INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although lung transplantation has been shown to be an effective therapy for end-stage lung disease, its widespread application has been limited by a shortage of suitable donors. We have hypothesized that the lung may remain viable after circulatory arrest and death for periods of time sufficient to retrieve the organ and transplant it (1). Thus, transplantation of lungs retrieved from non-heart-beating donors (NHBDs) could significantly expand the donor pool.

We have shown that 90% of parenchymal cells are viable in rat lungs ventilated with oxygen 4 h postmortem (2). In addition, we have demonstrated preservation of pulmonary ultrastructure (3) and a marked attenuation in the time-dependent decrement of lung high-energy phosphate stores in O₂-ventilated cadaveric rat lungs (2). These studies have focused on the state of cadaveric lung tissue before reperfusion, but they do suggest that the lung is viable for intervals of time after circulatory arrest.

Using an isolated perfused rat lung model, we investigated the relationship, after reperfusion, between postmortem lung viability and lung function, as measured by the capillary filtration coefficient (Kfc) (4). Lungs retrieved up to 1 h postmortem had normal Kfc values if cadavers were ventilated with O₂. Nonventilated cadavers had significantly increased Kfc values and decreased viability as the postmortem time interval increased. This study suggested that integrity of the pulmonary endothelial surface is maintained after short periods of ischemia in O₂-ventilated rat lungs, but that longer ischemic times result in increased microvascular permeability.

Elevated endothelial permeability has been associated with decreased levels of adenosine 3',5'-cyclic monophosphate (cAMP) (5). Studies have suggested that ischemia-reperfusion injury (IRI) to the lung can be attenuated by agents that increase intracellular concentrations of cAMP (8). Seibert and colleagues have shown a reversal of increased microvascular permeability associated with lung IRI when cAMP is directly or indirectly activated pharmacologically (11). Other groups have shown decreased lung wet-to-dry ratios and capillary permeability when lungs were reperfused with the cAMP analog dibutyryl cAMP (12).

Isoproterenol (iso) is a direct ₂-adrenergic receptor agonist that activates adenylate cyclase to increase intracellular cAMP. Rolipram (roli), an isozyme-selective cAMP phosphodiesterase inhibitor (type IV), prevents breakdown of cAMP. Several studies have demonstrated decreased pulmonary capillary IRI with phosphodiesterase inhibitors such as pentoxifylline, aminophylline, and roli, all of which theoretically increase intracellular cAMP levels (8, 12, 13). In addition, the use of roli in rat lung flush solution resulted in improved lung function after 6 h of hypothermic preservation (14). We have shown that the addition of iso alone to the reperfusate solution in an isolated perfused rat lung model significantly elevates tissue cAMP levels and reduces Kfc in rat lungs from NHBDs (15). The addition of roli in the same model had limited effects on cAMP but did significantly reduce Kfc in NHBDs after 2 h of postmortem ischemia (16). We reasoned that the combination of these pharmacologic agents may further augment tissue cAMP levels and further attenuate IRI-induced changes in Kfc.

We hypothesized that elevation and maintenance of tissue cAMP levels, using a combination of iso and roli, might act synergistically to ameliorate the increased pulmonary capillary permeability and IRI seen in NHBDs beyond the beneficial effects of administration of each agent separately. In addition, we sought to investigate the relationship between O₂ ventilation of the NHBD lung and augmentation of cAMP to Kfc.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolated Perfused Lung

The isolated perfused lung preparation was first used to measure the filtration coefficient in dogs by Gaar and coworkers (17) and later modified for use in rats by Fujimoto, Taylor, and colleagues (18, 19). This model provides a sensitive and reproducible method with which to assess alterations in permeability of the pulmonary microcirculation, the critical initial phase of lung IRI. The specific details of this preparation have been outlined previously (20).

Briefly, male Sprague-Dawley rats weighing 250-450 g were anesthetized intraperitoneally with pentobarbital sodium (35 mg/kg) (Abbott Laboratories, Chicago, IL). A small laparotomy incision was made, and 600 U of heparin (Elkins-Sinn, Cherry Hill, NJ) was injected intrahepatically under direct vision. The trachea was cannulated with 6-8 cm of p60 tubing. The rats were killed with an intrahepatic injection of pentobarbital sodium (120 mg/kg). The cadaver rats were either ventilated (Harvard rodent ventilator model 683; Harvard Apparatus, Millis, MA) with 100% O₂ at 60 breaths/min, a tidal volume of 3 cm³, and a positive end-expiratory pressure of 2 cm H₂O (O₂), or not ventilated (NV). The heart-lung block was left in situ in an effort to simulate the NHBD clinical situation as closely as possible.

At various intervals after death, the lungs were removed and reperfused in the isolated perfused rat lung model with a peristaltic pump (Minipuls 3; Gilson Medical Electronics, Middleton, WI) at a constant flow of 0.03 ml/g body weight. The perfusate was Earle's balanced salt solution (containing [in mmol] 2.4 CaCl₂·2H₂O, 0.4 MgSO₄ [anhydrous], 5.4 KCl, 116 NaCl, 0.88 NaH₂PO₄ [anhydrous], 5.5 D-glucose, and 0.3 phenol red) containing 0.21% NaHCO₃ and 4% bovine serum albumin (Sigma, St. Louis, MO) with or without the addition of iso (10 µM) and roli (2 µM). The initial 75 ml of perfusate, which contains residual red blood cells and plasma, was discarded. An additional 40 ml of perfusate placed in a water-jacketed reservoir was used for recirculation. The perfusate temperature was maintained between 35 and 38.5° C, and the perfusate pH was continuously monitored with a pH probe (Accumet; Fisher Scientific, Pittsburgh, PA) placed in the venous reservoir. The pH was maintained near 7.40 by adding dilute HCl or NaHCO₃ as necessary.

Kfc Determination

Pulmonary capillary pressure (Ppc) was estimated by the double-occlusion technique described by Townsley (21). Simultaneous occlusion of arterial and venous catheters results in equilibration of pulmonary artery pressure (Ppa) and pulmonary vein pressure (Ppv) to the same pressure. This equilibration equals the Ppc and also reflects the capillary pressure when the lung is not isogravimetric. After the lungs reached an isogravimetric state, the venous reservoir was rapidly elevated to increase Ppv by 6 to 8 cm H₂O for 15 min. The initial 3 to 5 min of weight gain represents vascular distention and recruitment and is not a reflection of capillary permeability. The w/t between minutes 6 and 15 represents increased transvascular fluid flux secondary to increased capillary permeability.

The Starling equation describes the role of Kfc in transvascular fluid flux (22): Jv = Kfc[(Pc  Pi)  (c  i)], where Jv is transvascular fluid flux, Kfc is the filtration coefficient, P is hydrostatic pressure in the capillary (Pc) and interstitium (Pi), is osmotic pressure in the capillary (IIc) and interstitium (IIi), and  is the osmotic reflection coefficient. At the extrapolated time 0, both capillary pressure and fluid flux are elevated to new steady states before the remaining factors can be effected. Therefore, Kfc can be calculated by the equation: Kfc = Jv/Pc.

Kfc is calculated by dividing w/t at time 0 by the change in Ppc that occurs after Ppv elevation. It is normalized using baseline wet lung weight and expressed as ml/min/cm H₂O/100 g lung tissue. At the end of perfusion, the right upper lobe was excised and immediately weighed. It was then dried in a 60° C oven for 48 h and reweighed to allow adenine nucleotide levels to be expressed per gram dry weight.

Lung Parenchymal Cell Viability

After excision of the right upper lobe, the right hilum was suture ligated. Right lung pieces were flash-frozen in liquid nitrogen and stored at 70° C. Thirty milliliters of a 500 mM trypan blue solution (Sigma), dissolved in Krebs-Heinseleit buffer (pH 7.4), was infused into the left pulmonary artery via the existing catheter. Trypan blue stains the nuclei of nonviable cells (23). The infusion reservoir was positioned 30 cm above the heart. After infusion of the trypan blue, 30 ml of fixative consisting of 2% glutaraldehyde, 2% paraformaldehyde in 0.1 M Sorenson's buffer was infused from the same reservoir. During both infusions, mechanical ventilation with 100% O₂ was performed. The left lung was then excised, placed in the same fixative, and stored at 4° C.

The left lung tissue was prepared for histologic analysis by standard techniques. Briefly, the tissue was dehydrated in ethanol, washed in xylene, and embedded in paraffin. Five-micron sections were cut, mounted on slides, and counterstained with eosin only. Cell viability was determined microscopically (Nikon, Melville, NY), using ×1,000 magnification with oil immersion and an ocular grid. The microscopist was blinded to the experimental group for the analysis. Twenty-five parenchymal cell nuclei were identified in each quadrant and recorded as either viable (pink) or nonviable (blue). Each slide was counted twice at different intervals. If a greater than 10% difference existed between counts, a third and final count was performed. Lung parenchymal cell viability was reported as the percentage of viable cells.

High-performance Liquid Chromatography

Tissue samples previously retrieved from the right lung for high-performance liquid chromatography (HPLC) were pulverized with a liquid nitrogen-cooled Bessman pulverizer and then homogenized with ice-cold 0.6 N perchloric acid (2-8 ml/g tissue), using a Tissue-Tearor (BioSpec Products, Bartlesville, OK) at 30,000 rpm for 30 s. After centrifugation for 2 min at 10,000 rpm, the supernatant was removed and neutralized with cold 1 M potassium phosphate dibasic (pH 12) to achieve a pH of 6.8. The supernatant was separated from precipitated salt by repeat centrifugation for 2 min at 10,000 rpm. The remaining solution was passed through a 0.45-µm pore size Acrodisc filter (Pall Gelman, Ann Arbor, MI).

Adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), and cAMP concentrations were determined by HPLC. Twenty microliters of each sample was analyzed with a Beckman System Gold apparatus (Beckman Instruments, Fullerton, CA). The column (15.0 cm × 4.6 cm id.; Supelco, Bellefonte, PA) was packed with LC-18-T 3-µm pore size silica particles and the analyses were performed under gradient elution with 0.15 M ammonium phosphate monobasic (pH 6.0) to 15% methanol-acetonitrile (50%, v/v) as elutants. A flow rate of 1.0 ml/min was used, and detection was performed by ultraviolet (UV, 254 nm) light absorbance. Chromatograms were analyzed on an IBM 486 DX computer with Gold Nouveau software (Beckman). Standard curves were made by analysis of serial dilutions for ATP, ADP, AMP, and cAMP (Sigma).

Specific Protocol

Eighty pairs of lungs were divided into 10 groups (n = 8/group; Table 1). Retrieval of lungs was performed 60 and 120 min postmortem with O₂ventilation (O₂) or without ventilation (NV). The same groups were reperfused with reperfusate with or without the addition of iso/ roli, as described.

Control lungs were extirpated immediately and reperfused with or without iso/roli within 5 min of death. All lungs were allowed to equilibrate, after initial reperfusion, for 15-20 min to achieve an isogravimetric state. Lungs that could not reach an isogravimetric state were discarded and the experiment was repeated. During equilibration, the Ppa, Ppv, and Paw were measured and recorded every minute. After equilibration, the Ppc was obtained, and the Kfc was measured. The left lung was then perfused with trypan blue to determine cell viability, and portions of the right lungs were frozen or weighed to calculate the wet-to-dry ratio as described above. All animals received humane care in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication 85-23, revised 1985).

Statistics

All results are expressed as means ± standard error. Comparisons between groups were made by analysis of variance with the Fisher post-hoc test for multiple comparisons. Significance was determined to be present when p < 0.05. Linear correlations were obtained with the Pearson correlation coefficient, and their significance was determined by paired t test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Capillary Filtration Coefficient Kfc

Changes in Kfc are shown in Figure 1. Kfc increased as the postmortem ischemic time increased. Sixty-minute cadaver lungs did not differ significantly from controls. Reperfusion with iso/roli did significantly attenuate the increase in Kfc in lungs retrieved after 120 min of postmortem ischemia in both O₂-ventilated and nonventilated cadavers (p < 0.03).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Capillary filtration coefficient related to postmortem ischemic time interval for each group (n = 8/group). Data are presented as means ± standard error of the mean. iso/roli = iso (10 µM) and roli (2 µM); *p < 0.05 iso/roli reperfusion versus non-iso/roli reperfusion.

Viability

Parenchymal cell viability was measured by the trypan blue exclusion technique. There were no significant differences in viability after 60 min of postmortem ischemia with or without O₂ ventilation or with the addition of iso/roli to the reperfusate. Reperfusion with iso/roli resulted in significantly improved viability in 120-minute cadaveric lungs that were not ventilated (p < 0.01) or O₂ ventilated (p = 0.013) compared with non-iso/ roli-reperfused lungs (Figure 2). There is a linear correlation in all groups between viability and Kfc (r = 0.92, p < 0.001).

View larger version (24K): [in this window] [in a new window]   Figure 2.   Percent viability of lung parenchymal cells at each postmortem ischemic time interval (n = 8/group). Data are presented as means ± standard error of the mean. iso/roli = iso (10 µM) and roli (2 µM); *p < 0.05 iso/roli reperfusion versus non-iso/roli reperfusion.

Hemodynamics

Hemodynamic data are shown in Table 2. There were no significant differences between groups with respect to Paw, Ppv, or Ppa. Pulmonary venous resistance (Rv) was not significantly different from baseline controls at any postmortem ischemic time intervals regardless of ventilation or the addition of iso/roli to the reperfusate. Conversely, pulmonary arterial resistance (Ra) was significantly increased above controls (postmortem time = 0) in 60-min O₂ and 120-min NV lungs that were reperfused with iso/roli.

Adenine Nucleotides and cAMP

Lung tissue levels of ATP, ADP, and AMP, and the total adenine nucleotide (TAN) level, are presented in Table 3. TAN levels were determined by the formula: TAN = ATP + ADP + AMP. TAN levels decreased with increasing postmortem ischemic times in all groups (r = 0.85). Iso/roli-reperfused lungs maintained significantly higher TAN levels than non-iso/roli-reperfused lungs at all time points beyond controls and had a strong correlation to Kfc (r = 0.97, p = 0.007; Figure 3).

View larger version (11K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 3.   The relationship between TAN and Kfc in iso-roli-reperfused non-heart-beating lungs is shown. (A) For group means ± standard error of the mean (r = 0.97, p < 0.01). (B) For all data pairs (r = 0.4, p < 0.01).

Similarly, ATP was significantly higher in iso/roli-reperfused groups, and analysis of TAN levels showed that attenuation of ATP levels was critical for maintaining elevated TAN levels. ATP levels remained near control levels after 60 min of postmortem ischemia when ventilated with O₂ (Table 3). In addition, ATP demonstrated a strong correlation between viability (r = 0.88) and Kfc (r = 0.92) for all groups (Figure 4).

View larger version (10K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 4.   Relationship between ATP and percent cell viability. (A) For group means ± standard error of the mean (r = 0.88, p < 0.002). (B) For all data pairs (r=0.63, p < 0.001).

Levels of cAMP showed a significant 2-fold increase in iso/ roli-reperfused cadaveric lungs compared with control and non-iso/roli-reperfused lungs (Table 3). Elevation of cAMP occurred regardless of the postmortem ventilation and was evident at all experimental ischemic times analyzed, including controls. cAMP correlated with Kfc for iso/roli-reperfused lungs (r = 0.93; Figure 5).

View larger version (10K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 5.   Relationship between cAMP and Kfc in iso/roli-reperfused non-heart-beating lungs is shown. (A) For group means ± standard error of the mean (r = 0.93, p = 0.015). (B) For all data pairs (r = 0.48, p < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Interest is growing in the potential use of organs from circulation-arrested cadavers or NHBDs to alleviate the critical shortage of available donors for lung transplantation. The finding that bronchial epithelial cells could be successfully cultured from lung tissue taken from morgue specimens was one of the first observations suggesting that tissue in lungs retrieved from NHBDs is viable (24). In NHBD, an obligatory period of warm ischemia and subsequent reperfusion will likely cause tissue damage. The time period of postmortem ischemia wherein lung function would be acceptable for subsequent transplantation is unknown. Defining this allowable time range and modifying preservation techniques may allow NHBD lungs to become a viable option for lung transplantation.

Although it has been shown that ischemic but nonreperfused lungs from non-heart-beating rat lungs have evidence of preserved viability and adenine nucleotide stores (2), we have only recently correlated lung parenchymal cell viability to lung function after reperfusion (20). We have found that agents that stimulate the production or reduce breakdown of tissue cAMP levels reduce IRI in NHBD rat lungs (15, 16). Both iso- and roli-enhanced reperfusate in NHBD rat lungs attenuated the IRI, as measured by Kfc, up to 120 min after death. In the iso-reperfused lungs, this beneficial effect was associated with a dramatic increase in cAMP. Roli alone had similar effects but did not elevate cAMP compared with controls (16). Postmortem ventilation was beneficial to retention of ATP levels and cell viability, but O₂ ventilation and the addition of iso or roli to the reperfusate was not synergistic (15, 16).

These promising studies were the impetus for the current study, in which reperfusate modification using iso and roli in combination was done to determine whether both agents provided a synergistic effect on capillary permeability and endothelial viability as the postmortem ischemic interval increased.

The isolated perfused rat lung model used in this study is different from other isolated perfused lung models. Lungs were left in situ after the animal was killed, instead of the usual protocol of harvest with immediate perfusion, followed by ischemia. This modification of the isolated perfused lung model most closely resembles the clinical scenario of the non-heart-beating donor, which is our interest.

Capillary permeability, assessed by Kfc, increased as the postmortem ischemic time interval increased. Whereas the increase was slightly attenuated by ventilating the cadaver with O₂, it was significantly ameliorated up to 120 min of postmortem ischemia when lungs were reperfused with a combination of iso and roli. In prior studies, either agent alone had a similar effect on Kfc. Interestingly, O₂ ventilation in addition to reperfusion with iso/roli was not synergistic with respect to further decreasing Kfc. After 60 min of postmortem ischemia, Kfc was lower in O₂-ventilated lungs compared with nonventilated lungs. However, this result did not reach significance (p = 0.59). Barnard and coworkers studied the effect of rolipram in a different model of ischemia/reperfusion rat lung injury (25). After initiating perfusion of a rat lung block with crystalloid, ischemia was induced for 45 min. In this setting, Kfc increased 2-fold, and this increase could be completely prevented if lungs were reperfused with 2 µM roli or roli (95 nM) and iso (0.9 nM) in combination. This period of ischemia was shorter than in our study and caused only a modest increase in Kfc.

Lung parenchymal cell viability was decreased at all postmortem ischemic times compared with controls. Preharvest O₂ ventilation did not appreciably increase the percentage of viable cells compared with nonventilated lungs after 60 or 120 min of postmortem ischemia. Similarly, iso/roli reperfusion had little effect on viability except for significant attenuation of cell death in 120-min ischemic lungs that were not ventilated. These findings are in contrast to lung viability studies before reperfusion in cadaver lungs. Prior studies have found that O₂ ventilation of the ischemic lung was a critical factor in delaying cell death in nonperfused lungs (2, 3).

In a prior study, trypan blue exclusion was used as a measure of lung cell viability in heparinized rats at intervals after circulatory arrest and death (26). In nonventilated lungs retrieved 12 h postmortem, 80% of the cells took up the dye, implying widespread distribution of the dye using this protocol. Thus, it is unlikely that "no reflow" to regions of lung could account for the observed differences in viability.

This study identified an important correlation between Kfc and total lung adenine nucleotides. Iso/roli-reperfused lungs also had increased TAN levels compared with non-iso/roli-reperfused groups, regardless of the preharvest ventilation status. These differences were mainly due to tissue levels of ATP. Irrespective of whether the lungs were reperfused with iso/roli, Kfc showed a strong linear correlation with TAN levels. This is intriguing because it implies that augmentation or retention of adenine nucleotide levels before or after ischemia may also directly affect capillary membrane integrity.

The current study demonstrated that iso/roli reperfusion markedly increased tissue cAMP levels and was inversely proportional to the capillary permeability. However, the addition of iso/roli to the reperfusate was not synergistic in terms of elevating cAMP beyond levels established with iso alone, nor did this combination reduce Kfc below those levels achieved by iso or roli alone after 120 min of postmortem ischemia. It is likely that iso and roli are acting by the same mechanisms to reduce Kfc while maintaining capillary integrity. We suspected from our previous studies that elevation of tissue cAMP with iso and subsequent inhibition of hydrolysis with roli could increase localized levels of cAMP. No additive response was observed beyond the similar reductions in Kfc observed with the use of each agent alone (15, 16). These data suggest that maintenance of threshold values of localized cAMP may potentially be more important for endothelial function and vascular homeostasis.

The actual mechanism of benefit by cAMP-elevating agents in this model of lung injury is unknown. Pulmonary vascular endothelial cells contain actomyosin fibrils (27), which contract when exposed to compounds and conditions that increase permeability. Disruption and rearrangement of these cytoskeletal elements may result in the formation of pores between endothelial cells and are thought to be a source of increased cell permeability (28). Relaxing the endothelial cytoskeletal contractile proteins increases cell-to-cell contact, surface area, and apposition. It may also stabilize focal adhesion contacts of actin fibrils, maintaining intracellular structure as well (7). Therefore, one action of cAMP may be related to microfilament-induced alterations in the endothelial cell cytoskeleton (11). Treatment of cultured endothelium with isoproterenol and other phosphodiesterase inhibitors enhances endothelial barrier function (6). It is possible that the decreased capillary permeability after iso/roli reperfusion seen in this study may involve a cAMP-dependent endothelial cell relaxation. Alternatively, Pinsky and colleagues have suggested that hypoxia-induced reduction in endothelium cAMP levels may affect the release and function of various cytokines (29). cAMP analogs have been shown to prevent tumor necrosis factor-induced decline in endothelial cell thrombomodulin expression. Ischemia/reperfusion lung injury has been shown to be associated with elevated levels of tumor necrosis factor  (TNF-) in the perfusate of an isolated perfused rat lung model, with amelioration of injury in rats pretreated with an anti-TNF- antibody (30), implicating this cytokine in the pathogenesis of capillary leak in response to ischemia and reperfusion. Further studies into exact mechanisms of action of cAMP and other agents that reduce capillary leak are necessary to develop a reliable method for preserving NHBD lungs for transplantation.

The duration of the elevated tissue cAMP levels after iso/ roli reperfusion was evaluated for only approximately 30 min, the time required to perform Kfc determination. The current study does not address whether these observations are irreversible or constant over time. This will be important to ascertain, because if cAMP levels begin to fall and capillary permeability increases, then treating the transplanted lung with iso, roli, or other agents may have merit. Finally, although the impact of the combined effects of reperfusion with iso/roli on Kfc was elucidated in this study, its effect on other measures of lung function after transplantation, such as oxygenation and ventilation, was not evaluated.

Increased capillary permeability of lungs retrieved from NHBDs can be ameliorated for up to 120 min of postmortem ischemic time if the lungs are reperfused with an iso/roli-modified solution. The mechanism through which this occurs in this model appears to involve an increase in intracellular cAMP concentrations, possibly through a cAMP-dependent mechanism that attenuates cell death and loss of ATP. This study provides further evidence that pulmonary capillary permeability due to ischemia-reperfusion injury is linked to lung tissue cAMP levels.