- Email: [email protected]
Immune-complex alveolitis in the rat: evidence for platelet activating factor and leukotrienes as mediators of the vascular lesions
European Journal of Pharmacology, 213 (1992) 63-70 ~_~ 1992 Elsevier Science Publishers B.V. All rights reserved 0014-2999/92/$05.0(I
63
EJP 52342
Immune-complex aiveolitis in the rat: evidence for platelet activating factor and leukotrienes as mediators of the vascular lesions W o t h a n T a v a r e s de L i m a , P i e r r e Sirois " a n d S o n i a J a n c a r Department of Immunology, hzstitute of Biomedical Sciences, Unit'ersity of S~o Pauh~, S~io Paulo, SP, Brasil aml " Department of Pharmacology, Unit,ersity of Sherbrooke, Sherbrtn~ke, Quebec, ('anada Received 29 August 1991, revised MS received 25 November 1991, 17 December 1991
In the present study we investigated thc involvement of lipid mediators in an experimental model of immune-complcx alveolitis induced in rat lungs by intrabronchial instillation of rabbit antibodies to ovalbumin followed by i.v. injection of the antigen• It was found that the reaction did not induce detectable oedema, as measured by the dry:wet weight ratio. A marked influx of neutrophils was observed in the bronchoalveolar lavage fluid, progressing from 6 to 24 h in parallel with the development of haemorrhagic lesions in lung parenchyma. The intensity of these lesions, evaluated by the conccntration of extravascular haemoglobin, was not significantly affected by pretreatment of the animals with a cyclo-oxygenase inhibitor (indomethacin), a thromboxane inhibitor (econazole) or a thromboxane antagonist (L-655,240). However, the antagonists of platelet activating factor (PAF), WEB-2086 and BN-52021, and the lipoxygenase inhibitors, nor-dihydroguaiaretic acid and L-663,536, all significantly inhibited the haemorrhagic lesions. A peptide leukotriene antagonist (L-660,711) had no effect. Furthermore, the PAF antagonists inhibited the levels of I,TB 4, but not of PGE 2 and thromboxane, released into the bronchoalveolar space 1 h after induction of the reaction. These results suggest that the haemorrhagic lesions in this model of immune-complex alveolitis are mediated by PAF and leukotrienes, possibly LTB 4. Immune-complex alvcolitis; Arthus reaction; PAF (platelet-activating factor, PAF-acether); Eicosanoids
1. Introduction
The triggering of an acute inflammatory reaction in the lungs by deposits of immune complexes results in a series of humoral and cellular events which culminate in acute lung injury. A number of occupational lung diseases, also known as allergic extrinsic alveolitis or immune-complex alveolitis, are triggered by deposition of immune complexes in the lungs (Pepys, 1969). Induction of a passive Arthus reaction in the lung was shown to be a suitable experimental model to study the mechanisms involved in this type of injury. In this model, as in the classical skin reaction, the tissue injury depends on the availability of both neutrophils and complement (Johnson and Ward, 1974)• Products released by activated neutrophils are responsible for the development of acute lesions at the site of deposition of immune complexes by inducing microvascular dam-
Correspondence to: S. Jancar, Departamento de Immunologia, lnstituto de Ci~ncias Biom6dicas, Universidade de S,~o Paulo, Avenida Prof. Lineu Prestes 2415, CEP 05508 S~o Paulo, SP, Brasil. Fax: 55.11.8130845.
agc, thrombosis, haemorrhagc (Johnson and Ward, 19741 and, in some cases, chronic lung injury with interstitial fibrosis (Brentjens et al., 1974). Although the participation of neutrophils in this acute lung injury is well documented, the factors involved in the recruitment of these cells to the site of injury and the mechanisms by which the neutrophils become activated to induce vascular damage are not completely understood. Eicosanoids and platclet activating factor (PAF) are potential mediators of the acute injury induced by immune complexes. Some of these lipid mediators are able to increase vascular pcrmcability, induce oedcma and cause cellular influx. Besides these actions, PAF may induce endothelial cell activation and damage, as well as activation of leucocytes to release lysosomal enzymes and to produce potent cytokines and oxygenderived radicals (reviewed by Braquet et al., 1987). At certain concentrations PAF can also prime neutrophils, which could lead to superoxide generation, clastase release, aggregation, adhesive glycoprotein expression and lysis or detachment of endothelial cells (Vcrcellotti et al., 1988). The participation of eicosanoids and PAF in Arthus-type reactions was previously demonstrated
64
in immune-complcx-induced pancreatitis (Jancar et al., 1988a), peritonitis (Jancar ct al., 1988b) and in the synovial fluid of rabbits with antigen-induced arthritis (Brum-Fernandes et al., 1989). In the classical skin reaction, PAF mediates the early increase in vasopermeability (Hellewell and Williams, 1986; Issekutz and Szpejda, 1986) and primes neutrophils (Warren et al., 1989). In an Arthus reaction induced in rat lungs, significant amounts of P G E 2, T X B . and LTB 4 were shown to be released into the bronchoalveolar space. The reaction was followed by systemic thrombocytopenia, which was completely reversed by PAF antagonists, suggesting that PAF was released (Tavares de Lima ct al., 1989). In the present study we investigated the role of cicosanoids and PAF in the pathological manifestations of immune-complex-induced alveolitis. We first characterized some of the pathological events in this model, which entails a passive reverse Arthus reaction induced in rat lungs. Subsequently, we investigated the role of eicosanoids and PAF on the development of the hacmorrhagic lesions which arc the most characteristic feature of this type of hypersensitivity reaction.
After drying (60°C for 24 h) the tissues were weighed again (dry weight). As positive control, one group received 1 m g / k g of lipopolysaccharide (endotoxin from Salmonella abortus) i.v. and the lung d r y : w e t weight ratio was determined 6 h later.
2.4. Bronchoah:eolar lacage The animals were killed by cervical dislocation and exsanguinated, whereafter the lungs were removed. A cannula was then inserted into one of the upper bronchi and the lavagc was performed with 4 × 5 ml of phosphate-buffered saline at 37°C. For cicosanoids analysis the lavagc fluid was stored at - 2 0 ° C until further treatment (see below).
2.5. Cell counting The bronchoalveolar lavage fluid was centrifuged at 5(X) × g for 10 min, the supernatant was removed and the cell pellet was resuspended in 1 ml of saline. The number of cells was determined by counting in a Neubauer chamber. Differential counts were performed on fixed and stained cell suspensions (0.05% crystal violet dissolved in 30% acetic acid).
2. Material and methods
2.1. Animals Male Wistar rats (160-200 g) from our own animal facilities were used throughout the experiments.
2.2. Induction of a passiee recerse Arthus reaction in rat lungs Rats were anaesthetized with chloral hydrate (400 m g / k g i.p.) and a cannula was inserted into the upper bronchus via tracheotomy. Rabbit IgG 2 antibodies to ovalbumin or rabbit antiserum to ovalbumin (100 p.l, containing 120-180 /zg of specific antibody protein) were instilled, intrabronchially, followed by i.v. injection of 10 mg of ovalbumin in a volume of 0.5 ml. As controls, groups of rats received either intrabronchial antiserum plus i.v. saline or intrabronchial normal rabbit serum plus i.v. ovalbumin.
2.3. Eealuation of oedema The d r y : w e t weight ratio of the lung tissue was determined at different times after induction of the Arthus reaction to evaluate oedema. Samples of tissue were taken from the area affected by the Arthus reaction, from the contralateral non-affected lungs and from lungs of untreated rats. The tissues were washed with cold saline and, after removal of excess liquid with blotting paper, were weighed immediately (wet weight).
2.6. Extraction and quantification of haemoglobin from lung tissue The concentration of haemoglobin (Hb) was determined colorimetricaily by the cyanmcthaemoglobin method (Van Kampcn and Zijlstra, 1961) using a commercially available kit. Since this method is used to determine Hb concentration in blood samples, wc first tested its adequacy to measure Hb in samples of lung tissue. To this end a known amount of Hb was injected into normal parenchyma from lungs pcrfused through the pulmonary artery with 50 ml of phosphate-buffered saline. The injected area was then removed and minced in 2 ml of potassium cyanide and hexacyanoferrate solution. After 24 h at room temperature in the dark, the tissue was removed, the sample was centrifuged and the optical density of the supernatant was determined spectrophotometrically at 546 nm. The concentration of Hb was calculated by comparison with a standard curve. This method allowed the recovery of 8 2 - 8 7 % of the injected Hb and was thus suitable for our purposes. The rats submitted to the Arthus reaction and their controls were killed at set times and samples of lung tissue, corresponding to the area of the reaction, were removed and processed as described above. The area of the reaction was visualized by a small amount of colloidal carbon instilled intrabronchially together with the antiserum. The concentration of Hb present in a fragment taken from the contralatcral lung of each
65 animal was subtracted from the value of Hb found in the affected lung and the results expressed as AHb.
P < 0.05 were considered as showing a significant difference between means.
2. 7. Drug treatments 3. Results
Groups of rats were treated with the following drugs 30 min before inducing the Arthus reaction: indomethacin, 4 m g / k g i.v.; nordihydroguaiaretic acid, 30 m g / k g i.p.; WEB 2086, 1 m g / k g i.v.; L-663,536, 10 m g / k g i.v.; L-655,240, 5 m g / k g i.v. and L-660,711, 10 m g / k g i.v. Econazole, 200 m g / k g was administered orally 18 and 2 h before induction of the reaction. BN-52021, 5 m g / k g i.v., was given 30 min before and 10 h after induction of the reaction.
2.8. Determination of PGE 2, TXB 2 and LTB 4 The cell-free supernatant from the bronchoalveolar lavage fluid was assayed for PGE z and TXB z by enzymatic immunoassay as described by Pradelles et al. (1985). Briefly, 100 t~l aliquots of each sample were incubated with the eicosanoids conjugated with acetylcholinesterase and the specific antiserum in 96-well microtitration plates, coated with anti-lgG immunoglobulins. After addition of the enzymatic substrate, the optical density of the samples was determined at 412 nm in a microplate ,-eader and the concentration of the eicosanoids calculated from standard curves. The concentration of LTB, was assayed by radioimmunoassay using a commercial kit.
2.9. Drugs used The following drugs were purchased from Sigma Chem. Co., St. Louis, USA; rabbit antibodies to egg albumin (whole serum); ovalbumin grade II; nor dihydroguaiaretic acid and indomethacin. The IgG fraction of rabbit antibodies to egg albumin was purchased from Cappel, Malvern, PA, USA; PAF (C~6) was from Bachem, Switzerland. The kit for determination of haemoglobin was from Boehringer-Mannheim, Germany; the reagents for enzyme immunoassays of PGE z and TXA 2 were from Cayman, Denver, Co, USA and the radioimmunoassay kit for LTB 4 was from Amersham Int., U.K. The following drugs were received as gifts: BN-52021 (IHB, France); WEB-2086 (Boehringer-lngelheim, Germany); econazole (Formil Quimica, SA, Brazil); compounds L-655,240; L-663,536, and L-660,711 (Merck Frosst Lab., Montreal, Canada).
2.10. Statistical analysis Data were compared with an analysis of variance. The means of the control and each treated group were compared with the Dunnett test (Zar, 1984). Values of
3.1. Characterization of the experimental model of immune-complex alt,eolitis 3.1.1. Ecaluation of lung oedema Lung oedcma was assayed as the dry:wet weight ratio of the lung tissue, at l-h intervals from 1 to 6 h and at 24 h after induction of the Arthus reaction. The reaction was induced by intrabronchial instillation of the antiserum in one of the lungs, followed by i.v. injection of ovalbumin. The contralateral lung of the same animal was used as control. The Arthus reaction did not cause detectable oedema. The dry:wet weight ratio of the lung tissue affected by the reaction was not significantly different from that of the contralateral, non-affected lung. Moreover, the ratio did not change as a function of time and did not significantly differ from that obtained from lungs of untreated animals (0.212 +_ 0.004 (mean _ S.D.); n = 5 - 6 in each group). In rats that had received 1 m g / k g of LPS (lipopolysaccharide), i.v., the weight ratio was 0.199 +_ 0.002 (mean +_ S.D. n = 9) 6 h after injection, indicating the presence of significant oedema. 3.1.2. Analysis of the cellular influx into the bronchoah'eolar space The number of cells in the bronchoalveolar lavage fluid was determined at 6 and 24 h after induction of Arthus reaction. The lavage was performed in the treated and control lungs of each animal. The number of cells present in the latter was similar to that in lungs from untreated animals and is presented as a single control value (C) in fig. 1. The total number of leucocytes in the lavage fluids collected at 6 h after antigen challenge (50.8 _+ 21.6 × 105/ml) was 3 times that present in control lavage fluid (16.7_+4.75× 10S/ml). However, as shown in fig. 1, this increase was mainly due to higher numbers of PMN (polymorphonuclear) leucocytes as the monocyte number did not change. This figure also shows that, at 6 h, rats receiving only the antibody intratracheally had a similarly increased number of PMN leucocytes in their [avage fluids. However, at 24 h after challenge, there was a clear difference between the two sets of lungs. Those with immune complexes showed almost 3-fold increases in PMNs over the antibody only group. There was also a smaller but significant increase in monocytes at 24 h. The other control group, which received normal rabbit serum intrabronchially and antigen intravenously, showed a pattern of cellular influx similar to that of the antiserum only group.
66 250" [ ]
•
3-
Sahe/Ab
A~Ab
1502
i N
E
iN
in_
*
I
¢r
¢r
0
6
I
24
11me (h)
Fig. 2. Lung tissue contents of haemoglobin after the Arthus reaction. The values for affecting lung were subtracted from the values for the contralateral, non-affected lung ( J l l b ) , 1, 6 and 24 h after induction of the Arthus reaction. Data represent the means + S.E. from 10 to 12 animals. ° P < 0.05.
m-
150-
N
3.2. Role of eicosanoids and PAF on the intensity of the haemorrhagic lesions
I00
|
,k 50
D c
6
24
Time (h)
Fig. I. Number of cells present in the bronchoalveolar lavage fluid of rats, 6 and 24 h after induction of the Arthus reaction ( A g / A b ) . As control, a group of animals received the antiserum intrabronchially and saline intravenously (saline/Ab). Control values (C) represented the number of cells in untreated lungs. Data represent the mean + S.E. of 8 to I 1 animals. '' P < 0.05 compared to the control group; ° P < 0.05 compared to the s a l i n e / A b group.
To investigate the role played by lipid mediators on the development of the lesions, groups of rats were treated with a cyclo-oxygenase inhibitor (indomethacin), a thromboxane synthetase inhibitor (cconazole), a thromboxane antagonist (L-655,240), two lipoxygenase inhibitors (nor-dihydroguaiaretic acid and L-663,536), a leukotrienc D 4 antagonist (L-660,711) and two PAFantagonists (BN-52021 and WEB-2086) prior to induction of thc Arthus reaction. The results obtained 24 h after the challenge (fig. 3) show that pre-treatment of the rats with indomethacin, cconazole or the compounds L-655,240 and L-660,711 did not significantly
! 1_
3.1.3. El,aluation of the haemorrhagic lesions The intensity of the haemorrhagic lesions in the lung was evaluated by measuring haemoglobin (Hb) extracted from a sample of lung parenchyma taken from the lobe where the Arthus reaction was induced and from an equivalent sample taken from the contralateral, non-affected lung. The difference between these two Hb values (AHb) was taken to represent the increase in Hb in the affected lung. The results obtained (fig. 2) show that the Arthus reaction increased thc amount of Hb both 6 and 24 h after the challenge. Control animals received the a n t i s e r u m intrabronchially and saline i.v. or normal rabbit serum intrabronchially and antigen intravenously and their lungs were analysed for Hb content at the same times as the experimental group. The AHb values were close to zero at all times examined.
.1
Indo
Eco
BN 52021
W E B NDGA 1,660 2~6 711
1.663 536
1.655 240
Fig. 3. Effects of pharmacological agents on the intensity of haemorrhagic lesions measured 24 h after induction of the Arthus reaction. Injury is expressed by the amount of haemoglobin present in the lung affected by the reaction, subtracted from the amount present in the contralateral lung (A Hb). Data represent the means _+S.E. from 6 to 10 animals; * P < [).[)5 compared to the non-treated group.
67
affect the zlHb values. Significant inhibition, however, was observed in the groups pretreated with BN-52021, WEB-2086, nor-dihydroguaiaretic acid and L-663,536, suggesting that PAF and leukotrienes could play a role in the development of the haemorrhagic lesions.
3.3. Effect of PAF-antagonists eicosanoids
on the release of
We have previously shown that eicosanoids arc released into the bronchoalveolar lavage fluid in this model of lung injury and that maximum release is observed in the first hour after challenge (Tavares de Lima et al., 1989). In the present experiments we investigated possible interactions between PAF and eicosanoids. Groups of rats were treated with the PAF
antagonists, BN-52021 and WEB-2086, before induction of the Arthus reaction. One hour later, the concentrations of P G E 2, TXB 2 and LTB 4 in the bronchoalveolar lavage fluid were determined in the treated and non-treated groups. The basal levels of these eicosanoids were determined in the lavagc fluids from untreated animals. The data presented in fig. 4 confirmed our previous results showing that the Arthus reaction induced release of cicosanoids into the bronchoalveolar space. Furthermore, they show that the PAF antagonists significantly inhibited the reicase of LTB 4 without decreasing the levels of PGE 2 and TXB 2. However, both antagonists increased P G E 2 levels to about twice their untreated values. Eicosanoid levels in the lavage fluid from rats receiving antibodies alone were not different from those of untreated rats.
4. Discussion
°
I
|.
12
1
z
j_
E
? I-
.
[]
i ~.~o 2110
I00
1'
c:~ m AIz/Ah All/All AII/Ah
Fig. 4. Effect of PAF antagonists on the concentration of eicosanoids in the bronchoalveolar lavage fluid of rat lungs 1 h after induction of the Arthus reaction. The animals were treated 30 min before induction of the reaction with 5 m g / k g of BN-52021 or 1 m g / k g of WEB-2086. Data represent the means + S.E. of four determinations. " P < 0.01 compared to the control group; " P < 0.001 compared to the A g / A b , non-treated group.
Our experiments show that an Arthus reaction locally induced in rat lung elicited inflammatory changes such as leucocyte infiltration and an increase in inflammatory mediators, but did not cause detectable oedema. We assessed leucocyte infiltration by measuring the ceils in bronchoalveolar lavagc, and the increases we observed are compatible with the earlier histopathological evidence of leucocyte infiltration in alveoli (Johnson and Ward, 1974) and of increased leucocytes in bronchoalveolar lavage (Brieland et al., 1987). Our results showed further that the leucocyte increase over 24 h was comprised chiefly of an increase in PMNs (over 100-fold), with a smaller increase in monocytes (about 4-fold). This differential effect may suggest a particular mixture of chemoattractant agents generated in this model. It was also clear that the influx of PMNs per se was not enough to produce the complete spectrum of inflammatory effects because instillation of antiserum alone induced an influx of neutrophils but no haemorrhagic lesions nor increases in inflammatory mediators in lung lavage. Either there is a minimum number of PMNs needed or the leucocytes attracted also need to be activated. The activation of leucocytes by immune complexes is well established (Cochrane, 1968; Fantone and Ward, 1982; Henson and Johnston, 1987) and could by itself serve to draw more leucocytes to thc site of deposition, apart from causing vascular damage after activation. Vascular damage in our model was assessed by the presence of haemorrhagic lesions. These were present 6 h after induction of the Arthus reaction and increased further at 24 h. The extent and intensity of these lesions were assessed from the amounts of haemoglobin remaining in the lung after perfusion of
68
the pulmonary circulation, i.e. extravascular haemoglobin and thus a measure of haemorrhagc. Thcrc could also be a contribution from blood not washed out of the pulmonary bed because of vasoconstriction and obstruction in the microvasculaturc due to thrombi. W h a t e v e r the exact causes of the increased haemoglobin content of the lung, the increase was restricted to the lung affected by the immune complex reaction. We therefore believe that this measure is a valid indicator of lung injury under these conditions. Increased levels of inflammatory lipid mediators in lung lavagc fluid have already been reported in this model of alveolitis (Tavares de Lima ct al., 1989). Our present work has further analysed the roles of these mediators by using specific inhibitors of thcir synthesis or antagonists of their action to alter the vascular damage induced in this model. Thc inhibitors and antagonists were used at doscs choscn from the relevant published reports of their activity. Thus the hacmorrhagic lesions appear to be mediated by PAF and leukotrienes as they were almost totally prevented by two antagonists of PAF action, BN-52021 (Braquet et al., 1987) and WEB-2086 (Casals-Stenzel et al., 1987), and by two lipoxygenasc inhibitors, nor-dihydroguaiaretic acid (Tappel ct al., 1953) and L-663,536 (Gillard ct al., 1988). Howcvcr, the LTD 4 receptor antagonist L-660,771 did not significantly affcct the intensity of the lesions. One possible interpretation of the lack of effect of L-660,711 is that the dose used (10 mg/kg) was not high enough to antagonize LTD a in vivo. However, much lower doses (0.5-1.(I mg/kg) inhibited LTD 4- and Ascaris sp-induced bronchoconstriction in guinea-pigs and squirrel monkeys, as well as antigen-induced bronchoconstriction in sensitized rats (Jones ct al., 1989). Although this compound is selective for LTD 4 in guinea-pig airway smooth muscle, in coronary vessels it antagonizes LTC 4 as well (Piper et al., 1990). Thus our result might reflect mediation by a non-peptide leukotriene, i.e. by LTB 4. We would favour this interpretation as LTB 4 was one of the eicosanoids released into the bronchoalvcolar space following induction of the reaction, it is a chemoattractant for PMNs and its level was markedly reduced by the PAF antagonists that prevented the haemorrhagic lesions. Our results also appear to rule out cyclo-oxygenase products as mediators of vascular damage, lndomethacin and an antagonist of thromboxane, L-655,240, both in adequate doses (Hall et al., 1987), were without any effect on the haemorrhagic lesions. The inhibitor of thromboxane synthesis, econazole (Jancar et al., 1991b), in a dose effective in preventing the thrombocytopenia accompanying this alveolitis (Tavarcs de Lima et al., 1989), did not significantly affect thc lung injury. This result was confirmed by additional data from our laboratory showing that pretreatment of rats with 10 m g / k g i.v. of another thromboxane inhibitor,
dazoxiben (Parry.et al., 1982), has no effect on the intensity of haemorrhagic lesions (data not shown). Furthermore, the PAF antagonists prevented lung injury but did not decrease the levels of TXB 2 released into the bronchoalveolar space. However, they did increase the levels of PGE 2, allowing the speculation that this action of PAF antagonists to increase prostaglandin production might contribute to their beneficial effect in this model. It is known that PGI 2 deactivates PMN leucocytes and that both PGE 2 and PG12 stimulate adenyl cyclase and increase intracellular cAMP (Boxer et al., 1980; Fantone et al., 1984; Rampart and Williams, 1986). However, this hypothesis cannot be applied here, since the thromboxane synthetase inhibitors, which are known to also increase the production of prostaglandins (De Freyn et al., 1982; Jancar et al., 1991b), did not show a protective effect. Although we did not measure PAF directly in our present experiments, its generation in this model is strongly indicated from the efficacy of PAF antagonists. This conclusion is in accordance with previous results showing the participation of PAF in immunecomplex hypersensitivity reactions in rats (Jancar et al., 1988a,b). In leucocytes and endothelial cells, both of which would be involved in our lung injury model, PAF formation precedes eicosanoid release (Stewart et al., 1990). In rats, injection of PAF into the peritoneal cavity induces the release of eicosanoids, including UI'B 4 (Jancar et al., 1989a). In our work, the release of LTB 4 appeared to be dependent of PAF action and, thus, on its formation; similar interactions between PAF and LTB 4 have been observed in rat peritoneal cavity stimulated by immune-complexes (Jancar et ai., 1989b) or by an anaphylactic reaction (Jancar et al., 1991a). Increased vascular permeability was an early event in this model as described by Johnson and Ward (1974) and therefore we expected to find oedcma in our experiments. However, the earlier work used higher amounts of antibody (250/xg protein) and 125i_albumin to estimate vascular permeability, whereas wc used less antibody (scc Methods) and assessed ocdema by changes in the dry:wet weight ratio. These differences and other factors, such as the magnitude of the permeability changes or the efficiency of lymphatic drainage preventing the accumulation of fluid in the lungs, could contribute to this discrepancy. Nevertheless, wc were able to measure altered dry:wet weight ratios after endotoxin, in accordance with Izumi and Bakhle (1988), demonstrating the ability of our method to detect pulmonary oedema. In summary, our results show that local deposition of immune-complexes in the lung caused microvascular damage exhibited as haemorrhagic lesions, together with an increased influx of leucocytes and an increased
69
level of inflammatory mediators in the bronchoalveolar space. The major mediator in this model of lung injury appears to be PAF, with a contribution from leukotrienes, probably LTB 4. Our results would suggest a place for PAF antagonists in the prevention or treatment of immune-complex alveolitis in occupational lung disease states.
Acknowledgements The authors gratefully acknowledge Dr. Y.S. Bakhle fi)r the careful revision of this manuscript and most helpful suggestions. This study was supported by grants from CNPq and FAPESP (Brasil) and MRC (Canada).
References Boxer, L.A., J.M. Allen, M. Schmidt, M. Yoder and R.L. Baehner, 1980, Inhibitkm of polymorphonuclear leukocyte adherence by prostacyclin, J. Lab. Clin. Med. 95, 672. Braquet, P., L. Touqui, Y.T. Shen and B.B. Vargaftig, 1987, Perspectives in platelet-activating factor research, Pharmacol. Rev. 39, 98. Brentjents, J.R., D.W. O'Connell, I.B. Pawlowski, K.C. llsu and G.A. Andres, 1974, Experimental immune complex disease of the lung. The pathogenesis of a laboratory model resembling certain human interstitial lung diseases, J. Exp. Med. 140, 105. Brieland, J.K., R.G. Kunkel and J.C Fantone, 1987, Pulmonary alveolar macrophage functkm during acute inflammatory lung injury, Am. Rev. Resp. Dis. 135. Brum-Fernandes, A.J., W. Cossermelli and S. Jancar, 1989, Slow-reacting substance of anaphylaxis (SRS-A) activity in the synovial fluids of rabbits with antigen-induced arthritis, Res. Commun. Chem. Pathol. Pharmacol. 63, 471. Casals-Stenzel, J., G. Muacevic and K.H. Weber, 1987, Pharmacological actions of WEB 21)86, a new specific antagonist of platelet activating factor, J. Pharmacol. Exp. Ther. 241,974. Cochrane, C.G., 1968, Immunologic tissue injury mediated by neutrophilic leukocytes, Adv. Immunol. 9, 97. De Freyn, G., H. Deckmyn and J. Vermylen, 1982, A thromboxane synthetase inhibitor reorients endoperoxide metabolism in whole blood towards prostacyclin and prostaglandin E2, Thromb. Res. 26. 389. Fantone. J.C. and P.A. Ward, 1982, Role of oxygen-derived free radicals and metabolites in leukocyte-dependent inflammatory reactions, Am. J. Pathol. 107, 397. Fantone, J.C., W.A. Marasco, L.J. Elgas and P.A. Ward, 1984, Stimulus specificity of prostaglandin inhibition of rabbit polymorphonuclear leukocyte lysosomal enzyme release and superoxide anion production, Am. J. Pathol. 115, 9. Gillard, J., A.W. Ford-Hutchinson, C. Chan, S. Charleson, D. Denis, A. Foster, R. Fortin, S. Leger, C.S. McFarlane, H. Mortom, H. Piechuta, D. Riendeau, C.A. Rouzer, J. Rokach, and R. Young, 1988, L-663.536 (MK-886)(3-[l-(4-chlorobenzyl)-3-t-butyl thio-5isopropylindol-2-yl]-2,2-dimethylpropanoic acid), a novel, orally active leukotriene biosynthesis inhibitor, Can. J. Physiol. Pharmacol. 67. 456. Hall, R.A., J. Gillard, Y. Guindon, G. Letts, E. Champion, D. Ethier, J. Evans, A.W. Ford-Hutchinson, R. Fortin, T.R. Jones, A. Lord, I[.E. Morton, J. Rokach and C. Yoakim, 1987, Pharmacology of L-655,240 (3-[ I-(4-chlorobenzyl)-5-fluom-3-methyl-indo-
2-yl]2,2-dimethyl propanoid acid); a potent, selective thromboxane/prostaglandin endoperoxide antagonist, Eur. J. Pharmacol. 1, 135. Hellewell, P.G. and T.J. Williams, 1986, A specific antagonist of platelet-activating factor suppresses oedema formation in an Arthus reaction but not oedema induced by leukocyte chemoattractants in rabbit skin. J. Immunol. 137. 3[)2. ttenson, P.M. and R.B. Johnston Jr., 1087, Tissue injury in inflammation. Oxidants, proteinases, and cationic proteins, J. Clin. Invest. 79, 669. Issekutz. A.C. and M. Szpejda, 1986, Evidence that platelet activating factor may mediate some acute inflammatory responses. Studies with the platelet-activating factor antagonist CV3988. Lab. Invest. 54. 275. Izumi, 1. and Y.S. Bakhle. 1988, Modification by steroids of pulmonary oedema and prostaglandin E 2 pharmacokineties induced by endotoxin in rats, Br. J. Pharmacol. 93, 955. Jancar, S., G. Dc Giacobbi, M. Mariano, P. Sirois and P. Braquet, 1988a, Immune-complex induced pancreatitis effect of BN-52021 a selective PAF-antagonist, Prostaglandins 35(5), 757. Jancar, S.. P. Sirois and P. Braquet, 1988b, Evidence for platelet activating factor (PAF) involvement in immune complex hypersensitivity reactions provided by BN 52021, in: Ginkgolides Chem. Biol. Pharm. Clin. Perspectives, ed. P. Braquet and J.R. Prous, p. 493. Jancar, S., P. Braquet and P. Sirois, 1989a, Release of eicosanoids in rat peritoneal cavity stimulated with platelct-activating factor (PAF). Effect of the PAF-antagonist BN 52021, Prostagl. Leukotr. Essential Fatty Acids 37, 23. Jancar, S., P. Braquet and P. Sirois, 1989b, Release of eicosanoids in rat peritoneal cavity during the Arthus reaction. Effect of the PAF-antagonist BN-52021 and indomethacin. Int. J. lmmunopharmacol. 11, 129. Jancar. S., M.G. Sirois, J. Carrier and P. Sirois, 1991a, PAF induces plasma extravasation and releases eicosanoids in the rat during anaphylaxis, Inflammation 15(5), 347. Jancar. S., C.F.P. Teixeira. W. Tavares de Lima. A. lloshikawaFujimura and P. Sirois. 1991b, Inhibitory effect of econazole nitrate on the release of thromboxane, Agents Actions 34, 387. Johnson, J.K. and P.A. Ward, 1974, Acute immunologic pulmonary alveolitis, J. CIin. Invest. 54, 349. Jones, T.R., R. Zamboni, M. Belley, E. Champion, L. Charlete, A.W. Ford-Hutchinson, R. Frenete. J.Y. Gauthier, S. Leger, P. Masson, C.S. McFarlane, H. Piechuta, J. Rokach, H. Williams and R.N. Young, 1989, Pharmacology of L-660,711 (MK-571): a novel potent and selective leukotriene D4 receptor antagonist, Can. J. Physiol. Pharmacol. 67, 17. Parry. M.J., M.J. Randall, E. ]tawkeswood, P.E. Cross and R.P. Dickindon, 1982, Enhanced production of prostacyclin in blood after treatment with selective thromboxane synthetase inhibitor, UK38485, Br. J. Pharmacol. 77, 547P. Pepys, J., 1969, Hypersensitivity diseases of the lungs due to fungi and organic dusts, Monogr. Allergy 4. I. Piper, P.J., A.P. Sampson, H. bin Yaacob and J.M. Mcleod, 1990, Leukotrienes in the cardiovascular system, in: Advances in Prostaglandin, Thromboxane, and Leukotriene Research, Vol. 20, ed. B. Samuelsson et al. (Raven Press, Ltd., New York) p. 146. Pradelles, P.. J. Grassi and J. Maclouf, 1985, Enzyme immunoassays of eicosanoids using acetylcholine esterase as label: an alternative to radioimmunoassay, Anal. Chem. 57, 1170. Rampart, M. and T.J. Williams, 1986, Polymorphonuelear leukocyte-dependent plasma leakage in rabbit skin is enhanced or inhibited by prostacyelin, depending on the route of administration, Am. J. Pathol. 124, 66. Stewart. A.G., P.N. Dubbin, T. llarris and G.J. Dusting, 1990. Platelet-activating factor may act as a second messenger in the
70 release of eicosanoids and superoxide anions from Icukocytes and endothelial cells, Proc. Natl. Acad. Sci. U.S.A. 87, 3215. Tappel, A.L.. W.O. Lindbcrg and P.D. Beyer, 1953, Effect of temperature and antioxidants upon the lipoxydase-catalyzed oxydation of sodium linoleate, Arch. Biochem. Biophys. 42, 293. Tavares de Lima, W., C.F.P. Teixeira, P. Sirois and S. Jancar, 1989, Involvement of eicosanoids and PAF in immune-complex alveolitis, Braziliian J. Mcd. Biol. Res. 22, 745. Van Kampen, E.J. and W.G. Zijlstra, 1961. Standardization of hemoglobinometry. I1. The hemoglobincyanide method, Clin. ('him. Acta, 6, 538.
Vercelloti, G.M., II.Q. Yin, K.S. Gustafson, R.D. Nelson, J.S. Jacob, 1988, Platelet-activating factor primes neutrophil responses to agonists: Role in promoting neutrophil-mediated endothelial damage, Blood 71, 1100. Warren, J.S., D.M. Mandel, K.J. Johnson and P.A. Ward, 1989, Evidence for the role of platelet-aetivating factor in immune complex vasculitis in the rat, J. Clin. Invest. 83, 669. Zar, J.H., 1984, Biostatistical Analysis, 2nd edn. (Prentice Hall, New Jersey) p. 718.
63
EJP 52342
Immune-complex aiveolitis in the rat: evidence for platelet activating factor and leukotrienes as mediators of the vascular lesions W o t h a n T a v a r e s de L i m a , P i e r r e Sirois " a n d S o n i a J a n c a r Department of Immunology, hzstitute of Biomedical Sciences, Unit'ersity of S~o Pauh~, S~io Paulo, SP, Brasil aml " Department of Pharmacology, Unit,ersity of Sherbrooke, Sherbrtn~ke, Quebec, ('anada Received 29 August 1991, revised MS received 25 November 1991, 17 December 1991
In the present study we investigated thc involvement of lipid mediators in an experimental model of immune-complcx alveolitis induced in rat lungs by intrabronchial instillation of rabbit antibodies to ovalbumin followed by i.v. injection of the antigen• It was found that the reaction did not induce detectable oedema, as measured by the dry:wet weight ratio. A marked influx of neutrophils was observed in the bronchoalveolar lavage fluid, progressing from 6 to 24 h in parallel with the development of haemorrhagic lesions in lung parenchyma. The intensity of these lesions, evaluated by the conccntration of extravascular haemoglobin, was not significantly affected by pretreatment of the animals with a cyclo-oxygenase inhibitor (indomethacin), a thromboxane inhibitor (econazole) or a thromboxane antagonist (L-655,240). However, the antagonists of platelet activating factor (PAF), WEB-2086 and BN-52021, and the lipoxygenase inhibitors, nor-dihydroguaiaretic acid and L-663,536, all significantly inhibited the haemorrhagic lesions. A peptide leukotriene antagonist (L-660,711) had no effect. Furthermore, the PAF antagonists inhibited the levels of I,TB 4, but not of PGE 2 and thromboxane, released into the bronchoalveolar space 1 h after induction of the reaction. These results suggest that the haemorrhagic lesions in this model of immune-complex alveolitis are mediated by PAF and leukotrienes, possibly LTB 4. Immune-complex alvcolitis; Arthus reaction; PAF (platelet-activating factor, PAF-acether); Eicosanoids
1. Introduction
The triggering of an acute inflammatory reaction in the lungs by deposits of immune complexes results in a series of humoral and cellular events which culminate in acute lung injury. A number of occupational lung diseases, also known as allergic extrinsic alveolitis or immune-complex alveolitis, are triggered by deposition of immune complexes in the lungs (Pepys, 1969). Induction of a passive Arthus reaction in the lung was shown to be a suitable experimental model to study the mechanisms involved in this type of injury. In this model, as in the classical skin reaction, the tissue injury depends on the availability of both neutrophils and complement (Johnson and Ward, 1974)• Products released by activated neutrophils are responsible for the development of acute lesions at the site of deposition of immune complexes by inducing microvascular dam-
Correspondence to: S. Jancar, Departamento de Immunologia, lnstituto de Ci~ncias Biom6dicas, Universidade de S,~o Paulo, Avenida Prof. Lineu Prestes 2415, CEP 05508 S~o Paulo, SP, Brasil. Fax: 55.11.8130845.
agc, thrombosis, haemorrhagc (Johnson and Ward, 19741 and, in some cases, chronic lung injury with interstitial fibrosis (Brentjens et al., 1974). Although the participation of neutrophils in this acute lung injury is well documented, the factors involved in the recruitment of these cells to the site of injury and the mechanisms by which the neutrophils become activated to induce vascular damage are not completely understood. Eicosanoids and platclet activating factor (PAF) are potential mediators of the acute injury induced by immune complexes. Some of these lipid mediators are able to increase vascular pcrmcability, induce oedcma and cause cellular influx. Besides these actions, PAF may induce endothelial cell activation and damage, as well as activation of leucocytes to release lysosomal enzymes and to produce potent cytokines and oxygenderived radicals (reviewed by Braquet et al., 1987). At certain concentrations PAF can also prime neutrophils, which could lead to superoxide generation, clastase release, aggregation, adhesive glycoprotein expression and lysis or detachment of endothelial cells (Vcrcellotti et al., 1988). The participation of eicosanoids and PAF in Arthus-type reactions was previously demonstrated
64
in immune-complcx-induced pancreatitis (Jancar et al., 1988a), peritonitis (Jancar ct al., 1988b) and in the synovial fluid of rabbits with antigen-induced arthritis (Brum-Fernandes et al., 1989). In the classical skin reaction, PAF mediates the early increase in vasopermeability (Hellewell and Williams, 1986; Issekutz and Szpejda, 1986) and primes neutrophils (Warren et al., 1989). In an Arthus reaction induced in rat lungs, significant amounts of P G E 2, T X B . and LTB 4 were shown to be released into the bronchoalveolar space. The reaction was followed by systemic thrombocytopenia, which was completely reversed by PAF antagonists, suggesting that PAF was released (Tavares de Lima ct al., 1989). In the present study we investigated the role of cicosanoids and PAF in the pathological manifestations of immune-complex-induced alveolitis. We first characterized some of the pathological events in this model, which entails a passive reverse Arthus reaction induced in rat lungs. Subsequently, we investigated the role of eicosanoids and PAF on the development of the hacmorrhagic lesions which arc the most characteristic feature of this type of hypersensitivity reaction.
After drying (60°C for 24 h) the tissues were weighed again (dry weight). As positive control, one group received 1 m g / k g of lipopolysaccharide (endotoxin from Salmonella abortus) i.v. and the lung d r y : w e t weight ratio was determined 6 h later.
2.4. Bronchoah:eolar lacage The animals were killed by cervical dislocation and exsanguinated, whereafter the lungs were removed. A cannula was then inserted into one of the upper bronchi and the lavagc was performed with 4 × 5 ml of phosphate-buffered saline at 37°C. For cicosanoids analysis the lavagc fluid was stored at - 2 0 ° C until further treatment (see below).
2.5. Cell counting The bronchoalveolar lavage fluid was centrifuged at 5(X) × g for 10 min, the supernatant was removed and the cell pellet was resuspended in 1 ml of saline. The number of cells was determined by counting in a Neubauer chamber. Differential counts were performed on fixed and stained cell suspensions (0.05% crystal violet dissolved in 30% acetic acid).
2. Material and methods
2.1. Animals Male Wistar rats (160-200 g) from our own animal facilities were used throughout the experiments.
2.2. Induction of a passiee recerse Arthus reaction in rat lungs Rats were anaesthetized with chloral hydrate (400 m g / k g i.p.) and a cannula was inserted into the upper bronchus via tracheotomy. Rabbit IgG 2 antibodies to ovalbumin or rabbit antiserum to ovalbumin (100 p.l, containing 120-180 /zg of specific antibody protein) were instilled, intrabronchially, followed by i.v. injection of 10 mg of ovalbumin in a volume of 0.5 ml. As controls, groups of rats received either intrabronchial antiserum plus i.v. saline or intrabronchial normal rabbit serum plus i.v. ovalbumin.
2.3. Eealuation of oedema The d r y : w e t weight ratio of the lung tissue was determined at different times after induction of the Arthus reaction to evaluate oedema. Samples of tissue were taken from the area affected by the Arthus reaction, from the contralateral non-affected lungs and from lungs of untreated rats. The tissues were washed with cold saline and, after removal of excess liquid with blotting paper, were weighed immediately (wet weight).
2.6. Extraction and quantification of haemoglobin from lung tissue The concentration of haemoglobin (Hb) was determined colorimetricaily by the cyanmcthaemoglobin method (Van Kampcn and Zijlstra, 1961) using a commercially available kit. Since this method is used to determine Hb concentration in blood samples, wc first tested its adequacy to measure Hb in samples of lung tissue. To this end a known amount of Hb was injected into normal parenchyma from lungs pcrfused through the pulmonary artery with 50 ml of phosphate-buffered saline. The injected area was then removed and minced in 2 ml of potassium cyanide and hexacyanoferrate solution. After 24 h at room temperature in the dark, the tissue was removed, the sample was centrifuged and the optical density of the supernatant was determined spectrophotometrically at 546 nm. The concentration of Hb was calculated by comparison with a standard curve. This method allowed the recovery of 8 2 - 8 7 % of the injected Hb and was thus suitable for our purposes. The rats submitted to the Arthus reaction and their controls were killed at set times and samples of lung tissue, corresponding to the area of the reaction, were removed and processed as described above. The area of the reaction was visualized by a small amount of colloidal carbon instilled intrabronchially together with the antiserum. The concentration of Hb present in a fragment taken from the contralatcral lung of each
65 animal was subtracted from the value of Hb found in the affected lung and the results expressed as AHb.
P < 0.05 were considered as showing a significant difference between means.
2. 7. Drug treatments 3. Results
Groups of rats were treated with the following drugs 30 min before inducing the Arthus reaction: indomethacin, 4 m g / k g i.v.; nordihydroguaiaretic acid, 30 m g / k g i.p.; WEB 2086, 1 m g / k g i.v.; L-663,536, 10 m g / k g i.v.; L-655,240, 5 m g / k g i.v. and L-660,711, 10 m g / k g i.v. Econazole, 200 m g / k g was administered orally 18 and 2 h before induction of the reaction. BN-52021, 5 m g / k g i.v., was given 30 min before and 10 h after induction of the reaction.
2.8. Determination of PGE 2, TXB 2 and LTB 4 The cell-free supernatant from the bronchoalveolar lavage fluid was assayed for PGE z and TXB z by enzymatic immunoassay as described by Pradelles et al. (1985). Briefly, 100 t~l aliquots of each sample were incubated with the eicosanoids conjugated with acetylcholinesterase and the specific antiserum in 96-well microtitration plates, coated with anti-lgG immunoglobulins. After addition of the enzymatic substrate, the optical density of the samples was determined at 412 nm in a microplate ,-eader and the concentration of the eicosanoids calculated from standard curves. The concentration of LTB, was assayed by radioimmunoassay using a commercial kit.
2.9. Drugs used The following drugs were purchased from Sigma Chem. Co., St. Louis, USA; rabbit antibodies to egg albumin (whole serum); ovalbumin grade II; nor dihydroguaiaretic acid and indomethacin. The IgG fraction of rabbit antibodies to egg albumin was purchased from Cappel, Malvern, PA, USA; PAF (C~6) was from Bachem, Switzerland. The kit for determination of haemoglobin was from Boehringer-Mannheim, Germany; the reagents for enzyme immunoassays of PGE z and TXA 2 were from Cayman, Denver, Co, USA and the radioimmunoassay kit for LTB 4 was from Amersham Int., U.K. The following drugs were received as gifts: BN-52021 (IHB, France); WEB-2086 (Boehringer-lngelheim, Germany); econazole (Formil Quimica, SA, Brazil); compounds L-655,240; L-663,536, and L-660,711 (Merck Frosst Lab., Montreal, Canada).
2.10. Statistical analysis Data were compared with an analysis of variance. The means of the control and each treated group were compared with the Dunnett test (Zar, 1984). Values of
3.1. Characterization of the experimental model of immune-complex alt,eolitis 3.1.1. Ecaluation of lung oedema Lung oedcma was assayed as the dry:wet weight ratio of the lung tissue, at l-h intervals from 1 to 6 h and at 24 h after induction of the Arthus reaction. The reaction was induced by intrabronchial instillation of the antiserum in one of the lungs, followed by i.v. injection of ovalbumin. The contralateral lung of the same animal was used as control. The Arthus reaction did not cause detectable oedema. The dry:wet weight ratio of the lung tissue affected by the reaction was not significantly different from that of the contralateral, non-affected lung. Moreover, the ratio did not change as a function of time and did not significantly differ from that obtained from lungs of untreated animals (0.212 +_ 0.004 (mean _ S.D.); n = 5 - 6 in each group). In rats that had received 1 m g / k g of LPS (lipopolysaccharide), i.v., the weight ratio was 0.199 +_ 0.002 (mean +_ S.D. n = 9) 6 h after injection, indicating the presence of significant oedema. 3.1.2. Analysis of the cellular influx into the bronchoah'eolar space The number of cells in the bronchoalveolar lavage fluid was determined at 6 and 24 h after induction of Arthus reaction. The lavage was performed in the treated and control lungs of each animal. The number of cells present in the latter was similar to that in lungs from untreated animals and is presented as a single control value (C) in fig. 1. The total number of leucocytes in the lavage fluids collected at 6 h after antigen challenge (50.8 _+ 21.6 × 105/ml) was 3 times that present in control lavage fluid (16.7_+4.75× 10S/ml). However, as shown in fig. 1, this increase was mainly due to higher numbers of PMN (polymorphonuclear) leucocytes as the monocyte number did not change. This figure also shows that, at 6 h, rats receiving only the antibody intratracheally had a similarly increased number of PMN leucocytes in their [avage fluids. However, at 24 h after challenge, there was a clear difference between the two sets of lungs. Those with immune complexes showed almost 3-fold increases in PMNs over the antibody only group. There was also a smaller but significant increase in monocytes at 24 h. The other control group, which received normal rabbit serum intrabronchially and antigen intravenously, showed a pattern of cellular influx similar to that of the antiserum only group.
66 250" [ ]
•
3-
Sahe/Ab
A~Ab
1502
i N
E
iN
in_
*
I
¢r
¢r
0
6
I
24
11me (h)
Fig. 2. Lung tissue contents of haemoglobin after the Arthus reaction. The values for affecting lung were subtracted from the values for the contralateral, non-affected lung ( J l l b ) , 1, 6 and 24 h after induction of the Arthus reaction. Data represent the means + S.E. from 10 to 12 animals. ° P < 0.05.
m-
150-
N
3.2. Role of eicosanoids and PAF on the intensity of the haemorrhagic lesions
I00
|
,k 50
D c
6
24
Time (h)
Fig. I. Number of cells present in the bronchoalveolar lavage fluid of rats, 6 and 24 h after induction of the Arthus reaction ( A g / A b ) . As control, a group of animals received the antiserum intrabronchially and saline intravenously (saline/Ab). Control values (C) represented the number of cells in untreated lungs. Data represent the mean + S.E. of 8 to I 1 animals. '' P < 0.05 compared to the control group; ° P < 0.05 compared to the s a l i n e / A b group.
To investigate the role played by lipid mediators on the development of the lesions, groups of rats were treated with a cyclo-oxygenase inhibitor (indomethacin), a thromboxane synthetase inhibitor (cconazole), a thromboxane antagonist (L-655,240), two lipoxygenase inhibitors (nor-dihydroguaiaretic acid and L-663,536), a leukotrienc D 4 antagonist (L-660,711) and two PAFantagonists (BN-52021 and WEB-2086) prior to induction of thc Arthus reaction. The results obtained 24 h after the challenge (fig. 3) show that pre-treatment of the rats with indomethacin, cconazole or the compounds L-655,240 and L-660,711 did not significantly
! 1_
3.1.3. El,aluation of the haemorrhagic lesions The intensity of the haemorrhagic lesions in the lung was evaluated by measuring haemoglobin (Hb) extracted from a sample of lung parenchyma taken from the lobe where the Arthus reaction was induced and from an equivalent sample taken from the contralateral, non-affected lung. The difference between these two Hb values (AHb) was taken to represent the increase in Hb in the affected lung. The results obtained (fig. 2) show that the Arthus reaction increased thc amount of Hb both 6 and 24 h after the challenge. Control animals received the a n t i s e r u m intrabronchially and saline i.v. or normal rabbit serum intrabronchially and antigen intravenously and their lungs were analysed for Hb content at the same times as the experimental group. The AHb values were close to zero at all times examined.
.1
Indo
Eco
BN 52021
W E B NDGA 1,660 2~6 711
1.663 536
1.655 240
Fig. 3. Effects of pharmacological agents on the intensity of haemorrhagic lesions measured 24 h after induction of the Arthus reaction. Injury is expressed by the amount of haemoglobin present in the lung affected by the reaction, subtracted from the amount present in the contralateral lung (A Hb). Data represent the means _+S.E. from 6 to 10 animals; * P < [).[)5 compared to the non-treated group.
67
affect the zlHb values. Significant inhibition, however, was observed in the groups pretreated with BN-52021, WEB-2086, nor-dihydroguaiaretic acid and L-663,536, suggesting that PAF and leukotrienes could play a role in the development of the haemorrhagic lesions.
3.3. Effect of PAF-antagonists eicosanoids
on the release of
We have previously shown that eicosanoids arc released into the bronchoalveolar lavage fluid in this model of lung injury and that maximum release is observed in the first hour after challenge (Tavares de Lima et al., 1989). In the present experiments we investigated possible interactions between PAF and eicosanoids. Groups of rats were treated with the PAF
antagonists, BN-52021 and WEB-2086, before induction of the Arthus reaction. One hour later, the concentrations of P G E 2, TXB 2 and LTB 4 in the bronchoalveolar lavage fluid were determined in the treated and non-treated groups. The basal levels of these eicosanoids were determined in the lavagc fluids from untreated animals. The data presented in fig. 4 confirmed our previous results showing that the Arthus reaction induced release of cicosanoids into the bronchoalveolar space. Furthermore, they show that the PAF antagonists significantly inhibited the reicase of LTB 4 without decreasing the levels of PGE 2 and TXB 2. However, both antagonists increased P G E 2 levels to about twice their untreated values. Eicosanoid levels in the lavage fluid from rats receiving antibodies alone were not different from those of untreated rats.
4. Discussion
°
I
|.
12
1
z
j_
E
? I-
.
[]
i ~.~o 2110
I00
1'
c:~ m AIz/Ah All/All AII/Ah
Fig. 4. Effect of PAF antagonists on the concentration of eicosanoids in the bronchoalveolar lavage fluid of rat lungs 1 h after induction of the Arthus reaction. The animals were treated 30 min before induction of the reaction with 5 m g / k g of BN-52021 or 1 m g / k g of WEB-2086. Data represent the means + S.E. of four determinations. " P < 0.01 compared to the control group; " P < 0.001 compared to the A g / A b , non-treated group.
Our experiments show that an Arthus reaction locally induced in rat lung elicited inflammatory changes such as leucocyte infiltration and an increase in inflammatory mediators, but did not cause detectable oedema. We assessed leucocyte infiltration by measuring the ceils in bronchoalveolar lavagc, and the increases we observed are compatible with the earlier histopathological evidence of leucocyte infiltration in alveoli (Johnson and Ward, 1974) and of increased leucocytes in bronchoalveolar lavage (Brieland et al., 1987). Our results showed further that the leucocyte increase over 24 h was comprised chiefly of an increase in PMNs (over 100-fold), with a smaller increase in monocytes (about 4-fold). This differential effect may suggest a particular mixture of chemoattractant agents generated in this model. It was also clear that the influx of PMNs per se was not enough to produce the complete spectrum of inflammatory effects because instillation of antiserum alone induced an influx of neutrophils but no haemorrhagic lesions nor increases in inflammatory mediators in lung lavage. Either there is a minimum number of PMNs needed or the leucocytes attracted also need to be activated. The activation of leucocytes by immune complexes is well established (Cochrane, 1968; Fantone and Ward, 1982; Henson and Johnston, 1987) and could by itself serve to draw more leucocytes to thc site of deposition, apart from causing vascular damage after activation. Vascular damage in our model was assessed by the presence of haemorrhagic lesions. These were present 6 h after induction of the Arthus reaction and increased further at 24 h. The extent and intensity of these lesions were assessed from the amounts of haemoglobin remaining in the lung after perfusion of
68
the pulmonary circulation, i.e. extravascular haemoglobin and thus a measure of haemorrhagc. Thcrc could also be a contribution from blood not washed out of the pulmonary bed because of vasoconstriction and obstruction in the microvasculaturc due to thrombi. W h a t e v e r the exact causes of the increased haemoglobin content of the lung, the increase was restricted to the lung affected by the immune complex reaction. We therefore believe that this measure is a valid indicator of lung injury under these conditions. Increased levels of inflammatory lipid mediators in lung lavagc fluid have already been reported in this model of alveolitis (Tavares de Lima ct al., 1989). Our present work has further analysed the roles of these mediators by using specific inhibitors of thcir synthesis or antagonists of their action to alter the vascular damage induced in this model. Thc inhibitors and antagonists were used at doscs choscn from the relevant published reports of their activity. Thus the hacmorrhagic lesions appear to be mediated by PAF and leukotrienes as they were almost totally prevented by two antagonists of PAF action, BN-52021 (Braquet et al., 1987) and WEB-2086 (Casals-Stenzel et al., 1987), and by two lipoxygenasc inhibitors, nor-dihydroguaiaretic acid (Tappel ct al., 1953) and L-663,536 (Gillard ct al., 1988). Howcvcr, the LTD 4 receptor antagonist L-660,771 did not significantly affcct the intensity of the lesions. One possible interpretation of the lack of effect of L-660,711 is that the dose used (10 mg/kg) was not high enough to antagonize LTD a in vivo. However, much lower doses (0.5-1.(I mg/kg) inhibited LTD 4- and Ascaris sp-induced bronchoconstriction in guinea-pigs and squirrel monkeys, as well as antigen-induced bronchoconstriction in sensitized rats (Jones ct al., 1989). Although this compound is selective for LTD 4 in guinea-pig airway smooth muscle, in coronary vessels it antagonizes LTC 4 as well (Piper et al., 1990). Thus our result might reflect mediation by a non-peptide leukotriene, i.e. by LTB 4. We would favour this interpretation as LTB 4 was one of the eicosanoids released into the bronchoalvcolar space following induction of the reaction, it is a chemoattractant for PMNs and its level was markedly reduced by the PAF antagonists that prevented the haemorrhagic lesions. Our results also appear to rule out cyclo-oxygenase products as mediators of vascular damage, lndomethacin and an antagonist of thromboxane, L-655,240, both in adequate doses (Hall et al., 1987), were without any effect on the haemorrhagic lesions. The inhibitor of thromboxane synthesis, econazole (Jancar et al., 1991b), in a dose effective in preventing the thrombocytopenia accompanying this alveolitis (Tavarcs de Lima et al., 1989), did not significantly affect thc lung injury. This result was confirmed by additional data from our laboratory showing that pretreatment of rats with 10 m g / k g i.v. of another thromboxane inhibitor,
dazoxiben (Parry.et al., 1982), has no effect on the intensity of haemorrhagic lesions (data not shown). Furthermore, the PAF antagonists prevented lung injury but did not decrease the levels of TXB 2 released into the bronchoalveolar space. However, they did increase the levels of PGE 2, allowing the speculation that this action of PAF antagonists to increase prostaglandin production might contribute to their beneficial effect in this model. It is known that PGI 2 deactivates PMN leucocytes and that both PGE 2 and PG12 stimulate adenyl cyclase and increase intracellular cAMP (Boxer et al., 1980; Fantone et al., 1984; Rampart and Williams, 1986). However, this hypothesis cannot be applied here, since the thromboxane synthetase inhibitors, which are known to also increase the production of prostaglandins (De Freyn et al., 1982; Jancar et al., 1991b), did not show a protective effect. Although we did not measure PAF directly in our present experiments, its generation in this model is strongly indicated from the efficacy of PAF antagonists. This conclusion is in accordance with previous results showing the participation of PAF in immunecomplex hypersensitivity reactions in rats (Jancar et al., 1988a,b). In leucocytes and endothelial cells, both of which would be involved in our lung injury model, PAF formation precedes eicosanoid release (Stewart et al., 1990). In rats, injection of PAF into the peritoneal cavity induces the release of eicosanoids, including UI'B 4 (Jancar et al., 1989a). In our work, the release of LTB 4 appeared to be dependent of PAF action and, thus, on its formation; similar interactions between PAF and LTB 4 have been observed in rat peritoneal cavity stimulated by immune-complexes (Jancar et ai., 1989b) or by an anaphylactic reaction (Jancar et al., 1991a). Increased vascular permeability was an early event in this model as described by Johnson and Ward (1974) and therefore we expected to find oedcma in our experiments. However, the earlier work used higher amounts of antibody (250/xg protein) and 125i_albumin to estimate vascular permeability, whereas wc used less antibody (scc Methods) and assessed ocdema by changes in the dry:wet weight ratio. These differences and other factors, such as the magnitude of the permeability changes or the efficiency of lymphatic drainage preventing the accumulation of fluid in the lungs, could contribute to this discrepancy. Nevertheless, wc were able to measure altered dry:wet weight ratios after endotoxin, in accordance with Izumi and Bakhle (1988), demonstrating the ability of our method to detect pulmonary oedema. In summary, our results show that local deposition of immune-complexes in the lung caused microvascular damage exhibited as haemorrhagic lesions, together with an increased influx of leucocytes and an increased
69
level of inflammatory mediators in the bronchoalveolar space. The major mediator in this model of lung injury appears to be PAF, with a contribution from leukotrienes, probably LTB 4. Our results would suggest a place for PAF antagonists in the prevention or treatment of immune-complex alveolitis in occupational lung disease states.
Acknowledgements The authors gratefully acknowledge Dr. Y.S. Bakhle fi)r the careful revision of this manuscript and most helpful suggestions. This study was supported by grants from CNPq and FAPESP (Brasil) and MRC (Canada).
References Boxer, L.A., J.M. Allen, M. Schmidt, M. Yoder and R.L. Baehner, 1980, Inhibitkm of polymorphonuclear leukocyte adherence by prostacyclin, J. Lab. Clin. Med. 95, 672. Braquet, P., L. Touqui, Y.T. Shen and B.B. Vargaftig, 1987, Perspectives in platelet-activating factor research, Pharmacol. Rev. 39, 98. Brentjents, J.R., D.W. O'Connell, I.B. Pawlowski, K.C. llsu and G.A. Andres, 1974, Experimental immune complex disease of the lung. The pathogenesis of a laboratory model resembling certain human interstitial lung diseases, J. Exp. Med. 140, 105. Brieland, J.K., R.G. Kunkel and J.C Fantone, 1987, Pulmonary alveolar macrophage functkm during acute inflammatory lung injury, Am. Rev. Resp. Dis. 135. Brum-Fernandes, A.J., W. Cossermelli and S. Jancar, 1989, Slow-reacting substance of anaphylaxis (SRS-A) activity in the synovial fluids of rabbits with antigen-induced arthritis, Res. Commun. Chem. Pathol. Pharmacol. 63, 471. Casals-Stenzel, J., G. Muacevic and K.H. Weber, 1987, Pharmacological actions of WEB 21)86, a new specific antagonist of platelet activating factor, J. Pharmacol. Exp. Ther. 241,974. Cochrane, C.G., 1968, Immunologic tissue injury mediated by neutrophilic leukocytes, Adv. Immunol. 9, 97. De Freyn, G., H. Deckmyn and J. Vermylen, 1982, A thromboxane synthetase inhibitor reorients endoperoxide metabolism in whole blood towards prostacyclin and prostaglandin E2, Thromb. Res. 26. 389. Fantone. J.C. and P.A. Ward, 1982, Role of oxygen-derived free radicals and metabolites in leukocyte-dependent inflammatory reactions, Am. J. Pathol. 107, 397. Fantone, J.C., W.A. Marasco, L.J. Elgas and P.A. Ward, 1984, Stimulus specificity of prostaglandin inhibition of rabbit polymorphonuclear leukocyte lysosomal enzyme release and superoxide anion production, Am. J. Pathol. 115, 9. Gillard, J., A.W. Ford-Hutchinson, C. Chan, S. Charleson, D. Denis, A. Foster, R. Fortin, S. Leger, C.S. McFarlane, H. Mortom, H. Piechuta, D. Riendeau, C.A. Rouzer, J. Rokach, and R. Young, 1988, L-663.536 (MK-886)(3-[l-(4-chlorobenzyl)-3-t-butyl thio-5isopropylindol-2-yl]-2,2-dimethylpropanoic acid), a novel, orally active leukotriene biosynthesis inhibitor, Can. J. Physiol. Pharmacol. 67. 456. Hall, R.A., J. Gillard, Y. Guindon, G. Letts, E. Champion, D. Ethier, J. Evans, A.W. Ford-Hutchinson, R. Fortin, T.R. Jones, A. Lord, I[.E. Morton, J. Rokach and C. Yoakim, 1987, Pharmacology of L-655,240 (3-[ I-(4-chlorobenzyl)-5-fluom-3-methyl-indo-
2-yl]2,2-dimethyl propanoid acid); a potent, selective thromboxane/prostaglandin endoperoxide antagonist, Eur. J. Pharmacol. 1, 135. Hellewell, P.G. and T.J. Williams, 1986, A specific antagonist of platelet-activating factor suppresses oedema formation in an Arthus reaction but not oedema induced by leukocyte chemoattractants in rabbit skin. J. Immunol. 137. 3[)2. ttenson, P.M. and R.B. Johnston Jr., 1087, Tissue injury in inflammation. Oxidants, proteinases, and cationic proteins, J. Clin. Invest. 79, 669. Issekutz. A.C. and M. Szpejda, 1986, Evidence that platelet activating factor may mediate some acute inflammatory responses. Studies with the platelet-activating factor antagonist CV3988. Lab. Invest. 54. 275. Izumi, 1. and Y.S. Bakhle. 1988, Modification by steroids of pulmonary oedema and prostaglandin E 2 pharmacokineties induced by endotoxin in rats, Br. J. Pharmacol. 93, 955. Jancar, S., G. Dc Giacobbi, M. Mariano, P. Sirois and P. Braquet, 1988a, Immune-complex induced pancreatitis effect of BN-52021 a selective PAF-antagonist, Prostaglandins 35(5), 757. Jancar, S.. P. Sirois and P. Braquet, 1988b, Evidence for platelet activating factor (PAF) involvement in immune complex hypersensitivity reactions provided by BN 52021, in: Ginkgolides Chem. Biol. Pharm. Clin. Perspectives, ed. P. Braquet and J.R. Prous, p. 493. Jancar, S., P. Braquet and P. Sirois, 1989a, Release of eicosanoids in rat peritoneal cavity stimulated with platelct-activating factor (PAF). Effect of the PAF-antagonist BN 52021, Prostagl. Leukotr. Essential Fatty Acids 37, 23. Jancar, S., P. Braquet and P. Sirois, 1989b, Release of eicosanoids in rat peritoneal cavity during the Arthus reaction. Effect of the PAF-antagonist BN-52021 and indomethacin. Int. J. lmmunopharmacol. 11, 129. Jancar. S., M.G. Sirois, J. Carrier and P. Sirois, 1991a, PAF induces plasma extravasation and releases eicosanoids in the rat during anaphylaxis, Inflammation 15(5), 347. Jancar. S., C.F.P. Teixeira. W. Tavares de Lima. A. lloshikawaFujimura and P. Sirois. 1991b, Inhibitory effect of econazole nitrate on the release of thromboxane, Agents Actions 34, 387. Johnson, J.K. and P.A. Ward, 1974, Acute immunologic pulmonary alveolitis, J. CIin. Invest. 54, 349. Jones, T.R., R. Zamboni, M. Belley, E. Champion, L. Charlete, A.W. Ford-Hutchinson, R. Frenete. J.Y. Gauthier, S. Leger, P. Masson, C.S. McFarlane, H. Piechuta, J. Rokach, H. Williams and R.N. Young, 1989, Pharmacology of L-660,711 (MK-571): a novel potent and selective leukotriene D4 receptor antagonist, Can. J. Physiol. Pharmacol. 67, 17. Parry. M.J., M.J. Randall, E. ]tawkeswood, P.E. Cross and R.P. Dickindon, 1982, Enhanced production of prostacyclin in blood after treatment with selective thromboxane synthetase inhibitor, UK38485, Br. J. Pharmacol. 77, 547P. Pepys, J., 1969, Hypersensitivity diseases of the lungs due to fungi and organic dusts, Monogr. Allergy 4. I. Piper, P.J., A.P. Sampson, H. bin Yaacob and J.M. Mcleod, 1990, Leukotrienes in the cardiovascular system, in: Advances in Prostaglandin, Thromboxane, and Leukotriene Research, Vol. 20, ed. B. Samuelsson et al. (Raven Press, Ltd., New York) p. 146. Pradelles, P.. J. Grassi and J. Maclouf, 1985, Enzyme immunoassays of eicosanoids using acetylcholine esterase as label: an alternative to radioimmunoassay, Anal. Chem. 57, 1170. Rampart, M. and T.J. Williams, 1986, Polymorphonuelear leukocyte-dependent plasma leakage in rabbit skin is enhanced or inhibited by prostacyelin, depending on the route of administration, Am. J. Pathol. 124, 66. Stewart. A.G., P.N. Dubbin, T. llarris and G.J. Dusting, 1990. Platelet-activating factor may act as a second messenger in the
70 release of eicosanoids and superoxide anions from Icukocytes and endothelial cells, Proc. Natl. Acad. Sci. U.S.A. 87, 3215. Tappel, A.L.. W.O. Lindbcrg and P.D. Beyer, 1953, Effect of temperature and antioxidants upon the lipoxydase-catalyzed oxydation of sodium linoleate, Arch. Biochem. Biophys. 42, 293. Tavares de Lima, W., C.F.P. Teixeira, P. Sirois and S. Jancar, 1989, Involvement of eicosanoids and PAF in immune-complex alveolitis, Braziliian J. Mcd. Biol. Res. 22, 745. Van Kampen, E.J. and W.G. Zijlstra, 1961. Standardization of hemoglobinometry. I1. The hemoglobincyanide method, Clin. ('him. Acta, 6, 538.
Vercelloti, G.M., II.Q. Yin, K.S. Gustafson, R.D. Nelson, J.S. Jacob, 1988, Platelet-activating factor primes neutrophil responses to agonists: Role in promoting neutrophil-mediated endothelial damage, Blood 71, 1100. Warren, J.S., D.M. Mandel, K.J. Johnson and P.A. Ward, 1989, Evidence for the role of platelet-aetivating factor in immune complex vasculitis in the rat, J. Clin. Invest. 83, 669. Zar, J.H., 1984, Biostatistical Analysis, 2nd edn. (Prentice Hall, New Jersey) p. 718.