INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung transplantation is the accepted therapy for end-stage pulmonary disease. In the immediate postoperative period; however, severe graft dysfunction occurs in 10 to 20% of recipients (1), resulting in the need for prolonged intensive care and accounting for 12 to 14% of early postoperative deaths. Furthermore, recent studies have shown that human leukocyte-associated antigen (HLA) matching has a significant effect on the outcome of lung transplant (2), and such matching would be more feasible if it were possible to preserve the lung safely for longer periods. Consequently, new strategies are required to reduce postischemic lung damage.

Concentrations of cyclic adenosine monophosphate (cAMP) in the lung decrease during ischemia (3), and measures that reduce this fall protect against ischemia-reperfusion injury (4). Intracellular cAMP concentrations are maintained by the methylxanthines, a class of drugs that operate by phosphodiesterase (PDE) inhibition (5). Of these, pentoxifylline protects isolated rat lungs against ischemia-reperfusion and human neutrophil-mediated injury (6, 7) and it reduces lung allograft reperfusion injury in dogs (8, 9). Moreover, pentoxifylline exerts significant anti-inflammatory activity even when present only in the cell-free flush solution (10). In addition, theophylline, a related methylxanthine, improves rat lung function in a dose-dependent manner (11) when added to bicarbonate buffer used for flushing and 6 h storage of rat lungs.

A variety of PDE isoenzymes have been identified and differential tissue activities of isoenzyme selective inhibitors described (12). The predominant PDE isoenzymes of therapeutic interest in the lung are the PDE III and IV subtypes (15). PDE III inhibitors selectively relax smooth muscle whereas PDE IV inhibitors reduce neutrophil activation. Pentoxifylline and theophylline are relatively nonselective PDE inhibitors, showing similar affinity for subtypes I, III, IV, and V (16) but are of lower potency than the newer, more selective agents. Rolipram, for example, is 20-fold more potent an inhibitor of PDE IV than pentoxifylline. Furthermore the beneficial effects of methylxanthines may be related to their action on other systems such as the adenosine receptor (17).

It has been reported that the nitric oxide donor, glyceryl trinitrate, which activates the cyclic guanosine monophosphate (cGMP) pathway, is more effective than prostacyclin, which activates cAMP-mediated pathways, in protecting rat lungs during 6-h hypothermic storage in EuroCollins solution (18). One potential approach to clarifying the relative roles of cAMP and cGMP in protection of the lung would be to compare the effect of the PDE inhibitors described previously with an agent such as zaprinast which selectively inhibits the cGMP-selective PDE V isoenzyme subtype (16).

We therefore hypothesized that comparative studies using specific PDE isoenzyme inhibitors might provide insights into the mechanisms of the beneficial effects of these drugs in ischemia-reperfusion injury and allow the identification of agents for improving lung preservation solutions. To test this proposal, we undertook studies in which we compared the nonselective PDE inhibitor, theophylline, to the specific PDE III subtype inhibitor, Milrinone, the specific PDE IV subtype inhibitor, rolipram, and the specific PDE V subtype inhibitor, zaprinast.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

Rolipram was a gift from Schering AG, Berlin, Germany. Milrinone was a gift from Sanofi Winthrop Ltd, Newcastle upon Tyne, UK. All other chemicals used in perfusate and storage solutions were supplied by BDH Ltd, Leicestershire, UK. Pentobarbitone was purchased from Rhone Merieux, Harlow, UK. Myeloperoxidase (MPO) was from Sigma, Poole, Dorset, UK.

Lung Preparation

Lungs were obtained from male Wistar rats weighing 250 to 330 g. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research (U.S.A.), the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources, published by the National Institutes of Health (U.S.A.) and with the "Guidance on the Operation of the Animals (Scientific Procedures) Act 1986" published by Her Majesty's Stationery Office, London, England.

Rats were anesthetized by intraperitoneal injection of pentobarbitone (2 ml/kg of a 60 mg/ml solution). After tracheal intubation, animals were ventilated at 80 breaths/min by positive pressure applied by a Harvard Small Animal Ventilator. After laparotomy, the diaphragm was removed and heparin (500 international units [IU]) injected into the vena cava. The animals were then exsanguinated by withdrawal of blood from the inferior vena cava and the thorax opened. A cannula was inserted into the pulmonary artery and held in place by tightening a ligature previously placed around the pulmonary artery and aorta. Next, a cannula was secured in the left atrium to receive the perfusate leaving the lungs. The lungs were then removed and suspended in a sealed, water-jacketed chamber maintained at 37° C. Perfusion was commenced with modified bicarbonate buffer (BB, composition mmol/L: NaCl 118.5, KCl 3.8, KH₂PO₄ 1.2, NaHCO₃ 25.0, CaCl₂ 2.0, MgSO₄ 1.2, glucose 10.0) containing whole rat blood (that drawn from the vena cava) mixed in a ratio of 4:1 to produce sanguineous BB (SBB). We have previously demonstrated the superiority of this sanguineous perfusate over asanguineous BB supplemented with bovine serum albumin used in an earlier study (11, 19). Despite the need for a second animal to provide blood for the reperfusion buffer, this model also retains the advantage of economy of animal usage over perfused lung systems employing a support animal where a third animal is normally employed to supply blood to prime the perfusion circuit. Additionally, the necessity of maintaining the support animal and the potential complication of humoral interactions between the isolated lung and the support animal are both avoided. The perfusate flow rate was adjusted to 15 ml/min and maintained at this value by a peristaltic pump (Watson Marlow 501; Falmouth, Cornwall, UK). The SBB was held in a heated reservoir (37° C) and gassed with 100% CO₂ whenever it exceeded pH 7.3; this was achieved by a feedback mechanism whereby the output of a pH meter controlled a solenoid-operated valve on tubing supplying CO₂ to the perfusate reservoir. In this way the pH of the perfusate was prevented from going above 7.4 as a result of the venting of this gas by the ventilated lung. The perfusate was also gassed with 100% nitrogen at this stage. Buffer leaving the lungs via the left atrial cannula was returned to the reservoir and recycled. Oxygenation of the perfusate was by the isolated lungs, which were ventilated with room air.

The peak tracheal pressure (Pt) applied by the ventilator was set to give a tidal volume (VT) of 2.0 to 2.3 ml (0.7 ml/100 g animal weight) and a positive end-expiratory pressure of 1 to 2 cm H₂O was applied. Pressure transducers were connected to the tracheal, arterial, and venous cannulae. Pt and VT, calculated by integration of the flow (dv/dt) through a pneumotachograph connecting the trachea to the ventilator, were measured. Pulmonary static compliance (Cstat) and airways resistance (Raw) were calculated by multiple linear regression according to the equation Pt = 1/Cstat · VT + Raw · dv/dt + I · d²v/dt² (20), where I = inertia and d²v/dt² is the differential of the flow. Pt, VT and dv/dt were sampled at 30 points in each breath and Cstat and Raw calculated over a 10-breath period for each timepoint examined. The difference in pressures between the pulmonary artery and venous cannulae divided by the perfusate flow rate, measured vascular resistance. The output of two flow-through oxygen electrodes (LazarLabs, Los Angeles, CA) placed in the perfusion circuit before the pulmonary artery cannula and after the left atrial cannula allowed determination of the gas-exchanging capacity of the perfused lung. All outputs from pressure transducers, pneumotachograph, and the oxygen and pH electrodes were recorded using a MacLab 8s analogue:digital converter connected to a PowerMacintosh computer employing the MacLab Chart software (AD Instruments Ltd, Hastings, UK).

Experimental Protocol

Lungs were perfused for an initial 20-min (control) period, during which time control lung function parameters were measured. By switching a three-way tap to open a reservoir, the lungs were then flushed with St. Thomas' Hospital cardioplegic solution No. 2 (STH; composition in mmol/L: NaCl 110, KCl 16.0, MgCl₂ 16.0, CaCl₂ 1.2, NaHCO₃ 10.0, pH 7.8 at 37° C) or STH containing rolipram (3, 10, 30, or 100 µM), milrinone (10, 30, or 100 µM), zaprinast (10, 30, or 100 µM), or theophylline (300, 1,000, or 3,000 µM), at a pressure of 30 cm H₂O. Where necessary the pH of the STH was adjusted to its original value after addition of PDE inhibitor. The doses of the various drugs were chosen so that the maximal dose studied was 30- to 100-fold greater than their published inhibitory concentration of 50% (IC₅₀) values against their respective PDE isoenzyme subtype (16), to ensure complete inhibition of the relevant subtype. Lungs were initially flushed with 10 ml of this solution at room temperature (20 to 25° C) to reduce cold-induced vasoconstriction caused by sudden infusion of storage solution at 4° C (21). After this initial 10-ml perfusion, flushing was continued with a further 20 ml of the same storage solution at 4° C. Control (no storage) lungs were perfused for a further 40 min after the initial 20-min period, making a total of 60 min continuous perfusion.

The flushed lungs were stored inflated with the vasculature open, immersed in the storage solution at 4 to 6° C. Lung inflation was achieved by attaching a syringe to the tracheal cannula after removal of the lungs from the perfusion chamber and injecting 2.5 ml of air; the trachea was then tied off to keep the lungs inflated.

After 8 h, lungs were removed from the storage solution, reattached to the perfusion circuit, and reperfusion with SBB (at 37° C) was instituted for a 40-min period. Reperfusion was at a rate of 15 ml/min, this rate being achieved within 1 min of reattachment of vascular cannulae. During reperfusion, lung function parameters were measured.

Determination of Wet:Dry Weight Ratios

At the end of the 40-min reperfusion period, the lungs were removed from the perfusion chamber and the left and right lungs separated. The right lung was frozen in liquid nitrogen and then stored at 80° C until subsequent determination of its MPO content; the left lung was blotted free of surface liquid, then weighed before being placed in an 80° C oven for 24 h, at which time the dry weight was obtained.

Determination of Lung MPO

The tissues were prepared for assay according to the method of Okabayashi and coworkers (8) and the assay performed as described by Bloomfield and coworkers (22). Briefly, 100 mg of tissue from the periphery of the lower right lobe of each lung was homogenized in 1 ml of 0.5% hydroxyethyltetraammonium bromide (HETAB) in 50 mM phosphate buffer (pH 6.2). The homogenate was then centrifuged for 20 min at 10,000 g and the supernatant assayed for MPO and protein content.

For MPO measurements, 50 µl of sample were added to 50 µl of o-dianisidine (0.025% in phosphate buffer with 0.5% HETAB) in the wells of a micrometer plate. The reaction was started by addition of 50 µl H₂O₂ (0.01% in phosphate buffer) and the increase in optical density at 510 nm measured over 3 min. This value was used to calculate the number of units of MPO present by interpolation with a standard curve established using dilutions of a commercially available preparation of human neutrophil MPO (Sigma Myeloperoxidase: 1 unit is defined as causing an increase in optical density of 1 per minute at pH 7 at 25° C with guaiacol substrate). Protein in the supernatant was then determined according to the method of Smith and coworkers (23) and MPO concentration expressed as units/mg protein.

Determination of Lung Cyclic Nucleotide Content

To determine whether the PDE inhibitors studied were effective in maintaining tissue cAMP or cGMP concentrations during hypothermic storage, we examined the concentrations of these nucleotides in lungs immediately after 20 min perfusion with SBB and 8 h hypothermic storage in STH or STH containing rolipram (100 µM), milrinone (100 µM), zaprinast (100 µM), or theophylline (3,000 µM). For comparison, cyclic nucleotide concentrations were determined in lungs after 20 min perfusion in SBB. After each experimental period, lungs were immediately frozen in liquid nitrogen and then stored at 80° C until assay. Just prior to assay, frozen lung samples were thawed and homogenized in ice-cold phosphate buffer containing 5 to 10% trichloroacetic acid (TCA) (0.1 g lung tissue/ml buffer). The protein precipitate was then removed by centrifugation (1,500 × g, for 10 min) and the supernatant extracted 3 times with water-saturated ether. The resulting extract was dried and resuspended in 1 ml phosphate buffer. Lung cAMP or cGMP contents were then measured using a commercially available immunoassay kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer's instructions. After solubilization with 1 N sodium hydroxide, the protein content of the TCA precipitated pellet was determined according to the method of Smith and coworkers (23) and cyclic nucleotide concentration expressed as pmol/mg protein.

Statistics

Data are displayed as mean ± SEM, with 5 to 6 animals/group. To compare the effects of the various treatments on lung function over the time course of reperfusion, trapezoid integration was used to calculate the area under the time-response curve for each parameter for each animal. These individual values were then employed for statistical comparisons of the various groups. Comparisons between groups were carried out by one-way analysis of variance (ANOVA) and, if this revealed significant differences, Dunnett's test was used to compare multiple values with control lungs, unless initial testing revealed the data deviated significantly from normality in which case a nonparametric ANOVA and Dunn's test were applied. Where several values were to be compared with one another, Student-Newman- Keuls test was employed. In all tests, a value of p < 0.05 was taken as indicating significance. The power of the performed statistical tests was calculated as greater than 0.95 for comparisons of Cstat, oxygenation of the perfusate and tissue cyclic nucleotides, and MPO. However, the power fell below 0.80 for comparison of vascular resistance (0.78) and wet:dry weight ratios (0.20).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

After 20 min perfusion in control (no storage) lungs, Cstat was 0.25 ± 0.04 ml · cm H₂O¹, Raw 0.15 ± 0.03 cm H₂O · s · ml¹, vascular resistance 1.27 ± 0.10 cm H₂O · min · ml¹, and transpulmonary difference in perfusate PO₂ 41.7 ± 6.7 mm Hg. None of the other experimental groups differed significantly from these values at this timepoint.

Lung Function during Reperfusion

Cstat and Raw. Storage in STH alone for 8 h resulted in a marked decrease in Cstat from 0.21 ± 0.03 to 0.07 ± 0.01 ml · cm H₂O¹ after 40 min of reperfusion (Figure 1). Addition of rolipram or theophylline to the STH solution improved the post-storage Cstat in a dose dependent manner (Figures 1A and 1D). This was the case when data were assessed as area under the curve over the 40 min reperfusion time and as the final values measured at the end of the reperfusion period. Neither milrinone or zaprinast at any concentration considered exerted any beneficial effect on post-storage lung function (Figures 1B and 1C), although the highest dose of milrinone (100 µM) appeared to produce a transient increase in compliance in the first 15 min of reperfusion (Figure 1B). Pulmonary compliance of lungs stored for 8 h in STH containing 3,000 µM theophylline or 100 µM rolipram was similar to that seen in lungs that had not been stored (Figure 1).

View larger version (28K): [in this window] [in a new window]   Figure 1.   Cstat during 40 min reperfusion with sanguineous BB in isolated rat lungs after no storage (closed squares), 8 h storage in STH solution alone (closed circles), or containing (A) rolipram 3 (inverted triangles), 10 (open squares), 30 (open circles), 100 (open triangles) µM, (B) milrinone 10 (open squares), 30 (open circles), 100 (open triangles) µM, (C) zaprinast 10 (open squares), 30 (open circles), 100 (open triangles) µM, or (D) theophylline 300 (closed triangles), 1,000 (open diamonds), or 3,000 (closed diamonds) µM. Data are expressed as mean ± SEM, n = 5-6 lungs per group. Some error bars have been omitted for clarity. *Significantly different from tissues stored for 8 h in STH alone by area under curve analysis (Dunnett's test p < 0.05).

Storage in STH alone for 8 h resulted in an increase in Raw from 0.16 ± 0.03 to 0.58 ± 0.23 cm H₂O · s · ml¹ after 40 min of reperfusion. A statistically significant difference between the two groups was seen for the area under the curve values for the reperfusion period (p < 0.05). Although addition of rolipram (100 µM) or theophylline (3,000 µM) to STH reduced Raw after 40 min reperfusion to 0.18 ± 0.04 and 0.21 ± 0.04 cm H₂O · s · ml¹, respectively, appearing to mirror their effects on Cstat, a significant improvement compared with 8 h control storage was seen only for the area under the curve value for 30 µM rolipram (Raw after 40 min reperfusion: 0.09 ± 0.03), probably as a result of the high variability in this parameter. Interestingly, milrinone (100 µM) and zaprinast (30 µM) also reduced Raw after 40 min reperfusion to 0.24 ± 0.03 and 0.21 ± 0.08 cm H₂O · s · ml¹, respectively, although these effects were not statistically significant.

Vascular resistance. Vascular resistance was raised after storage, although this effect was not great. No statistically significant difference was found between vascular resistance in unstored lungs and those reperfused after 8 h storage in STH (Figure 2). It should be noted that the power of the study for this parameter was 0.72 and because this is less than the minimum desirable level of 0.8, some real differences may not have been detected. It was therefore not possible to determine the effects of the PDE inhibitors on this parameter with any clarity. However, all the PDE inhibitors appeared to reduce vascular resistance during reperfusion at the higher dose levels.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Pulmonary vascular resistance during 40 min reperfusion with sanguineous BB in isolated rat lungs after no storage (closed squares), 8 h storage in STH solution alone (closed circles), or containing (A) rolipram 3 (inverted triangles), 10 (open squares), 30 (open circles), 100 (open triangles) µM, (B) milrinone 10 (open squares), 30 (open circles), 100 (open triangles) µM, (C) zaprinast 10 (open squares), 30 (open circles), 100 (open triangles) µM, or (D) theophylline 300 (closed triangles), 1,000 (open diamonds) or 3,000 (closed diamonds) µM. Data are expressed as in Figure 1. *Significantly different from tissues stored for 8 h in STH alone by area under curve analysis (Dunn's test p < 0.05).

Because vascular resistance in lungs that had been stored for 8 h in STH solution was raised to a greater extent at the initial timepoint in reperfusion, we also tested for significant differences at this time. Vascular resistance in lungs stored in STH alone was not greater than that in control lungs. However, it is worth noting that vascular resistance was significantly lower in lungs stored in the presence of 3,000 µM theophylline than in control (8 h) storage lungs at this time (Figure 2).

The increase in post-storage vascular resistance seen in lungs stored with either 10 µM rolipram or 10 µM zaprinast is accounted for by two lungs in each of these groups which, immediately upon reperfusion, exhibited very high vascular resistance accompanied by the rapid development of edema.

Gas exchange. Passage through the pulmonary vasculature increased perfusate PO₂ in control, no-storage, lungs at time 0, immediately after set-up (41.5 ± 5.8 mm Hg), this increase was well maintained after 20 min (41.7 ± 6.7 mm Hg) and 60 min (46.6 ± 2.8 mm Hg) perfusion.

Oxygenation of the perfusate by lungs was significantly (p < 0.05) impaired after 8 h storage in STH alone (Figure 3) being only 4.7 ± 2.2 mm Hg after 40 min reperfusion. This effect was only partially alleviated by the highest concentration (3,000 µM, Figure 3D) of theophylline, the transpulmonary increase in perfusate PO₂ after 40 min reperfusion being 25.2 ± 2.3 mm Hg, an approximately 50% recovery of function. Although there was a trend toward improved gas exchange (20.7 ± 4.9 mm Hg) by rolipram (100 µm), this failed to reach significance (Figure 3A). Milrinone and zaprinast were without effect (Figures 3B and 3C).

View larger version (27K): [in this window] [in a new window]   Figure 3.   Difference in PO2 between perfusate entering and leaving the lung during 40 min reperfusion with sanguineous BB in isolated rat lungs after no storage (closed squares), 8 h storage in STH solution alone (closed circles), or containing (A) rolipram 3 (inverted triangles), 10 (open squares), 30 (open circles), 100 (open triangles) µM, (B) milrinone 30 (open circles), 100 (open triangles) µM, (C) zaprinast 10 (open squares), 30 (open circles), 100 (open triangles) µM, or (D) theophylline 300 (closed triangles), 1,000 (open diamonds), or 3,000 (closed diamonds) µM. Data are expressed and treated as in Figure 2.

Lung MPO Content

In lungs stored for 8 h in STH and then reperfused for 40 min with SBB, MPO concentration was 0.62 ± 0.10 units/mg protein compared with 0.39 ± 0.11 U/mg protein in lungs perfused for 60 min in SBB without storage (Figure 4). This apparent increase in tissue MPO content was, however, not significant, nor was it attenuated by addition of PDE inhibitors to the storage solution (data not shown).

View larger version (10K): [in this window] [in a new window]   Figure 4.   MPO content of rat removed directly from the animals (control lungs) or removed after flush with bicarbonate buffer (BB flush) or sanguineous bicarbonate buffer (SBB flush) or in lungs after 60 min perfusion with SBB without storage (SBB perfusion) or after 8 h hypothermic storage in STH and 40 min reperfusion (8 h storage). Bars represent mean + SEM. *Significantly different from BB flush only and significantly different from SBB flush only (Student-Newman-Keuls test, p < 0.05).

To elucidate the underlying causes of this phenomenon of lack of significant effect of storage on lung MPO concentrations, three additional groups of lungs were studied. In the first of these, lungs were removed from rats and frozen immediately in liquid nitrogen without flushing or perfusion. In a second group, lungs were flushed with 30 ml blood-free BB at room temperature (20 to 25° C) immediately before freezing but not otherwise perfused; and in the third of these additional groups, lungs were flushed with SBB (30 ml at room temperature, 20 to 25° C) but not otherwise perfused before freezing. From Figure 4 it can be seen that flushing of lungs with BB or SBB significantly lowers MPO content compared with native (unflushed, unperfused) lungs. However, 60 min perfusion with SBB resulted in an increase in MPO to concentrations seen in control lungs. Against this background, it is therefore difficult to detect the effects of storage and reperfusion. Presumably, MPO accumulation in control lungs reflects an accumulation of SBB-derived neutrophils during perfusion. Furthermore, the MPO content of lungs that were not flushed, or perfused is probably largely due to neutrophils in the blood.

Lung Cyclic Nucleotide Content

Storage for 8 h in STH alone resulted in a significant decrease in lung cAMP concentration compared with perfusion for 20 min alone (13.0 ± 3.7 versus 26.4 ± 7.7 pmol/mg protein, p < 0.05). This decrease in tissue cAMP content was reversed by addition to the storage solution of the PDE inhibitors (Figure 5A), rolipram (100 µM), theophylline (3,000 µM) or Milrinone (100 µM), which resulted in post-storage cAMP concentrations of 44.1 ± 5.3, 31.3 ± 5.5, and 29.0 ± 2.5 pmol/mg protein, respectively. Zaprinast (100 µM) was without effect (14.9 ± 4.1 pmol/mg protein).

View larger version (15K): [in this window] [in a new window]   Figure 5.   Cyclic nucleotide content (A: cAMP and B: cGMP) of rat lungs after (I) 20 min control perfusion or (II) 8 h hypothermic storage in STH or after 8 h hypothermic storage in STH containing either (III) 100 µM rolipram, (IV) 3,000 µM theophylline, (V) 100 µM milrinone, or (VI) 100 µM zaprinast and 40 min reperfusion. Bars represent mean + SEM. *Significantly different from 8 h control storage only and significantly different from 20 min control perfusion (Student-Newman- Keuls test, p < 0.05).

In contrast, cGMP concentrations after storage were not reduced compared with those seen after 20 min control perfusion (8.2 ± 1.0 versus 7.8 ± 1.2 pmol/mg protein, respectively). Of the PDE inhibitors tested, only zaprinast (100 µM) and theophylline (3,000 µM) increased tissue cGMP concentrations, although the variability of this parameter, particularly in the presence of theophylline, meant that these changes were not significant (Figure 5B).

Wet:Dry Weight Ratios

The wet:dry weight ratios in lungs perfused for 60 min in SBB (6.3 ± 0.5; n = 6) were not significantly different from those of lungs removed directly from the animals (5.5 ± 0.4; n = 9). Although wet:dry weight ratios were increased after 8 h storage in STH (8.3 ± 1.6 versus 6.3 ± 0.5), this increase was not statistically significant. In addition, no statistically significant differences were found between the various treatment groups (data not shown). The negative findings with this parameter should, however, be treated cautiously because the performed statistical tests had, in this case, a very low power (0.30).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There is considerable evidence that maintaining intracellular cAMP concentrations during prolonged hypothermic storage and reperfusion is protective. The cAMP analogue, dibutyryl-cAMP, was protective in a lung transplant model (4) and several studies have demonstrated beneficial effects of nonselective PDE inhibitors such as theophylline (11) and pentoxifylline (8). PDE inhibition may also be protective by virtue of enhancing cGMP levels (18). Our aim in this study was to compare three different isoenzyme-selective PDE inhibitors, together with the nonselective theophylline, as additives to STH when used for 8 h hypothermic storage of rat lungs.

Storage in STH for 8 h led to significantly reduced values of Cstat and gas exchange and increased values of Raw during reperfusion compared with those in unstored, control lungs. Vascular resistance in this model was not significantly raised by storage alone. We have previously shown that this is a beneficial effect of employing STH solution, compared with higher potassium solutions such as EuroCollins (19). Nonetheless, there was a degree of vasoconstriction in the reperfusion period which all the PDE inhibitors inhibited to some degree. The wet:dry weight ratios of the various groups of lungs failed to reflect the changes seen in Cstat. This might seem surprising because Cstat, is a measure of the elasticity of the lung, and acute decreases in a model such as that employed here are usually the result of interstitial edema. However, it should be borne in mind that Cstat is a much more sensitive measure and will therefore decline before gross changes in lung weight are detected, as others have reported (20). This consideration, combined with the low statistical power associated with the measurements of wet:dry weight ratios in this study, probably accounts for the discordance with the Cstat measurements.

In this study, lung cAMP content, and the decrease induced by storage, were of similar magnitude to those previously reported by others (3). The lack of effect of storage on cGMP may suggest this nucleotide is more susceptible to the effects of reperfusion rather than storage. Other investigators have found that nitric oxide and consequently cGMP begin to fall upon reperfusion in rat lungs stored in EuroCollins solution; however, cGMP concentrations prior to storage were not reported (24).

Of the various PDE inhibitors employed, only theophylline and rolipram were able to reverse the decline in Cstat due to storage, seen during reperfusion. Gas exchange was significantly improved only by the highest dose of theophylline, although the highest dose of rolipram also showed a tendency to improve this parameter. The PDE III inhibitor, milrinone and the PDE V inhibitor, zaprinast were without significant effect. These data suggest that PDE IV inhibition plays an important role in protection of the lung during hypothermic storage.

These findings are in agreement with previous studies examining the effect of nonselective and PDE IV-selective inhibitors. Rolipram has been reported to attenuate lung ischemia- reperfusion injury (25, 26) and the efficacy of pentoxifylline has been demonstrated in a number of transplant models (6, 8). However, Pinsky and coworkers (24) found that PDE V inhibition in a rat lung transplant model causes improved arterial oxygenation and reduced pulmonary vascular resistance, responses not seen in this study. We are unaware of any other studies examining the effects of selective PDE III inhibition on prolonged storage of lungs. Although the improvement in oxygenation of perfusate by theophylline and rolipram was not as marked as the effects of these agents on Cstat, it is nonetheless encouraging, particularly in the light of the fact that ventilation in these experiments was with room air. Other studies reporting improved oxygenation in various in vivo transplant models typically employed ventilation with 100% oxygen (8, 10, 24, 27). In ex vivo ischemia-reperfusion lung injury models little data are available for comparison since gas exchange data are not reported (6, 28) or the methodology employed differs with respect to species used, concentration of blood in the perfusate, or in the degree of deoxygenation of perfusate (29, 30).

The highest doses of the three selective PDE inhibitors used were at least an order of magnitude greater than their known in vitro IC₅₀ values against their respective PDE isoforms, but not greater than their IC₅₀ values against the isoforms which they do not selectivity inhibit (16). It is likely, therefore, that selective inhibition of the various isoenzyme subtypes was achieved. This assumption is supported by the lung cyclic nucleotide concentrations observed. Milrinone, rolipram, and theophylline, at the highest doses employed, prevented a storage-induced decline in cAMP and neither milrinone nor rolipram affected cGMP levels. Zaprinast (which selectively inhibits the cGMP-selective PDE V isoenzyme subtype) did not alter cAMP concentrations. Although cGMP concentrations were not reduced by hypothermic storage, zaprinast, and the nonselective inhibitor theophylline tended to increase this nucleotide. If, as has been suggested, cGMP can improve lung preservation and the decline in tissue cGMP occurs primarily during reperfusion (24), strategies such as those employed in this study that lead to increased cGMP during storage and, therefore, at the beginning of reperfusion would still be protective, because the decline in the nucleotide would be from a higher baseline. Increasing lung cGMP concentrations has been suggested to be beneficial in a transplantation model as a result of its ability to modulate proinflammatory processes, such as leukocyte adherence to the endothelium, in addition to its vasodilatory properties (24, 31). The data presented here suggest that this may not be the case.

Milrinone (100 µM), despite increasing cAMP to a degree similar to that seen with 3,000 µM theophylline or 100 µM rolipram, failed to protect the lung against the adverse effects of storage on pulmonary compliance and gas exchange. This may be due to the differential effects of the selective PDE inhibitors on different pools of cAMP within the lung; PDE III, inhibited by milrinone, may occur in vascular and other smooth muscle but not in inflammatory cells such as alveolar macrophages and neutrophils (16). Thus, our use in this study of a preservation solution (STH) that minimizes post-storage vasoconstriction (19) might have the consequence that those agents that act primarily by vasodilation, such as the PDE III and V inhibitors, will have less effect. In contrast, rolipram, a PDE IV inhibitor, is anti-inflammatory (32, 33) and this may attenuate ischemia-reperfusion injury mediated by neutrophils and other inflammatory cells. However, the effect of the 60-min control (no storage) perfusion on lung MPO concentrations has the consequence that it is not possible to define, in this model, the contribution of neutrophils to ischemia-reperfusion injury. This finding is consistent with reports from other investigators who showed an increase in tissue MPO caused by perfusion alone using an isolated, blood-perfused rat lung preparation (28).

Given the efficacy of rolipram compared with milrinone and zaprinast in this study, it was surprising that the nonselective agent, theophylline, provided the greatest improvement in post-storage lung function, significantly improving gas exchange and Cstat. At least three possible explanations exist for this finding: First, benefit may be achieved from "blanket" PDE isoenzyme subtype inhibition. For example, while inhibition of PDE III or V alone is without effect, when occurring in combination with inhibition of PDE IV, they enhance its effect. Second, although not tested in this study, inhibition of type I or II PDE may have an important effect. We believe this is unlikely, however, given their cellular distributions, which are predominantly restricted to neuronal and cardiac tissue respectively (16). Third, the adenosine receptor antagonist activity of theophylline (5) may contribute significantly to its protective effects. At least one other study has suggested that adenosine receptor antagonism during ischemia reduces the damage seen in lungs after reperfusion (34).

In conclusion, it is clear from these data that PDE inhibitors may be useful additives to solutions intended for long-term preservation of lungs. Of the selective agents examined, those inhibiting the PDE IV isoenzyme subtype are the most effective. However, the apparent superiority of the nonselective agent theophylline merits further investigation.