Expression of Proinflammatory Cytokines by Hepatic Macrophages in Acute Classical Swine Fever

J. Comp. Path. 2005, Vol. 133, 23–32

www.elsevier.com/locate/jcpa

Expression of Proinflammatory Cytokines by Hepatic Macrophages in Acute Classical Swine Fever A. Nu´n˜ez, J. C. Go´mez-Villamandos, P. J. Sa´nchez-Cordo´n, M. Ferna´ndez de Marco, M. Pedrera, F. J. Salguero* and L. Carrasco Departamento de Anatomı´a Patolo´gica Comparada, Facultad de Veterinaria, Universidad de Co´rdoba. Edifico de Sanidad Animal, Campus Universitario de Rabanales, 14014, Co´rdoba and *Centro de Investigacio´n en Sanidad Animal—Instituto Nacional de Investigaciones Agrarias (CISA—INIA), Valdeolmos, Madrid, Spain

Summary Fourteen pigs were inoculated with the ‘Alfort 187’ strain of classical swine fever (CSF) virus and killed in pairs at 2, 4, 7, 9, 11, 14 or 17 days post-inoculation for histopathological, ultrastructural and immunohistochemical examination. For the latter method, the antibodies used were those against viral antigen Gp55, porcine myeloid marker SWC3, IL-1a, IL-6, TNF-a and Factor VIII-related antigen. Activation and increase in the number of hepatic macrophages was observed following viral detection in liver, as well as an increase in IL-1a and IL-6 production, mainly by Kupffer cells. Maximum detection of viral antigen was observed in the middle stage of the experiment coinciding with overexpression of the three cytokines studied, with IL-6 production by interstitial macrophages prominent at the end. Additionally, the labelling of platelets for Factor VIII-related antigen and the ultrastructural study of the sinusoids revealed activation and aggregation of thrombocytes close to Kupffer cells at the beginning of the infection. The liver seems to play a prominent role in the origin of the thrombocytopenia that occurs in CSF and contributes to the overexpression of proinflammatory cytokines considered responsible for the disorders observed during the course of the disease. q 2005 Elsevier Ltd. All rights reserved. Keywords: classical swine fever; cytokines; liver; macrophages; pig; viral infection

Introduction Classical swine fever (CSF) is a fatal disease caused by a small-enveloped RNA virus belonging to the genus Pestivirus, family Flaviviridae (Van Regenmortel et al., 2000). The acute disease is characterized by haemorrhage, thrombocytopenia and lymphoid depletion (Cheville and Mengeling, 1969; Trautwein, 1988). These changes were initially attributed to the effects of viral replication in platelets, endothelial cells and lymphoid cells (Weiss et al., 1973; Narita et al., 1996). Recent studies, however, seem to rule out a direct viral effect on these cells as the main cause of the changes (Summerfield et al., 1998; Go´mez-Villamandos Correspondence to: L. Carrasco. 0021-9975/$ - see front matter

doi:10.1016/j.jcpa.2005.01.003

et al., 2000; 2001; Sa´nchez-Cordo´n et al., 2002); instead, it seems possible that activated monocytemacrophages, in providing an early source of proinflammatory and immune response mediators, are responsible for the typical lesions of CSF (Knoetig et al., 1999; Carrasco et al., 2001; Bautista et al., 2002; Sa´nchez-Cordo´n et al., 2002). Undoubtedly, the main target of the virus is the monocyte– macrophage (Trautwein, 1988; Knoetig et al., 1999) and, although CSF virus does not exert a cytopathogenic effect in cell culture (Korn and Zoeth, 1971), these cells undergo major changes, including phagocytic and biosynthetic activation, during the course of the disease (Carrasco et al., 2001; Go´mezVillamandos et al., 2001). The liver, in contrast to the kidneys and lymphoid organs, has not been considered an q 2005 Elsevier Ltd. All rights reserved.

24

A. Nu´n˜ez et al.

important organ for CSF replication, since hepatic lesions are somewhat inconsistent (Van der Molen and van Oirschot, 1981). However, viral antigen has been detected in Kupffer cells (KCs), interstitial macrophages (IMøs), epithelial cells of the biliary ducts and infiltrating mononuclear cells (Martı´n de las Mulas et al., 1997), and has been associated with the hypertrophy of KCs and mononuclear infiltrates that characterize the hepatic lesions (Liess, 1987). Hepatic macrophages, which represent an important component of the mononuclear phagocyte system, are composed mainly of fixed macrophages in the sinusoids (KCs, and, to a lesser extent, IMøs). Slight disturbances in the function of these immunomodulatory cells may lead to serious disorders. KCs are activated by a variety of particles and substances, including viruses (Sendai virus, Newcastle disease virus and African swine fever virus), bacterial lipopolysaccharide, muramyl dipeptide, gamma interferon (IFN) and tumour necrosis factor (TNF) (Decker, 1990; Go´mez-Villamandos et al., 1995). These induce the release of interleukins (ILs), IFN, TNF and nitric oxide, substances that trigger and modulate inflammatory processes (Marianneau et al., 1999). Infection and activation of mononuclear cells play an important role in CSF (Knoetig et al., 1999; Carrasco et al., 2001; Go´mez-Villamandos et al., 2001; Sa´nchezCordo´n et al., 2002), but the role of hepatic macrophages as producers of inflammatory mediators has not yet been investigated. TNF-a, IL-1a and IL-6 are three proinflammatory cytokines that form part of a complex defence network that protects the host against inflammatory agents, microbial invasion and injury (Laskin and Pendino, 1995). However, overproduction or aberrant regulation of these cytokines may harm the host, by inducing tissue injury or alteration of the immune system (Laskin and Pendino, 1995). In this way, the pulmonary lesions induced by African swine fever (ASF) virus have been attributed to the release of IL-1a and TNF-a by pulmonary intravascular macrophages activated by the virus (Carrasco et al., 2002); moreover, apoptosis of lymphocytes in the spleen and lymph nodes in ASF has been associated with an increase in IL-1a, IL-6 and TNF-a expression by macrophages (Salguero et al., 2002). Recently, the vascular disorders and thymocyte apoptosis in CSF have been related to an increase in the production of these same cytokines by activated macrophages (Knoetig et al, 1999; Sa´nchez-Cordo´n et al., 2002). Understanding the complex interaction of cytokines in immunomodulatory diseases

would seem to be the next step in elucidating the pathogenesis of CSF. This paper evaluates the changes that occur in the hepatic population of macrophages in CSF, and the expression of TNF-a, IL-1a and IL-6 by these cells.

Materials and Methods Animals, Virus and Experimental Design Sixteen 4-month-old Large White ! Landrace pigs of either sex, weighing approximately 30 kg at the start of the study, and serologically negative for CSF, ASF, porcine respiratory and reproductive syndrome and Aujeszky’s disease were employed. Fourteen animals were inoculated intramuscularly with 105 TCID50 of the virulent CSF virus isolate ‘Alfort 187’ (Wensvoort et al., 1989). Two pigs used as uninfected controls received only phosphatebuffered saline (PBS), pH 7$2. After inoculation, clinical signs and rectal temperature were monitored daily. The inoculated pigs were anaesthetized and killed in pairs at 2, 4, 7, 9, 11, 14 or 17 days postinoculation (dpi) with thiopental-sodium (Thiovetw, Vet Limited, Leyland, England), following sedation with azaperone (Stresnilw, Janssen Animal Health, Beerse, Belgium). The remaining two pigs (controls) were killed at the end of the experiment. This experiment was performed in the Centro de Investigacio´n en Sanidad Animal (Valdeolmos, Madrid, Spain), in accordance with the Code of Practice for the Housing and Care of Animals used in Scientific Procedures, approved by the European Economic Community Union in 1986 (86/609/EEC). Processing of Specimens for Histopathology and Transmission Electron Microscopy (TEM) Samples from liver were fixed by immersion in 10% buffered formalin solution (pH 7$2) and Bouin’s solution for structural and immunohistochemical studies and in glutaraldehyde 2$5% in 0$1 M phosphate buffer (pH 7$4) for ultrastructural study. Samples for structural and immunohistochemical examination were dehydrated through a graded series of alcohol to xylol and embedded in paraffin wax by routine techniques for light microscopy. Wax-embedded sections (4 mm) were cut and stained with haematoxylin and eosin for structural study. For TEM, samples were postfixed in 2% osmium tetroxide, dehydrated in acetone and embedded in epoxy resin (Durcupanw ACM, Fluka Chemie AG,

25

Hepatic Cytokines in Classical Swine Fever

Buchs, Switzerland). Sections (50 nm) of liver for ultrastructural examination were counterstained with uranyl acetate and lead citrate, and viewed through a Philips CM-10 transmission electron microscope.

Excel 97w and differences from controls were tested for significance (P!0$05) by Student’s t-test.

Results Clinical Signs and Gross and Histopathological Findings

Immunohistochemistry The avidin—biotin—peroxidase complex (ABC) technique was performed on serial sections (3 mm) of liver, according to Salguero et al. (2001). The primary antibodies employed in this study, their specific dilutions and antigen retrieval procedures are summarized in Table 1. Cell Counts and Statistical Analysis To calculate the number of immunolabelled KCs and IMøs present and compare the results obtained with the various antibodies employed, two paraffinwax blocks from the liver of each animal were selected and serial sections were used for the various immunohistochemical studies. From these serial sections the number of immunolabelled KCs was determined from a cell count in 25 consecutive fields (0$20 mm2) of hepatic parenchyma per block, while the IMø count was determined from 25 fields (0$20 mm2) in the interlobular spaces centred in a periportal region. The distinction between cells was made by virtue of location, size and morphology. The results were expressed in cells/mm2 at various dpi. Means and standard deviations (SD) were calculated with Microsoft

Inoculated animals showed non-specific signs (loss of appetite, apathy and digestive disturbances) from 2 dpi, coinciding with the onset of pyrexia (39$5–41$2 8C). Liquid to pasty, greenish diarrhoea was present from 9 dpi. Post-mortem examination of animals killed from 2 dpi revealed petechial haemorrhages in submandibular lymph nodes and tonsils; petechiae were also found in urinary bladder, kidneys (cortex and medulla) and mediastinal lymph nodes from 4 dpi, and in mesenteric lymph nodes, serous membranes, respiratory and intestinal mucosa and pulmonary parenchyma from 9 dpi. Necrotic foci were observed in the tonsils from 4 dpi and in the ileocaecal valve and colon from 11 dpi. No macroscopical lesions were found in the liver, but histopathological study revealed periportal foci of mononuclear infiltrate from 4 dpi onwards. They were mainly composed of monocytes/macrophages and increased in size throughout the experiment. At 14 and 17 dpi, similar foci were also observed in the hepatic parenchyma. The uninoculated animals remained healthy throughout the experiment, showing no gross or microscopical lesions.

Table 1 Details of the immunohistochemical reagents

Primary antibody

Class*

Chemical fixative

Antigen or cell detected

SWC3

mAb

Bouin’s solution

WH303

pAb

Bouin’s solution

Anti-Factor VIII-related antigen IL-1a IL-6 TNF-a

pAb

10% Buffered formalin solution

Porcine myeloid marker Viral glycoprotein gp55 (E2) Factor VIII-related antigen

pAb pAb mAb

Bouin’s solution Bouin’s solution Bouin’s solution

Anti-human IL-1a Anti-porcine IL-6 Anti-human TNF-a

Antibody dilution†

Antigen retrieval procedure ‡

1 in 10

Citrate buffer

1 in 50

None

1 in 800

Protease¶

1 in 100 1 in 10 1 in 25

Tween 20§ Tween 20§ Tween 20§

Commercial source Biovet-UCO, Cordoba, Spain Veterinary Laboratories Agency, Weybridge, UK Dako, Glostrup, Denmark Endogen, Woburn, USA Endogen, Woburn, USA Biosource, Nivelles, Belgium

The avidin–biotin complex (ABC) method (Vector Laboratories, Burlingame, CA, USA) was used with 3,3 0 -diaminobenzidine tetrahydrochloride (Sigma-Aldrich Chemie, Steinheim, Germany) as chromogen. * Monoclonal (mAb) or polyclonal (pAb) antibodies. † In PBS containing normal goat serum 10%. ‡ Tri-sodium citrate dihydrate (Merck, F.R., Germany), 0.01molar (pH 3$2), microwaves 10 min. ¶ Protease type XIV (Sigma-Aldrich Chemie, Steinheim, Germany) 0$1%, 8 min. § Tweenw 20 (Merck, Mu¨nchen, Germany) 0$1%, 10 min.

26

A. Nu´n˜ez et al.

Immunohistochemistry Myeloid cells, such as neutrophils, circulating monocytes, KCs and IMøs, were immunolabelled by the anti-SWC3 monoclonal antibody, in both control and infected animals. From 2 dpi, the number of immunopositive KCs showed a marked increase, which was maintained throughout the experiment. The highest number of immunolabelled cells was reached at 7 dpi, progressively decreasing thereafter until 17 dpi (Fig. 1). Immunoreactive IMøs also increased in number from 2 dpi, and from 4 dpi were associated with the appearance of periportal foci of mononuclear infiltrate composed of cells reacting with the myeloid marker (Fig. 2). Specific reaction against glycoprotein 55 (E2) of CSF virus was observed in the cytoplasm of neutrophils, circulating monocytes, KCs and IMøs in all the infected animals, but only at 11 dpi were endothelial cells, biliary duct epithelial cells and fibroblasts immunolabelled. The number of immunopositive KCs remained steady at first, rising significantly at 9 and 11 dpi to a maximum at 11 dpi, falling thereafter to numbers resembling those in the first week of the disease. IMøs behaved similarly, except that the counts at 9 and 11 dpi did not reach such high numbers (Fig. 3). In the periportal infiltrates, only a few cells were labelled for viral antigen. The polyclonal antibody against Factor VIIIrelated antigen reacted only with the endothelial cells and not with the platelets. By contrast, all the infected animals showed immunolabelled platelets (and aggregates thereof) in sinusoids and portal vessels throughout the experiment, being most numerous at 2, 4 and 7 dpi, but especially at 4 dpi. The sinusoidal platelet aggregates were usually found close to KCs. In addition, some KCs showed SWC3 75

*

Cells/mm2

60 *

*

45

*

*

*

**

*

14

17

30 *

15

*

*

*

*

0 UI

2

4

7

9

11

dpi

Fig. 1. Counts (GSD) of KCs (,) and IMøs (&) labelled against the myeloid marker SWC3. *Difference from uninfected (UI) controls statistically significant, P!0$05.

Fig. 2. Numerous Kupffer cells, with sinusoidal location and oval nuclei with non-condensed chromatin (arrows), and periportal infiltrate of mononuclear cells (arrowheads) immunolabelled with porcine anti-SWC3; 17 dpi. ABC. !400.

cytoplasmic reactivity against Factor VIII-related antigen throughout the experiment but most abundantly at 2, 4 and 7 dpi (Fig. 4). Some circulating neutrophils were immunolabelled by the anti-IL-1a antibody in control animals. All the infected pigs, however, showed positive labelling in neutrophils, monocytes, KCs and IMøs (Fig. 5). The numbers of KCs expressing IL-1a, which were significantly different from those of the controls, progressively rose from 2 dpi to peak at 11 dpi, decreasing thereafter during the final days of the experiment (Fig. 3). IL-1a-positive IMøs showed statistically significant increases at 7, 11, 14 and 17 dpi, reaching their highest value at 17 dpi (Fig. 3). IL-6 expression by liver macrophages could not be detected in control animals. However, immunolabelled KCs and IMøs were observed in infected pigs from 2 dpi, both populations presenting similar cell count values and being significantly different from those of uninoculated animals at 2, 4

27

Hepatic Cytokines in Classical Swine Fever

25

35

20

Cells/mm2

Cells/mm2

Gp55 45

20 15 10

IL-1α

*

15

*

10

* *

5

5

0

0

*

UI

2

4

7

9

11

14

17

UI

dpi

7

9

11

14

17

dpi

25 20

*

*

Cells/mm2

20

4

*

IL-6

TNF-α

25

2

*

*

*

*

Cells/mm2

*

15 10

15 * 10 * 5

5

**

* 0

0 UI

2

4

7

9

11

14

17

dpi

UI

2

*

*

4

**

** 7

9

11

* 14

* 17

dpi

Fig. 3. Counts (GSD) of KCs (,) and IMøs (&) immunolabelled against viral antigen (Gp55), IL-1a, TNF-a and IL-6. *Difference from controls statistically significant, P!0$05.

and 7 dpi. No positive immunolabelling could be detected at 9 dpi, but it reappeared at 11 dpi, the counts of labelled cells being similar to those in the initial stage of infection. At 14 dpi, while the number of positive KCs remained unchanged, IMøs showed a large increase, diminishing thereafter, whereas labelled KCs rose slightly at 17 dpi (Fig. 3). TNF-a labelling was found only in a few KCs in infected animals from 4 dpi, (Fig. 6). It persisted throughout the remainder of the experiment, but only at 9 and 11 dpi were there statistically significant differences from the controls. IMøs were immunolabelled from 7 dpi, reaching significant values at 9 dpi (Fig. 3). Additionally, some circulating monocytes expressed TNF-a at 9, 11 and 17 dpi. TEM In the infected animals, the cytoplasm of some KCs and neutrophils showed intracytoplasmic vesicles containing electron-dense material, or membranous structures, or both. The membranous structures, which were usually close to the vesicle wall, were spherical (45–50 nm in diameter) and, electron-dense nucleoids (virus-like particles) were occasionally present. These intracytoplasmic vesicles and related structures were similar to those

described by others (Go´mez-Villamandos et al., 2000; Carrasco et al., 2001). At 2 and 4 dpi, KCs showed an increase in size and proliferation of primary and secondary lysosomes. Additionally, some KCs contained either phagocytized apoptotic cells, apoptotic bodies, or both, or platelets. In the lumina of sinusoids, platelets showing signs of activation were seen. These included enlargement and deformation of the cells, protrusion of one or more pseudopodia, decrease in granule numbers, and dilatation of the open canalicular system. Multiple platelets formed aggregations of variable numbers of cells, usually three to five. These clusters were seen as platelets retaining their morphological characteristics, or with membrane fusion (evident as highly electrondense lines), or as mosaic-like clusters consisting of fully degranulated platelets with marked fusion of membranes. From 7 dpi, some phagocytized apoptotic cells were observed in KCs, but a higher proportion of KCs showed an increase in size associated with hyperplastic Golgi complexes and expanded rough endoplasmic reticulum (Fig. 7), signs considered to indicate biosynthetic activation. In contrast to observations at previous timepoints, hepatic sinusoids appeared free of platelet clusters or signs of platelet activation.

28

A. Nu´n˜ez et al.

Fig. 4. Sinusoidal platelet aggregates showed positive reaction against Factor VIII-related antigen; 4 dpi. ABC. !630.

No signs of endothelial activation were observed throughout the infection.

Discussion The entry of CSF virus into the liver induced significant changes in the hepatic macrophages. These included phagocytic and biosynthetic activation, increases in the number and in the expression of different proinflammatory cytokines, and the appearance of mononuclear infiltrates. These were closely related to the presence of viral antigen in KCs and monocytic cells of sinusoids, and in the inflammatory infiltrates. The simultaneous expression of both IL-1a and TNF-a was detected at the peak of viral antigen production. Additionally, activation of platelet aggregates was observed in the early stages of the infection. CSF viral antigen was classically detected by immunofluorescence (Cheville and Mengeling, 1969; Mengeling and Packer, 1969) and by immunohistochemical labelling against viral Gp55 (E2)

Fig. 5. IL-1a inmmunolabelling of Kupffer cells (large cells located in the sinusoidal lumen with oval nuclei containing non-condensed chromatin) (arrows), interstitial macrophages (large arrowheads) and neutrophils (rounded cells containing segmented nuclei with condensed chromatin) (small arrowheads); 11 dpi. ABC. !400.

in experimental infections (Martı´n de las Mulas et al., 1997; Bautista et al., 2002). The present study, in contrast to the previous studies, detected an earlier infection in KCs and IMøs and a less consistent presence of viral antigen in biliary duct epithelial cells and endothelium, probably because of the lower virulence of the strain used. Although phagocytosis of cell debris from apoptotic infected cells may play a role in the infection of scavenging cells (Go´mez-Villamandos et al., 2001), viral replication in non-phagocytic cells supports the hypothesis that pestiviruses use a specific membrane receptor to enter target cells (Leyssen et al., 2000). In liver, CSF viral infection of myeloid cells occurs before that of non-myeloid cells (Martı´n de las Mulas et al., 1997), which suggests that infection of cell populations, apart from Møs, requires the expression of non-constitutive specific surface

Hepatic Cytokines in Classical Swine Fever

Fig. 6. Kupffer cells, with sinusoidal location and oval nuclei (arrows), and neutrophils (rounded cells with segmented nuclei) (arrowheads), expressing TNF-a; 11 dpi. ABC. !400.

receptors; these may be induced by cytokines released from activated Møs, as has been proposed for ASF, in which the main target cell is the M-Mø (Go´mez-Villamandos et al., 1999). In the same way, high concentrations of cytokines or virus, or both, would be required for non-myeloid cell infection; in the present study, such concentrations were demonstrated in the middle stage of the experiment. Evaluation of IL-1a, TNF-a and IL-6 expressed by the hepatic macrophages showed significant differences between infected and uninoculated animals. The detection of viral antigen in liver was accompanied by high expression of IL-1a and IL-6. The progression of the disease was paralleled by an increased expression of these cytokines by Møs, peaking at 9-11 dpi, coinciding with maximum viral replication and marked TNF-a detection, and subsequently falling. The increased production of these three cytokines coincided with signs of secretory activation of KCs in the ultrastructural study. The most significant cytokine changes were observed in KCs, except for IL-6, the highest

29

Fig. 7. Kupffer cell with ultrastuctural changes indicative of biosynthetic activation. Note the increase in cell size and the presence of numerous Golgi complexes (arrowheads); 9 dpi. TEM. !4600. Inset: proliferation and dilation of Golgi complexes indicative of secretory activity; 9 dpi. TEM. !17500.

expression of which occurred in IMøs. Similar observations were made by Carrasco et al. (2002) in relation to ASF, being associated with the variable response of Mø populations to stimuli and to their different anatomical locations (Ogle et al., 1994). The highest cytokine expression in liver was that of IL-1a. The enhanced production of this cytokine by KCs coincided with the onset of the clinical signs and haemorrhagic changes found in the infected animals, and its progressive increase paralleled the severity of the haemorrhages. Elevated production of IL-1 in CSF has been described in monocytes in vitro (Knoetig et al., 1999) and thymic macrophages in vivo (Sa´nchez-Cordo´n et al., 2002), but it could not be related to the concentration found in serum (Sa´nchez-Cordo´n et al., 2002). In the present study, however, IL-1 serum concentration was related to expression by KCs. Nevertheless, no dramatic increase in the number of leucocytes adhering to

30

A. Nu´n˜ez et al.

the hepatic endothelial lining was associated with enhanced IL-1 expression, although macrophage activation was observed. This suggests that the secretion of IL-1 by KCs does not reach high levels in the hepatic microvasculature, as low concentrations of IL-1 in the sinusoids do not significantly increase the adherence of leucocytes or the appearance of vascular changes but would be sufficient to induce macrophage activation in the liver (Nishida et al., 1999) and the appearance of systemic responses and clinical signs. In contrast to the in-vitro infection of monocytes with CSF virus (Knoetig et al., 1999), an increase in the expression of TNF-a and IL-6 was observed in the hepatic macrophages in vivo, as previously described in the thymus by Sa´nchez-Cordo´n et al. (2002). These differences may be due to the role of Møs in the more complex in-vivo interactions between cytokines and the immune system cells. TNF-a is responsible for the induction of apoptosis of lymphocytes in CSF (Sa´nchez-Cordo´n et al., 2002); however, the production of TNF-a in liver occurs in the middle stage of infection, after the main occurrence of lymphocyte death by apoptosis, but would contribute from this timepoint to the increased TNF-a concentration found in serum (Sa´nchez-Cordo´n et al., 2002). IL-6 expression was significantly increased from the start of the infection, inducing effects that act synergistically with the systemic responses to IL-1 (Hirano, 1998). Apart from its role in the regulation of the immune system, IL-6 stimulates the production of acute phase proteins by hepatocytes, the increase of haptoglobin and the decrease of albumin synthesis (Yoshioka et al., 1999). Thus, the progressive reduction of the albumin:globulin ratio that occurs in acute CSF (A. Nu´n ˜ ez, personal observation) may be induced in great part by the IL-6 expressed by hepatic macrophages. Moreover, the increase in IL-6 expression observed in the late stage of infection may be responsible for the hypergammaglobulinaemia observed in the chronic stage of CSF (Mengeling and Packer, 1969). In the present study, both KCs and IMøs doubled in number; in the late stage of infection, however, the KCs decreased while the IMøs increased. Such changes were described in lung, spleen and thymus of CSF-affected pigs, associated with activation and apoptosis of M-Møs induced by the virus (Carrasco et al., 2001; Go´mez-Villamandos et al., 2001; Sa´nchez-Co´rdo´n et al., 2002). As the cellular increase was not accompanied by increased mitosis, the elevated number of hepatic Møs was probably due to recruitment from the blood, related to

the higher expression of proinflammatory cytokines, especially IL-1. In this way, the chemotactic effect of the proinflammatory cytokines studied may have induced the mononuclear infiltrates found in liver. Although such infiltrates have been associated with viral replication (Liess, 1987), in our study they were mainly composed of activated uninfected Møs that expressed IL-1 and IL-6. It is suggested that they resulted from a mechanism employed by the virus to concentrate cells susceptible of being infected. The appearance of these infiltrates coincides with the detection of infiltrates in other organs, such as the perivascular cuffs of mononuclear cells found in the central nervous system (Ferna´ndez de Marco et al., 2003). Recent studies ruled out a direct viral effect as the cause of the thrombocytopenia in CSF, suggesting that activation and aggregation of platelets in the initial period of the infection and their phagocytosis by splenic macrophages and KCs were responsible (Bautista et al., 2002). Similar changes occurred in the hepatic thrombocytes in the present study, despite the comparatively low virulence of the strain employed. Additionally, no endothelial damage was observed in the infected animals and the location of aggregates was usually associated with KCs; this supports the hypothesis that platelet activation is due to soluble mediators, most likely secreted by M-Møs (Carrasco et al., 2001; Bautista et al., 2002). As the early increase in expression of IL-1 by KCs coincided with the onset of platelet aggregation, this cytokine may play a role in the origin of the thrombocytopenia, for IL-1 is known to induce thrombocyte activation (Reale et al., 1996; Bar et al., 1997). The lack of platelet plugs at later timepoints, despite a high expression of IL-1a, may have been due to the low number of circulating platelets, or to the role of chemical mediators such as platelet activation factor in the early phase of infection. This study revealed that liver has systemic implications in the pathogenesis of CSF, playing a role in the thrombocytopenia and in the increased expression of cytokines, primarily IL-1.

Acknowledgments The authors thank Dr D. Llanes and Dr A. Moreno from the Department of Genetics of the University of Co´rdoba, who kindly provided the antibody against SWC3, and M. Nu´n ˜ ez for assistance with English. This work was supported by grants from Plan Andaluz de Investigacio´n (AGR-137) and DGES (PB98-1033).

Hepatic Cytokines in Classical Swine Fever

References Bar, J., Zosmer, A., Hod, M., Elder, M. G. and Sullivan, M. H. (1997). The regulation of platelet aggregation in vitro by interleukin-1beta and tumor necrosis factor-alpha: changes in pregnancy and in preeclampsia. Thrombosis and Haemostasis, 78, 1255–1261. Bautista, M. J., Ruiz-Villamor, E., Salguero, F. J., Sa´nchezCordo´n, P. J., Carrasco, L. and Go´mez-Villamandos, J. C. (2002). Early platelet aggregation as a cause of thrombocytopenia in classical swine fever. Veterinary Pathology, 39, 84–91. Carrasco, L., Nu´n ˜ ez, A., Salguero, F. J., Dı´az San Segundo, F., Sa´nchez-Cordo´n, P., Go´mez-Villamandos, J. C. and Sierra, M. A. (2002). African swine fever: expression of interleukin-1 alpha and tumour necrosis factor-alpha by pulmonary intravascular macrophages. Journal of Comparative Pathology, 126, 194–201. Carrasco, L., Ruiz-Villamor, E., Go´mez-Villamandos, J. C., Salguero, F. J., Bautista, M. J., Macia´, M., Quezada, M. and Jover, A. (2001). Classical swine fever: morphological and morphometrical study of pulmonary intravascular macrophages. Journal of Comparative Pathology, 125, 1–7. Cheville, N. F. and Mengeling, W. L. (1969). The pathogenesis of chronic hog cholera (swine fever). Histologic, immunofluorescent and electron microscopic studies. Laboratory Investigation, 20, 261–274. Decker, K. (1990). Biologically active products of stimulated liver macrophages (Kupffer cells). European Journal of Biochemistry, 192, 245–261. Ferna´ndez de Marco, M., Garcı´a de Leaniz, I, Sa´nchezCordo´n, P.J., Gutierrez, J., Nu´n ˜ ez, A., and Go´mezVillamandos, J.C. (2003). Proinflammatory cytokine expression in central nervous system of pigs with classical swine fever. In: Proceedings of the 21st meeting of the European Society of Veterinary Pathology, p. 176. Go´mez-Villamandos, J. C., Carrasco, L., Bautista, M. J. and Sierra, M. A. (1999). Pathogenesis of African swine fever. The role of monokines. Recent Research and Development in Virology, 1, 7–17. Go´mez-Villamandos, J. C., Herva´s, J., Me´ndez, A., Carrasco, L., Villeda, C. J., Wilkinson, P. J. and Sierra, M. A. (1995). A pathological study of the perisinusoidal unit of the liver in acute African swine fever. Research in Veterinary Science, 59, 146–151. Go´mez-Villamandos, J. C., Ruiz-Villamor, E., Bautista, M. J., Quezada, M., Sa´nchez, C. P., Salguero, F. J. and Sierra, M. A. (2000). Pathogenesis of classical swine fever: renal haemorrhages and erythrodiapedesis. Journal of Comparative Pathology, 123, 47–54. Go´mez-Villamandos, J. C., Ruiz-Villamor, E., Bautista, M. J., Sa´nchez, C. P., Sa´nchez-Cordo´n, P. J., Salguero, F. J. and Jover, A. (2001). Morphological and immunohistochemical changes in splenic macrophages of pigs infected with classical swine fever. Journal of Comparative Pathology, 125, 98–109.

31

Hirano, T. (1998). Interleukin 6 and its receptor: ten years later. International Reviews of Immunology, 16, 249–284. Knoetig, S. M., Summerfield, A., Spagnuolo-Weaver, M. and McCullough, K. C. (1999). Immunopathogenesis of classical swine fever: role of monocytic cells. Immunology, 97, 359–366. Korn, G. and Zoeth, B. (1971). The reproduction of swine fever virus in a lymphocytic-phytocell-line and a line of monocytic cells. Zentralblatt fuer Bakteriologie A, 218, 407–416. Laskin, D. L. and Pendino, K. J. (1995). Macrophages and inflammatory mediators in tissue injury. Annual Reviews of Pharmacology and Toxicology, 35, 655–677. Leyssen, P., De Clercq, E. and Neyts, J. (2000). Perspectives for the treatment of infections with Flaviviridae. Clinical Microbiology Reviews, 13, 67–82. Liess, B. (1987). Pathogenesis and epidemiology of hog cholera. Annales de Recherches Ve´terinaires, 18, 139–145. Marianneau, P., Steffan, A. M., Royer, C., Drouet, M. T., Jaeck, D., Kirn, A. and Deubel, V. (1999). Infection of primary cultures of human Kupffer cells by Dengue virus: no viral progeny synthesis, but cytokine production is evident. Journal of Virology, 73, 5201–5206. Martı´n de las Mulas, J., Ruiz-Villamor, E., Donoso, S., Quezada, M., Lecocq, C. and Sierra, M. A. (1997). Immunohistochemical detection of hog cholera viral glycoprotein 55 in paraffin-embedded tissues. Journal of Veterinary Diagnostic Investigation, 9, 10–16. Mengeling, W. L. and Packer, R. A. (1969). Pathogenesis of chronic hog cholera: host response. American Journal of Veterinary Research, 30, 409–417. Narita, M., Kawashima, K. and Shimizu, M. (1996). Viral antigen and B and T lymphocytes in lymphoid tissues of gnotobiotic piglets infected with hog cholera virus. Journal of Comparative Pathology, 114, 257–263. Nishida, J., McDonnell, D., McCuskey, M. K., Ekataksin, W. and McCuskey, R. S. (1999). In vivo and electron microscopic observations of the responses of the hepatic sinusoid to interleukin-1. Bulletin of Tokyo Dental College, 40, 139–148. Ogle, C. K., Wu, J. Z., Mao, X., Szczur, K., Alexander, J. W. and Ogle, J. D. (1994). Heterogeneity of Kupffer cells and splenic, alveolar, and peritoneal macrophages for the production of TNF, IL-1, and IL-6. Inflammation, 18, 511–523. Reale, M., Barbacane, R. C., Frydas, S., Anogianakis, G., Trakatellis, A., Dimitriadou, D., Vacalis, D., Placido, F. C., De Fazio, P., Porreca, E., Di Febbo, C. and Conti, P. (1996). Human recombinant interleukin-1 beta induces thromboxane A2 release in polymorphonuclear leukocytes, macrophages and platelets: effect on IL-1 receptor antagonist. Molecular and Cellular Biochemistry, 159, 163–168. Salguero, F. J., Mekonnen, T., Ruiz-Villamor, E., Sa´nchezCordo´n, P. J. and Go´mez-Villamandos, J. C. (2001). Detection of monokines in paraffin-embedded tissues of pigs using polyclonal antibodies. Veterinary Research, 32, 601–609.

32

A. Nu´n˜ez et al.

Salguero, F. J., Ruiz-Villamor, E., Bautista, M. J., Sa´nchezCordo´n, P. J., Carrasco, L. and Go´mez-Villamandos, J. C. (2002). Changes in macrophages in spleen and lymph nodes during acute African swine fever: expression of cytokines. Veterinary Immunology and Immunopathology, 90, 11–22. Sa´nchez-Cordo´n, P. J., Romanini, S., Salguero, F. J., Nu´n ˜ ez, A., Bautista, M. J., Jover, A. and Go´mezVillamandos, J. C. (2002). Apoptosis of thymocytes related to cytokine expression in experimental classical swine fever. Journal of Comparative Pathology, 127, 239–248. Summerfield, A., Kno¨etig, S. M. and McCullough, K. C. (1998). Lymphocyte apoptosis during classical swine fever: implication of activation-induced cell death. Journal of Virology, 72, 1853–1861. Trautwein, G. (1988). Pathology and pathogenesis of the disease. In: Classical Swine Fever and Related Viral Diseases, B. Liess, Ed., Martinus Nijhoff Publishing, Boston, pp. 27–54. Van der Molen, E. J. and van Oirschot, J. T. (1981). Congenital persistent swine fever (hog cholera) I. Pathomorphological lesions in lymphoid tissues, kidney and adrenal. Journal of Veterinary Medicine B, 28, 89–101. Van Regenmortel, M.H.V., Fauquet, C.M., Bishop, D.H.L., Carstens, E.B., Estes, M.K., Lemon, S.M.,

Maniloff, J., Mayo, M.A., McGeoch, D.J., Pringle, C.R., and Wicknes, R.B. (2000). Virus Taxonomy, classification and nomenclature of viruses. In: Proceedings of Seventh Report of the International Committee on Taxonomy of Viruses, Academic Press, London, pp. 867-872. Weiss, E., Teredsal, A., Hoffmann, R. and HoffmannFezer, G. (1973). Volume distribution and ultrastructure of platelets in acute hog cholera. Thrombosis and Diathesis Haemorrhagica, 30, 371–380. Wensvoort, G., Terpstra, C., De Kluijver, E. P., Kragten, C. and Warnaar, J. C. (1989). Antigenic differentiation of pestivirus strains with monoclonal antibodies against hog cholera virus. Veterinary Microbiology, 21, 9–20. Yoshioka, M., Mori, Y., Miyazaki, S., Miyamoto, T., Yokomizo, Y. and Nakajima, Y. (1999). Biological functions of recombinant bovine interleukin 6 expressed in a baculovirus system. Cytokine, 11, 863–868.



Received; July 14th; 2004 Accepted; January 10th; 2005