INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

POLYMORPHONUCLEAR LEUKOCYTES become primed by inflammatory agents released at sites of injury or circulating in the vasculature (4, 22, 33). On attachment to vascular endothelial cells and/or after emigration into the underlying tissue, polymorphonuclear leukocytes become activated, resulting in the release of highly toxic reactive oxygen species such as hydrogen peroxide (H₂O₂) and cytoplasmic granules (6). H₂O₂ generated extracellularly can freely enter the endothelial cell where it has been shown to target and affect various cellular components (6, 11, 24, 25). These cellular changes ultimately result in an increase in the passage of protein and water across the vascular endothelium, leading to edema, abnormalities in gas exchange, and finally pulmonary insufficiency or the acute respiratory distress syndrome (3). We have shown that an infusion of H₂O₂ can increase the transvascular clearance of ¹²⁵I-labeled albumin in an isolated, perfused rabbit lung under isogravimetric conditions (31). However, the effect of H₂O₂, or any other oxidant, on transvascular protein clearance under nonisogravimetric conditions, where both the diffusive and convective clearances of protein can be affected, has not been evaluated in an intact vasculature.

Previous studies have demonstrated the permeability-decreasing activity of cAMP-enhancing agents using a variety of inflammatory agents, diseases, and syndromes (1, 7, 15, 18-21, 26, 28, 29). However, the ability of cAMP-enhancing agents to prevent a H₂O₂ or any oxidant-induced increase in transvascular protein clearance in an intact vasculature has not been evaluated. Previous studies in the isolated lung have relied on the measurements of the capillary filtration coefficient (K_f) and lung weight to assess the efficacy of cAMP-enhancing agents (7, 26). These two parameters, however, are profoundly affected by changes in vascular surface area; therefore, it is difficult to make definitive conclusions about changes in endothelial barrier function induced by vasoactive agents such as H₂O₂ and cAMP-enhancing agents by using measurements of K_f and lung weight.

Drug therapy to increase intracellular levels of cAMP can take on many forms, such as stimulation of the ₂-adrenergic receptor by isoproterenol, stimulation of adenylate cyclase by forskolin, the use of a cell-permeable analog such as 8-bromo-cAMP, or inhibition of phosphodiesterases that metabolize cAMP to 5'-AMP. Recent developments in the literature would suggest that H₂O₂ and/or its metabolite, the hydroxyl radical (· OH), may adversely affect ₂-adrenergic receptors (5). Thus the use of agonists that stimulate ₂-adrenergic receptors may not be as effective as compared with, for example, drugs that inhibit cAMP metabolism in preventing an oxidant-induced increase in transvascular protein clearance.

The first objective of the present study was to determine whether H₂O₂ increases the diffusive and convective clearances of albumin across an intact vasculature under nonisogravimetric conditions. The second objective was to determine the efficacy of posttreatment with cAMP-enhancing agents that function to either stimulate production or inhibit metabolism of cAMP on prevention of the increased transvascular albumin clearance induced by H₂O₂. The isolated, perfused rat lung was utilized because it represents an intact vascular endothelium. Furthermore, it allows for the measurement of ¹²⁵I-albumin clearance and the subsequent calculation of the permeability-surface area product (PS), as a measure of diffusive albumin permeability, and of the reflection coefficient (), a measure of convective albumin clearance. The measurement of  is independent of changes in vascular surface area and, unlike the measurements of PS, K_f, and lung weight, is not affected by the vasoactive properties of H₂O₂ and cAMP-enhancing agents.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolated, perfused rat lung preparation. Protocols for animal care and experimental design were approved by the Institutional Animal Care and Use Committee at Albany Medical College. Male Sprague-Dawley rats (250-350 g) were injected intraperitoneally with pentobarbital sodium (0.65 mg/kg) and administered heparin (1,000 U/kg) via the penile vein before surgery. Lungs were isolated as described previously (15, 30, 34). Briefly, lungs and heart were suspended from a beam balance, and the pulmonary artery and left atrium were cannulated and perfused with a Ringer phosphate solution containing 2% (wt/vol) BSA. The perfusate (125 ml) was recirculated at 28 ml/min (~50% of the normal cardiac output of a rat), maintained at 38°C and a pH of 7.4, and continuously oxygenated. Lungs were maintained in zone 3 of West with a venous pressure of 4 cmH₂O and an alveolar pressure of 2 cmH₂O. Lung weight was measured by a sensitive displacement transducer attached to a beam balance. A change of 200 mg was equal to a 1-cm deflection. Wet weight of the lungs was normalized to dry weight to allow for comparison of lung weights among rats. Pulmonary arterial and left atrial pressures and weight were continuously recorded on a multichannel strip-chart recorder (model 7D polygraph, Grass Instruments). Pulmonary arterial and left atrial cannulas had in-line solenoids, allowing for rapid simultaneous occlusion to determine capillary pressure. Pulmonary vascular resistances were determined from the differences between the capillary pressure and the arterial pressure or venous pressure divided by the flow rate.

Measurement of ¹²⁵I-albumin clearance. Albumin clearance (J_s/C_p, where J_s is transvascular ¹²⁵I-albumin flux and C_p is the perfusate concentration of ¹²⁵I-albumin) was determined at t = 60 min after a 3-min exposure of the lungs to ¹²⁵I-albumin at different rates of fluid filtration (J_v) by using the method of Kern et al. (13). Albumin clearance was calculated by using the equation

(1)

where A is the amount of ¹²⁵I-albumin in the tissue in counts per minute per gram dry lung, t is the exposure time in minutes, and C_p is in counts per minute per milliliter. Albumin clearance was used to calculate PS and  for albumin, as previously described (15, 30, 34), by using the nonlinear flux equation of Patlack et al. (23)

(2)

where J_s and J_v are in microliters per minute per gram dry lung, C_i is the interstitial concentration of ¹²⁵I-albumin, and Pe is the Péclet number equal to J_v (1  )/PS, where PS (in µl · min¹ · g dry lung¹) is the albumin PS. The initial tracer concentration in the interstitium (C_i) was considered negligible during the 3-min infusion of the radiolabeled tracer; thus C_i was deleted from Eq. 2. The C_p was then divided into each term of the equation leaving

(3)

For control and experimental rats, albumin clearance was measured either under isogravimetric conditions or during a weight gain (increased J_v) induced by transiently raising the venous outflow cannula and thus venous pressure. The J_v following the increase in venous pressure was calculated from the slow component of the weight-gain curve, as determined in the calculation of K_f. Values for PS and  were calculated from a nonlinear regression analysis of J_s/C_p vs. J_v by using Eq. 3 (Statistica). In some experiments, PS was determined from tissue counts or J_s/C_p obtained at a J_v of 0.

Experimental protocols for H₂O₂. These studies consisted of two protocols: 1) the effect of iron on the H₂O₂-induced increase in lung weight and 2) the effect of an infusion of H₂O₂ on transvascular clearance of ¹²⁵I-albumin and the calculated parameters, PS and .

Desferrioxamine (DFO) chelation of free iron in albumin perfusate. The artificial perfusate used in this isolated lung model was found to contain ~90 µM iron (Sigma Chemical quality control sheet), which could convert extracellularly the infused H₂O₂ to the · OH via the Fenton reaction (8, 16). Because an objective of these studies was to assess the effect of H₂O₂ on transvascular albumin clearance, it was important to maximize the delivery of H₂O₂ to the pulmonary endothelial cells. To achieve this end, DFO, an iron chelator, was used to bind extracellular iron in the perfusate. Two types of DFO were used, low molecular weight (i.e., 400) DFO, which is reported to be cell permeable (16), and DFO conjugated to hydroxyethyl starch (HES-DFO; 40,000-70,000 mol wt), which is impermeable to cells and thus binds only extracellular iron. The ability of DFO to lower the concentration of iron in the perfusate was analyzed by inductively coupled plasma emission spectroscopy (ICP 2.5, Leeman Labs). Nontreated and DFO-treated perfusates (5-ml samples) were dialyzed against 0.9% saline (2 liters; 3×) to remove DFO and DFO-iron from the samples. Aliquots (1 ml) of the samples were dried at 60°C for 8 h and ashed at 450°C for 12 h. The ashed samples were reconstituted in 1.5 ml of 10% HNO₃, weighed, and brought to a near boil for 20 min. Samples were cooled and brought back to their original weight with distilled water. An additional 1.5 ml of distilled water were added before analysis of the iron concentration. The effect of DFO on the H₂O₂-induced gain in lung weight was assessed by infusing H₂O₂ at 40 nmol · ml¹ · min¹ into the pulmonary arterial line of the isolated rat lung perfused without or with 150 µM DFO or HES-DFO. The experiments were run as a time course study and concluded when lung weight increased by ~1.5 g, which was a doubling of the wet lung weight.

H₂O₂ infusion and measurements of pulmonary hemodynamics, lung weight, and transvascular albumin clearance. The effect of H₂O₂ on transvascular albumin clearance was assessed in the presence of 150 µM HES-DFO. The rate of infusion of H₂O₂ was 80 nmol · ml¹ · min¹ for 15 min, followed by 60 nmol · ml¹ · min¹ for 10 min, 40 nmol · ml¹ · min¹ for 15 min, and 20 nmol · ml¹ · min¹ for 10 min (see Fig. 2). Control lungs were infused with Ringer phosphate solution at the same infusion rate. The measurement of ¹²⁵I-albumin clearance was performed at 60 min. To demonstrate that the infused H₂O₂ was taken up continuously by the lungs and did not increase in concentration in the perfusate, the concentration of H₂O₂ in the perfusate entering and exiting the lungs was determined by the xylenol orange assay (12). Samples (900 µl) of perfusate were incubated at 27°C for 45 min with 100 µl of a 10× xylenol orange stock solution [10 µM xylenol orange, 2.5 µM Fe(NH₄)₂(SO₄)₂ · 6H₂O, 25 mM H₂SO₄; Sigma Chemical] after which absorbance was read at 590 nm and the concentration of H₂O₂ was determined from a standard curve (0.1, 0.5, 1.0, 2.5, 5.0, 7.5, 10.0, and 25 µM H₂O₂) (12). An equal concentration of DFO was present in the standards and experimental samples. A fresh stock solution of H₂O₂ was made each day, and its concentration was determined by spectrophotometric analysis at 240 nm with an extinction coefficient of 43.6 M¹ · cm¹.

Experimental protocol for cAMP-enhancing agents. On the day of the experiment, 25 mg of isoproterenol (Sigma Chemical) were dissolved in 10 ml of perfusate, resulting in a 10 mM stock solution. Thirty minutes after the start of the H₂O₂ infusion, 100 µl of the isoproterenol stock solution were injected into the 125-ml recirculating reservoir, resulting in a final concentration of 8 µM. CP-80633 (5 mg), a gift from Pfizer, was dissolved in 2 ml of 100% DMSO so that its stock concentration was 15 mM. Samples were aliquoted and stored at 4°C until needed. Thirty minutes after the start of the H₂O₂ infusion, 10 µl of the CP-80633 stock were diluted with 490 µl of perfusate (300 µM CP-80633), of which 400 µl were injected into the 125-ml recirculating reservoir, resulting in a final concentration of 1 µM (<0.1% DMSO). The drug vehicle, either perfusate or perfusate with DMSO (0.1%), was administered in control studies.

Statistics. Lung weight and hemodynamics were analyzed by using a two-way ANOVA with repeated measures. A Newman-Keuls multiple-range test was used to evaluate significant differences between groups and times (32). Nonlinear regression analysis was performed on the ¹²⁵I-albumin clearance data as a function of J_v to calculate PS and . Student's t-tests with Bonferroni correction for multiple comparisons were performed to determine differences in PS and  between groups (32). Statistical significance was set at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DFO chelation of free iron in albumin perfusate. While developing a protocol to measure the levels of H₂O₂ in the recirculating perfusate of the isolated rat lung, we observed that the H₂O₂ concentration decreased rapidly in perfusate that contained albumin as opposed to perfusate without albumin. Perfusate with albumin contained 73.3 ± 0.2 µM free iron as measured by inductively coupled plasma emission spectroscopy. To minimize the content of free iron in the perfusate and the potential extracellular conversion of H₂O₂ to an iron-dependent metabolite such as · OH, DFO (150 µM) was added to the perfusate and was found to significantly reduce the free iron to 8.0 ± 0.3 µM (Table 1). It has been reported that DFO is cell permeable and can affect the intracellular conversion of H₂O₂ to · OH (16). Therefore, DFO coupled to HES (HES-DFO, 150 µM) was also used.

1 · 

1

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Temporal measurements of lung weight. Lung weight was determined by hanging the lungs from a balanced force transducer. Time 0 denotes commencement of H2O2 infusion at a rate of 40 nmol · ml1 · min1 into pulmonary arterial line. Without desferrioxamine (DFO), lung weight remained fairly constant for 30 min until fulminant alveolar edema developed by 40 min, whereas with DFO or DFO conjugated to hydroxyethyl starch (HES-DFO), lung weight increased steadily for 80 min until fulminant alveolar edema developed by 90 min. , Change. Values are means; n, no. of rats in each group.

H₂O₂ infusion and measurements of pulmonary hemodynamics, lung weight, and transvascular albumin clearance. H₂O₂ was infused at a rate of 80 nmole · ml¹ · min¹ for 15 min, 60 nmol · ml¹ · min¹ for 10 min, 40 nmol · ml¹ · min¹ for 15 min, and 20 nmol · ml¹ · min¹ for 10 min into the pulmonary arterial line. This infusion scheme was chosen because it did not produce fulminate alveolar edema but did significantly increase pulmonary arterial and capillary pressures (Table 2), pulmonary arterial and venous resistances (Table 2), lung weight (Fig. 2), and the transvascular clearance of ¹²⁵I-albumin at 60 min (Table 3 and Fig. 3). Nonlinear regression fit of the data representing ¹²⁵I-albumin clearance (J_s/C_p) as a function of J_v allowed for the calculation of PS (denoted by the y-intercept) and  (denoted by the slope of the line, 1  ). Nonlinear regression analysis revealed that H₂O₂ decreased  from a control value of 0.93 ± 0.02 to 0.38 ± 0.05. H₂O₂ alone did not change PS from control values (Fig. 3 and Table 3). These findings demonstrate that H₂O₂ can increase the convective clearance of albumin across the intact vasculature of the isolated, perfused rat lung.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Temporal measurements of lung weight. Lung weight was determined by hanging the lungs from a balanced force transducer. Time 0 denotes commencement of infusion of H2O2 at a rate of 80 nmol · ml1 · ml1 for 15 min, 60 nmol · ml1 · ml1 for 10 min, 40 nmol · ml1 · ml1 for 15 min, and 20 nmol · ml1 · ml1 for 10 min or vehicle control into pulmonary arterial line. In designated rats, 8 µM isoproterenol or 1 µM CP-80633 was injected (arrow) into perfusate reservoir 30 min after start of H2O2 infusion. Posttreatment with isoproterenol increased rate of lung weight gain due presumably to a combination of an isoproterenol-induced vasodilation and a H2O2-induced increase in fluid filtration. Posttreatment with CP-80633 prevented H2O2-induced increase in lung weight due, in part, to significant decrease in pulmonary capillary pressure (see Table 2). Values are means ± SE; n, no. of rats in each group. * P < 0.05 compared with control value.

View this table: [in this window] [in a new window]   Table 3.   Albumin reflection coefficient () and permeability-surface area product (PS) for isolated rat lungs treated with H2O2 and cAMP-enhancing agents

View larger version (22K): [in this window] [in a new window]   Fig. 3.   125I-labeled albumin clearance (Js/Cp, where Js is transvascular 125I-albumin flux and Cp is the perfusate concentration of 125I-albumin) as a function of transvascular fluid filtration (Jv). Curves represent nonlinear regression analysis of Js/Cp vs. Jv using nonlinear flux equation: Js/Cp = Jv (1  ) (1  ePe), where  is the reflection coefficient and Pe is the Péclet number. Infusion of H2O2, as described in Fig. 2 legend, decreased  from a control value of 0.93 ± 0.02 to 0.38 ± 0.05 at 60 min. A 30-min posttreatment with 8 µM isoproterenol partially reduced the decrease in  to 0.65 ± 0.10, whereas 1 µM CP-80633 reduced the decrease in  more effectively to 0.79 ± 0.06. n, No. of rats in each group. * P < 0.05 compared with control value. t P < 0.05 compared with H2O2 value.

Posttreatment with isoproterenol. To determine the effects of stimulation of cAMP production on the H₂O₂-induced increases in hemodynamics, weight, and albumin clearance of the lung, 8 µM isoproterenol was injected as a bolus into the recirculating reservoir 30 min into the infusion of H₂O₂. The H₂O₂-induced increases in pulmonary arterial pressure and pulmonary arterial resistance were reduced within 10 min by isoproterenol (Table 2) in association with a further increase in lung weight, presumably due to an increase in vascular surface area (Fig. 2). Isoproterenol partially reduced the H₂O₂-induced decrease in  from 0.38 ± 0.05 to 0.65 ± 0.10 (Fig. 3 and Table 3) and increased PS from the H₂O₂ value of 32 ± 47 to 82 ± 38 µl · min¹ · g dry lung¹ (Fig. 3 and Table 3). A similar value for PS (91 µl · min¹ · g dry lung¹) was determined when PS was directly measured at J_v = 0 in one rat infused with H₂O₂ and posttreated with isoproterenol. These results demonstrate that posttreatment with isoproterenol was partially effective in reducing the H₂O₂-induced decrease in  but was ineffective in stabilizing lung weight.

Posttreatment with CP-80633. To determine the effects of inhibition of cAMP metabolism on the H₂O₂-induced increases in hemodynamics, weight, and albumin clearance of the lung, 1 µM CP-80633 was injected as a bolus into the recirculating reservoir 30 min into the infusion of H₂O₂. CP-80633 also reduced the H₂O₂-induced increases in pulmonary arterial pressure and pulmonary arterial resistance, as did isoproterenol, but, in addition, CP-80633 also reduced pulmonary capillary pressure and pulmonary venous resistance within 10 min (Table 2). This significant decrease in postcapillary hemodynamics was associated with a complete block of the H₂O₂-induced increase in lung weight (Fig. 2). CP-80633 significantly reduced the H₂O₂-induced decrease in  from 0.38 ± 0.05 to 0.79 ± 0.06 (Fig. 3 and Table 3). H₂O₂ in the presence of CP-80633 increased PS from the H₂O₂ value of 32 ± 47 to 67 ± 36 µl · min¹ · g dry lung¹. Similar values for PS (34 ± 5 to 54 ± 6 µl · min¹ · g dry lung¹) were obtained when PS was directly measured at J_v = 0 in two control rats and in two rats infused with H₂O₂ and posttreated with CP-80633 (Fig. 3). These results demonstrate that posttreatment with CP-80633 was effective in attenuating the H₂O₂-induced increases in convective clearance of albumin and lung weight. Furthermore, an increase in PS was evident when H₂O₂ was delivered in the presence of isoproterenol or CP-80633.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The first objective of the present study was to determine which parameters of increased transvascular protein permeability were altered by an infusion of H₂O₂ into the isolated, perfused rat lung. The  and PS were calculated from nonlinear regression analysis of ¹²⁵I-albumin clearance (J_s/C_p) as a function of J_v. An infusion of H₂O₂ into the pulmonary artery increased lung weight and decreased . There was no change in PS, which was masked by the vasoconstriction induced by H₂O₂. The second objective was to compare the efficacy of posttreatment with cAMP-enhancing agents that function either by inhibition of the type 4 phosphodiesterase or by activation of the ₂-adrenergic receptor. Stimulation of cAMP production by posttreatment with isoproterenol attenuated the decrease in  and augmented the increase in lung weight induced by H₂O₂. Inhibition of cAMP metabolism by CP-80633, a phosphodiesterase 4 inhibitor, more effectively attenuated the decrease in  compared with isoproterenol and prevented the increase in lung weight. The vasodilation induced by these cAMP-enhancing agents unmasked the increased PS induced by H₂O₂. Therefore, H₂O₂ increased the convective and diffusive clearances of albumin in an intact pulmonary vasculature.

While developing a model of a H₂O₂-induced increase in transvascular clearance of albumin in the isolated, perfused rat lung, we determined that the artificial perfusate contained 73 µM iron. Because the availability of only one coordination site in ferrous iron (Fe²⁺) is needed for the reduction of H₂O₂ to · OH (8, 16), it would be essential to minimize the content of free iron in the perfusate to maximize the delivery of H₂O₂ to the pulmonary vascular endothelium. To this end, DFO or HES-DFO was added to the perfusate at a dose of 150 µM to obtain a 2:1 ratio of iron chelator to perfusate iron; each DFO molecule can wrap around an iron molecule and successfully chelate all four coordination sites (8, 16). Both forms of DFO were successful in delaying the onset of fulminant alveolar edema induced by an infusion of 40 nmol · ml¹ · min¹ of H₂O₂ into the pulmonary arterial line. Another iron chelator, U-74500A, has also been shown to be effective in attenuating the increased capillary permeability induced by ischemia-reperfusion or tert-butyl hydroperoxide (9). There are reports in the literature that low-molecular-weight DFO is permeable to endothelial cells and can chelate intracellular free iron and protect cells against oxidant injury (16). To prevent the blockade of this potentially important cellular pathway of oxidant injury, high-molecular-weight HES-DFO, which is not permeable to cells, was used, allowing H₂O₂ to freely enter the cell and interact with any available free iron. The concentration of H₂O₂ infused into the pulmonary arterial line and that sampled in the perfusate 80 cm downstream at the entrance to the lung were identical as measured by the xylenol orange assay; no H₂O₂ was detected in the perfusate exiting the lungs (data not shown). Thus the addition of HES-DFO to the perfusate minimized the extracellular conversion of H₂O₂ to · OH. Because the concentration of H₂O₂ did not change in the perfusate entering the lungs and no H₂O₂ was detected in the eluate, we conclude that H₂O₂ diffused into the lungs. However, it is possible that some of the H₂O₂ could have been converted by myeloperoxidase, released by any remaining polymorphonuclear leukocytes, into hypochlorous acid, although this conversion would have had to occur in close proximity to the endothelial cells as albumin in the perfusate could scavenge the hypochlorous acid. The fact that we saw an identical delay in lung weight gain induced by H₂O₂ in both groups treated with DFO and HES-DFO would suggest that DFO is not permeable and, therefore, like HES-DFO only binds extracellular free iron. Alternatively, any amount of DFO in the perfusate not chelated to iron that may permeate the endothelial cells and bind intracellular free iron may have an insignificant effect on the generation of intracellular · OH, because the conversion of H₂O₂ to · OH requires very little free iron (8). These findings emphasize the importance of ensuring that H₂O₂, and not an iron-dependent metabolite, is delivered to cells when the effects of H₂O₂ are studied in a perfusate requiring albumin.

In vitro studies using endothelial cell monolayers have demonstrated that H₂O₂ increases K_f and diffusive and convective clearances of protein (27, 28) and that cAMP-enhancing agents prevent and reverse the increase in K_f and protein clearance (21, 28). In situ studies using the isolated, perfused lung have shown that H₂O₂, administered as a bolus or by infusion into the recirculating reservoir, increases pulmonary arterial pressure, K_f, lung weight, and transvascular protein clearance. In addition, pretreatment with cAMP-enhancing agents attenuates the increases in pulmonary arterial pressure, K_f, and lung weight (7, 26). We also chose the isolated, perfused lung model to assess whether H₂O₂ and cAMP-enhancing agents alter transvascular protein clearance, an initial cause of acute respiratory distress syndrome. This model represents the intact vascular endothelium of a whole lung, an improvement on studies that use endothelial cell monolayers. Furthermore, the measurement of transvascular protein clearance can be obtained in this model (15, 30, 34), and two parameters, PS and , that define the restrictive properties of the vascular endothelium to the diffusive and convective clearances of protein, respectively, can be calculated. The latter parameter, , is independent of vascular surface area, unlike the parameters PS, K_f, and lung weight, which can be affected by changes in vascular surface area induced by the vasoactive properties of H₂O₂ and cAMP-enhancing agents.

One of the significant findings of the present study was that  was decreased by an infusion of H₂O₂ from a control value of 0.93 to 0.38. This result demonstrates that H₂O₂ increases the convective clearance of albumin across an intact vascular endothelium. Although  was markedly decreased by H₂O₂, diffusive clearance (PS), calculated as the y-intercept by using nonlinear regression analysis of J_s/C_p at different J_v, was not affected by an infusion of H₂O₂ (Table 3). H₂O₂ constricted the vasculature upstream from the capillary bed. Thus an increase in diffusive permeability with a coinciding decrease in vascular surface area due to precapillary constriction could cancel each other and result in no change in PS. In contrast, PS increased when H₂O₂ was posttreated with either cAMP-enhancing agent, most likely due to the vasodilatory effects of intracellular cAMP that unmasked the precapillary constriction. This increase in PS induced by H₂O₂ in the presence of isoproterenol and CP-80633, however, is most likely underestimated for two reasons. First, these cAMP-enhancing agents would be expected to decrease the diffusive permeability of albumin as indicated by the elevation in  and as demonstrated in the literature using endothelial cell monolayers (18, 19). Second, the increase in J_s, which is accentuated under nonisogravimetric conditions, would increase the C_i during the 3-min infusion of the tracer. This would result in a diminished concentration gradient from the vascular space to the interstitium during the infusion period compared with that in the control group under isogravimetric conditions. This increase in C_i would also create a greater concentration gradient from the interstitium to the vascular space during the 3-min washout period, causing a greater back flux of tracer from the interstitium. These findings demonstrate that H₂O₂ increases the convective clearance of albumin across the vascular endothelium and that diffusive clearance induced by H₂O₂ can be masked by a concomitant increase in precapillary constriction. A previous publication in which an arachidonic acid-induced increase in albumin clearance was significantly attentuated with isoproterenol supports the above findings and interpretations (15). We demonstrated that arachidonic acid, like H₂O₂, significantly decreased . Arachidonic acid was infused in combination with indomethacin to minimize changes in capillary pressure and vascular resistances, and, as a result, PS increased. Isoproterenol attenuated the decrease in  and the increase in PS induced by arachidonic acid. Capillary pressure and resistances were also unchanged when arachidonic acid was infused in the presence of isoproterenol, demonstrating that isoproterenol can modify the parameters of albumin clearance,  and PS, independently of changes in hemodynamics.

H₂O₂ also increased lung weight as a result of increases in capillary pressure and the transvascular clearance of albumin. Posttreatment with isoproterenol augmented the H₂O₂-induced increase in lung weight, whereas a posttreatment with CP-80633 prevented the increase in lung weight. Although both agents reduced the H₂O₂-induced increases in pulmonary arterial pressure and pulmonary arterial resistance, CP-80633 also reduced pulmonary capillary pressure and pulmonary venous resistance and, hence, decreased hydrostatic pressure in the lung, which would result in a decrease in lung weight. Therefore, CP-80633 prevented the H₂O₂-induced increase in lung weight via a hemodynamic mechanism, as well as by reducing the H₂O₂-induced increase in transvascular clearance of albumin.

We, as well as others, have demonstrated the permeability-decreasing activity of various cAMP-enhancing agents using a variety of inflammatory agents, diseases, and syndromes (1, 7, 15, 18-21, 26, 28, 29). For example, both rolipram, a phosphodiesterase 4 inhibitor like CP-80633, and isoproterenol reversed the increased pulmonary capillary permeability induced by ischemia-reperfusion in the isolated, perfused rat lung (1). However, in the present study, the significant finding was that CP-80633 reduced the H₂O₂-induced decrease in  more effectively than did isoproterenol. This finding suggests that the use of ₂-adrenergic receptor agonists, which stimulate the production of intracellular cAMP, may not be as effective in counteracting the increase in transvascular protein clearance induced by reactive oxygen species such as H₂O₂ and · OH (5). These oxidants have been reported to influence the components involved in the production of cAMP. These proposed effects are 1) a decrease in the affinity of the ₂-adrenergic receptor by · OH (17); 2) a decrease in intracellular ATP (10); 3) a decrease in the activity of adenylate cyclase (17); and/or 4) an increase in the metabolism of cAMP (14). Another possibility is that H₂O₂ may directly alter the structure and activity of isoproterenol. Because production of cAMP may be attenuated by reactive oxygen species, we also treated lungs with the phosphodiesterase 4 inhibitor, CP-80633, as a means of decreasing catabolism of cAMP and increasing intracellular levels of cAMP. CP-80633, compared with isoproterenol, attenuated the H₂O₂-induced decrease in  more effectively. In addition, CP-80633 decreased capillary pressure. These findings suggest that CP-80633 decreases the catabolism of intracellular cAMP in both vascular smooth muscle and endothelial cells and that CP-80633 would be more effective at decreasing the development of pulmonary edema because it affects both the barrier properties and hemodynamics of the vasculature that lead to increased transvascular protein and fluid fluxes.

In conclusion, the results of the present study demonstrate that an inhibitor of cAMP metabolism was more effective, compared with a stimulator of cAMP production, in reducing the transvascular clearance of albumin and lung weight induced by H₂O₂. Furthermore, the addition of DFO to the perfusate of the isolated lung minimized the conversion of H₂O₂ to an iron-dependent metabolite, thus maximizing the delivery of H₂O₂ to the vascular endothelium.