High-Dose Borna Disease Virus Infection Induces a Nucleoprotein-Specific Cytotoxic T-Lymphocyte Response and Prevention of Immunopathology

Abstract

Experimental Borna disease virus (BDV) infection of rats and natural infection of horses and sheep leads to severe central nervous system disease based on immunopathological pathways. The virus replicates slowly, and the cellular immune response results in immunopathology. CD8⁺ T cells exert effector cell functions, and their activity results in the destruction of virus-infected cells. Previously, Oldach and colleagues (D. Oldach, M. C. Zink, J. M. Pyper, S. Herzog, R. Rott, O. Narayan, and J. E. Clements, Virology 206:426–434, 1995) have reported protection against Borna disease after inoculation of high-dose cell-adapted BDV. Here we show that the outcome of the infection, i.e., immunopathology versus protection, is simply dependent on the amount of virus used for infection. High-dose BDV (10⁶ FFU) triggers an early virus-specific reaction of the immune system, as demonstrated by strong cellular and humoral responses. In particular, the early presence and function of nucleoprotein-specific CD8⁺ T cells could be demonstrated in the brain. We present evidence that in a noncytolytic and usually persistent virus infection, high-dose input virus mediates early control of the pathogen due to an efficient induction of an antiviral immune mechanism. From these data, we conclude that immune reactivity, in particular the cytotoxic T-cell response, determines whether the virus is controlled with prevention of the ensuing immunopathological disease or whether a persistent infection is established.

CD8⁺ T cells are important in the control of many intracellular pathogens, where they function as primary effector cells. Whereas an early and efficient induction of CD8⁺ T cells is crucial after infection with highly cytolytic viruses to eliminate the agent before viral replication produces viral progeny, the role of CD8⁺ T cells in infections with noncytolytic viruses appears to be more complicated. Noncytolytic viruses are mostly defined as such because they do not cause overt tissue destruction in vitro. However, in vivo this situation might change considerably if this type of virus encounters an intact immune system and induces an antiviral immune response. Although the virus is mainly innocuous, the induced immune response produces immunopathological pathways, often resulting in severe disease. During the initial encounter with a virus, CD8⁺ T cells bearing T-cell receptors specific for the given antigen are selected to undergo clonal expansion. In the case of rapidly replicating viruses, it can be assumed that antigen is produced in an amount that triggers a vigorous immune response that either suffices to eliminate virus-infected cells or is not efficient enough to control virus infection and results in disease and/or early death. In viral infections in which only comparably low doses of infectious virus are passed to a new host or in infections with slowly replicating, noncytolytic and persistent virus, the concept of early action of the immune system might not be valid simply due to an insufficiently strong trigger for the immune system. In this case, only the increasing number of infected cells over time provides a stimulus to the immune system; however, this stimulus is too late to eliminate the virus early after infection and/or to prevent persistence.

Borna disease virus (BDV) is an example of a noncytolytic persistent virus. In recent years, this viral infection of the central nervous system (CNS) has been diagnosed in a wide variety of animals including cattle, cats, dogs, and birds (4, 6, 15, 16, 38). Furthermore, virus, nucleic acid, and antibodies have been detected in the blood of patients with psychiatric diseases (1, 5, 13, 18, 26, 31, 37; N. Nowotny and J. Kolodziejek, Letter, Lancet 355:1462–1463, 2000). However, so far no direct correlation between BDV as the causative agent and any of these human disorders has been demonstrated. BDV causes a persistent infection of the CNS and induces Borna disease (BD), an immune-mediated encephalomyelitis originally described in horses and sheep (14, 19, 30). The infiltrating immune cells have been characterized as CD4⁺ CD8⁺ T cells and macrophages (2, 8, 29). CD8⁺ T cells represent the effector cell population, exhibiting antigen specificity for the nucleoprotein p40, specifically for the peptide ASYAQMTTY, in the Lewis rat (23–25, 27, 32). No evidence has been presented that antibodies might contribute to neuropathology, although neutralizing antibodies apparently control virus tropism and can prevent the spread of virus from peripheral infection sites to the CNS (9, 11, 34).

After experimental BDV infection of rats, protection against the immune-mediated brain disease has been achieved by adoptive transfer of CD4⁺ T-cell lines, resulting in the loss of virus from the CNS (20, 24, 28). The underlying mechanisms responsible for virus elimination have been extensively investigated, and strong evidence for a role of CD8⁺ effector cells induced by virus-specific CD4⁺ T-cell lines has been provided (20). In addition to this T-cell-mediated protection, Oldach et al. have reported protection against disease after infection with high-dose (HD) cell-attenuated BDV (21); however, the mechanism of virus control and protection from disease has not been investigated, and therefore this interesting phenomenon remains to be elucidated. To determine which effector mechanism might be responsible for the elimination of BDV from the host after infection with HD virus, we used HD virus obtained from two different cell types and determined the immunological nature of this phenomenon by demonstrating enhanced kinetics of an anti-BDV CD8⁺ T-cell response.

MATERIALS AND METHODS

Viruses and infection. (i) BDV-BR.

The fourth rat passage of BDV originally obtained from R. Rott and S. Herzog was used for the infection experiments (19).

(ii) BDV-CRL.

CRL1405 cells (guinea pig cell line) were infected with the fourth rat passage of BDV-BR. Then the virus was isolated from these cells after two to eight passages by sonication.

(iii) BDV-MDCK.

Persistently infected BDV-MDCK cells were lysed by sonication (12). Five-week-old rats were infected intracerebrally (i.c.) in the left brain hemisphere with 0.05 ml of the different virus isolates corresponding to 10⁶, 10⁴, or 10² focus-forming units (FFU). The vaccinia virus (VV)-BDV recombinants (VV-p40, VV-p24, VV-gp18, and VV-gp94) were created by J. C. de la Torre, Scripps Research Institute, La Jolla, Calif.

Experimental animals and immunosuppression.

Male and female Lewis rats were bred in the animal-breeding facilities at the Bundesforschungsanstalt für Viruskrankheiten, Tübingen, Germany. For immunosuppression, rats thymectomized as adults were treated with 2 mg of purified mouse monoclonal antibody directed against rat CD8⁺ T cells (OX 8) 1 day before and 1 day after infection. Furthermore, the rats were thymectomized as newborns within 24 h of birth. Successful T-cell depletion was controlled by fluorescence-activated cell sorter (FACS) analysis.

Clinical evaluation.

All experimental animals were examined daily and weighed. Disease symptoms were scored using a scale from 0 to 3 based on the general state of health and the appearance of neurological signs (score 1, slight incoordination and vigilance; score 2, distinct ataxia or slight paresis; score 3, marked pareses or paralyses).

Infectivity assay and viral antigen detection.

Virus infectivity of brain homogenates from BDV-infected rats was determined on CRL1405 cells. Titer determinations were carried out in flat-bottom 96-well microtiter plates. The CRL1405 cells were cultured for 7 days in the presence of brain homogenates from infected rats. Thereafter the cells were fixed with 4% formaldehyde–phosphate-buffered saline (PBS) and treated with 1% Triton X-100–PBS, and viral antigen was demonstrated in an immunocytochemical reaction using anti-BDV-specific mouse monoclonal antibodies. Nonspecific binding of immunological reagents was blocked by incubation of plates with 10% fetal calf serum–PBS. The reaction of monoclonal antibodies with cells was detected by a secondary anti-mouse biotin-labeled antibody (Dianova, Hamburg, Germany) and by a streptavidin-peroxidase conjugate (Dianova). The reaction was visualized with ortho-phenylendiamine and H₂O₂ (Sigma, Munich, Germany). Additionally, tissue homogenates were used as antigens in Western blot analysis, and the presence of virus-specific antigen was detected by BDV-specific monoclonal antibodies.

Detection of BDV-specific antibodies.

All rat sera were tested in a solid-phase enzyme-linked immunosorbent assay (ELISA) using a 1:1,000 dilution of a brain homogenate from BDV-infected rats as coating antigen and by Western blot analysis with a 10% brain homogenate from BDV-infected rats. The tests were performed as described previously (20).

Histology and immunohistochemistry.

At different time points after infection, brain samples were obtained and immediately either frozen in isopentane at −150°C or fixed in buffered formalin. Cryostat sections were fixed in isopropanol. All tissue sections were stained with hematoxylin-eosin. Encephalitic infiltrates were scored on an arbitrary scale ranging from 0 to 3 based on the number of infiltrates per section and the number of cell layers in each infiltrate (score 1, up to 5 small infiltrates per section; score 2, more than 5 small infiltrates per section or more than 3 infiltrates with multiple layers; score 3, more than 10 small infiltrates per section or more than 5 infiltrates with multiple layers). Immunohistochemistry was carried out on cryostat sections for the presence of BDV-specific antigen, using an anti-nucleoprotein-specific monoclonal antibody (38/17C1) (36).

In situ hybridization.

Digoxigenin-labeled RNA probes complementary to BDV nucleoprotein p40, phosphoprotein p24, or matrix protein gp18 mRNAs were used. Brains from experimental animals were frozen in isopentane to −150°C. Sections (5 μm) were mounted on slides and fixed in 4% formaldehyde–PBS. After treatment with 0.1 N HCl and acetic acid, hybridization was carried out overnight at 65°C with 20 ng of probe per slide. The slides were washed with 5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) followed by 2× SSC, both at 56°C. They were then incubated with an alkaline phosphatase-labeled anti-digoxigenin antibody and then placed overnight in a 5-bromo-4-chloro-3-indolylphosphate–Nitro Blue Tetrazolium (BCIP/NBT) solution.

Isolation of effector cells.

Lymphocytes from the brain were isolated by a method described previously (23). Briefly, animals were anesthetized and perfused with balanced salt solution. The brain was homogenized carefully through a stainless steel mesh and collected in balanced salt solution containing 0.05% collagenase D, 0.1 μg of trypsin inhibitor per ml, 10 μg of DNase I per ml, and 10 mM HEPES. The cell suspension was stirred at room temperature for 1 h and allowed to settle for 25 min. The supernatant was pelleted at 200 × g for 10 min and resuspended in 10 ml of Ca- and Mg-free PBS. Then 5 ml of the suspension was layered on top of 10 ml of a modified RPMI-Ficoll gradient and centrifuged at 500 × g for 30 min. The pellet containing the lymphocytes was resuspended in Iscove modified Dulbecco medium with 2% fetal calf serum and cultivated overnight at 37°C. Finally, the cells were counted for further use as effectors in cytotoxicity assays.

⁵¹Cr release assay.

Lymphocytes isolated from the brains of BDV-infected rats were tested at different time points after the infection in a cytotoxicity assay with BDV-infected or recombinant VV-infected target cells. The test was performed as described previously (27). The synthetic peptide ASYAQMTTY was dissolved in dimethyl sulfoxide, and target cells were labeled as described previously (25).

RESULTS

HD virus from different sources prevents BD.

HD-attenuated BDV from MDCK cells protects against Borna disease (21). To define whether this is due to virus properties after passages in tissue culture or to a protective antiviral immune response to BDV, we investigated the pathogenic potential of different doses of BDV after i.c. inoculation in rats. BDV was obtained either from rat brain (BDV-BR; 10⁴ and 10² FFU) or from MDCK (BDV-MDCK; 10⁶, 10⁴, and 10² FFU) or CRL1405 (BDV-CRL; 10⁶, 10⁴, and 10² FFU) cells. An infectious dose of 10⁶ BDV-BR was not applicable simply due to insufficiently high titers in the brain.

After infection with 10² FFU, disease symptoms were seen around day 24 (BDV-BR or BDV-CRL) or day 37 (BDV-MDCK) (Fig. 1A). At later time points, e.g., day 60 after infection, no difference between the severity of clinical symptoms after infection with BDV-BR, BDV-CRL, or BDV-MDCK was found.

FIG. 1.

Appearance of BDV-specific clinical symptoms. The health status of six rats was monitored daily, and disease symptoms were scored as described in Materials and Methods. (A and B) The rats were infected with 10² FFU (A) or 10⁴ FFU (B) of either BDV-BR (■), BDV-CRL (□), or BDV-MDCK (■). (C) For infection with 10⁶ FFU, only BDV-CRL and BDV-MDCK were used. After infection with 10⁴ FFU (B) of BDV-MDCK, two animals developed severe BD (■) while four animals showed no clinical symptoms ().

When 10⁴ FFU was used for infection, the first clinical symptoms were seen 12 to 15 days after infection with BDV-BR or BDV-CRL whereas no symptoms were observed within 40 days after infection when BDV-MDCK was used (Fig. 1B). However, thereafter two BDV-MDCK rats came down with severe BD, whereas four rats infected with the same dose had no disease when the observation period was ended on day 60 (Fig. 1B).

Rats infected with HD (10⁶ FFU) BDV-CRL displayed only slight nonspecific and transient symptoms between days 9 and 12; those infected with HD BDV-MDCK showed no symptoms (Fig. 1C).

In addition to the daily examination, the body weight was determined as a reflection of the overall health status. All infected animals gained weight within the first 12 to 15 days after BDV infection. However, at this time point and in correlation with the appearance of clinical symptoms, rats infected with 10² and 10⁴ FFU of BDV-BR or BDV-CRL lost about 40% of their body weight until day 35. In contrast, no weight loss was found in rats (n = 4) infected with 10⁴ FFU of BDV-MDCK until day 60 (data not shown).

Early presence of BDV-specific antibodies after infection with 10⁶ FFU of BDV.

The BDV-specific antibody response was monitored every third day. As described previously (19), the first nucleoprotein (p40)-specific antibodies were found around day 15 postinfection (p.i.) and phosphoprotein (p24)-specific antibodies were found around day 18 p.i. in rats infected with 10⁴ FFU of BDV-BR or BDV-CRL (Fig. 2B). Infection with 10² FFU of BDV-BR or BDV-CRL resulted in a delayed production of BDV-specific antibodies (Fig. 2A).

FIG. 2.

BDV-specific antibodies after infection with BDV from different sources. (A) Rats were infected with the standard infectious dose of 10⁴ FFU of either BDV-BR (■), BDV-CRL (□), or BDV-MDCK (■). (B and C) Furthermore, rats were infected with 10² FFU (B) of either BDV-BR, BDV-CRL, or BDV-MDCK or with 10⁶ FFU (C) of BDV-CRL or BDV-MDCK. Sera were taken as indicated, and the BDV-specific antibody titer was measured by ELISA as −log₂ starting with a 1:40 dilution. The standard error of the mean of all titers was <0.20 (n = 6).

When rats were infected with either 10² or 10⁴ FFU of BDV-MDCK, antibody production was significantly delayed (Fig. 2; Table 1). Most interestingly, at a dose of 10² FFU, no phosphoprotein p24-specific antibodies were detected in sera, irrespective of the virus used for infection (Table 1). Remarkably, HD BDV-MDCK and HD BDV-CRL triggered nucleoprotein p40-specific antibodies and phosphoprotein p24-specific antibodies by day 9 and day 15 p.i., respectively (Table 1; Fig. 2C), and BDV-MDCK-infected rats initially developed higher titers of BDV-specific antibodies than did BDV-CRL-infected rats (Fig. 2C).

TABLE 1.

BDV-specific antibodies after infection with different BDV isolates

Virus	Infection (log10 FFU)	Appearance of nucleoprotein (p40)- or phosphoprotein (p24)-specific antibodies on day (p.i.)a:
		6	9	12	15	18	21	24	27	31	37
BDV-BR
	4				p40	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24
	2					p40	p40	p40	p40	p40	p40
BDV-CRL
	6		p40	p40	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24
	4					p40	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24
	2						p40	p40	p40	p40	p40
BDV-MDCK
	6		p40	p40	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24	p40/p24
	4									p40	p40
	2							p40	p40	p40	p40

^aAnalysis was performed by Western blotting. 

Detection of antigen, nucleic acid, and infectious virus.

After infection with 10² or 10⁴ FFU of BDV-BR or BDV-CRL, Western blot analysis revealed the presence of the nucleoprotein in brain homogenates by day 14 or 9 p.i. and the phosphoprotein by day 16 or 12 p.i., respectively (data not shown). Infectious virus could be isolated from the brains of rats infected with 10⁴ FFU with a titer of 5.7 log₁₀ units around day 10 p.i., reaching maximal titers of 6.3 log₁₀ units on day 21. Similar titers of infectious virus were found when 10² FFU was used for infection; however, there was a delay of 4 to 5 days. BDV-specific nucleic acid and BDV-specific nucleoprotein could be demonstrated in the cortex and hippocampus of infected rats using in situ hybridization and immunohistochemistry (data not shown).

Infection with 10⁴ FFU of BDV-MDCK resulted in the presence of viral antigen and infectious virus on day 18 p.i.; on day 37 p.i., however, little antigen and no infectious virus were found. In situ hybridization revealed the presence of mRNA for the nucleoprotein p40, the phosphoprotein p24, and the matrix protein gp18 in the hippocampus but not in the neocortex. After infection with 10² FFU of BDV-MDCK, no viral antigen or infectious virus was detectable before day 37 p.i. Thereafter, BDV-specific antigen could be detected by Western blotting and an infectious titer of 6.8 log₁₀ units was found in the brain (day 37). In situ hybridization confirmed these results by detecting mRNA for the BDV-specific proteins p40, p24, and gp18 (data not shown).

When infection was carried out with HD BDV-MDCK or BDV-CRL, either no or only traces of nucleic acid and antigen and no infectious virus were detectable on day 6 p.i.. By day 12, viral antigen, but no infectious virus, was found in HD BDV-MDCK-infected animals. By day 18, only single BDV-infected cells could be found by immunohistochemistry and in situ hybridization in rats infected with HD BDV-MDCK (Fig. 3A and B), whereas neither viral antigens nor infectious virus, with one exception, could be detected by Western blot analysis and virus titer determination (Table 2). In contrast, significant amounts of nucleic acid, antigen, and infectious BDV were detectable in the brains of HD BDV-CRL-infected rats (Table 2). By day 28, no nucleic acid or antigen was found in HD BDV-MDCK-infected rats, whereas nucleic acid and antigen were found in the cortex in HD BDV-CRL-infected rats, although at reduced levels (Fig. 3C and D). In HD BDV-CRL-infected animals, antigen could be detected in the hippocampus by immunohistochemistry and nucleic acid could be detected by in situ hybridization at levels comparable to those in control animals (Fig. 3E and F). At later time points (day 37), the viral titer was significantly reduced in the brains of HD BDV-CRL-infected rats and no antigen or infectious virus was found in the brains of HD BDV-MDCK-infected rats (Table 2).

FIG. 3.

Presence of BDV in rat brains. Immunohistochemistry (A, C, and E) and in situ hybridization (B, D, and F) reveal that 18 days after infection with 10⁶ FFU of BDV-MDCK (A and B), only a few BDV-infected cells are found, whereas by day 28, the reduction of BDV in the brains of rats infected with 10⁶ FFU of BDV-CRL (C and D) is visible compared to controls (E and F). Magnification, ×48.5.

TABLE 2.

Viral antigen and infectious virus after HD BDV infection

Day (p.i.)	Virus	na	Antigen present		BDV detection (nucleic acid in situ hybrid)	Infectious virus titer (log10)b
			Western blotting	Histology
6	BDV-CRL	2	−/−	s.c.c/−	s.c./−	<1.8/<1.8
	BDV-MDCK	6	−/−/−/−/−/−	−/−/−/−/−/−	−/−/−/−/−/−	<1.8/<1.8/<1.8/<1.8/<1.8/<1.8
12	BDV-CRL	2	+/+	+/+	+/+	5.0/5.3
	BDV-MDCK	6	−/−/−/−/−/−	s.c./−/s.c.d	−/−/−/−/−/−	<1.8/<1.8/<1.8/<1.8/<1.8/<1.8
18	BDV-CRL	2	+/+	s.c./NDe	+/+	3.7/4.2
	BDV-MDCK	6	−/+/−/−/−/−	−/1/s.c.d	−/+/−/−/−/−	<1.8/3.4/<1.8/<1.8/<1.8/<1.8
28	BDV-CRL	2	+/+	s.c./s.c.	+/+	4.2/4.0
37	BDV-CRL	2	−/+	ND	+/+	<1.8/3.8
	BDV-MDCK	2	−/−	ND	−/−	<1.8/< 1.8

^aNumber of rats (experiments were repeated at least three times with the same results). 

^bDetection limit, <1.8. 

^cs.c., single cell. 

^dOnly three brains were tested in immunohistochemistry. 

^eND, not done. 

Outcome of HD BDV infection in immunosuppressed rats.

To investigate whether the immune system is responsible for the absence of BD symptoms in rats infected with HD BDV-CRL or HD BDV-MDCK, immunosuppressed animals were infected. For immunosuppression, animals were thymectomized as adults and then treated with anti-CD8-specific antibodies to deplete CD8⁺ T cells. Other rats were thymectomized as newborns and infected as adults. Successful thymectomy and T-cell depletion were confirmed by FACS analysis (data not shown). As shown in Table 3, no clinical disease symptoms or encephalomyelitis could be detected in thymectomized rats depleted of CD8⁺ T cells and infected with HD BDV-CRL or HD BDV-MDCK, but viral antigen and infectious virus were present in the brain. Comparably, rats thymectomized as newborns and infected with the same viruses did not show any clinical signs and both BDV-specific antigen and infectious virus were detectable in the brain. However, in contrast to animals thymectomized as adults and infected with HD BDV-MDCK, no antiviral antibodies were found in those thymectomized as newborns. Interestingly, in those thymectomized as newborns, infectious titers were higher in the brains after infection with HD BDV-CRL than in those after infection with HD BDV-MDCK.

TABLE 3.

Infection of immunosuppressed Lewis rats with 10⁶ FFU of BDV-CRL or BDV-MDCK

Virus	Treatmenta	Day p.i. of sacrifice	BDV in brain		BDV antibodies present	Encephalitis present
			Antigen	Infectious virus (log10)
BDV-CRL	None	21	+/+/+	5.5/4.9/3.7	+/+/+	+b/+b/NDc
BDV-MDCK	None	21	−/−/−	3.4/<1.8/<1.8	+/+/+	+b/−/−
BDV-CRL	Newborn/TX	21	+/+/+	8.6/8.7/7.7	−/−/−	−/−/−
BDV-MDCK	Newborn/TX	18	+/+/+	6.2/6.2/6.8	−/−/−	−/−/−
BDV-CRL	Adult/TX	21	+/+/+	6.7/6.6/6.9	−/−/−	−/−/−
	Anti-CD8
BDV-MDCK	Adult/TX	18	+/+/+	6.0/6.5/6.3	+/+/−	−/−/−
	Anti-CD8

^aEfficency of treatment was controlled by FACS analysis. TX, thymectomized. 

^bOnly in the hippocampus. 

^cND, not done. 

Presence of BDV-specific cytotoxic T cells.

In anti-CD8-treated rats infected with HD BDV, high titers of virus were found in the brains. Therefore, the cytolytic activity of CD8⁺ T cells isolated from the brains of rats infected with HD BDV-MDCK or HD BDV-CRL was compared to the activity of cells isolated from the brains of rats infected with 10⁴ FFU of BDV-BR on day 9 p.i. As target cells, syngeneic fibroblasts from Lewis rats infected with BDV were used. As shown in Fig. 4A, lymphocytes isolated from the brains of rats infected with BDV-BR showed no cytolytic activity, correlating with previous results showing that the cytolytic activity of CD8⁺ T cells isolated from the brains of BDV-BR-infected rats starts around day 18 (20). In contrast, in rats infected with HD BDV-MDCK or HD BDV-CRL cytolysis of target cells was found as early as day 9 after infection (Fig. 4A).

FIG. 4.

Cytotoxicity assay with lymphocytes isolated from the brains of rats infected with 10⁶ FFU of BDV-MDCK or BDV-CRL and from rats infected with 10⁴ FFU of BDV-BR. Lymphocytes were taken 9 days after infection. As target cells, BDV-infected syngeneic Lewis fibroblasts were used. (A) No lysis was found on noninfected target cells. (B and C) Lymphocytes isolated from the brains of rats infected with HD BDV-MDCK (B) or HD BDV-CRL (C) were cultivated with Lewis cells infected with recombinant VV expressing different BDV-specific proteins as indicated. (D) Lewis cells pulsed with the nucleoprotein peptide ASYAQMTTY were used as target cells for lymphocytes, isolated from the brains of rats infected with HD BDV-MDCK.

Furthermore, we questioned which BDV-specific protein represents the target of the CD8⁺ T-cell response. To investigate this, Lewis cells were infected with wild-type VV (VV-WT) or with recombinant VV expressing either the nucleoprotein (VV-p40), the phosphoprotein (VV-p24), the matrix protein (VV-gp18), or the glycoprotein (VV-gp94) of BDV. These target cells were either incubated with effector cells isolated from the brains of BDV-MDCK-infected (Fig. 4B) or BDV-CRL-infected (Fig. 4C) rats. As shown in Fig. 4B and C, only target cells infected with recombinant VV expressing the BDV nucleoprotein (VV-p40) were recognized by CD8⁺ T cells. In addition, Lewis cells pulsed with BDV nucleoprotein peptide ASYAQMTTY were recognized by effector cells isolated from the brains of BDV-MDCK-infected rats (Fig. 4D).

DISCUSSION

In a previous publication, Oldach et al. (21) reported on virus control and protection against disease after infection of rats with HD BDV from tissue culture. In the present study, we have identified the underlying mechanisms as being immunological in nature, demonstrating that a persistent virus infection can be abrogated after infection with a strong trigger for an efficient cellular antiviral immune response by high-input virus titers. As well as an early and strong cell-mediated immune reaction, an early humoral immune response was observed. In this study we used virus obtained from different cell types (BDV-CRL and BDV-MDCK) or from the brains of BDV-infected rats (BDV-BR) to assess whether characteristics of particular virus isolates, e.g., adaptation to tissue culture, contribute to control of the virus from the host or whether this effect is simply triggered by the virus and, in fact, is dependent on the efficiency of the antiviral immune response.

Infection with 10² FFU of BDV-CRL, BDV-BR, or BDV-MDCK in general resulted in late onset of disease. This finding indicates that infection even with low doses of BDV from different preparations causes disease with a comparable clinical picture.

Similarly, using the standard dose (10⁴ FFU) for infection, we were able to demonstrate that BDV-CRL infection results in BD in a comparable time range, with comparable symptoms and severity of disease, as well as onset, quality, and antibody titers in BDV-CRL- and BDV-BR-infected rats. This indicates that BDV obtained from CRL1405 cells (BDV-CRL) and BDV harbored in the brain (BDV-BR) have very similar biological properties as far as the above-mentioned parameters are concerned. In contrast, BDV obtained from persistently infected MDCK cells (BDV-MDCK) and used for infection at 10⁴ FFU apparently differed from BDV-CRL and BDV-BR insofar as only one-third of the animals developed disease within the observation period while the others stayed healthy. These results show that 10² FFU of BDV-MDCK induces disease while infection with 10⁴ FFU apparently represents a breaking point, where some animals develop disease and others control the virus and do not develop BD. This finding also indicates differences among BDV isolates that possibly depend on their passage history in vitro; BDV has been cultured in MDCK cells for more than 20 years (12), whereas it has been passaged in CRL1405 cells only for a short time (9). Infection of rats with HD BDV-CRL or BDV-MDCK (10⁶ FFU) resulted in an early induction of the humoral immune response. The detection of viral antigen and nucleic acid in the hippocampus of BDV-MDCK-infected rats as well as in the cortex and the hippocampus of BDV-CRL-infected rats points to a productive viral infection. Since in both BDV-MDCK- and BDV-CRL-infected rats only traces of virus were detectable, this demonstrates that the infection is controlled, although gradually differently. This is further demonstrated by the detection of infectious virus in BDV-CRL-infected rats but not in BDV-MDCK-infected rats. Furthermore, the quantity of antigen detected in the brains of BDV-MDCK-infected rats was below the detection limit of immunoblotting but antigen was easily detectable in BDV-CRL-infected rats. Encephalitic reactions were seen in the hippocampus of all BDV-CRL-infected rats but only one BDV-MDCK-infected rat. Finally, we injected rats with a low dose of BDV-CRL i.c. and at the same time with HD BDV-CRL subcutaneously and/or intravenously. Unfortunately, rats that were only infected with HD BDV-CRL subcutaneously and/or intravenously developed disease (data not shown). Therefore, this experiment was not suitable to investigate the balance between the local infection in the brain and the induction of a systemic immune response. Control of a virus infection usually is due to the activity of the humoral and/or cellular arm of the immune system. Only recently has it been demonstrated that neutralizing glycoprotein-specific antibodies can have a prophylactic effect in experimental BD (9). However, apart from the early induction of nonneutralizing antibodies against the nucleoprotein and the phosphoprotein, no neutralizing antibodies were found in sera from BDV-CRL- or BDV-MDCK-infected rats. Therefore, and in agreement with previous findings which underline the role of the cellular immune system in both immunopathology and immunoprotection in the experimental model of BD in rats (reviewed in references 3, 33, and 35), we investigated the cellular immune response in more detail.

In BD, T cells represent the major pathway of immunopathology; CD4⁺ and, particularly, CD8⁺ T cells are involved in severe encephalitis and the ensuing degenerative encephalopathy (2, 10, 23, 24, 29). On the other hand, immunoprotection is mediated by T cells (28), where CD8⁺ cytotoxic T cells were again demonstrated as the relevant effector cells (20). Therefore, on the basis of having demonstrated that HD BDV-MDCK-infected rats do not show disease and do not have neutralizing antibodies, we scrutinized the impact of the cellular immune response in the HD phenomenon. Since CD8⁺ T cells are the most likely effector cells after BDV infection, we concentrated on their role by using different strategies. Rats thymectomized at birth and tested for the absence of T cells in the peripheral blood by flow cytometry did not show disease symptoms or antibody synthesis, whether infected with HD BDV-MDCK or HD BDV-CRL. However, whereas untreated BDV-MDCK-infected rats showed only a very limited presence of viral antigen in the brain and had no detectable amounts of virus, thymectomized rats had high virus titers in the brain. This finding was even more pronounced in immunocompromised HD BDV-CRL-infected rats, which had an extraordinary high virus titer in their brains.

To further define the mechanism of the cellular immune response that is operative in controlling the virus, adult thymectomized rats were infected with virus after they had been treated with an antibody against CD8⁺ T cells. Neither HD BDV-MDCK- nor HD BDV-CRL-infected rats showed disease or encephalitis, but, again, high virus titers were found in their brains. This result clearly demonstrated that the presence of CD8⁺ T cells is indispensable and sufficient for virus control in HD BDV-infected rats.

The result of cytotoxicity assays convincingly showed that HD BDV-CRL induces a BDV-specific CD8⁺ T-cell response as early as day 9 p.i. and that BDV-MDCK triggers a substantial cytotoxic T-cell response, whereas BDV-BR does not cause the lysis of target cells at this early time point. Moreover, we directly demonstrated the presence of specific cytotoxic T cells directed against the nucleoprotein by the use of recombinant VV (27). Finally, a naturally processed Lewis rat major histocompatibility complex class I-associated peptide, ASYAQMTTY, of the BDV nucleoprotein had been identified previously as the relevant target structure (25). These present and earlier results show that a fast and strong immune response directed against the peptide ASYAQMTTY is necessary for the control of BDV and prevention of disease. These data also show that HD BDV does not cause cytotoxic T-lymphocyte exhaustion as has been demonstrated in lymphocytic choriomeningitis virus infection with the consequence of a persistent infection (17). The role of CD8⁺ T-cell-derived cytokines has not been addressed in this paper in association with protection from disease. These studies are extremely difficult to perform since only very few, if any, of the cytokines produced during an antiviral immune response are uniquely produced by CD8⁺ T cells.

Our data support previous results observed after infection with other noncytopathic viruses. In lymphocytic choriomeningitis virus infection of mice and hepatitis B virus infection in humans, damage of virus-infected CNS or liver cells is caused by cytotoxic T cells (7, 22, 40). If the host immune response is efficient and quick, the replication and, more important, the spread of the virus can be controlled very early after infection. Therefore, tissue damage is limited, virus is eliminated, and functions are restored within a few weeks after infection (22, 40). If there is no immune response, the virus will spread. Since the virus is noncytopathic, the infection does not cause any damage and the host will become a virus carrier. If there is a low and slow T-cell response, the virus can infect many cells, and as a consequence of the delayed immune response, severe tissue damage occurs due to immunopathology (reviewed in reference 39). This finding is comparable to the situation after BDV infection of rats. If there is no T-cell response, HD BDV-infected rats develop a carrier status. If the T-cell response is excellent, CD8⁺ T cells prevent a disseminated viral encephalitis and can even control the virus and prevent immunopathology. If the immune response is delayed after BDV infection, the animals develop encephalitis and severe brain cell destruction.

The present work clearly demonstrates that BDV infection does not necessarily result in immunopathology and underlines the role of CD8⁺ T cells in mediating and preventing disease. Nevertheless, even though it seems unlikely that antibodies, e.g., neutralizing antibodies directed against the glycoprotein (9), play a major role in mediating and preventing disease, they cannot be excluded because they might function at a level that is below detection.

Our data support the hypothesis that CD8⁺ T-cell-mediated lysis mechanisms are required for immunity to nonlytic viruses. It is not known how deeply gamma interferon, as a product of CD4⁺ or CD8⁺ T cells, is involved in the clearance of persistent BDV infection of the brain since cytokine-deficient rats are not available. Therefore, at present the only valid explanation for the control of BDV infection in the brain involves the presence and action of CD8⁺ T cells, although the question cannot be answered whether these cells exert a noncytolytic mechanism that controls the viral genome with only little tissue damage. On the other hand, a very early and potent cytolytic T-cell response that is capable of controlling virus-infected cells in the brain before the virus disseminates would be suited to limit cellular destruction to only very few cells, resulting in no or only transient and slight disturbances of organ functions.