INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Platelet-activating factor (PAF) is a proinflammatory lipid mediator that is thought to contribute to a variety of inflammatory lung diseases, such as asthma or acute respiratory distress syndrome (ARDS). In addition, PAF has been shown to contribute to pulmonary injury from extrapulmonary lesions, such as intestinal ischemia-reperfusion (1). Direct administration of PAF causes bronchoconstriction, pulmonary hypertension, pulmonary edema, mucus secretion, and reduced ciliary beat frequency (2, 3). The mechanisms of the PAF-induced alterations in lung functions are only partly understood. Thromboxane derived from cyclooxygenase-1 (4) and to a lesser extent also leukotrienes are responsible for the PAF-induced bronchoconstriction and vasoconstriction (5). The mechanisms of the PAF-induced edema formation as well as the changes in mucus production or ciliary beating frequency are largely unknown.

Pulmonary edema caused by increased vascular permeability is an important clinical problem that is difficult to treat. PAF has been shown to play an important role in many models of pulmonary edema such as those induced by lipopolysaccharides (LPS) (6), antigen (7), interleukin-2 (IL-2) (8), or pulmonary (9) as well as intestinal ischemia-reperfusion (1). PAF causes edema by increasing endothelial permeability by an unknown mechanism. The few drugs that have been identified to ameliorate PAF-induced alterations in pulmonary vascular permeability are steroids (10), cyclic adenosine monophosphate (cAMP)-raising agents (11), vitamin D₃ (12), and copolymer of polyinosinic and polycytidylic acids (polyIC) (13). Two other agents, i.e., heparin sulfate (14) and dextrane sulfate (14) probably act by direct binding of PAF. Thus, any new agent that prevents PAF-induced edema is of both mechanistic and clinical interest.

Quinolines such as quinine, quinidine, or chloroquine are best known as antimalaria drugs or antiarrhythmics. Beyond this, a number of investigations support the hypothesis that quinolines possess anti-inflammatory properties. For instance, hydroxychloroquine was studied as an antiasthma drug in humans with promising results (15). To further investigate the antiphlogistic properties of quinolines and to gain more insight into the mechanism of PAF-induced edema formation, we have examined the effects of quinine, its diastereomer quinidine, and the antimalaria drug chloroquine on the PAF-induced responses in the perfused rat lung and on PAF- as well as LPS-induced edema in mice in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

Male Balb/c mice (25 to 30 g) or female Wistar rats (220 to 250 g; Hannover, Germany) were used. Pentobarbital sodium (Nembutal) was purchased from the Wirtschaftsgenossenschaft Deutscher Tierärzte (Hannover, Germany), bovine albumin (fraction V) from Serva (Heidelberg, Germany); acetylsalicylic acid (ASA), PAF, quinine and chloroquine from Sigma (Deisenhofen, Germany); quinidine from Sigma-Aldrich (Steinheim, Germany); SQ29548 (5-heptenoic acid, 7-[3-[[2-[(phenylamino)carbonyl]hydrazino]methyl]-7-oxabicyclo[2.2.1]hept-2-yl]-, [1S-[1,2(Z),3,4]]-) from Cayman Chemical Company (Ann Arbor, MI); AA861 (2,3,5-trimethyl-6(12-hydroxy-5,10-dodecadynyl)-1,4-benzochinone) from Biomol (Hamburg, Germany). Iloprost was a kind gift from Schering (Berlin, Germany).

Isolated Perfused Rat Lung Preparation (IPL)

The rat lungs were prepared and perfused essentially as described recently (16, 17). All equipment was obtained from Hugo Sachs Electronics (March-Hugstetten, Germany). Briefly, lungs were perfused at constant hydrostatic pressure (12 cm H₂O) through the pulmonary artery, which resulted in a flow rate of approximately 25 ml/min. As a perfusion medium we used Krebs-Henseleit buffer (37° C) that contained 2% albumin, 0.1% glucose, and 0.3% HEPES. The total amount of recirculating buffer was 100 ml. The lungs were suspended by the trachea and were ventilated by negative pressure ventilation with 80 breaths/min and a tidal volume of approximately 2 ml. Every 5 min a hyperinflation (20 cm H₂O) was performed. Artificial thorax chamber pressure was measured with a differential pressure transducer (Validyne DP 45-14), and air flow velocity with a pneumotachograph tube (Fleisch Type 0000) connected to a differential pressure transducer (Validyne DP 45-15). The perfusate flow (Narcomatic RT-500) and the arterial and venous pressure (Statham P23BB) were continuously monitored. The pH of the perfusate before entering the lung was kept at 7.35 by automatic bubbling of the buffer with CO₂ as soon as the pH exceeded this value. A weight transducer was integrated into the chamber lid and allowed continuous assessment of lung weight (18). Data were recorded on a Graphtec Linearrecorder WR 3310 and were also transmitted to a computer. For lung mechanics, the data were analyzed by applying the following formula:

where P is chamber pressure, C pulmonary compliance, VT tidal volume, and Raw airway resistance. The capillary filtration coefficient (K_f,c) was measured by the gravimetric method introduced by Drake and coworkers (19) and consisted of raising the perfusate pressure by 5 cm H₂O for 10 min. The resulting weight transient of both the increase as well as the decrease in lung weight was analyzed by a bi-exponential equation as recently described (20). Capillary pressure was measured by the double-occlusion method (21), i.e., the arterial and the venous line where simultaneously occluded for 30 s by electronically controlled solenoid pinch valves and the resulting arterial and venous pressures (which were similar) were taken as the capillary pressure.

Measurement of Tumor Necrosis Factor (TNF) and Interleukin-6 (IL-6) by ELISA

ELISAs were performed in flat-bottomed high-binding polystyrene microtiter plates (Greiner, Nürtingen, Germany), using specific rat anti-mouse monoclonal antibody pairs (biotinylated detecting mAb) that were purchased from Pharmingen (San Diego, CA). For detection of TNF, a protein G-plus-purified sheep anti-mouse TNF capture polyclonal Ab (protein solution 20 mg/ml) was used instead of the Pharmingen Ab. Streptavidin-peroxidase was from Jackson Immuno Research (West Grove, PA) and the peroxidase-chromogen BM blue (3,3'-5,5' tetramethyl-benzidine) was from Boehringer Mannheim (Mannheim, Germany). The detection limits were 10 pg/ml TNF and 10 pg/ml IL-6.

Measurement of Thromboxane

Samples taken from the perfusate were stored at 20° C. Thromboxane A₂ (TXA₂) was assessed as the stable by-product thromboxane B₂ (TXB₂) by enzyme immunoassay (EIA) (Cayman, Ann Arbor, MI). The crossreactivity of the detecting antibody was TXB₂ 100%, 2,3-dinor TXB₂ 8.2%, prostaglandins (e.g., PGD₂, PGE₂, 6-keto PGF₁) < 0.5%. Recovery was > 95%.

Experimental Design of the Perfused Lung Studies

PAF was always injected as a bolus of 5 nmol directly into the perfusate after 40 min of perfusion. All other agents were added to the buffer reservoir. ASA was prepared in bicarbonate solution; AA861 was prepared as a stock solution in ethanol, SQ29548 as a stock solution in ethanol; 10 µl of these stock solutions were added 10 min before PAF. Quinine, quinidine, and chloroquine were added directly to the perfusate 10 min before PAF.

Experimental Design of the In Vivo Studies

To study the effects of PAF, mice were intravenously (tail vein) injected with Evans Blue (20 mg/kg) and PAF simultaneously (22). To determine edema formation after intraperitoneal injection of LPS, Evans Blue was injected 6 h after LPS treatment. Seven minutes after Evans Blue injection, the animals were killed by intravenous injection (tail vein) of Nembutal (4.8 mg/mouse)/heparin (100 U/mouse) in a volume of 100 µl and edema formation was assessed as detailed subsequently. Quinine, quinidine, or ASA/AA861 were administered intravenously 30 min before PAF or LPS, WEB2086 1 h before LPS. ASA, quinidine, quinine, and WEB2086 were prepared as NaCl solutions, AA861 was dissolved in Sandimmune Placebo (Sandoz, Basel, Switzerland) 1/40 in NaCl.

After the animals were killed, lungs were perfused free of blood by perfusion with ice-cold phosphate-buffered saline (PBS) through the pulmonary artery. Then the lung tissue was separated into two parts, one of which was used to determine the wet to dry ratio. The other part was minced and Evans Blue was extracted in dimethylformamide for 24 h at 60° C. The samples were centrifuged and extinction of the supernatant was measured at 620 nm in a spectrophotometer. Evans Blue content of lung slices was calculated by Evans Blue standard curves and referred to the dry weight of the lung tissue.

Histology

At the end of the experiment representative slices of lung parenchyma were immersed in 4% buffered paraformaldehyde and processed according to standard procedures. Dewaxed paraffin sections were counterstained with hematoxylin-eosin. The histological preparations were examined by an experienced pathologist who was blinded to the experimental groups.

Statistics

Data were analyzed by unpaired Student's t test, either one-sided or two-sided as indicated. In case of repeated measurements (time courses) the maximal value of each curve was taken for analysis. The -error resulting from multiple comparisons was adjusted by the method of Hommel (23). In case of heteroscedasticity data were log-transformed prior to analysis. A value of p < 0.05 was considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The responses to PAF in isolated rat lungs perfused by constant pressure are characterized by a rapid decrease in perfusate flow, VT, and airflow velocity, by a rapid increase in Raw, and by slowly increasing weight gain (Figure 1). The increase in Raw and vascular resistance (Rv) is largely mediated by thromboxane (5). Accordingly, the thromboxane receptor antagonist SQ29548 prevented the PAF-induced increase in Raw (Figure 2A) and Rv (Figure 2B). In the presence of cyclooxygenase blockade, arachidonic acid metabolites appear to be shifted toward the lipoxygenase pathway (5); in line with this, the cyclooxygenase inhibitor ASA was less effective than SQ29548 in preventing PAF-induced broncho- and vasoconstriction (Figures 2A and 2B). The lipoxygenase inhibitor AA861 had no effect alone, but given together with ASA or the thromboxane receptor antagonist SQ29548 almost completely prevented both pressor responses (Figures 2A and 2B). Because of problems with weight measurement during negative pressure ventilation in our previous study, the edema formation in response to PAF was difficult to analyze at that time (5). However, we recently have developed a new weight transducer that also works in an environment with oscillating negative pressures (18). Using this novel weight transducer, Figure 2C shows that a small part of the PAF-induced weight gain formation can be blocked by administration of ASA or ASA/ AA861, whereas pretreatment with AA861, SQ29548, or with AA861/SQ29458 was without any effect. In line with this, U46619 did not alter lung weight ([18], see Figure 5). In addition, neither U46619 (100 nM) nor the stable prostacyclin analogue iloprost (300 µM) increased K_f,c, i.e., the difference in K_f,c before and 30 min after infusion of U46619 or iloprost was 0.00 ± 0.03 and 0.13 ± 0.18 ml · min¹ · 100 g¹ · cm H₂O¹, respectively (n = 3 each).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Pulmonary responses to perfusion of rat lungs with PAF. Shown are the original traces from one single experiment where a bolus of 5 nmol PAF was injected into the tubing leading to the pulmonary artery. Shown from top to bottom are perfusate flow velocity, VT, airflow velocity, relative lung weight, Raw, and the pressure in the artificial thorax chamber as generated by the ventilator. Every 5 min a sigh (deep breath) was initiated to open atelectatic areas.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Pulmonary responses to PAF in perfused rat lungs in the presence of eicosanoid antagonists. Shown are the maximal Raw (A) and Rv (B) that were obtained about 1.5 min after injection of PAF as well as the weight gain after 30 min (C ). 5 nmol PAF were injected as a bolus into the tubing leading to the pulmonary artery 40 min after beginning the perfusion. 500 µM ASA or 10 µM AA861 were given 10 min before injection of PAF. 500 µM ASA and 10 µM SQ29548 were administered 15 min before PAF when given together with 10 µM AA861 (always 10 min) in the same experiments. The number of experiments is shown in parentheses. Data are mean ± SD. *Larger (one-sided t test) than control (p < 0.05); * smaller (one-sided t test) than PAF (p < 0.05).

View larger version (23K): [in this window] [in a new window]   Figure 5.   Pretreatment with quinidine (Qd) in lungs perfused with U46619. Shown are Raw (left axis) and lung weight (right axis). 100 nM U46619 was added to the buffer reservoir at time point 41 min. 330 µM Qd were administered 10 min before injection of U46619. Raw: U46619 (squares), Qd + U46619 (triangles); Weight: U46619 (circles), QD + U46619 (inverted triangles). Data are mean ± SEM. There was no significant effect of Qd (p > 0.05, two-sided t test).

These results suggested that a small part of the PAF-induced edema formation is caused by a cyclooxygenase metabolite, but that the major part is mediated by a different mechanism. Next we investigated whether quinine, a substance whose anti-inflammatory properties have been known for long, prevents the PAF-induced edema formation. Figure 3C shows that approximately two-thirds of the weight gain caused by PAF was blocked in the presence of quinine, quinidine, or chloroquine. In addition, all three quinolines mitigated the increase in Raw and Rv caused by PAF (Figures 3A and 3B). The PAF-induced release of thromboxane was reduced in the presence of quinine (Figure 4). Quinidine, however, had no effect against U46619-induced increases in Raw (Figure 5) or Rv (not shown). In addition, quinidine in combination with U46619 did not alter lung weight (Figure 5).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Pulmonary responses to PAF in perfused rat lungs in the presence of quinolines. Shown are the maximal Raw (A) and Rv (B) that were obtained approximately 1.5 min after injection of PAF as well as the weight gain after 30 min (C ). 5 nmol PAF were injected as a bolus into the tubing leading to the pulmonary artery 40 min after beginning the perfusion. 100 µM quinine (100 Q), 330 µM quinine (330 Q), 100 µM quinidine (100 Qd), 330 µM quinidine (330 Qd), or 100 µM chloroquine (100 Q) were administered 10 min before injection of PAF. The number of experiments is shown in parentheses. Data are mean ± SD. *Larger (one-sided t test) than control (p < 0.05); * smaller (one-sided t test) than PAF (p < 0.05).

View larger version (16K): [in this window] [in a new window]   Figure 4.   Effect of quinine on PAF-induced thromboxane formation. 5 nmol PAF (squares) were injected as a bolus into the tubing leading to the pulmonary artery 40 min after beginning the perfusion. Thromboxane formation was followed by measuring the concentrations of the stable metabolite TXB2 in the venous effluate of the lungs. 100 µM quinine (circles) or 330 µM quinine (diamonds) were administered 10 min before injection of PAF. The number of independent experiments is shown in parenthesis. Data are mean ± SEM of three experiments each. Analyzed by the maximum of each curve, the TXB2 concentrations in experiments with 330 µM quinine/PAF were significantly different from both other curves (p < 0.05, two-sided t test).

Because both aspirin and quinine partly blocked PAF-induced edema formation, we investigated whether the effects of ASA and quinine were additive (Figures 6 and 7). To entirely eliminate all possible interactions between PAF-induced edema formation and vasoconstriction, we also added AA861 in these experiments (Figures 7A and 7B). Figure 6 shows that simultaneous treatment with ASA/AA861 and quinine prevented PAF-induced edema formation completely. To further demonstrate that the effects of ASA and quinine depend on reducing vascular permeability, in a subset of experiments K_f,c was measured 30 min before and 30 min after administration of PAF. Figure 6B shows that ASA as well as quinine both reduced the increase in K_f,c caused by PAF and that both together completely prevented the PAF-induced increase in K_f,c. Figure 6C shows that PAF causes a relatively stronger increase in postcapillary than in precapillary resistance and that all treatments affected both pre- and postcapillary resistance in a similar way. The capillary pressure in cm H₂O for the five groups shown in Figure 6 measured 2 min after PAF administration was controls 6.83 ± 1.5 (n = 14), PAF 7.83 ± 1.99 (6), ASA/AA861 + PAF, 7.52 ± 0.41 (5), quinine + PAF 7.85 ± 1.12 (4), ASA/AA861/quinine + PAF 7.64 ± 0.35 (5). There were no differences between any of these groups. In addition to quinine, also quinidine blocked the PAF-induced weight gain if lungs were additionally pretreated with ASA/AA861 (Figure 7C).

View larger version (31K): [in this window] [in a new window]   Figure 6.   PAF-induced weight gain (A), filtration coefficient (B), and segmental Rv (C ) in perfused lungs in the presence of eicosanoid antagonists and/or quinolines. PAF was always given as a 5-nmol bolus injection 40 min after beginning the perfusion. 500 µM ASA were given 15 min, 10 µM AA861 10 min, and 330 µM quinine 9 min before injection of PAF. (A) Weight gain: PAF (open circles, n = 14); ASA, AA861, and PAF (open squares, n = 9); quinine and PAF (open triangles, n = 6); ASA, AA861, quinine, and PAF (closed diamonds, n = 5). Data are mean ± SEM. All four curves are significantly different from each other (p < 0.05, two-sided t test). (B) Kf,c (expressed as ml · min1 · 100 g1 · cm H2O1): mean change in Kf,c (± SD) from 30 min before to 30 min after injection of PAF. (C) Segmental Rv: mean precapillary (Rvpre) and postcapillary resistance (± SD) 2 min after injection of PAF. *Different from control (two-sided t test, p < 0.05); #different from PAF (two-sided t test, p < 0.05).

View larger version (14K): [in this window] [in a new window]   Figure 7.   Pulmonary responses to PAF in perfused rat lungs in the presence of quinolines and eicosanoid block. Shown are the maximal Raw (A) and Rv (B) that were obtained approximately 1.5 min after injection of PAF as well as the weight gain after 30 min (C ). 5 nmol PAF were injected as a bolus into the tubing leading to the pulmonary artery 40 min after beginning of the perfusion. In addition, all lungs were treated with 500 µM ASA and 10 µM AA861 15 and 10 min before injection of PAF, respectively. 100 µM quinine (100 Q), 330 µM quinine (330 Q), 33 µM quinidine (33 Qd), or 100 µM quinidine (100 Qd) were administered 9 min before injection of PAF. The number of experiments is shown in parentheses. Data are mean ± SD. #Smaller (one-sided t test) than PAF (p < 0.05).

To investigate whether these mechanisms also apply in vivo, mice were injected with PAF and edema formation was assessed by extraction of Evans Blue from lung tissue. Figure 8 shows that pretreatment of mice with ASA/AA861 slightly, though not significantly, reduced the PAF-induced edema formation in vivo. Pretreatment with quinine or quinidine largely protected mice from pulmonary edema caused by PAF (Figure 8). LPS-induced pulmonary edema formation has been ascribed to PAF (6, 24). In line with this, we found that the PAF receptor antagonist WEB2086 prevented the LPS-induced formation of pulmonary edema. Thus, because quinine acted against PAF, we hypothesized that it should also protect against LPS-induced edema. Figure 9 shows that quinine as well as quinidine prevented LPS-induced Evans Blue extravasation into lung tissue. Additional evidence is presented in Figure 10, showing representative micrographs of lungs form untreated (3 analyzed), LPS-treated (5), and quinidine/LPS-treated mice (3). Compared with control lungs (Figures 10A and 10B), lungs from LPS-treated mice showed a prominent perivascular (Figure 10C) and alveolar septal edema formation (Figure 10D). Both perivascular edema and alveolar septal edema were apparently reduced in animals pretreated with quinidine before injection of LPS (Figure 10E and 10F), but not completely absent in comparison with control lungs (Figures 10A and 10B). Additionally, the leukocyte content in the lungs from mice pretreated with quinidine before injection of LPS (Figure 10F) was diminished. Moreover, circulating TNF and IL-6 concentrations were measured in LPS-treated mice. Here we found that quinine and quinidine prevented the LPS- induced formation of TNF, but not that of IL-6 (Table 1). Similarly, in isolated rat lungs quinine also prevented the TNF release elicited by endotoxin (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 8.   PAF-induced edema formation in mice in vivo. Pulmonary protein extravasation was measured as the amount of Evans Blue sequestered in the lung tissue 7 min after PAF treatment. Animals were pretreated either with ASA (100 mg/kg) and AA861 (50 mg/kg), quinine (115 mg/kg) or with quinidine (32.5 mg/kg) 30 min before injection of 0.1 µg PAF together with Evans Blue (20 mg/kg). Control (cont, n = 29), PAF (n = 25), ASA/AA861 (n = 6), quinine (n = 9), quinidine (n = 5). Data are mean ± SD. # Smaller (one-sided t test) than PAF (p < 0.05).

View larger version (14K): [in this window] [in a new window]   Figure 9.   LPS-induced edema formation in mice in vivo. Pulmonary protein extravasation was measured as the amount of Evans Blue sequestered in the lung tissue. Evans Blue (20 mg/kg) was injected intravenously 6 h after LPS (10 mg/kg) intraperitoneal injection. Seven minutes after Evans Blue injection, the animals were killed and the pulmonary Evans Blue content analyzed. WEB2086 (5 mg/kg), quinine (115 mg/kg), or quinidine (32.5 mg/kg) were injected intraperitoneally 30 min before injection of LPS. Control (n = 29), LPS (n = 15), WEB2086 (WEB, n = 6), quinine (n = 8), quinidine (n = 8). Data are mean ± SD. # Smaller (one-sided t test) than PAF (p < 0.05).

View larger version (100K): [in this window] [in a new window]   Figure 10.   Micrographs showing lungs from control (A, B), LPS-treated (C, D), and quinidine/LPS-treated (E, F ) lungs. The scale of graphs A, C, E is shown in graph E, and the scale of graphs B, D, F in graph F. In LPS-treated (10 mg/kg intraperitoneally) animals a marked perivascular (C ) and alveolar septal edema formation (D) as well as accumulation of leukocytes (D) were observed. Pretreatment with quinidine (32.5 mg/kg) markedly reduced the LPS-induced edema formation (E and F ), but did not completely prevent it in comparison with control lungs (A and B). Leukocyte content was also diminished by pretreatment with quinidine.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study suggests that quinolines form a new class of antiedematogenous drugs. We have shown that PAF-induced edema formation is composed of two separate mechanisms, a smaller part dependent on an unknown cyclooxygenase metabolite and a larger part that is blocked by quinolines. Quinolines were effective against PAF-and LPS-induced edema in mice in vivo as well as against PAF-induced edema in perfused rat lungs. In addition, we also demonstrated that quinolines suppress LPS-induced TNF formation as well as PAF- induced thromboxane formation and the vasoconstriction and bronchoconstriction evoked by PAF.

PAF-induced Edema Formation

PAF-induces edema by increasing vascular permeability. This conclusion is based on experiments in perfused lungs showing that PAF-induced edema is accompanied by an increased K_f,c in the presence of unchanged capillary pressure (5, 20, this study). This conclusion is further supported by findings reported here and elsewhere (25, 26) that PAF increases Evans Blue extravasation in vivo, because Evans Blue binds to albumin that cannot cross the intact endothelial barrier easily. In addition, in our perfused lung experiments the lungs were perfused by constant pressure, thereby largely minimizing the possibility of hydrostatic edema formation. The mechanisms, however, by which PAF increases vascular permeability are largely unknown.

Only few drugs have previously been shown to attenuate PAF-induced edema formation in perfused lungs, i.e., dextrane sulfate (14) and heparin sulfate (14) in guinea pigs, polyIC in rabbits (13) and steroids (our unpublished observations) as well as cAMP-raising agents (27) in rats. While the protective effect of dextrane and heparin sulfate was ascribed to direct PAF binding, the mechanistic basis for the protective effects of polyIC and steroids is not known. The anti-edematogenous properties of cAMP elevation have recently been discussed (28). There are also experiments which show that phosphodiesterase inhibitors (29), thromboxane receptor antagonists (25), lipoxygenase inhibitors (25, 26), cyclooxygenase inhibitors (25, 26), or steroids (10) reduce PAF-induced Evans Blue extravasation in vivo. However, such in vivo data are difficult to interpret because all of these agents also reduce PAF-induced hypertension that may exacerbate PAF-induced edema formation (25). Such a mechanism, however, is unlikely in our model, because the lungs are perfused by constant hydrostatic pressure and capillary pressure is not changed by PAF (5, this study). In fact, in our perfused lung model edema occurred also in the absence of any pressor responses (Figure 7).

The present study suggests that the PAF-induced edema formation proceeds by two distinct mechanisms: (1) two-thirds are mediated by a mechanism that is sensitive to quinolines; (2) one-third may be attributed to a cyclooxygenase metabolite. If both quinolines and a cyclooxygenase inhibitor were given, the PAF-induced edema was completely abolished. That the concentrations of ASA and the quinolines were sufficient, i.e., that the effect of cyclooxygenase inhibition and quinolines was additive, is concluded from the fact that ASA at the concentrations used completely blocks PAF- and also LPS-induced thromboxane formation (4, 5), as well as from the fact that in most instances when two different quinoline concentrations were employed the higher concentration was not significantly more effective than the lower one.

Quinolines are a long known family of therapeutic drugs; however, their mode of action is still only poorly understood. In the present study quinolines were active in two different species (rats and mice) as well as in vivo and in perfused lungs, suggesting that a more general type of mechanism is involved. Among the activities described for quinolines are blockade of ion channels such as potassium or sodium channels (30), blockade of the IP₃ receptor (31), and binding to the P glycoprotein (32). Which of these actions, if any, is involved in the reduction of PAF-induced edema formation is the subject of ongoing studies. The most likely explanation for the ability of quinolines to ameliorate PAF-induced edema is blockade of the inositol 1,4,5-trisphosphate (IP₃) receptor. This hypothesis is supported is supported by the following findings: First, heparin sulfate, another compound that blocks IP₃ receptors, also reduced PAF-induced edema (14). Second, increases in intracellular calcium that in the case of PAF depend on activation of the IP₃ receptor (31), may be important for edema formation in general (28), although this was not yet shown for PAF directly.

It has also been suggested that quinolines may block phospholipase A₂ (PLA₂) (33), a finding which is supported by the present study in which quinine pretreatment attenuated PAF-induced thromboxane release. However, from the present study the contribution of PLA₂ to PAF-induced edema remains unclear. Of the obvious arachidonic acid metabolites, here we have excluded leukotrienes (derived from 5-lipoxygenase) and thromboxane. In addition, the additive effects of ASA and quinolines suggest that the major anti-edematogenous effect of quinolines is independent form cyclooxygenase metabolites. The possibility that PLA₂ products other than arachidonic acid and/or arachidonic acid metabolites other than cyclooxygenase or 5-lipoxygenase products contribute to PAF-induced edema remains to be investigated.

The cyclooxygenase metabolite that appears to be responsible for one-third of the PAF-induced edema formation was not identified in the present study. The most likely candidate, however, is prostaglandin E₂ (PGE₂), a compound that is well known for its edematogenous properties. It was shown, for instance, that PGE₂ antibodies attenuated carageenan-induced paw edema (34). However, in preliminary experiments we have been unable to detect PGE₂ in the effluent of PAF-treated lungs. Metabolites that we could detect, such as thromboxane and prostacyclin, are generally not thought to be edematogenous and both substances failed to increase vascular permeability, e.g., K_f,c, when given directly. In addition, pretreatment with a thromboxane receptor antagonist, though effective against the PAF-induced pressor responses, had no effect on PAF-induced edema formation.

Responses Induced by LPS

We here report the novel finding that quinine prevents pulmonary edema during endotoxic shock. Because PAF is an important mediator of LPS-induced edema (6) and because quinolines were active also against PAF-induced edema, it appears likely that the protective effect of quinine depends at least partly on its action against PAF. However, part of the protection against LPS-induced edema may also depend on the reduction in LPS-induced TNF release by quinine. Although the evidence for a role of TNF in LPS-induced pulmonary edema is limited (35) and controversial (36, 37), this mechanism clearly seems possible as suggested by studies with direct TNF infusion (38) or TNF antibodies in bacteremic animals (39). The mechanism by which quinine interferes with LPs-induced TNF release is unknown at present, but has been ascribed to blockade of potassium channels (40). Interestingly, we found that IL-6 was not suppressed by quinine which suggests that the effect on TNF was specific and not due to an unspecific cytotoxic action of quinine. We would also like to point out that the present study complements our previous study on TNF release in the galactosamine/endotoxin model (41) where only 1 µg/kg LPS was used compared with a 10,000 times higher dose in the present study. Thus, quinine appears to be a very potent suppressor of LPS-induced TNF release.

PAF-induced Pressor Responses

Besides their antiedematogenous action, quinolines also attenuated the PAF-, but not the U46619-induced bronchoconstriction and vasoconstriction, suggesting that this is not caused by a direct smooth muscle relaxant effect. The fact that quinine reduced the amount of thromboxane formed in response to PAF adequately explains the potential of quinine to prevent PAF-induced pressor responses, because thromboxane is the major metabolite responsible for the induced pressor responses ([5], Figures 2A and 2B). The most likely explanations for the reduced thromboxane release in the presence of quinolines are inhibition of PLA₂ (33) and/or blockade of IP₃ receptors by quinolines (31). IP₃ receptors appear important because they are activated upon stimulation of the PAF receptor (42), leading to increased intracellular calcium concentrations, which in turn is necessary for activation of many other cellular processes including activation of PLA₂. Thus, whether the effect of quinolines on PLA₂ is direct or indirect remains to be established. In addition, it is important to note that quinolines possess the ability to interfere with production of mediators whose signal transduction is not dependent on the IP₃ pathway or on phospholipases such as TNF.

The present study has given new insight into the mechanisms of PAF-induced pulmonary edema. In addition, we have described novel anti-inflammatory properties of quinolines and identified them as a new class of antiedematogenous drugs. They have antiedematogenous properties in vivo and in vitro, and they interfere with the production of inflammatory mediators such as thromboxane and TNF. These properties of quinolines may be of therapeutic interest for the treatment of inflammatory lung diseases.