrane potential changes, as detected with fluorescent dyes, was evaluated by perfusing the lungs with nonselective K⁺-channel blockers BaCl₂ and TEA. These treatments led to an increase in lung surface bis-oxonol fluorescence during perfusion, indicating membrane depolarization via K⁺-channel blockade (Table 2). Ischemia in lungs treated with the K⁺-channel blockers did not induce any significant further change in bis-oxonol fluorescence. Similar results were obtained by depolarizing the lungs with high K⁺ in the perfusate (Table 2). Pulmonary perfusion pressure demonstrated a stable increase of approximately 1-2 cm H₂O in lungs treated with high K⁺ or K⁺-channel inhibitors.

ATP content in O₂-ventilated lungs after 1 h ischemia did not differ significantly from control perfused lungs (Table 3), indicating that membrane depolarization was not due to ATP depletion during the ischemic period. In these dextran-perfused lungs, the lack of ischemic effect on ATP content was similar to the values reported previously for BSA-perfused lungs (5). As further evidence against a role for hypoxia in the response, the increase in bis-oxonol fluorescence during ischemia with air-versus-oxygen ventilation was similar (130.4 ± 6.8% of baseline with air versus 134.7 ± 3.3% with O₂ ventilation; mean ± SE, n = 3).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The major goal of our studies is to evaluate the mechanisms for oxidant generation in ischemic but ATP-sufficient lungs. Tissue metabolism in these lungs is maintained through ventilation with oxygen-containing gas. The hypothesis is that diminished shear stress to the pulmonary endothelium with ischemia results in endothelial cell membrane depolarization which, in turn, leads to oxidant generation. Our previous studies have demonstrated oxidant generation with K⁺-induced depolarization of bovine pulmonary artery endothelial cells in culture and with the isolated perfused lung (10, 11). An association between membrane depolarization and oxidant generation has been shown previously for neutrophils and macrophages (17, 18). The present results provide evidence for membrane depolarization with cessation of perfusate flow to the isolated rat lung.

Membrane depolarization with ischemia was demonstrated in the present study using fluorescent dyes. The ability of bis-oxonol surface fluorescence to reflect membrane depolarization in the perfused lung was shown previously using varying perfusate K⁺-concentrations (11) and confirmed in this study by perfusion with high K⁺ and nonspecific K⁺-channel inhibitors. Alternative explanations for increased bis-oxonol fluorescence with ischemia are not likely because fluorescence of bis-oxonol is not sensitive to possible physiologic changes in pH and does not change due to interaction with oxidants (11). Potentially, the fluorescence signal from bis-oxonol could increase due to ischemia-induced changes in the area of lung surface as seen by the light-collection device. However, this potential artifact should result in increased fluorescence for any probe that is membrane-localized, as well as for lung tissue intrinsic fluorescence. The lack of effect of I/R on intrinsic lung fluorescence or on TMA-DPH fluorescence and the decrease in JC-1 fluorescence suggest that the increase in bis-oxonol fluorescence with ischemia did not arise from altered lung-surface area. Although fluorescence anisotropy with TMA-DPH has been used to evaluate membrane fluidity, such changes would be unlikely to occur during a brief (15-min) episode of ischemia. H₂O₂ treatment of isolated cells required 30 min to induce membrane fluidity changes (19) and we have found a similar time course for lipid peroxidation associated with lung ischemia (5). Therefore, we conclude that the fluorescence changes observed with lung ischemia are a result of membrane depolarization.

The phenomenon of cell membrane depolarization with ischemia is not unique to the lung and has been shown to occur in the brain, myocardium, and skeletal muscle (20- 22). The mechanism for depolarization in these organs, although poorly understood, may be related to altered energy state (decreased ATP) due to inhibition of oxidative metabolism. In contrast, lung ischemia with continued oxygen ventilation does not lead to ATP depletion (Table 3) and results in membrane depolarization. The evidence from this study is that depolarization in the lung (and, by inference, possibly in other organs) is attributable to the absence of flow, i.e., altered shear stress.

While we have postulated that decreased shear is responsible for membrane depolarization, it is important to exclude subtle membrane "damage" as the precipitating event. Generation of partially reduced O₂ species with subsequent oxidation of tissue components could be the initiator rather than the result of membrane depolarization. However, analysis of the time course of tissue changes indicates that membrane depolarization precedes both the onset of oxidant generation and the oxidation of lipids and proteins. In the present study, membrane depolarization was observed within 1-2 min of the onset of ischemia. Previous observations indicated a lag of approximately 2-3 min for oxidant generation, whereas products of oxidative injury (tissue thiobarbituric acid-reactive substances and protein carbonyls) were not detected until 30-60 min of ischemia (5, 14, 23). Rapid reversibility of membrane potential changes with reperfusion and the reproducibility of subsequent cycles of fluorescence changes with ischemia and reperfusion provide additional evidence that membrane damage was not responsible for the observed depolarization.

The bis-oxonol surface-fluorescence results suggest that one possible localization for the electrical potential change is the endothelial cell membrane. First, endothelial cells in situ are directly exposed to shear related to the mechanical effects of flow/no-flow. Second, flow-dependent changes in bis-oxonol fluorescence are compatible with shear stress-induced hyperpolarization observed with endothelial cells in culture (24). Third, the rapid and almost complete removal of lung-associated bis-oxonol with high-affinity ligand (BSA) indicates its ready availability to the perfusate, suggesting localization primarily to the endothelium. Finally, a major role of the endothelial cell in contributing to the bis-oxonol fluorescence-signal change with ischemia was demonstrated by our intravital microscopic study of subpleural microvessels in the intact perfused lung. Although we cannot exclude a contribution from other cell types, it would appear that the capillary endothelium is the major source of the observed fluorescence signals. It is also possible that other cell types undergo ischemia-mediated membrane depolarization but are not detected by our fluorescence system.

Our previous observation that maintaining a flow rate of 1 ml/min in the perfused rat lung was protective against ischemia-mediated lipid peroxidation (5) suggested the presence of a threshold of flow rate, and hence of depolarization, for oxidant generation by the endothelial cell. For the rat pulmonary artery (assuming an internal diameter of 1 mm), a flow rate of 7 ml/min used as the control in this study would generate shear stress of about 9 or 16 dyn/cm² for a solution of 3% albumin or 3% dextran, respectively. The shear stress is calculated for laminar flow in idealized cylindrical tubes using the following equation (25) and parameters:

where  represents shear stress in dyn/cm²; µ is the coefficient of viscosity in Poise (in these calculations, µ for water at 37°C, 0.006915 Poise, times the viscosity of 3% albumin or 3% dextran [26] has been used); Q is the flow rate in ml/s; and r is the internal radius of the tube in centimeters. For a capillary of 25-µm diameter and dispensing as little as 0.001% of the flow, the shear-stress magnitude would be 6 and 10 dyn/cm² for albumin and dextran perfusates, respectively. Endothelial shear-stress sensitive K⁺-channels (I_K,S) have been described and were shown to have a broad activation range of 0.2-17 dyn/cm² (13). Hence, even a flow rate of 1 ml/min in this model would generate enough force (estimated at 0.9 and 1.4 dyn/cm² for albumin and dextran perfusates, respectively) to keep these capillary endothelial channels in the activated state, and thus maintain endothelial cell membrane potential. Our experiments show that maximum increase in bis-oxonol fluorescence occurred when flow rate was reduced from 1.8 to 0 ml/min as opposed to a modest increase (28% of maximum) with a reduction of flow from 13 to 1.8 ml/min. This suggests a threshold of shear stress for membrane depolarization that lies below a flow rate of 1.8 ml/ min in the rat lung (~ 1.5 g wet weight). This flow rate represents about 10% of the normal resting cardiac output for a 170-g rat. With low flow rates, it is possible that flow through some capillaries is below the threshold, resulting in focal depolarization and submaximal change in fluorescence. The present results indicating a threshold for depolarization might explain the previously observed protection from oxidative damage with low-flow (1 ml/min) rat lung perfusion (5).

In summary, implication of mechanotransduction forms a novel approach for our understanding of the pathophysiology of lung ischemic injury. We have shown that lung ischemia leads to pulmonary endothelial cell membrane depolarization, and have shown previously that ischemia leads to oxidant generation. This suggests that oxidant generation is initiated as a consequence of membrane depolarization, although a cause-and-effect relationship and the mechanism for coupling remain speculative.