Biotherapy 2: 227-234, 1990. © 1990 Kluwer Academic Publishers. Printed in the Netherlands,

Preventive effect of several drugs against Pseudomonas aeruginosa infection and the toxicity of combined tumor necrosis factor with lipopolysaccharide: Relationship between lethality and the arachidonic cascade Nobuko Satomi, Akiko Sakurai, Fumio Iimura, Ruriko Haranaka 1 & Katsuyuki Haranaka

Department of Internal Medicine, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo 108, Japan; 1Department of Biochemistry, Nihon University School of Medicine, Itabashi-ku, Tokyo 173, Japan Received 12 September 1989; accepted 20 November 1989

Key words: leukotriene (LT), lipopolysaccharide (LPS), platelet activating factor (PAF), Pseudomonas aeruginosa, tumor necrosis factor (T'NF) Abstract

The participation of tumor necrosis factor (TNF) and lipopolysaccharide (LPS) in Pseudomonas aeruginosa (Pa) infection was examined. The lethal challenge of Pa or TNF and LPS injection could be prevented by pretreatment with anti-TNF antibody, polymyxin B, ONO 1078, or Shosaiko-to. The combined effects of TNF and LPS may be deeply related to the lethality of Pa infection. The activities of leukotriene(LT) C4/D4/E4 or platelet activating factor (PAF) were also related to the lethality of Pa infection, probably due to the subsequently produced TNF which acts in combination with LPS. Activating the host defence mechanism with biological response modifiers like Chinese medicines was effective against Pa infection. One mechanism could involve an activity as an LT inhibitor or PAF antagonist. Following the administration of TNF and/or LPS, the serum levels of arachidonic cascade products underwent various changes. With a combination of TNF and LPS, there was a synergistic increment of prostaglandins, thromboxane, and LT. Following pretreatment with Shosaiko-to, suppression of LTs was dominant even with the combination of TNF and LPS, which might be related to the lethality of the infection or combined TNF with LPS.

Introduction

We have investigated the effectiveness of antibody against Pseudomonas aeruginosa (P. aeruginosa) antigen (OEP: original endotoxin protein of P. aeruginosa). Anti-OEPAgG treatment was found to be very effective against the infection of mice, even under immunosuppressive conditions [ 1, 2, 3]. Zweerink et al. reported a protective efficacy of P. aeruginosa-specific monoclonal antibod-

ies using X-linked immunodeficient mice [4, 5]. Most of such antibodies react with the O-specific carbohydrate on the lipopolysaccharide (LPS) of the seven Fisher's immunotype of P. aeruginosa. Although the effectiveness of antibodies and immunization has been demonstrated, some other factors might be intimately related with the pathogenesis of P. aeruginosa infection. The contribution of LPS from gram negative bacteria to the pathogenesis cannot be ignored. Tumor

228 necrosis factor (TNF) exerts strong antitumor activities against transplanted tumors [6]. We must also consider the strong toxicity associated with the combined effect of TNF and LPS [7]. Kiger et al. [8] reported TNF production in B.C.G.-pretreated mice by P. aeruginosa. Administration of anti-TNF antibody improved the survival rate in mice infected with P. aeruginosa [9]. In the present experiments, the participation of TNF and LPS in P. aeruginosa infection was examined from the standpoint of preventing the lethality and elucidating the relation with the arachidonic cascade. Furthermore, the effects of pretreatment with leukotriene (LT) inhibitor or platelet activating factor (PAF) antagonist against the P. aeruginosa infection or TNF and LPS injection were examined.

Materials and methods

Animals: DDY strain female mice weighing 28-35 g (Japan SLC Co., Shizuoka, Japan) were used. Bacteria: P. aeruginosa NC-5 strain was provided by the Type Culture Collection Room at our Institute. Drugs: Astromycin (ASTM) was provided by Kyowa Hakko Kogyo (Tokyo, Japan), cefbuperazone (CBPZ) by Kakenseiyaku (Tokyo, Japan), and polymyxin B sulfate (PMB) was purchased from Sigma Chemical Co. (Mo, USA). ONO 1078 was provided by Ono Pharmaceutical Co., Ltd. (Osaka, Japan), WEB 2086 BS by Boehringer Ingelheim KG (Ingelheim, FRG), and Shosaikoto (Xiao-chai-hu-tang), Juzen-taiho-to (Shi-quan-da-bu-tang) and Hochu-ekki-to (Bu-zhong-yi-qi-tang) by Tsumura & Co. (Tokyo, Japan). Recombinant human TNF (rhTNF) was supplied by Dainippon Pharmaceutical Co., Ltd. (Osaka, Japan). The specific activity was 3 x 106U/mg protein, and the endotoxin content was below 3 pg/ 106U as determined by Limulus assay. Es-

cherichia coil O 111 :B 4 w LPS was purchased from Difco Lab. (Mich, USA). Assay kits for prostaglandin (PG), thromboxane (TX), and leukotriene (LT) were purchased from Amersham International plc (Amersham, UK). Mode of drug administration: ASTM (300 pg/mouse) and/or CBPZ (1.5 mg/mouse) were administered subcutaneously at the same time and at 12 hrs after the inoculation of P. aeruginosa. PMB (0.5 mg/mouse) was administered subcutaneously at 1 hour prior to bacterial challenge. ONO 1078 (0.1 rag/ mouse) or WEB 2086 (1 or 10 #g/mouse) was injected intraperitoneally at 1 hour prior to bacterial challenge. Spray-dried aqueous extracts of 50 mg/mouse Shosaiko-to, Juzentaiho-to, or Hochu-ekki-to were given to mice in their drinking water for 2 weeks. Bacterial challenge: P. aeruginosa NC-5 strain was incubated in heart infusion broth for 18 hrs. The bacteria were washed with physiological saline, suspended in 2.5% bacteriological mucin (Difco) and inoculated intraperitoneally. Each experimental group consisted of 5 mice. Measurement of PG, TX, and LT: PG D2, PG E2, TX B2, LT B4, and LT C4/D4/E 4 were measured using a [3H] assay system. Briefly, the assay is based on the competition between the unlabelled substance and a fixed quantity of the tritium-labelled substance for binding to a limited quantity of an antibody which has a high specificity and affinity for the substance. Antitumor activity: Meth A sarcoma cells (4 x 105 cells) were transplanted into the flank of Balb/c mice (supplied by our Institute). The tumors and extent of necrosis were measured with Vernier calipers. Statistics: Analysis of significance was carried out by Student's t-test. The LDso was calculated using Probit analysis. Results

Antitumor effect and disadvantageous action in the case of combined use of rhTNF and

229 LPS: With a combination of rhTNF and LPS, the antitumor activity was synergistically enhanced. However, with this combination, the lethality was also enhanced (Table 1). Several chemicals and drugs were able to prevent such lethal activity by pretreatment. ONO 1078 and WEB 2086 revealed a preventive effect against the lethality of rhTNF and LPS by pretreatment (p < 0.005, p < 0.0001). Pretreatment with Shosaiko-to could also prevent against the lethality (p < 0.005) (Fig. 1). Fig. 2 summarizes the changes in relative tumor volume (RTV) following combined use of rhTNF and LPS in mice pretreated with ONO 1078, WEB 2086, and Shosaikoto. Even such treatment did not inhibit the antitumor activity of rhTNF and LPS. Protective effect of drugs against P. aeruginosa infection: ASTM or CBPZ revealed no therapeutic effect against P. aeruginosa infection at the present dosage. However, with a combination of these antibiotics, synergistic effects were observed (p < 0.005). PMB exhibited a protective effect against experimental P. aeruginosa infection (p

60 u~

40 ~o IX

20

2 Days

3

4

after administration

lOG

80 >

>6G c

~40.

,,

20"

Table I. Combined toxicity of rhTNF and LPS.

1 LDso of TNF combined with LPS LPS ( - ) LPS 50 lag LPS 100 ng

2 Days

over 2.0 x 105 3.3 × 104 2.5 x 104

LPS

1//g

8.1 × 10 3

LPS

2ttg

4.3 x 103

Several dosages of rhTNF and/or LPS were administered intravenously to Meth A sarcoma-bearing mice. The LDso was calculated and expressed in units of rhTNF.

3

4

5

after a d m i n i s t r a t i o n

Fig. 1. Preventive effect against the lethality of rhTNF and LPS obtained with drugs. ONO 1078 (0.1 rag/mouse) or WEB 2086 (10 #g/mouse) was injected s.c. at 1 hr prior to rhTNF and LPS administration. Shosaiko-to (50 rag/mouse) was administered p.o. for 2 weeks before the administration of rhTNF and LPS. rhTNF (1.0 or 2.0 x 104 U/mouse) and LPS ( 100 ng/mouse) were administered i.v. to Meth A sarcoma-bearing mice. Upper column, 1.0 x 104 U; lower column, 2.0 x 104 U. , Control; ,'~,,,~o¢, ONO 1078; ~ , WEB 2086; .:..:-:., Shosaiko-to.

230

/

Table 2. Therapeutic effects of several treatments against P. aeruginosa infection. LDso of P. aeruginosa challenge Exp. 1 Control ASTM CBPZ ASTM + CBPZ Exp. 2 Control TNF Anti-TNF antibody Exp. 3 Control Polymyxin B ONO 1078 WEB 2086 Shosaiko-to Exp. 4 Control Shosaiko-to Juzen-taiho-to Hochu-ekki-to

1.0

!--

2

4

6 8 10 12 14 16 18 Days after administration

Fig. 2. Changes in relative tumor volume following administration of rhTNF and LPS to pretreated mice. The tumors and extent of necrosis were measured with Vernier calipers. E3, Control; It, r h T N F + L P S ; O, ONO 1078 (rhTNF + LPS); ©, WEB 2086 (rhTNF + LPS); A, Shosaiko-to (rhTNF + LPS) (mean, n = 10).

against P. aeruginosa infection. Pretreatment with Shosaiko-to demonstrated a protective effect against P. aeruginosa infection (p < 0.005). Pretreatment with Juzen-taihoto, and Hochu-ekki-to also exhibited a protective effect against P. aeruginosa infection (p < 0.005, p < 0.005) (Table 2). Changes in serum levels of arachidonic cascade products following administration of rhTNF and/or LPS: The serum levels of arachidonic cascade products at 1 hr after the administration of rhTNF and/or LPS underwent various changes as illustrated in Fig. 3. With a combination of rhTNF and LPS, there was a synergistic increment in PGs, TX and LTs. On pretreatment with Shosaiko-to,

7.5 x 7.4 x 5.5 x 5.0 x

103 103 103 104

9.1 x 103 1.8 x 103 2.1 x 105 1.3 x 2.1 x 4.7 x 1.3 x 1.1 x

104 I04 104 103 105

1.0 x 1,7 x 4.5 x 4.2 x

104 105 105 105

Astromycin (ASTM; 300/~g/mouse) and/or cefbuperazone (CBPZ; 1.5 mg/mouse) were administered s.c. at the same time as and at 12 hrs after inoculation of P. aeruginosa. Recombinant human tumor necrosis factor (TNF; 1000 U/mouse), or anti-mouse-TNF antibody (300U/mouse) was administered s.c. at the same time as the bacterial challenge. Polymyxin B (0.5mg/mouse), ONO 1078 (0.1mg/mouse) or WEB2086 (10 #g/mouse) was injected s.c. at 1 hr prior to the bacterial challenge. Shosaiko-to (50 mg/mouse), Juzen-taiho-to (50 mg/ mouse) or Hochu-ekki-to (50mg/mouse) was administered p.o. for 2 weeks before the bacterial challenge. P. aeruginosa NC-5 strain was suspended in 2.5% mucin and inoculated i.p.. Evaluations of the therapeutic effect were made on the basis of Probit analysis.

suppression of the release of PGs and LTs was observed. Suppression of LT B4 and LTC4/D4/E4 was dominant as shown in Fig. 3. Inhibition of LT C4/Da/E4 release following administration of rhTNF and LPS by pretreatment with drugs: The serum levels of LT C4/D4/E4 at 1 hr after administration of rhTNF and LPS in mice pretreated with ONO 1078, WEB 2086, and Shosaiko-to are shown in Fig. 4. Increase of LTC4/D4/E4 was inhibited by pretreatment with ONO1078, WEB 2086 and Shosaiko-to. The degree of inhibition of LT C4/D4/E 4 release b y

231 100

200

I

300

I

400 pg/ml

I

I

Control

PG D2 Shosai ko-to

l::'_" .'__' .+_.-.7--.-.:::5~._ ---:~-:_.:_-:_............................... :_:.:.:.~ .:~.:~~.z.: :.::.,z.z. :.:.:.;.:.,...:,..........;.:.:.;.:...:..:::.~~:.~:~+....x.:~v;.>::..,.,:......,.:,.:.,:......,-:::.~..,~..#..:::", : : . : : : ~:~:: : :': : : : >:~::.'.":.:.,.:-::. !

i 10 i Con t rol

20 ng/ml i

i::=====;;~i [:--[l[~l:I>:~I>:'>>>:

['[I>:--:I:IF>:" > : ' > : ' : ]

'I

PG E2 S hos a i k o - t o

10i Control

:--=-=---=-'-'-'-'-----:-:::;1 ;.::.:,:.:.:.::,:.:.:,:e:.:,:,:.:+::.::.:+:.:.z.z.:.:.:.::.:.:.:.z.:,:-:,:::::::::::::::::::::::::::: :.:~.~:.z.::.:.: .::.::.:.~:.:.:l

Shosa iko-to

f'i'-'-'?~''-''4

TX B2

:1

loo C o n t ro I

Shosai ko-to

2 0 ng/ml

200

i

.............

4()aO

pg/ml

|

L.,~=.~,~_=..~........~.-.==.=,....

I

--

Control

Shosaiko-to

~

~-=-~-"-=-=

1I

2|

I ng/ml

3

-"-"-= =-~ -=-=-=-=-=~-..=.".-~-=-=-=~:1

i........................................... ;~

,

LT

C4/D4/E4

Fig. 3. Changes of PG, TX, and LT in mice treated with Shosaiko-to.

r h T N F (1.0 x 104 U/mouse) and/or LPS (100 ng/mouse) were injected into Meth A sarcoma-bearing mice. Blood was collected after 2 hours, and the serum release of PG, TX, and LT were measured by radioimmunoassay (mean, n = 5). ..................., - ; ~' , LPS; ............. , rhTNF; +:.:~:+:.:.:~.:,LPS + rhTNF.

Shosaiko-to pretreatment in mice administered with rhTNF and LPS was found to be dominant.

Discussion

PMB has been shown to prevent mortality due to experimental endotoxemia, and to decrease the incidence of endotoxin-induced

disseminated intravascular coagulation [10, 11]. P. aeruginosa frequently displays resistance to antibiotics (Table 2). Nevertheless, combinations of antibiotics (Table 2) or of antibiotics and specific antibodies [ 1, 2, 3] have been found to demonstrate therapeutic effects. P. aeruginosa itself has resistance to PMB. However, PMB does exert a protective effect in experimental P. aeruginosa infection (Table 2). The reason for this effectiveness is

232 LT

C41D4/E 4 1 1

2

nglml

1

Control

0,3 mg O N O 10 7 8 3 m~

O.03u~ WEB 2086 0.3 ~g

Shosaiko-to

Fig. 4. Changes o f LT C4/D,/E 4 in mice treated with several kinds o f drugs. rhTNF and LPS were administered at 2 hours prior to blood collection (mean, n = 5). ONO 1078 (0.1 or I rag/mouse) or WEB 2086 ( 10 or 100 pg/mouse) was injected s.c. at 1 hr prior to rhTNF and LPS administration. Shosaiko-to (50 rag/mouse) was administered p.o. for 2 weeks. II, _ ; [3, rhTNF + LPS.

thought to be related to a decrement in the disadvantageous action of LPS. Kiger et al. [8] observed that P. aeruginosa treatment had a capacity to induce TNF release. TNF itself exhibits very little toxicity, but in combination with LPS, a strong toxicity is observed in the murine model (Table 1). Administration of rhTNF increased the lethality of P. aeruginosa. In contrast, administration of anti-TNF antibody or administration of PMB by neutralizing the LPS activity exhibited a strong protective activity against the lethal bacterial challenge (Table 2). Combined administration of TNF with LPS revealed a synergistic increment in the production of PGs and LTs (Fig. 3). Since drugs which affect the arachidonic cascade or oxygen scavengers prevent the lethal effects of TNF combined with LPS [12, 13], we speculate that LTs or subsequently produced reactive oxygen species (ROS) might be deeply related to the toxicity associated with the combined action of TNF and LPS. Injection of PAF can induce similar symptoms to endotoxin shock [14]. Since TNF is a potent inducer of both PAF synthesis and release [15, 16], we examined the

effects of LTs or PAF against P. aeruginosa infection in relation to TNF. As shown in Table 2, ONO 1078 (an LT C4/Dg/E4 inhibitor) could exert a protective effect against the lethality. ONO 1078, 8-[p-(4-phenylbutyloxy)benzoyl]amino-2-(tetrazol-5-yl)-4-oxo-4H- 1-benzopyran hemihydrate, has been identified as an end organ antagonist for LTs in the guinea pig ileum and trachea in vitro [ 17]. Our previous experiments suggested that TNF might activate the arachidonic cascade. We speculate that PGs and TX might induce tumor necrosis. The side effects of a combination of TNF and LPS may be closely related to subsequently produced ROS. If the subsequently produced ROS or LTs could be effectively removed, TNF could be employed safely and yield the expected antitumor activities, and the same may hold true for P. aeruginosa infection. TNF is a potent inducer of both PAF synthesis and release [15, 16]. TNF-treated rat peritoneal macrophages and polymorphonuclear cells, and human vascular endothelial cells synthesize and release PAF. PAF receptor antagonists are known to

233 inhibit or reverse endotoxin-induced hypotension and prolong the survival of animals. WEB 2086, 3-[4-(2-chlorophenyl)-9-methyl6H-thieno[3,2-f][ 1,2,3]-triazolo-[4,3-a][ 1,4]diazepin - 2- yl] - 1 - (4- morpholinyl) - 1- propanone, is known to be an antagonist of PAF, effective on PAF-induced anaphylaxis but not on active anaphylaxis [18]. WEB2086 failed to reveal any protective effect against P. aeruginosa infection (Table 2), but it did exhibit a preventive effect against the lethal challenge of combined TNF and LPS without impairing the antitumor activity (Figs. 1 and 2). We have found that traditional Chinese medicines are potent agents which c a n prevent the lethality of T N F combined with LPS [12] and also prevent against P. aeruginosa infection (Table 2). In the present experiments, we examined the effect of pretreatment with Shosaiko-to in relation to PGs, TX and LTs. Baicalein may exert an anti-allergic action by inhibiting 5-1ipoxygenase, which produces LTs [19]. Baicalein is a constitutent of Ogon (a component of Shosaiko-to). Other examples of agents of Chinese herb origin are kadsurenon (a PAF inhibitor) isolated from harfenteng (Piperfutokazura) [20], as well as ginkgolide B (BN 52021) and a ginkgolide mixture (BN 52063) isolated from the leaves of Ginko biloba (PAF inhibitors) (21). Pretreatment with Shosaiko-to inhibits the release of LT C 4 / D 4 / E 4 t o a greater extent than that observed in the case of ONO 1078 or WEB 2086 pretreatment (Fig. 4). Such inhibitions are observed not only with combined TNF and LPS but also with each individual agent (Fig. 3). These results suggest that pretreatment with Chinese medicines affords potent inhibition not at all inferior to that of chemical compounds, involving not only protection from infection or lethal challenge with T N F combined with LPS but also inhibition of LT production. In conclusion, LTs or PAF may be related to the lethality of P. aeruginosa infection or the toxicity of combined administration of

TNF and LPS. The lethality of P. aeruginosa infection can be improved by pretreatment with inhibitors of LPS or T N F activity, LT inhibitors, or Chinese medicines (as biological response modifiers).

References 1. Haranaka K, Sugane K, Mashimo K. Combination therapy of anti-endotoxin antibody and gentamicin in the immuno-suppressed mice with P. aeruginosa infection. Jpn J Exp Med 1975; 207-13. 2. Haranaka K, Matsuo M, Mashimo K. The enhancement of phagocytosis and intracellular killing of Pseudomonas aeruginosa and its common antigen (OEP) coated latex particles by mouse spleen macrophages to which antiOEP-IgG and gentamicin have been added. Jpn J Exp Med 1977; 47: 35-40. 3. Haranaka K, Satomi N, Sakurai A, Kunii O. The combined use of antibiotics and specific antibodies against mouse Pseudomonas aeruginosa infection in vivo and the phagocytosis of peritoneal exudate cells in vitro. In: Eickenberg HU et al., eds. The influence of antibiotics on the host-parasite relationship I. Heidelberg: Springer-Verlag, 1982: 161-7. 4. Zweerink H J, Gammon MC, Hutchison CF, Jackson J J, Pier GB, Puckett JM, Swell T J, Sigal NH, X-linked immunodeficient mice as a model for testing the protective efficacy of monoelonal antibodies against Pseudomonas aeruginosa. Infect Immun 1988; 56: 1209-14. 5. Pier GB, Thomas D, Small G, Siadak A, Zweerink H. In vitro and in vivo activity of polyclonal and monoclonal human immunoglobulins G, M, and A against P. aeruginosa lipopolysaccharide. Infect Immun 1989; 57: 174-9. 6. Haranaka K, Satomi N, Sakurai A. Antitumor activity of murine tumor necrosis factor (TNF) against transplanted murine tumors and heterotransplanted human tumors in nude mice. Int J Cancer 1984; 34: 263-7. 7. Sakurai A, Satomi N, Iimura F, Haranaka R, Haranaka K. Synergistic effects of combining endotoxin-free tumor necrosis factor with lipopolysaccharide in vivo and in vitro: Correlation with the arachidonic cascade. J Biol Response Modif, in press. 8. Kiger N, Khalil A, Mathe G. Tumor-necrotizing serum production by administration of BCG + Pseudomonas: Its application in treatment of fibrosarcoma in mice. RRCR 1980; 75: 220-3. 9. Satomi N, Sakurai A, Haranaka R, Haranaka K. Traditional Chinese medicines and drugs in relation to the host-defence mechanism. In: Peters G e t al., eds. The influence of antibiotics on the host-parasite relationship III. Heidelberg: Springer-Verlag, 1989; 77-86. 10. Bannatyne RM, Harnett NM, Cheung R. Protective effect of polymyxin B sulfate in experimental meningococcal infection in mice. Can J Microbiol 1977; 23: 1526-8.

234 11. Bannatyne RM, Cheung R. Protective effect of polymyxin B sulfate in experimental enterobacterial infection in mice. Can J Microbiol 1979; 996-8. 12. Satomi N, Sakurai A, Iimura F, Haranaka R, Haranaka K. Japanese modified traditional Chinese medicines as preventive drugs of the side effects induced by tumor necrosis factor and lipopolysaccharide. Motec Biother 1989; 1: 155-62. 13. Satomi N, Sakurai A, Haranaka R, Haranaka K. Preventive effect of several chemicals against lethality of recombinant human tumor necrosis factor. J Biol Response Modif 1988; 7: 54-64. 14. Braquet P, Paubert-Braquet M, Bessin P, Vargaftig BB. Platelet-activating factor: A potential mediator of shock. In: Samuelsson B e t al., eds. Advances in prostaglandin, thromboxane and leukotriene research. New York: Raven Press, 1987; 17: 822-7. 15. Camussi G, Bussolino, F, Salvido G, Baglioni C. Tumor necrosis factor/cachectin stimulates peritoneal macrophages, polymorphonuclear neutrophils, and vascular endothelial cells to synthesize and release platelet activating factor. J Exp Med 1987; 166: 1390-1404. 16. Valone FH, Epstein LB. Biphasic platelet-activating factor synthesis by human monocytes stimulated with IL-l-fl, tumor necrosis factor, or IFN-~,. J Immunol 1988; 141: 3945-50.

17. Obata T, Namba F, Kitagawa T, Terashima H, Toda M, Okegawa T, Kawasaki A. ONO-1078: An antagonist of leukotrienes. In: Samuelsson B e t al., eds. Advances in prostaglandin, thromboxane and leukotriene research. New York: Raven Press, 1987; 17: 540-3. 18. Casals-Stenzel J, Muacevic G, Weber KH. Pharmacological actions of WEB 2086, a new specific antagonist of platelet activating factor. J Pharmac Exp Ther 1987; 241: 974-81. 19. Sekiya K, Okuda H. Selective inhibition of platelet lipoxygenase by Baicalein. Biochem Biophys Res Comm 1982; 105: 1090-5. 20. Shen TY, Hwang SB, Chang MN, Doebber TW, Lain MH, Wu MS, Wang X, Han GQ, Li RZ. Characterization of platelet-activating factor receptor antagonist isolated from haifenteng (Piper futokazura): Specific inhibition of in vitro and in vivo platelet-activating factor-induced effects. Proc Natl Acad Sci USA 1985; 82: 672-6. 21. Roberts NM, Page CP, Chung KF, Barnes PJ. Effect of a PAF antagonist, BN52063, on antigen-induced, acute, and late-onset cutaneous responses in atropic subjects. J Allergy Clin Immunol 1988; 82: 236-41. Address for correspondence: K. Haranaka, Dept. of Internal Medicine, Institute of Medical Science, University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108, Japan.