Pathogenesis of Porcine Actinobacillus Pleuropneumonia - Europe PMC

Pathogenesis of Porcine Actinobacillus Pleuropneumonia: Part I. Effects of Surface Components of Actinobacillus pleuropneumoniae In Vitro and In Vivo Hongsheng Huang, Andrew A. Potter, Manuel Campos, Frederick A. Leighton, Philip J. Willson, and William D.G. Yates

ABSTRACT To understand the role of nonsecreted components of Actinobacillus pleuropneumoniae in virulence, we investigated in vitro cytotoxicity and in vivo pulmonary changes in pigs due to various A. pleuropneumoniae (serotype 1) fractions. Following 1.5 h incubation, lipopolysaccharide (LPS), 2 crude extracts and bacterial culture supernatant (BCS) at high concentrations were cytotoxic to porcine pulmonary alveolar macrophages (PAM), peripheral blood mononuclear leucocytes, neutrophils and a cultured porcine bone marrow cell line. Heat-killed bacteria were cytotoxic to PAM after 24 h incubation. The 2 crude extracts were prepared by shaking either intact bacteria after removing culture supernatants (crude surface extract, CSE), or whole bacterial culture (crude surface plus culture supernatant extract, CSSE) with glass beads in saline at 60°C. Further experiments showed that proteins from the bacterial membrane were partially involved in cytotoxicities of these 2 extracts. Both BCS and CSSE caused multivocal hemorrhage and neutrophil infiltration when inoculated into porcine lungs, but CSE did not. The lung:whole body weight ratios of the pigs treated with CSSE were significantly higher (P < 0.05) than those of pigs treated with BCS, CSE, or control solution. It is concluded that beside the secreted proteins, bacterial surface

components including LPS and non-secreted proteins were cytotoxic in vitro; and secreted and nonsecreted components act synergistically to cause lung lesions.

RESUME La cytotoxicite in vitro et les changements pulmonaires in vivo causes par differentes fractions d'Actinobacillus pleuropneumoniae (serotype 1) furent investigues afin d'etudier le r8le de composantes non-secretees d'A. pleuropneumoniae dans la virulence des isolats. Apres 1,5 h d'incubation, le lipopolysaccharide (LPS), deux extraits bruts et le surnageant de culture bacterienne (SCB) etaient cytotoxiques pour des macrophages alveolaires pulmonaires de porc (MAP), pour des polymorphonucleaires du sang peripherique, pour des neutrophiles ainsi qu'une lignee celulaire porcine de cellules de la moelle osseuse. Des bacteries tuees par la chaleur etaient cytotoxiques pour des MAP apres 24 h d'incubation. Les deux extraits bruts furent prepares en melangeant soit des bacteries intactes apres avoir enleve le surnageant de culture (extrait brut de surface, EBS), soit une culture bacterienne complete (extrait brut de surface plus surnageant, EBSS) a l'aide de billes de verre dans de la saline a 60 'C. Des experiences subsequentes ont demontre que des proteines provenant de la paroi bacterienne etaient partiellement

impliquees dans la cytotoxicite de ces deux extraits. Le SCB et le EBSS, mais non le EBS, ont tous deux causes des hemorragies multifocales et une infiltration de neutrophiles lorsqu'inocules au niveau pulmonaire chez les porcs. En conclusion, il semble qu'en plus des proteines secretees, des composantes de la surface bacterienne comme le LPS et des proteines non-secre tees sont cytotoxiques in vitro; et que les composantes secretees et non-secretees agissent en synergie pour causer des lesions pulmonaires. (Traduit par docteur Serge Messier)

INTRODUCTION Porcine Actinobacillus pleuropneumonia, caused by Actinobacillus (Haemophilus) pleuropneumoniae, is a highly contagious, fibrinous, hemorrhagic and necrotizing pneumonia that causes high mortality in acutely infected pigs or localized lung lesions in those that are chronically infected (1,2). The pathogenesis of porcine Actinobacillus pleuropneumonia, particularly the factors inducing pulmonary lesions, has not been well understood. Sonicated A. pleuropneumoniae or bacterial culture supernatant induces lung lesions similar to those in natural infections (3), indicating that A. pleuropneumoniae possesses both secreted and non-secreted virulence factors. Three different secreted protein cytotoxins ApxI, II and III have been

Department of Veterinary Pathology (Huang, Leighton), University of Saskatchewan, Saskatoon, Saskatchewan S7N 5B4; Veterinary Infectious Disease Organization (VIDO) Saskatoon, Saskatchewan S7N 5E3 (Potter, Willson); Animal Diseases Research Institute, Canadian Food

Inspection

Agency, Lethbridge, Alberta T1J 3Z4 (Yates); Immunology and Pathogenesis, Central Research Division, Pzifer Inc., Eastern Point Road, Groton, Connecticut 06340 USA (Campos). Correspondence to Dr. H. Huang, Microbiology Section, Animal Diseases Research Institute, Canadian Food Inspection Agency, 3851 Fallowfield Road, Nepean, Ontario K2H 8P9. Published with the consent of the Director of the Veterinary Infectious Disease Organization as Journal Series No. 226. Received December 3, 1996.

Can J Vet Res 1998; 62: 93-101

93

studied extensively and considered to be important virulence factors (4). They are either cytotoxic and hemolytic (Apx I and II) or cytotoxic (Apx III) for porcine cells (4-7), and may induce pulmonary lesions (8). A low molecular weight protein hemolysin and proteases have been also identified in vitro (9-11), and they may play destructive roles in vivo. An in vitro study has also shown that lipopolysaccharide (LPS) from A. pleuropneumoniae is cytotoxic to porcine endothelial cells (12), indicating that endotoxin may directly cause the damage to the parenchymal tissues. Because of the in vitro cytotoxicity of A. pleuropneumoniae toxins on various types of porcine cells, it has been commonly thought that direct cytotoxic effects could play a major role in the development of hemorrhagic and necrotic lung lesions (4). However, it is also believed that host factors may play a role in development of the lung lesions (13,14), which was further investigated in our subsequent study (15). Current commercial bacterins or experimental vaccines containing one or a combination of secreted bacterial toxins only reduce the severity of the lung lesions and clinical disease; vaccinated pigs can still be chronic carriers and suffer acute and chronic disease (16,17). However, pigs which recover from natural infection are resistant to subsequent infection by A. pleuropneumoniae (18). Therefore, it is likely that vaccination against one or several secreted virulence factors may not be enough to induce full immunity (19). It is also possible that some bacterial virulence factors have not been identified. Pigs immunized with either surface extract from A. pleuropneumoniae or different fractions of this extract separated by isoelectric focusing are protected from death after challenge (20,21), suggesting that there might be an important virulence factor(s) present in these preparations. It is hypothesized that beside the secreted components, some surface components on this bacterium may have direct cytotoxic effects. Therefore, the cytotoxicity of various A. pleuropneumoniae fractions including extracts either from bacterial surface or from bacterial surface plus culture supernatant, killed bacteria, and LPS in vitro, and

94

the effects of these preparations on 1-1.5, 1.6-3, 3.1-5, 5.1-6 and 6.1-10. pulmonary tissues in vivo were inves- The collected solutions were dialysed against phosphate buffered saline tigated in the present study. (PBS, 0.01 M, pH 7.2) for 48 h. To examine if bacterial surface compoMATERIALS AND METHODS nents would stabilize the cytotoxic activities of secreted protein cytotoxBACTERIA ins during treatment at 60°C (this Actinobacillus pleuropneumoniae temperature inactivates the cytotoxic serotype 1 strain, AP37, was isolated activities of secreted protein cytotoxfrom the lung of a pig submitted to the ins), CSSE was prepared using the Western College of Veterinary same method as CSE except the whole Medicine, University of Saskatchewan bacterial culture including the super(22,23). The bacteria were grown on natant was diluted 10 fold in saline PPLO agar plates or in PPLO broth and shaken with glass beads at 60°C (DIFCO Laboratories, Detroit, Michi- for 1 h, followed by centrifugation to gan, USA) supplemented with 1% remove bacteria and debris. The super(vol/vol) IsoVitaleX (BBL Microbiol- natant was sterilized by filtration. Lipopolysaccharide (LPS) was ogy Systems, Becton Dickinson & Co., Cockeysville, Maryland, USA) either A. pleuropneumoniae LPS (a smooth type of serotype 1, donated by (24). National Research Council of BACTERIAL PREPARATIONS Canada) (25) or Escherichia coli LPS Bacterial culture supernatants (serotype 0111:B4, DIFCO Labora(BSC), heat-killed bacteria, and tory, Detroit, Michigan, USA). 2 bacterial extracts were prepared from A. pleuropneumoniae serotype 1 PROTEIN AND LPS DETERMINATIONS Protein concentration was deterstrain, AP37, at the mid-log phase of growth. Bacterial culture supernatants mined by Lowry assay (26) using (BCS) were obtained following cen- bovine serum albumin (Bio-Rad Labtrifugation of whole bacterial cultures oratories Ltd, Mississauga, Ontario) at 10 000 X g for 20 min and filtration as the standard. Protein concentrathrough 0.22 ,um pore size Nalgene tions in CSSE and BSC were defined filter with cellulose nitrate membrane as the total proteins measured minus (Nalge Nunc International, Rochester, the amount of the proteins in the New York, USA). The supernatant PPLO medium. The LPS concentrawas stored at -20°C. Bacterial cells tion was determined by Limulus amecollected by centrifugation were bocyte lysate (BioWhittaker, Inc., resuspended in an appropriate cell Walkersville, Maryland, USA) using culture medium as mentioned in cell purified A. pleuropneumoniae LPS preparation. Heat-killed bacteria were (25) as the standard. prepared by incubating the resuspended bacteria at 600C for 1 h, and ANIMALS The animal trials were conducted kept at -70°C until use. Bacterial extracts, crude surface extract (CSE) using outbred pigs of either sex, and crude surface plus culture super- approximately 8 to 10 wk of age, from natant extract (CSSE), were prepared an A. pleuropneumoniae-free herd in based on the method of Bielefeldt Saskatchewan without a history of Ohmann and co-workers (21). To pre- A. suis infection. Animals arrived at pare CSE, the bacterial pellet was least 2 d before the beginning of the resuspended in 1/10th volume of experiments. They were randomly 0.85% NaCl and shaken with glass allocated to 4 groups (3 pigs/group) beads at 60°C for 1 h. After centrifu- and housed in an isolation room with gation at 10 000 X g for 20 min, the controlled temperature and ventilasupernatants were sterilized through tion, on vinyl-covered metal flooring 0.22 ,um Nalgene filter as mentioned with free access to water and commerabove. In one experiment, the CSE cially prepared feed. They were cared was further fractionated by isoelectric for in accordance with the principles focusing using a Rotofor apparatus outlined in the Guide to the Care and (Bio-Rad laboratories, Richmond, Use of Experimental Animals (27). To verify that the animals used did California, USA) to obtain fractions with pI's (isoelectric pH value) of not have detectable levels of antibodies

into the trachea through a thin catheter, and usually a quarter of this volume was recovered by vacuum. Lung lavages were also done on lungs from freshly slaughtered pigs (Intercontinental Packers, Saskatoon). The lungs without gross lesions from healthy market-weight pigs (over 55 kg) were selected by Agriculture Canada inspectors. One litre of collecting medium was gently poured into the trachea via a funnel. After gentle shaking, the lung lavage fluids were collected through a funnel by gravity. Lavage fluids were centrifuged on Ficol-paque to remove neutrophils and debris. The cell concentrations were adjusted to 5 X 106 mL-' in culture medium, minimum essential medium (MEM) (GIBCO Laboratories), and 100 ,uL added to each well of 96-well microtiter plate. CELL PREPARATIONS The cells were allowed to adhere to Citrated porcine blood was col- the plates for 2 h, then non-adherent lected from clinically healthy pigs. cells were removed by rinsing with Blood was centrifuged at 675 X g for the culture medium. Non-specific 30 min. Neutrophils (i.e., polymor- esterase staining for macrophages was phonuclear leukocytes or "PMN" with conducted to determine the purity of a nucleus having more than 2 lobes cell preparations (31). and fine neutrophilic granules in the A cultured porcine bone marrow cytoplasm) were prepared from the cell line (PBMCL) of unknown cell red blood cell (RBC) pellet by hypo- type, established at Veterinary Infectonic lysis of the RBC (29). Periph- tious Disease Organization (VIDO), eral blood mononuclear leucocytes Saskatoon, Saskatchewan was also (PBML, i.e., cells with an ovoid used in cytotoxicity assays. The cell nucleus and basophilic cytoplasm) concentrations were adjusted to 2 X were obtained from the buffy coat by 106 mL' and cultured in RPMI 1640 centrifugation at 1050 .X g for 45 min medium with 10% FBS. followed by additional separation on Ficol-paque (Pharmacia LKB, Bio- CYTOTOXICITY ASSAYS One hundred ,uL of 2 X 106 mL' technology AB, Uppsala, Sweden) (30). The cells were adjusted to a final PBML, PMN, PBMCL or 5 X concentration of 2 X 106 mL-1 and 106 mL-' PAM were incubated with cultured in RPMI 1640 medium with 100 pL of bacterial preparations seri10% fetal bovine serum (FBS) ally diluted in cell culture medium for (GIBCO Laboratories, Burlington, either 1.5, 8, 24 or 72 h at 37°C. The cytotoxicity was determined either by Ontario). Porcine pulmonary alveolar macro- measuring lactate dehydrogenase phages (PAM, i.e., large cells with a (LDH) release, trypan blue dye excluround or indented nucleus) were col- sion, or MTT [3-(4,5-dimethylthiazollected by lung lavage from either 2-yl)-2,5-diphenyl tetrazolium brohealthy pigs or pig lungs from a mide] conversion. The cytotoxicity slaughter house. For live pigs, 60 mL was presented as a percentage of of lung lavage collecting medium killing. In the LDH assay (32), perconsisting of Hanks' Balanced Salt centage of killing was calculated by Solution (HBSS) (GIBCO Laborato- the following formula: killing% = 100 ries) with 1% Antibiotic and Antimy- X (test well - cells alone well - toxin cotic Solution (containing 100 units alone well) / (complete killing well of penicillin, 100 p,g of streptomycin cells alone well), where the meaand 250 ng of amphotericin B mL-') surement was optical density (450 nm) (SIGMA Chemical Company, St. and the complete killing well was the Louis, Missouri, USA) were instilled well in which cells were completely against A. pleuropneumoniae prior to the experiment, sera from all the pigs were tested using an enzyme-linked immunosorbent assay (ELISA) for detection of antibodies against ethylenediaminetetracetic acid-extracted surface antigens of A. pleuropneumoniae serotypes 1, 2, 5 and 7, which are the most common serotypes in Saskatchewan (28). Animals used were all negative (titer ' 100) for antibodies against A. pleuropneumoniae. Furthermore, to confirm that pulmonary changes following treatment with sterilized A. pleuropneumoniae preparations were not due to live A. pleuropneumoniae in the bacterial preparations or by possibly natural exposure, the pulmonary tissues taken after euthanatization were cultured for A. pleuropneumoniae.

destroyed by sonication. Due to the contamination of RBC in PBML and PMN preparations, trypan blue dye exclusion assay using a hemocytometer was used to ensure that the LDH assay did measure cytotoxic effects on PBML and PMN. In this assay, percentage of killing was calculated as killing% = 100 X (number of dead cells in test well - number of dead cells in control well) / (number of total cells in control well). Due to the difficulty of removing the adherent PAM and of completely destroying them for the total release of LDH, the MTT assay, based on the modified method of Hansen and coworkers (33), was used to measure the cytotoxicity of above bacterial preparations on PAM. After incubation with bacterial preparations in 96-well microtiter plates, 100 ,uL of the medium was removed from each well of the 96-well microtitre plates and 25 ,uL of MTT (5 mg mL'1 in PBS) (SIGMA Chemical Company, St. Louis, Missouri, USA) was added to each well of the plates. After 2 h incubation, 100 ,uL of extraction buffer was added. Following 16 h incubation, the plates were read in a microtiter plate reader (Bio-Rad, Model 3550). The cytotoxicity was calculated by the following formula: killing% = 100 - [(Test-Blank) / (Control-Blank)] X 100, where the measurement was optical density (595 nm with a reference wavelength at 655 nm). IDENTFCATION OF CYTOTOXIC COMPONENTS IN TWO BACTERIAL EXTRACTS

For trypsin and heat treatments, bacterial preparations were either incubated with trypsin (100 ,ug mL-) at 37°C for 15 min for trypsin treatments, or incubated at 60°C, or 800 or 100°C for 15 min in sealed tubes for heat treatments, then tested for cytotoxic effects. In neutralization tests, bacterial preparations were incubated with antisera for 60 min at 37°C. Rabbit antiserum against the cell-free culture supernatant of A. pleuropneumoniae serotype 1 strain, AP37, was used (diluted 1:100). This rabbit antiserum was produced against bacterial culture supernatant after removing bacteria, with Freund's Complete Adjuvant for the 1st immunization and Freund's Incomplete for the 95

TABLE L. Cytotoxicity of heat-killed bacteria and LPS on PAM (as measured by the MTT assay) and other types of porcine cells (as measured by the LDH assay) in vitro Concentrations and effects of killed A. pleuropneumoniae and A. pleuropneumoniae LPS Heat-killed Bacteria (bacteria mL-1) 1 X 106 1 X 107 1 x 103 1 x 102 1 X 105 1 x 104 24 PAM 42.5 ± 3.Ob 34.4 t 5.6 28.6 ± 3.7 12.7 ± 4.7 10.8 ± 2.2 0.2 ± 2.9 72 47.8 ± 2.9 39.0 ± 3.9 32.2 ± 5.3 20.3 ± 7.9 10.5 ± 1.9 12.2 ± 1.0 LPS (p,g mL-') 10 100 5 2.5 1.25 0.63 5.7 ± 3.5 6.9 ± 4.1 4.4 ± 2.6 2.6 ± 1.9 2.2 ± 0.8 PAM 1.5 44.7 ± 9.2 1.5 PBML 34.3 ± 10.5 19.3 ± 6.8 6.6 ± 3.0 2.4 ± 1.6 1.7 ± 1.3 0.1 ± 0.2 PMN 1.5 12.0 ± 2.3 3.0 ± 1.7 0.4 ± 0.4 0.2 ± 0.1 0.3 ± 0.2 0.3 ± 0.5 a PAM, PBML and PMN are pulmonary alveolar macrophages, peripheral blood mononuclear leukocytes and polymorphonuclear leukocytes (neutrophils) respectively bKilling% presented as mean ± standard deviation (SD), n = 4 Incubation Cellsa time (h)

boost. Immune sera from 2 pigs 46 d after infection with A. pleuropneumoniae serotype 1 strain, AP37 (named as anti-PAPI and 2) and antiserum from a pig 7 d after 1 injection with recombinant Apx II from serotype 7 (named as anti-CytoII) were also used (diluted 1:50). Sera from healthy rabbit and pig, and irrelevant rabbit antiserum against Streptococcus suis cell membrane antigen (VIDO) with similar protein concentrations as the above specific polyclonal antisera were used as controls. ANIMAL TREATMENT

experienced veterinarian and experienced animal health technician, and mean scores of the 2 assessments were used. Rectal temperatures were recorded during clinical evaluation. LUNG LESION EVALUATION

Gross changes in the lungs were observed 24 h after treatment. In order to examine objectively the total amount of edema, cell infiltration and hemorrhage in whole lungs, whole pigs and their lungs (removed from thoracic cavity) were weighed after killing. Lung to whole body weight ratios were calculated. To examine microscopic changes in the lungs, tissues in the affected and normal areas were fixed in formalin (10%, pH 7.0) and embedded in paraffin, and then sectioned tissues were stained with hematoxylin and eosin. The tissues in the affected and normal areas were also collected for A. pleuropneumoniae culture as mentioned above.

The bacterial preparations used were BCS, CSE and CSSE. Bacterial cultural medium (PPLO broth with 1% IsoVitaleX) containing 0.85% NaCl was used as a control solution. Following anesthesia, 10 mL of each different bacterial preparation and of the control solution was instilled intratracheally into the lungs of each pig (3 in each group). The pigs were euthanatized 24 h later. STATISTICAL ANALYSIS CLINICAL EVALUATION

Animals were evaluated clinically 8 and 24 h after treatment. Clinical scores were based on the severity of respiratory distress, the degree of general clinical illness and depression (34). Five grades were used: 0 was normal; 1 was slight increase in respiratory rate and effort with slight depression; 2 was marked increase in respiratory rate and effort with marked depression; 3 was severe increase in respiratory rate and effort with severe depression, mouth breathing and/or cyanosis. Pigs that died peracutely were graded 4. Clinical evaluations were performed by an 96

Analysis of variance (ANOVA) was used on group data for kinetic study (MGLH procedure of Systat,

1990, Systat Inc., Evanston, Illinois, USA), and a paired Student's t-test (GraphPad Instat, 1993, GraphPad Software, San Diego, California, USA) was used for comparing data collected at the same time point. The effect was considered significant at P ' 0.05. Killing percentages were transformed using arcsin Y and subjected to ANOVA. Linear correlation analysis was conducted to determine the correlation between trypan blue dye exclusion and LDH assays (GraphPad Instat).

RESULTS PURITY OF CELL PREPARATIONS

Both PBML and PMN preparations contained over 95% of PBML or PMN respectively in total leukocytes as judged by morphology. Both PBML and PMN preparations, especially PMN preparations, contained large numbers of erythrocytes ranging from 20 to 90% of the total cell population. Based on morphology and nonspecific esterase staining, PAM preparations contained 93.92 t 4.89% (range from 85% to 100%) macrophages with or without small numbers of other leukocytes and erythrocytes. CYTOTOXIC EFFECTS OF BACTERIAL PREPARATIONS IN VITRO

Due to the contamination by RBC of the PBML and PMN preparations, trypan blue dye exclusion assay using a hemocytometer was compared with the LDH assay to ensure that the LDH assay did measure cytotoxic effects on PBML and PMN. Trypan blue dye exclusion and LDH assays showed a similar pattern in measuring cytotoxic effects of all bacterial preparations on PBML and PMN cell preparations (data not shown, linear correlation analysis, r > 0.85 and 0.9, respectively). The cytotoxic effects on PBML and PMN as measured by the trypan blue dye exclusion assay were not significantly different from those on the total cell population as measured by the LDH assay (ANOVA, P > 0.09 and 0.1, respectively). The LDH assay did show the trend of cytotoxic effects on PBML and PMN although it did not measure exactly the cytotoxic effects on PBML and PMN due to contamination by erythrocytes, and therefore this assay was used throughout the study. The heat-killed bacteria were not cytotoxic to PAM (ANOVA, P = 0.25), PBML (P = 0.28) or PMN (P = 0.39) after 1.5 h incubation (data not shown). However, they had cytotoxic effects on PAM after 24 and 72 h incubation (P < 0.001 for both time points) (Table I). At high concentration (100 ,uL mL-'), A. pleuropneumoniae LPS had cytotoxicity on PAM, PBML and PMN, and these cytotoxic effects on PAM and PBML were stronger than that on PMN (P < 0.05 for all assays) (Table I). The

cytotoxicity decreased with the reduction of the LPS concentrations (Table I). In addition, E. coli LPS had cytotoxic effects similar to those of A. pleuropneumoniae LPS (data not shown; 37). Bacterial culture supernatant (BCS) showed marked cytotoxic effects on PAM (Fig. 1, P < 0.0001), PBML (P < 0.0001), PMN (P = 0.0003) and PBMCL (P < 0.0001) (37). LPS determination showed that a suspension containing 1 X 107 mL-' A. pleuropneumoniae immediately after heat-killing contained about 100 pLg mL-' LPS, and approximately 4 ,ug mL' in 4 h bacterial culture supernatant. Both CSSE and CSE preparations showed dosedependent cytotoxic effects on PAM, PBML, PMN and PBMCL (ANOVA, P < 0.001) (Table II). Heat over 60°C completely inhibited the cytotoxic effect of BCS on PAM, and heat over 80°C significantly inhibited cytotoxicity of CSSE and CSE (P < 0.001) (Fig. lA). Trypsin treatment completely inhibited the cytotoxic activities of BCS and CSSE, and partially those of CSE (P < 0.001) (Fig. IB). Rabbit antiserum against A. pleuropneumoniae culture supematant almost completely inhibited the cytotoxic activities of BCS and CSSE, but not those of CSE (Fig. IC). Sera from 2 pigs that had recovered from A. pleuropneumoniae infection (46 d after infection) either completely or partially neutralized the cytotoxic effects of CSE, CSSE and BCS (Table III). Porcine antiserum against one of the cytotoxins (antiCytoll) partially neutralized the cytotoxicity of CSSE and BCS, but had no effect on CSE cytotoxicity (Table III). Control sera including normal and irrelevant immune sera did not show neutralization effects. To further test whether the cytotoxicity of CSE was mainly due to protein(s), the cytotoxicities of different protein fractions of CSE separated with the isoelectric focusing method were measured. The different fractions had different degrees of cytotoxicity, namely, the lower the pI, the higher the cytotoxicity (Table IV). Both protein and LPS determinations (Table IV) showed higher concentrations in the collections at the 2 extreme pl zones and lower in the middle pI zones; hence the LPS and protein distributions in different CSE

A

n;r

A 'v>

6000 -teatZ H

38

_

24 6

CD

v

14 2,1 -1 O -

_~ ~ 800C

CSE

CSSE

BCS

c. 1O0C

Treatment

B

80 - Trypsin

3

+ Trypsin

60 1

40 20 0

CSSE Treatment

BCS

C

CSE

80 -

Antibody E

+

Antibody

60 F

40

F

20 O

B CO

CSSE Treatment

CSE

Figure 1. Identification of cytotoxic components in 2 bacterial extracts (CSSE and CSE) by comparing the effects of heat, trypsin and rabbit antiserum against A. pkuropneumoniae culture supernatant on the cytotoxicities of bacterial culture supernatant (BCS) and the 2 extracts as measured by the MTT assay on PAM after 8 h incubation. A) Effects of heat. 60°C, 80°C and 100°C on BCS, 80°C and 100°C on CSSE and CSE (CSSE and CSE were extracted at 60°C); B) Effects of trypsin (100 ,g mL-', 370C, 15 min) on BCS, CSSE and CSE; C) Neutralizing effects of rabbit antiserum against bacterial culture supernatant on BCS, CSSE and CSE. BCS was 1:4 diluted, and CSSE and CSE were 1:16 diluted from the original preparations for the final protein concentrations of 60 .g mL-', 34.4 pg mL-1 and 19.35 p,g mL-', respectively. Bars are means ± SD, n = 4. The data presented in A, B and C of Figure 1 were obtained from different experiments.

97

TABLE II. Cytotoxicity of CSE and CSSE on PAM (as measured by the MTT assay) and other types of porcine cells (as measured by the LDH assay) in vitro Dilutions and effects of CSE and CSSE 1:8 1:32 1:64 1:128 1:512 Effects of CSEb 19.3 ± 6.4 PAM 21.8 ± 7,7b 23.1 ± 8.4 16.9 ± 9.0 16.4 ± 8.8 13.3 ± 8.6 3.7 ± 3.7 PBML 20.6 ± 14.9 21.1 ± 14.9 22.6 ± 15.9 19.7 ± 15.3 16.6 ± 12.6 6.4 ± 5.4 0.0 PMN 18.4 ± 7.8 16.6 ± 7.7 15.9 ± 6.2 15.9 ± 6.8 8.9 ± 6.1 6.2 ± 7.4 0.0 PBMCL 45.4 ± 1.5 56.7 ± 11.7 45.9 ± 8.3 24.9 ± 7.6 15.9 ± 0.8 13.4 ± 1.1 4.2 ± 5.9 Effects of CSSEb PAM 41.4 ± 37.1 36.9 ± 31.2 7.1 ± 5.3 3.1 ± 2.7 16.4 ± 14.3 0.0 0.0 PBML 38.8 ± 12.1 35.4 ± 9.0 6.7 ± 4.3 25.8 ± 5.5 15.8 ± 4.2 2.15 ± 2.1 0.0 PMN 33.0 ± 9.9 25.1 ± 6.3 14.2 ± 4.6 5.7 ± 3.9 2.3 ± 2.1 1.3 ± 1.4 0.0 PBMCL 39.5 ± 6.3 28.4 ± 3.1 18.1 ± 2.0 5.6 ± 3.0 0.0 0.0 0.0 a PAM, PBML, PMN and PBMCL are pulmonary alveolar macrophages, peripheral blood mononuclear leukocytes, polymorphonuclear leukocytes (neutrophils) and porcine bone marrow cell line respectively bThe original preparations of CSE contained 215 ,ug mL-' protein and 433.5 ,ug mL-' LPS, and CSSE contained 499.3 i,g mL-' protein and 382.72 ,ug mL-' LPS c Killing% (Mean ± SD, n = 4)

Cellsa

1:2

1:4

TABLE IlI. Neutralizing effects of pig antisera on the cytotoxicities of 2 bacterial extracts and culture supernatant for PAM as measured by the MTT assay Bacterial preparationsb PAM with various antisera treatment, BCS CSSE CSE Bacterial preparation only 58.5 ± 3.3c 23.1 ± 4.9 31.1 ± 3.8 Bacterial preparation with irrelevant antiserum 51.4 ± 4.3 22.5 ± 1.5 29.3 ± 2.0 Bacterial preparation with anti-PAP1 0.0 ± 4.0 14.6 ± 3.8 11.3 ± 1.9 Bacterial preparation with anti-PAP2 0.0 ± 3.2 0.0 ± 7.2 10.9 ± 1.9 Bacterial preparation with anti-CytoIl 20.5 ± 11.3 6.8 ± 6.4 30.5 ± 3.3 2"Bacterial preparation only" is the positive control in which the PAM were incubated with different A. pleuropneumoniae preparations for 8 h. "Bacterial preparation with irrelevant antiserum," "Bacterial preparation with anti-PAPl," or "Bacterial preparation with anti-PAP2" or "Bacterial preparation with anti-CytoII" is the treatment in which different bacterial preparations were incubated with either control swine serum, anti-PAPl, or anti-PAP2 or anti-CytoIl swine antiserum (1:50 dilution) for 1 h at 37°C, followed by incubation with PAM for 8 h. bBCS (bacterial culture supernatant, 60 ,ug mLr' protein), CSSE and CSE were 1:8 dilution of original preparations (Table II) c Killing% (Mean ± SD, n = 4)

fractions did not follow the pattern of ferences in the ratios between the pigs cytotoxicity. treated with BCS or CSE and control pigs (Table V). EFFECTS OF BACTERIAL PREPARATIONS Microscopically, the lungs from the IN VIVO control pigs injected with PPLO Two pigs treated with CSSE had medium with 0.85% NaCl showed slightly increased temperatures (40.5 congestion and marked mononuclear and 40.9°C, respectively) about 8 h cell accumulation both in alveolar after treatment. One pig from the con- spaces and septa, but very little neutrol group also had a temperature of trophil infiltration. Occasionally, 40.5°C. None of the other animals individual dead cells could be seen. In exhibited clinical signs. all treatments, multiple epithelial Lungs were examined 24 h after cells in the bronchioles showed injection. Grossly, the lungs in all swelling with clear cytoplasm. The pigs examined, including control pigs, lungs from both CSSE and BCSwere diffusely reddened with the most treated pigs showed multiple foci of obvious change on the dorsal sur- interstitial mononuclear inflammatory faces. None of pigs had lesions typi- cells and neutrophil infiltration in cal of PAP. The lungs from the pigs alveoli and bronchioles. Some of treated with CSSE had multifocal these neutrophils were degenerative. areas of hemorrhage, but the ratios of Multiple foci of hemorrhage also were lung:whole body weight in pigs present (37). These changes were treated with CSSE were significantly most pronounced in CSSE-treated higher than those in control pigs (P < pigs. The pigs treated with CSE 0.05). There were no significant dif- showed changes similar to those in 98

the controls, without obvious neutrophil infiltration or hemorrhage. Actinobacillus pleuropneumoniae was not isolated from the lungs of any pigs in the in vivo experiment.

DISCUSSION The present study has shown that, in addition to secreted protein cytotoxins, A. pleuropneumoniae possesses cytotoxic protein components on its surface. In order to accomplish its rapid multiplication and invasion, live A. pleuropneumoniae secretes several virulence factors including heat-labile high molecular weight RTX protein cytotoxins, a low molecular weight protein hemolysin and proteases (4-11,17). Heat-killed bacteria did not show significant cytotoxicity within a short time in this and other studies (36,37), but they killed PAM after 24 h incubation (Table I). The LPS of A. pleuropneumoniae or E. coli at the concentrations of 100 p,g mL-' had in vitro cytotoxic effects on porcine PAM within 1.5 h (Table I) or porcine endothial cells within 5 h (12) or human macrophages (38) within 24 h (38). The present study showed that a suspension containing 1 X 107 CFU A. pleuropneumoniae immediately after heat-killing had over 100 ,ug mL-' LPS. Thus, it is possible that LPS released from heat-killed bacteria is involved in the cytotoxicity of killed bacteria on PAM. Pigs immunized with CSE, a crude surface extract from A. pleuropneumoniae, or with low pI fractions of

CSE separated from isoelectric focusing on ion exchange were protected from death after challenge (20,21), suggesting that there might be an important virulence factor(s) present in these preparations. The present study has further shown that CSE and another extract, CSSE (crude surface plus culture supernatant extract), are cytotoxic to several types of porcine cells in vitro (Table II). Although the cytotoxic activities of CSE and CSSE were resistant to 60°C during preparation, which could inhibit the cytotoxic effect of culture supernatant (Fig. 1A), their cytotoxic effects could be reduced significantly by higher temperatures (80 and 100°C) (Fig. 1A) and trypsin treatment (Fig. iB), indicating that proteins were involved in their cytotoxicities. The concentrations of LPS were over 300 ,ug mL' in the original preparations of the extracts (Table II). Therefore, the cytotoxic effects in these assays may have been due partially to LPS. Further experiments showed that the active components in CSE and CSSE are different. Rabbit antiserum against BCS blocked the cytotoxicities of BCS and CSSE, but not those of CSE (Fig. IC). By sodium dodecyl sulphate-polyacrylamide gel (10%) electrophoresis, CSSE showed a strong band with molecular weight about 104 kilodaltons (kDa), but CSE did not show a visible band with molecular weights between 104-110 kDa (unpublished data). These results indicate that secreted components were involved in CSSE cytotoxicity while bacterial surfaceassociated components were involved in CSE activity. Further study is needed to determine the exact location of the cytotoxic components of CSE. CSSE was extracted at 60°C for 1 h which could completely inactivate the cytotoxicity of BCS. It is not clear why this temperature (600C) did not inactivate the cytotoxicity of secreted components in CSSE during the extraction procedure. Our preliminary experiment showed that CSSE preparations extracted with or without saline demonstrated a similar cytotoxic effect on PAM (37), indicating that saline is not the factor stabilizing the cytotoxic activity of secreted components in CSSE. One recent experiment by Beaudet and co-workers (10) showed that bovine serum albumin

TABLE IV. Cytotoxicity of fractions of CSE separated by isoelectrical focusing on PBML in vitro as measured by the LDH assay LPS and protein concentrations of CSE fractions CSE fractions (dilutions and killing %) (,ug mL-')b 1:2 pIRangea 1:8 1:32 1:128 LPS Protein 1-1.5 73.3 ± 19.4c 27.4 ± 22.0 36.5 ± 24.1 0.0 738.7 1072.4 1.6-3 65.2 ± 12.3 35.4 ± 20.3 21.7 ± 23.0 0.5 ± 1.0 358.7 505.5 3.1-5 53.9 ± 19.7 33.7 ± 19.9 5.9 ± 7.8 0.0 288.8 612.5 42.2 ± 14.9 8.2 ± 6.8 5.1-6 31.9 ± 8.8 3.9 ± 3.2 472.2 1056 6.2 ± 6.4 6.1-10 27.9 ± 19.5 18.0 ± 7.2 0.0 547.9 1404.7 CSE 20.6 ± 14.9 21.1 ± 14.9 19.7 ± 15.3 5.4 ± 6.4 395.4 215 a The values in the pl column are the pH values at which the different fractions were collected; the CSE is the original preparation b LPS and protein concentrations in various pl ranges of original preparations c Killing% (Mean ± SD, n = 4)

TABLE V. Effects of different bacterial preparations in vivo: Lung:whole body weight ratio Lung:Body weight Treatmenta Lung weight (kg ± SD) Body weight (kg ± SD) (Ratio ± SD) CSSE 0.37 ± 0.05b 20.16 ± 4.16 0.018 ± 0.001* BCS 0.30 ± 0.03 18.41 ± 1.88 0.016 ± 0.000 CSE 0.36 ± 0.03 24.26 ± 2.33 0.015 ± 0.001 Control 0.33 ± 0.01 20.53 ± 0.49 0.016 ± 0.000 a Intratracheal treatments with 10 mL of bacterial preparations: CSSE, BCS (bacterial culture supernatant) and CSE. Control was PPLO medium with 0.85% NaCl. Lungs and bodies were weighed 24 h after treatment bMean ± SD, n =3 * Significantly different from control group, P ' 0.05

and dithiothreitol (DTT) stabilized a low molecular weight protein hemolysin. Perhaps a bacterial cellassociated factor(s) extracted by glass beads stabilized the toxins secreted into the culture supernatants. The different fractions of CSE separated by isoelectric focusing showed different degrees of cytotoxicity (Table IV), and the patterns of such cytotoxic effects did not follow the pattern of LPS distribution in the different fractions (Table IV). These results suggest that LPS is not the main factor in the cytotoxicity of CSE and that several protein components are involved in CSE cytotoxicity. This is in agreement with the experiment by Bielefeldt Ohmann et al (21) which showed that 4 different fractions of CSE separated by isoelectric focusing induced different secondary B-cell responses in vitro. The nature of cell-surface associated protein components in CSE is not clear. Actinobacillus actinomycetemcomitans produces a non-secreted protein leukotoxin which is located on the bacterial surface (39). This cytotoxin belongs to the RTX cytotoxin family which includes those found in many gram-negative bacteria such as A. pleuropneumoniae, P. haemolytica

and E. coli (4,40). So far it has not been reported that RTX cytotoxins from A. pleuropneumoniae and other members of the RTX cytotoxin family are located on the bacterial envelope. In addition, the rabbit antiserum against A. pleuropneumoniae (serotype 1) culture supernatant (containing Apx toxins) and porcine antiserum against one of the Apx toxins (antiCytoll, Table III) did not have a neutralizing effect on the cytotoxicity of CSE. Therefore, it seems that A. pleuropneumoniae (serotype 1) can produce bacteria surface-associated cytotoxic proteins, which are unlikely to be the RTX toxins. The antisera from 2 pigs collected 47 d after infection with A. pleuropneumoniae (serotype 1, strain AP37) partially inhibited the activities of CSE and CSSE (Table III), and immunization with CSE and one fraction of low pI from isoelectric focusing separated CSE protected the pigs from death after A. pleuropneumoniae infection (21), indicating that at least some components in CSE and CSSE were recognized by the host. After intratracheal inoculation, CSSE induced significant increases in lung:whole body weight ratios, but CSE and BCS did not (Table V). The

99

increase in such ratios indicates that the weight of lungs after CSSE treatment was increased. Histologically, CSSE induced more severe hemorrhage and neutrophil infiltration than that induced by BCS and CSE. The increase in the weight of lungs after CSSE treatment could be caused by the accumulation of extravascular blood due to hemorrhage and inflammatory exudates. The lesions induced by CSSE and BCS were similar to those of experimental disease caused by live A. pleuropneumoniae except that there was little necrosis. It is interesting that CSE was toxic to cells in vitro, but did not induce changes such as hemorrhage and neutrophil infiltration in vivo. It is possible that components in CSE may enhance the effects of BCS, but that the amount of CSE used was not enough to induce in vivo effects alone. In addition, it is not clear how CSSE and BCS induced the pulmonary hemorrhage and inflammation. Based on the in vitro cytotoxic effects of LPS on porcine endothelial cells (12) and the secreted protein toxins on other porcine cells (present results), as well as elsewhere (36), it is possible that the pulmonary hemorrhage could be due to the direct cytotoxic effects on vascular endothelium by a large amount of LPS contained in CSSE, and secreted protein toxins in CSSE and BCS. It is also possible that the secreted protein toxins and LPS may induce pulmonary inflammation by triggering the production of proinflammatory mediators including cytokines (15). In conclusion, besides the well known cytotoxins/hemolysins and LPS, A. pleuropneumoniae (serotype 1) produces cell-surface associated protein components which may either be cytotoxic or stabilize the cytotoxic components in culture supernatant. The CSSE or BCS, but not CSE alone, were able to induce hemorrhagic pneumonia with marked neutrophil infiltration in alveoli and bronchioles. Based on the fact that CSSE containing BCS and CSE caused more profound changes than that induced by BCS or CSE alone, it is concluded that cell-surface associated and secreted components act synergistically in the pathogenesis of porcine Actinobacillus pleuropneumonia. Identification of the active components in CSSE and CSE may enhance 100

our understanding of the pathogenicity of A. pleuropneumoniae and improve vaccines.

8. SMITS M, KAMP E, BRIAIRE J, JANSEN R, VAN LEENGOED L, VAN DIJK J. Induction of pneumonic lesions by recombinant cytolysins of Actinobacillus pleuropneumoniae. Proc Int Pig Vet Soc 1992: 187. 9. KILIAN M, MESTECKY J, SCHROACKNOWLEDGMENTS HENLOHER RE. Pathogenic species of genus Haemophilus and Streptococcus Dr. Huang was financially suppneumoniae produce immunoglobulin Al ported by a Scholarship from the protease. Infect Immun 1979; 26: 143-149. College of Graduate Studies and 10. BEAUDET R, MCSWEEN G, BISAILLON JG. Production and purification of a Research, and by funds from Agricullow molecular weight haemolysin proture and Agri-Food Canada and the duced by Actinobacillus pleuropneumoDepartment of Veterinary Pathology, niae serotype 1. Res Vet Sci 1993; 54: 45-51. University of Saskatchewan. The present experiments were sponsored by a 11. NEGRETE-ABASCAL E, TENORIO VR, SERRANO JJ, GARCIA CC, DE fund from Alberta Agricultural LA GARZA M. Secreted proteases from Research Institute through VIDO and Actinobacillus pleuropneumoniae serotype were carried out at VIDO. We sin1 degrade porcine gelatin, hemoglobin and immunoglobulin A. Can J Vet Res 1994; cerely thank Drs. A. Rossi-Campos, 58: 83-86. Janet Lees and Trent Watts, and the 12. SEREBRIN S, ROSENDAL S, VALanimal care staff at VIDO for obtainDIVIESO-GARCIA A, LITTLE PB. ing porcine cells, and Dr. G. Gerlach Endothelial cytotoxicity of Actinobacillus for helping with bacterial culture. pleuropneumoniae. Res Vet Sci 1991; 50: 18-22. 13. BERTRAM TA. Pathobiology of acute pulmonary lesions in swine infected with REFERENCES Haemophilus (Actinobacillus) pleuropneuCan Vet J 1988; 29: 574-577. 1. SHOPE RE. Porcine contagious pleurop- 14. moniae. BERTRAM TA. Actinobacillus pleuropneumonia. I. Experimental transmission, neumoniae: Molecular aspects of virulence etiology and pathology. J Exp Med 1964; and pulmonary injury. Can J Vet Res 1990; 119: 357-368. 54: S53-S56. 2. SEBUNYA TNK, SAUNDERS JR. HS, POTTER AA, CAMPOS Haemophilus pleuropneumoniae infection 15. HUANG M, LEIGHTON FA, WILLSON P, in swine: A review. J Am Vet Med Assoc HAINES D, YATES WDG. Pathogenesis 1983; 182: 1331-1337. of porcine Actinobacillus pleuropneumo3. ROSENDAL S, MITCHELL WR, WILnia: Part II. Roles of Proinflammatory SON WR, ZAMAN MR. Hemophilus cytokines (submitted to Can J Vet Res). pleuropneumonia. Lung lesions induced by sonicated and sterile culture supernatant. 16. MACINNES JI, ROSENDAL S. Prevention and control of Actinobacillus Proc Int Pig Vet Soc 1980: 221. (Haemophilus) pleuropneumoniae infec4. FREY J, BOSSE JT, CHANG Y-F, tion in swine: A review. Can Vet J 1988; CULLEN MJ, FENWICK B, GER29: 572-574. LACH GF, GYGI D, HAESEBROUCK F, INZANA R, JANSEN R, KAMP EM, .17. FEDORKA-CRAY PJ, ANDERSON GA, CRAY WC Jr, GRAY JT, BREISCH MACDONALD J, MACINNES JI, MITSA. Actinobacillus (Haemophilus) pleuTAL KR, NICOLET J, RYCROFT AN, ropneumoniae. Part II. Virulence factors, SEGERS RPAM, SMITS MA, STENimmunity, and vaccines. Compend Contin BEAK E, STRUCK DK, VAN DEN Educ Pract Vet 1994; 16: 117-125. BOSCH JF, WILLSON PJ, YOUNG R. Actinobacillus pleuropneumoniae RTX- 18. NIELSEN R. Haemophilus pleuropneumoniae: Serotypes, pathogenicity, and toxins: Uniform designation of hemolycross immunity. Nord Vet Med 1979; 31: sins, cytolysins, pleurotoxin and their 407-413. genes. J Gen Microbiol 1993; 139: 19. FEDORKA-CRAY PJ, STINE DL, 1723-1728. GREENWALD JM GRAYJT, HUETHER 5. FREY J, NICOLET J. Regulation of MJ, ANDERSON GA. The importance of hemolysin expression in Actinobacillus secreted virulence factors in Actinobacillus pleuropneumoniae serotype 1 by Ca2+. pleuropneumoniae bacterin preparation: A Infect Immun 1988; 56: 2570-2575. comparison. Vet Microbiol 1993; 37: 6. KAMP EM, POPMA JK, ANAKOTTA 85-100. J, SMITS MA. Identification of hemolytic and cytotoxic proteins of Actinobacillus 20. WILLSON PJ, ROSSI-CAMPOS A, POTTER AA. Tissue reaction and immupleuropneumoniae by use of monoclonal nity in swine immunized with Actinobacilantibodies. Infect Immun 1991; 59: lus pleuropneumoniae vaccines. Can J Vet 3079-3085. Res 1995; 59: 299-305. 7. RYCROFT AN, WILLIAMS D, CULLEN JM, MACDONALD J. The cytotoxin of 21. BIELEFELDT OHMANN H, MCDOUGALL L, POTTER A. Secondary in vitro Actinobacillus pleuropneumoniae (pleuroB lymphocyte (antibody) response to toxin) is distinct from the haemolysin and microbial antigens: Use in appraisal of is associated with 120 kDa polypeptide. J vaccine immunogenicity and cytokine Gen Microbiol 1991; 137: 561-568.

22.

23.

24.

25.

26.

27.

immunoregulation. Vaccine 1991; 9: 171-176. WILLSON PJ, OSBORNE AD. Comparison of common antibiotic therapies for Haemophilus pleuropneumoniae in pigs. Can Vet J 1985; 26: 312-316. ANDERSON C, POTTER A, GERLACH G. 1991. Isolation and Molecular characterization of spontaneously occurring cytolysin-negative mutants of Actinobacillus pleuropneumoniae serotype 7. Infect Immun 1991; 59: 4110-4116. GERLACH GF, ANDERSON C, POTTER A, KLASHINSKY A, WILLSON PJ. Cloning and expression of a transferrin-binding protein from Actinobacillus pleuropneumoniae. Infect Immun 1992; 60: 892-898. ALTMAN E, BRISSON JR, PERRY M. Structure of the 0-chain of the lipopolysaccharide of Haemophilus pleuropneumoniae serotype 1. Biochem Cell Bio 1986; 64:1317-1325. STOSCHECK CM. Quantitation of Proteins. In: Deutscher MP, ed. Methods in Enzymology: Guide to Protein Purification. New York: Academic Press, Inc, 1990: 57-60. CANADIAN COUNCIL ON ANIMAL CARE. Guide to the Care and Use of Experimental Animals. 2 vols. Ottawa, Ontario: CCAC, 1980-1984.

28. WILLSON PJ, SCHIPPER C, MORGAN ED. The use of an enzyme-linked immunosorbent assay for diagnosis of Actinobacillus pleuropneumoniae infection in pigs. Can Vet J 1988; 29: 583-585. 29. FIELD TR, WILLIAMS MR, BUNCH KJ. An improved method for the isolation of a neutrophil-rich fraction from porcine blood. Br Vet J 1985; 141: 355-361. 30. CAMPOS M, ROSSI CR. Inability to detect a K cell in bovine peripheral blood leukocytes. Vet Immunol Immunopathol 1985; 8: 351-362. 31. KOSKI IR, POPLACK DG, BLAESE RM. A non-specific esterase stain for the identification of monocytes, and macrophages. In: Bloom BR, David JR, eds. In Vitro Methods of Cell-Mediated and Tumor Immunity. New York: Academic Press, Inc, 1976: 359. 32. BERGMEYER HU, GRAVEHN K. Methods of enzymatic analysis. New York: Academic Press, Inc, 1978: 574-579. 33. HANSEN MB, NIELSEN SE, BERG K. Re-examination and further development of a precise and rapid dye method for measuring cell growth/cell kill. J Immunol Meth 1989; 119: 203-210. 34. ROSSI-CAMPOS A, ANDERSON C, GERLACH G, KLASHINKY S, POTTER A, WILLSON PJ. Immunization of pigs against Actinobacillus pleuropneuno-

niae with two recombinant protein preparations. Vaccine 1992; 10: 512-518. 35. HUANG HS. A study of the pathogenesis of acute Actinobacillus pleuropneumoniae infection in swine (PhD dissertation). Saskatoon,University of Saskatchewan, 1995. 204 p. 36. BENDIXEN P, SHEWEN PE, ROSENDAL S, WILKIE BN. Toxicity of Haemophilus pleuropneumoniae for

porcine lung macrophages, peripheral 37.

38.

39.

40.

blood monocytes, and testicular cells. Infect Immun 1981; 33: 673-676. VAN LEENGOED LA, KAMP EM, POL JMA. Toxicity of Haemophilus pleuropneumoniae to porcine lung macrophages. Vet Microbiol 1989; 19: 337-329. DAVIS WB, BARSOUM IS, RAMWELL PW, YEAGER H. Human alveolar macrophages: Effects of endotoxin in vitro. Infect Immun 1980; 30: 753-758. OHTA H, HARA H, FUKUI K, KURIHARA H, MURAYAMA Y, KATO K. Association of Actinobacillus actinomycetemcomitans leukotoxin with nucleic acids on the bacterial cell surface. Infect Immun 1993; 61: 4878-4884. WELCH RA. Pore-forming cytolysins of Gram-negative bacteria. Mol Microbiol 1991; 5: 521-528.

101