A Central Role for CD4⁺ T Cells and RANTES in Virus-Induced Central Nervous System Inflammation and Demyelination

Abstract

Infection of C57BL/6 mice with mouse hepatitis virus (MHV) results in a demyelinating encephalomyelitis characterized by mononuclear cell infiltration and white matter destruction similar to the pathology of the human demyelinating disease multiple sclerosis. The contributions of CD4⁺ and CD8⁺ T cells in the pathogenesis of the disease were investigated. Significantly less severe inflammation and demyelination were observed in CD4^−/− mice than in CD8^−/− and C57BL/6 mice (P ≤ 0.002 and P ≤ 0.001, respectively). Immunophenotyping of central nervous system (CNS) infiltrates revealed that CD4^−/− mice had a significant reduction in numbers of activated macrophages/microglial cells in the brain compared to the numbers in CD8^−/− and C57BL/6 mice, indicating a role for these cells in myelin destruction. Furthermore, CD4^−/− mice displayed lower levels of RANTES (a C-C chemokine) mRNA transcripts and protein, suggesting a role for this molecule in the pathogenesis of MHV-induced neurologic disease. Administration of RANTES antisera to MHV-infected C57BL/6 mice resulted in a significant reduction in macrophage infiltration and demyelination (P ≤ 0.001) compared to those in control mice. These data indicate that CD4⁺ T cells have a pivotal role in accelerating CNS inflammation and demyelination within infected mice, possibly by regulating RANTES expression, which in turn coordinates the trafficking of macrophages into the CNS, leading to myelin destruction.

Demyelination is a complex neuropathological process in which the myelin sheath that insulates and protects axons is damaged or destroyed. Several animal models of demyelination have been developed that have provided valuable contributions to the understanding of the immunopathological events that may drive human demyelinating diseases such as multiple sclerosis (MS) (22, 31). Among these is the neurotropic mouse hepatitis virus (MHV) model of virus-induced demyelination (12, 18). MHV is a positive-strand RNA virus that causes a variety of clinical diseases in susceptible strains of mice (23). Neurovirulent strains of MHV cause an acute encephalomyelitis that may ultimately progress to demyelinating disease characterized clinically by abnormal gait and hind-limb paralysis. Histologically, affected animals exhibit mononuclear cell infiltration and myelin destruction. Early studies suggested that the demyelination observed in MHV-infected mice was the result of virus-induced damage or destruction of oligodendrocytes (9, 36). However, more recent reports have indicated that MHV-induced demyelination is more complex and may also involve immunopathologic responses against viral antigens expressed in infected tissues (5, 35).

As T cells are considered central to the development of demyelinating lesions in animal models of demyelination as well as MS, it is imperative to better understand the mechanisms by which these cells exert their pathological effect (24, 25). We sought to evaluate the contributions of CD4⁺ and CD8⁺ T cells in MHV-induced central nervous system (CNS) disease in order to provide insight into the role(s) of these cells in the development of demyelination. To this end, we have taken advantage of the availability of transgenic knockout (ko) (−/−) mice to evaluate the roles of individual T-cell subsets in protecting against and contributing to CNS disease in MHV-infected mice (6, 29). We have infected CD4^−/− and CD8^−/− mice and compared the outcomes of infection, i.e., viral clearance and the development of clinical and histologic disease, with those in wild-type (wt) C57BL/6 mice. Our results indicate that while both CD4⁺ and CD8⁺ T cells are required for optimal host defense and clearance of virus from the CNS, CD4⁺ T cells are key contributors to the amplification of demyelination. One possible mechanism by which CD4⁺ T cells accomplish this is by producing or influencing the production of the C-C chemokine RANTES, which acts to attract macrophages into the CNS during viral infection. Indeed, treatment of mice with anti-RANTES antibody resulted in a significant reduction in both macrophage infiltration and demyelination.

MATERIALS AND METHODS

Virus and mice.

The MHV strain V5A13.1 (referred to henceforth as MHV) was derived from wild-type MHV-4 as previously described (4). Age-matched (5 to 7 weeks), male wt C57BL/6 mice (H-2^b background) and homozygous CD4^−/− (29) and CD8^−/− (6) mice (5 to 7 weeks, on the C57BL/6 H-2^b background) were used for studies described. CD4^−/− and CD8^−/− mice were purchased from Jackson Laboratories (Bar Harbor, Maine). Following anesthetization by inhalation of methoxyflurane (Pitman-Moore Inc., Washington Crossing, N.J.), mice were injected intracranially (i.c.) with 10 PFU of MHV suspended in 30 μl of sterile saline. Control (sham) animals were injected with sterile saline alone. Animals were sacrificed by methoxyflurane inhalation followed by cardiac puncture at days 7, 12, and 21 postinfection (p.i.), at which point brains and spinal cords were removed. One-half of each brain was used for plaque assay on the DBT astrocytoma cell line to determine viral burden (10, 17, 20). The remaining half of each brain and the spinal cords were either fixed for histologic analysis or stored at −80°C for RNA isolation.

Histology.

Brains and spinal cords were directly embedded in OCT (Sakura Finetek, Torrance, Calif.) or fixed by immersion overnight in 10% normal buffered formalin, after which the tissues were embedded in paraffin. The severity of inflammation was determined by staining tissue sections with hematoxylin and eosin, while demyelination was scored on slides stained with Luxol fast blue. Slides were coded and read blind by three investigators. Inflammation was evaluated as follows: 0, no inflammation; 1, one cell layer of inflammation; 2, two cell layers of inflammation; 3, three cell layers of inflammation; and 4, four or more layers of inflammation (25). Demyelination was scored as follows: 0, no demyelination; 1, mild inflammation accompanied by loss of myelin integrity; 2, moderate inflammation with increasing myelin damage; 3, numerous inflammatory lesions accompanied by significant increase in myelin stripping; and 4, intense areas of inflammation accompanied by numerous phagocytic cells engulfing myelin debris (11, 19). Scores were averaged and are presented as means ± standard deviations (SD).

Clinical disease.

Following infection with virus, mice were evaluated for signs of clinical disease by using a previously described scale (11, 19). Scoring was as follows: 0, no abnormality; 1, limp tail; 2, waddling gait and partial hind-limb weakness; 3, complete hind-limb paralysis; 4, death. Scores are presented as means ± SD.

Mononuclear cell preparation and flow cytometry.

Cells were obtained from brains and spinal cords of MHV-infected mice at days 7 and 12 p.i. based on a previously described protocol (28). In brief, brains and spinal cords were removed and a single cell suspension was obtained by grinding the tissue and then mincing it with a razor blade. All of these techniques were performed within sterile tissue culture plates placed on ice; the plates contained Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum. Cell suspensions were transferred to 15-ml conical tubes and Percoll (Pharmacia, Uppsala, Sweden) was added for a final concentration of 30%. One milliliter of 70% Percoll was underlaid and the cells were spun at 1,300 × g for 30 min at 4°C. Cells were removed from the interface and washed twice. Fluorescein isothiocyanate-conjugated rat anti-mouse F4/80 (C1:A3-1; Serotec, Oxford, England) was used to detect activated macrophages/microglial cells. As a control, an isotype-matched fluorescein isothiocyanate-conjugated antibody was used. Cells were incubated with antibodies for 30 min at 4°C, washed, fixed in 1% paraformaldehyde, and analyzed on a FACStar (Becton Dickinson, Mountain View, Calif.).

Immunohistochemistry.

Primary antibodies (diluted in phosphate-buffered saline containing 2% normal goat serum [NGS]) used for immunohistochemical detection of cellular antigens were as follows: rat anti-mouse CD4 (GK1.5; PharMingen, San Diego, Calif.) at 1:200, rat anti-mouse CD8a (53-6.7; PharMingen) at 1:100, and rat anti-mouse F4/80 (C1:A3-1; Serotec) at 1:50. In all cases, a biotinylated secondary antibody was used (1:300, Vector Laboratories, Burlingame, Calif.). Staining was performed on 8-μm-thick frozen sections fixed in 95% ethanol for 10 min at −20°C. The ABC Elite (Vector Laboratories) staining system was used according to the manufacturer's instructions, and diaminobenzidine was used as a chromogen. All slides were counterstained with hematoxylin, dehydrated, and mounted. Staining controls were (i) omission of primary antibodies from the staining sequence and (ii) treatment of sham-infected mice with primary and secondary antibodies.

RPA.

Total RNA was extracted from brains and spinal cords of MHV-infected and sham-infected mice by using the TRIzol reagent as previously described (17). Chemokine transcripts were analyzed using a previously described multitemplate probe set containing antisense riboprobes specific for 10 chemokine transcripts (1, 17). The probes targeted the following chemokines: the C-C chemokines macrophage inflammatory protein 2 (MIP-2) and interferon-inducible protein 10/cytokine responsive gene 2 (IP-10/CRG-2); the C-X-C chemokines C10, RANTES, macrophage chemoattractant proteins 1 and 3 (MCP-1 and MCP-3), MIP-1α and MIP-1β, and T-cell activation 3 (TCA-3); and the C chemokine lymphotactin (LT) (17). A probe for L32 was included in the probe set to verify consistency in RNA loading and assay performance (17). Ribonuclease protection assay (RPA) analysis was performed with 10 μg of total RNA using a previously described protocol (1, 17). For quantification of signal intensity, autoradiographs were scanned and the individual chemokine bands were normalized as the ratio of band intensity to the L32 control (1, 17, 19). Analysis was performed with NIH Image 1.61 software (1, 17, 19).

RT-PCR.

The antisense riboprobe used to detect RANTES mRNA was derived by reverse transcription-PCR (RT-PCR) amplification of cDNA generated from total RNA isolated from the brain of an MHV-infected mouse at day 7 p.i. Oligonucleotide primers for RANTES amplification were as follows: forward, 5′-TTT GCC TAC CTC TCC CTA GAG CTG-3′, and backward, 5′-ATG CCG ATT TTC CCA GGA CC-3′ (7). PCR amplification was performed with an automated Perkin-Elmer (Norwalk, Conn.) model 480 DNA thermocycler and the following profile: step 1, initial denaturation at 94°C for 45 s; step 2, annealing at 55°C for 45 s; and step 3, extension at 72°C for 2 min. Steps 1 to 3 were repeated 34 times for a total of 35 cycles. The expected PCR amplicon (300 bp) was cloned into the pCR Script SK+ vector (Stratagene, San Diego, Calif.), and sequence analysis showed >95% nucleotide identity with mouse RANTES (26).

In situ hybridization.

The protocol for in situ hybridization of brain and spinal cord sections has been previously described in detail (17, 20). The ³⁵S-UTP-radiolabeled RANTES antisense RNA probe was derived by in vitro transcription with an RNA transcription kit (Stratagene). Upon completion of the in situ procedure, the slides were dehydrated and dried. Next, slides were dipped in a Kodak NTB2 nuclear emulsion at 46°C and exposed at 4°C for 2 to 4 weeks in a desiccator. The slides were developed and fixed with Kodak D-19 developer and fixer, counterstained with hematoxylin and eosin Y solutions, dehydrated, and mounted.

RANTES enzyme-linked immunosorbent assay (ELISA).

RANTES was quantitated in brain and spinal cord samples obtained from MHV-infected mice by using the Quantikine M mouse RANTES immunoassay kit (R&D Systems, Minneapolis, Minn.). Tissue samples were homogenized in 1 ml of sterile phosphate-buffered saline and spun at 400 × g for 10 min at 4°C (16). Duplicate supernatant samples were used to determine RANTES levels present within the tissues according to the manufacturer's instructions. Following the enzymatic color reaction, samples were read at 450 nm and RANTES levels were quantitated in comparison to a standard curve (supplied by the manufacturer); the results are presented as picograms per milliliter. The limit of sensitivity of RANTES detection was approximately 8.0 pg/ml. The reagents used for these experiments do not cross-react with other mouse chemokines or cytokines.

Anti-RANTES treatment.

Neutralizing anti-RANTES antibodies were generated by immunizing goats with a peptide corresponding to 14 amino acid residues at the carboxy terminus of the RANTES protein (3, 16). This antiserum reacts with murine and human RANTES and no other identified cytokine or chemokine (3, 16). Mice were injected intraperitoneally with 0.5 ml of antiserum (titer, >10⁶, containing between 1.0 and 1.5 mg depending upon lot) on days 3, 5, and 8 p.i. Antibody-treated mice were evaluated only until day 12, as after this time the efficacy of the antibody is diminished due to accelerated decay of the antibody within the mouse. Control MHV-infected mice were treated with NGS. Experimental and control mice were sacrificed at days 7 and 12 p.i.

Statistics.

Statistically significant differences between the groups of mice were determined by the Mann-Whitney rank sum test for nonparametric samples, using Sigma Stat 2.0 software. P values of ≤0.05 were considered significant.

RESULTS

MHV infection and viral clearance.

Following i.c. infection of C57BL/6 mice with MHV, acute viral encephalitis developed, with 20% of the animals dying between days 8 and 10 p.i. The majority of surviving animals developed the clinical characteristics of MHV-induced demyelination, e.g., awkward gait and hind-limb paralysis, by day 12 p.i. Virus could not be isolated from infected mice at this time point (Fig. 1). In contrast to C57BL/6 mice, approximately 30% of CD4^−/− and 70% of CD8^−/− mice succumbed between days 7 and 10 p.i. Both CD4^−/− and CD8^−/− mice displayed, on average, higher titers of virus at day 7 p.i. Consistent with earlier studies, neither strain cleared virus from the brains by 12 days p.i. (Fig. 1) (11, 21, 27, 33, 37, 38).

FIG. 1.

Viral titers in brains. Mice were infected i.c. with 10 PFU of MHV and sacrificed at days 7 and 12 p.i., and viral titers were determined by plaque assay using one-half of the brain. By day 12 p.i., C57BL/6 mice cleared virus below the limit of detection (100 PFU/g), while CD4^−/− and CD8^−/− mice as well as C57BL/6 mice treated with anti-RANTES antibody still had detectable levels of virus present. Numbers of mice examined were as follows: C57BL/6, eight for day 7 and five for day 12; CD8^−/−, eight for day 7 and three for day 12; CD4^−/−, seven for day 7 and five for day 12; and anti-RANTES-treated C57BL/6, four for day 7 and six for day 12. Data are presented as means ± SD.

Disease severity in MHV-infected mice.

Infected mice were evaluated for clinical disease and histologic disease. Clinical scoring of the mice at day 12 p.i. revealed that CD4^−/− animals displayed less severe symptoms (score = 1.6 ± 0.4; n = 7), i.e., waddling gait and hind-limb paralysis, than did C57BL/6 mice (score = 2.8 ± 0.2; n = 5) and CD8^−/− mice (score = 2.4 ± 0.4; n = 4), although this difference was not significant. To evaluate the severity of histological disease, brains and spinal cords were removed at scheduled time points p.i. and stained with either hematoxylin and eosin or Luxol fast blue to assess inflammation or demyelination, respectively. There was extensive perivascular inflammation at all time points examined within the CNSs of C57BL/6 and CD8^−/− mice (Table 1 and Fig. 2). In contrast, CD4^−/− mice displayed significantly less severe perivascular inflammation at 7 and 12 days p.i. than did CD8^−/− and C57BL/6 mice (Table 1 and Fig. 2). In addition to decreased inflammation, CD4^−/− mice displayed significantly less severe demyelination than did the other mice at both days 12 and 21 p.i. (Table 1 and Fig. 2).

TABLE 1.

Disease severity in MHV-infected mice^a

Mice	Day(s) p.i.	Perivascular inflammation	Demyelination
C57BL/6	7	2.8 ± 0.3	ND
	12–15	2.8 ± 0.4	2.6 ± 0.3
	21	ND	3.7 ± 0.3
CD8−/−	7	3.0 ± 0.2	ND
	12–15	2.5 ± 0.5	2.6 ± 0.7
	21	ND	ND
CD4−/−	7	1.2 ± 0.2b	ND
	12–15	1.4 ± 0.3c	1.2 ± 0.2d
	21	ND	2.1 ± 0.5e
Anti-RANTES treated	7	2.2 ± 0.5	ND
	12	1.4 ± 0.6c	1.1 ± 0.3d
	21	ND	ND

^aMice were infected i.c. with 10 PFU of MHV. Animals were sacrificed at days 7, 12, and 21 p.i., and brains and spinal cords were removed and stained with either hematoxylin and eosin or Luxol fast blue to determine severity of inflammation or demyelination, respectively. Scores were assigned on a scale from 0 to 4 for inflammation and demyelination as described in Materials and Methods. ND, not done. 

^bP ≤ 0.005 compared to C57BL/6 and CD8^−/− mice. Numbers of mice examined: C57BL/6, three; CD4^−/−, three; and CD8^−/−, three. 

^cP ≤ 0.05 compared to C57BL/6 mice. Numbers of mice examined: C57BL/6, six; CD4^−/−, six; CD8^−/−, four; and anti-RANTES treated, five. 

^dP < 0.001 compared to C57BL/6 mice and P ≤ 0.002 compared to CD8^−/− mice. Numbers of mice examined: C57BL/6, eight; CD4^−/−, nine; CD8^−/−, five, and anti-RANTES treated, four. 

^eP ≤ 0.05 compared to C57BL/6 mice. Numbers of mice examined: C57BL/6, three, and CD4^−/−, four. 

FIG. 2.

Inflammation and demyelination. Brains and spinal cords were obtained from mice and the severity of histologic disease was evaluated. (Top row) Representative hematoxylin and eosin staining from brains of mice at day 7 p.i. Note the limited perivascular cuffing around the vessels of the CD4^−/− mouse and anti-RANTES-treated mouse compared to C57BL/6 and CD8^−/− mice. (Bottom row) Representative spinal cord sections from mice at day 12 p.i. stained with Luxol fast blue to assess severity of demyelination. In C57BL/6 and CD8^−/− mice there is extensive inflammation and myelin destruction (outlined with arrows), in marked contrast to the spinal cords from the CD4^−/− mouse and the mouse treated with anti-RANTES antiserum. Magnifications, ×324 for hematoxylin and eosin and ×162 for luxol fast blue.

Analysis of infiltrating mononuclear cells.

The reduction in severity of both inflammation and demyelination in MHV-infected CD4^−/− mice suggested that CD4⁺ T cells may be important in accelerating the severity of virus-induced CNS disease by promoting the entry of specific cells into the CNS. Because macrophages have been shown to be important contributors to demyelination in mice with experimental allergic encephalomyelitis (EAE) (2, 16), we focused our attention on this cell population. Cells were obtained from the brains and spinal cords of MHV-infected mice at days 7 and 12 p.i., and fluorescence-activated cell sorter (FACS) analysis was performed in an attempt to determine if macrophage infiltration, determined by F4/80 antigen expression, was affected by the absence of the CD4 compartment. The data shown in Fig. 3 reveal that at day 7 p.i. both C57BL/6 and CD8^−/− mice had, on average, approximately twice as many activated macrophages/microglial cells present within the CNS as did CD4^−/− mice. By day 12 p.i., macrophage/microglial cell infiltration remained compromised in the CD4^−/− mice, with fewer cells present than in C57BL/6 and CD8^−/− animals. Immunohistochemistry was performed on brain and spinal cord sections in an attempt to determine if lower numbers of macrophages/microglial cells were present within spinal cord white matter tracts of CD4^−/− mice than in those of C57BL/6 and CD8^−/− mice. The data presented in Fig. 4A show intense F4/80 staining in white matter tracts of both C57BL/6 and CD8^−/− mice at day 12 p.i. In marked contrast, only limited numbers of F4/80-positive cells were found in CD4^−/− mice at the same time. Quantification of positive cells revealed that there were significantly fewer activated macrophages/microglial cells in spinal cord white matter tracts of CD4^−/− mice than in those of C57BL/6 and CD8^−/− mice (Fig. 4B).

FIG. 3.

FACS analysis of F4/80-positive cells infiltrating the CNS. Single-cell suspensions were obtained from brains and spinal cords of infected mice at 7 and 12 days p.i., and F4/80 antigen expression was evaluated. Two mice were used from each group, with the exception of only one CD8^−/− mouse being examined at day 12 p.i. Data are presented as means ± SD.

FIG. 4.

(A) F4/80 staining in spinal cord white matter tracts. Spinal cords were removed from mice at day 12 p.i., and immunohistochemical staining for F4/80 antigen was performed. Representative sections from mice are shown. Both C57BL/6 and CD8^−/− mice exhibited numerous cells positive for F4/80 (cells stained brown) within white matter tracts at this time. In marked contrast were the white matter tracts of spinal cords obtained from CD4^−/− mice and mice treated with RANTES antiserum, in which very few positive cells were detected. In all cases, control sections were negative (not shown). Magnification, ×400. (B) Enumeration of F4/80-positive cells in spinal cord white matter tracts. A minimum of six fields (40× objective) were counted per mouse at day 12 p.i. (three mice from each group were tested). Data are presented as means ± SD.

Chemokine expression in the CNS of MHV-infected mice.

The reduced mononuclear cell infiltration suggested that CD4⁺ T cells may participate in MHV-induced CNS disease via the release of soluble molecules that enhance the severity of inflammation and demyelination by attracting effector cells, e.g., macrophages, to the CNS. To investigate this possibility, we evaluated the chemokine mRNA profiles of MHV-infected CD4^−/−, CD8^−/−, and C57BL/6 mice in an attempt to correlate specific mRNA signals with the presence or absence of inflammation and demyelination. Similar levels of transcripts for up-regulated chemokines were detected in all three strains of mice examined. However, the band intensity for the C-C chemokine RANTES appeared to be lower for CD4^−/− mice at days 7 and 12 than for C57BL/6 and CD8^−/− mice (Fig. 5A). Quantitation of RANTES signal intensity indicated that C57BL/6 and CD8^−/− mice had approximately 1.5 times higher levels of mRNA transcripts at both time points than did CD4^−/− mice (Fig. 5B). In order to determine if reduced mRNA correlated with lowered protein levels, the concentration of RANTES protein in the brains and spinal cords of infected mice at day 7 p.i. was determined by ELISA. Consistent with the mRNA profiles of infected mice, RANTES protein levels were reduced by approximately 40 to 50% within the CNS of CD4^−/− mice compared to protein levels of C57BL/6 and CD8^−/− mice (Fig. 5C).

FIG. 5.

(A) RPA showing kinetics of chemokine mRNA expression in brains of MHV-infected C57BL/6, CD4^−/−, and CD8^−/− mice. Ten micrograms of total RNA obtained from the brains of infected mice was hybridized with a probe set designed to detect 10 different chemokine transcripts as well as an internal L32 control. Mouse strains used and the days of examination are indicated below the lanes. A sample consisting of a set of sense RNAs complementary to the chemokine probe set for use in standardization of fragment size and assay integrity was included and is shown on the right margin of the autoradiograph. These sense RNAs contain cloning sequences and consequently run slightly higher than protected fragments from brains. Up-regulated chemokine transcripts and L32 are indicated in the left margin of the autoradiograph. Each lane contains a sample from an individual mouse at the indicated time point. Sham controls included brains at day 7 from mice receiving sterile saline alone. The results presented are from one experiment and are representative of the chemokine profiles from three separate experiments. (B) Quantitative analysis of chemokine mRNA transcripts expressed in brains of mice at days 7 and 12 p.i. Densitometric analysis of each lane representing a brain sample from an individual mouse was performed on the scanned autoradiograph (A) using NIH Image 1.61 software. The levels are based on normalized units that allow comparison of chemokine mRNA transcript levels. (C) Analysis of RANTES protein levels in the CNS of mice. RANTES protein levels in brains and spinal cords obtained from MHV-infected C57BL/6, CD8^−/−, and CD4^−/− mice at day 7 p.i. were determined by ELISA as described in Materials and Methods. Two mice from each group were examined. Data are presented as means ± SD.

Localization of RANTES expression.

To determine the cellular source of RANTES expression within the CNS of infected mice, in situ hybridization analysis was performed using a RANTES-specific riboprobe. The representative data shown in Fig. 6 demonstrate cells surrounding a perivascular cuff within the brain of a CD8^−/− mouse expressing RANTES mRNA, whereas no such signal was detected surrounding a cuff from a CD4^−/− mouse. These data suggest that inflammatory cells, in part, are responsible for production of RANTES during MHV infection of the CNS.

FIG. 6.

In situ hybridization detection of RANTES expression in tissue from MHV-infected CD4^−/− and CD8^−/− mice at day 12 p.i. Shown is a representative perivascular cuff from a CD8^−/− mouse with RANTES expression (represented by overlaying silver grains) by cells which morphologically appear to be inflammatory leukocytes (indicated by arrows). In contrast, no signal is detected in cells surrounding a vessel within the brain of a CD4^−/− mouse. No signal was detected in sections probed with the RANTES sense probe (not shown). Magnification, ×352.

Anti-RANTES treatment.

To assess the potential role of RANTES in MHV-induced CNS disease, infected C57BL/6 mice were treated with neutralizing RANTES antiserum. Such treatment resulted in a delay in viral clearance from the CNS at day 12 p.i. compared to clearance in both infected mice and infected mice treated with NGS (data not shown and Fig. 1). Examination of the brains and spinal cords at days 7 and 12 p.i. revealed less severe perivascular inflammation in anti-RANTES-treated mice (similar to that observed in CD4^−/− mice) than in both C57BL/6 mice and CD8^−/− mice (Table 1). Immunohistochemical staining for CD4, CD8, and F4/80 antigens indicated very little infiltration of cells positive for these antigens into the parenchyma of mice treated with anti-RANTES compared to that in control mice (Fig. 7). To correlate the reduction in viral clearance and cellular infiltration with the development of demyelination, spinal cords from anti-RANTES-treated mice were stained with Luxol fast blue. The severity of demyelination in anti-RANTES-treated animals was significantly reduced compared to that of demyelination in control mice (Table 1 and Fig. 2). In addition, anti-RANTES-treated mice had significantly lower levels of F4/80-positive cells located within the spinal cord at day 12 p.i., which correlated with the reduction in the severity of demyelination (Fig. 4B).

FIG. 7.

Comparison of cellular infiltration in MHV-infected C57BL/6 mice and mice treated with anti-RANTES antiserum. Shown are representative sequential sections of a perivascular cuff in the brains of an MHV-infected C57BL/6 mouse (B6) and a mouse treated with anti-RANTES (α RANTES) at day 12 p.i. Sections were stained with the indicated antibodies. Note the much higher intensity of staining for CD4, CD8, and F4/80 in the C57BL/6 mouse (brown cells) than in the mouse treated with anti-RANTES antiserum. Control sections were negative in all cases (not shown).

DISCUSSION

To evaluate the role of CD4⁺ and CD8⁺ T cells in virus-induced CNS disease, CD4^−/− and CD8^−/− mice were infected with the demyelinating strain MHV-V5A13.1 and examined at various stages p.i. Infection of CD4^−/− and CD8^−/− mice resulted in higher mortality rates and higher levels of virus at all time points examined than those for wt C57BL/6 mice. These observations confirm studies by others that have demonstrated that both subsets of T cells are required for host defense and optimal clearance of virus from the CNS (11, 21, 27, 33, 37, 38). Previous studies have indicated that neither T-cell subset is required for demyelination to occur in MHV-infected mice (11, 34). However, the data presented in this study indicate that CD4⁺ T cells are important contributors to the development of both CNS inflammation and demyelination. These data are similar to other models of demyelination, such as models of EAE and Theiler's virus, which have demonstrated that CD4⁺ T cells are required for demyelination to occur (24, 25). Sutherland et al. (34) have reported that thymectomized mice and mice depleted of either CD4⁺ or CD8⁺ T cells develop demyelination following infection with MHV. However, the authors state that their experimental results do not exclude the possibility that T cells may be important in initiating demyelination early in the infection, which is consistent with what we have presented. Houtman and Fleming (11) have reported that a percentage of either A_β ko (lacking major histocompatibility complex class II and having low numbers of CD4⁺ T cells) or β₂-microglobulin ko mice (lacking major histocompatibility complex class I and having low numbers of CD8⁺ T cells) developed demyelination following infection with MHV strain JHM, indicating that neither T-cell subset is required for demyelination to occur. Differences between these studies and the results presented in this report may be due to differences in MHV strains as well as the genetic backgrounds of the animals tested. In addition, demyelination was determined only at day 12 p.i.; therefore, it is possible that a significant population of the A_β ko mice may have been found to have had less severe demyelination than wt controls if mice had been examined at later time points.

To determine why CD4^−/− mice exhibited less severe CNS disease, the cellular infiltrate within the CNS was analyzed by FACS analysis and immunohistochemical staining. A marked reduction in the number of activated macrophages/microglial cells within the brains and spinal cords of CD4^−/− mice at days 7 and 12 p.i. compared to the numbers present within the CNS of C57BL/6 and CD8^−/− mice was found. The contributions of macrophages to demyelination have been documented in other models of MS. Inhibition of infiltration of these cells into the CNS resulted in a decrease in the severity of both clinical and histologic disease, indicating an important role in the pathogenesis of demyelinating disease for this cell population (2, 14, 16). The data presented in this paper indicated that macrophages contributed to demyelination in MHV-infected mice. Furthermore, these findings point to a central role for CD4⁺ T cells in the early stages of disease following viral infection in the amplification of inflammation and, ultimately, demyelination by promoting the entry of macrophages into the CNS.

Several possible mechanisms by which CD4⁺ T cells contribute to the entry of inflammatory mononuclear cells into the CNS following viral infection can be proposed. One possibility is that activated, virus-specific CD4⁺ T cells present in the brain release chemokines that serve to recruit and retain infiltrating mononuclear cells within the CNS. We have recently shown there is differential expression of chemokine genes following MHV infection (17). RANTES was among the chemokines prominently expressed during both the acute and chronic stages of disease, suggesting an important role for this protein in the disease process (17). Data presented in this report indicate that expression of RANTES is compromised in CD4^−/− mice, as demonstrated by lower levels of mRNA transcripts and protein than those of CD8^−/− and C57BL/6 mice. Furthermore, in situ hybridization suggested that inflammatory leukocytes were a source of RANTES transcripts. The fact that inflammatory cells were the prevalent source of RANTES expression was not surprising given that members of our group previously demonstrated that in vitro infection of astrocytes with MHV resulted in a chemokine profile identical to that detected in vivo, with the notable exception of RANTES (17). These data suggested that CD4⁺ T cells are the predominant source of RANTES following MHV infection of the CNS. It is also possible that CD4⁺ T cells influence expression of RANTES by other cell populations through the release of cytokines and/or chemokines. Additional cellular sources, e.g., cytokine-activated glial cells, must be considered as sources of RANTES due to the fact that RANTES mRNA transcripts and protein were detected, albeit at lower levels, within the CNS of CD4^−/− mice. In light of the fact that RANTES exerts a potent chemotactic effect on both T cells and monocytes, these data suggest that the reduction in macrophage infiltration and the severity of demyelination in the CNS of CD4^−/− mice is, in part, the result of reduced RANTES levels (30).

In a direct test of the importance of RANTES in MHV-induced CNS inflammation and demyelination, MHV-infected C57BL/6 mice were treated with anti-RANTES antibodies and the severity of disease was evaluated. Treatment led to a disease in wt C57BL/6 mice similar to the phenotype observed in CD4^−/− mice, with delayed viral clearance from the brain and decreased cellular infiltration as well as a significant reduction in the severity of demyelination. The decreased capacity to clear virus from the brains is explained by the limited infiltration of CD4⁺ and CD8⁺ T cells into the brain during the acute stage of disease. Furthermore, the decrease in macrophage infiltration correlated with the reduced severity of demyelination, which supports the observations with MHV-infected CD4^−/− mice. These observations reinforce the functional significance of RANTES expression during virus-induced CNS disease, indicating that this chemokine has a prominent role in recruitment of both CD4⁺ and CD8⁺ T cells as well as macrophages into the CNS following MHV infection.

In support of the argument that chemokines play a central role in inflammatory CNS disease are recent studies demonstrating that chemokines are crucial to the development of inflammation and demyelination in EAE. Production of chemokines within the CNS correlates with the development of both acute and relapsing EAE (7, 8, 15, 16). Furthermore, administration of neutralizing antibodies against selected chemokines has been shown to reduce the severity of both clinical and histologic disease by limiting the entry of selected populations of inflammatory cells, such as macrophages (14, 16).

Recent studies have examined chemokine expression in patients with MS (13, 31). Hvas and colleagues (13) were able to demonstrate RANTES expression by infiltrating T lymphocytes surrounding MS plaque lesions. Elevated levels of the C-X-C chemokine IP-10 and the closely related chemokine MIG (monokine induced by gamma interferon) as well as RANTES were found in the cerebrospinal fluid of MS patients during periods of attack (31). Recent work has demonstrated a direct correlation between clinical progression in the severity of MS with CNS infiltration, suggesting that production of IP-10, MIG, and RANTES may contribute to the pathogenesis of MS by recruiting inflammatory leukocytes into the CNS (31). The studies presented within this report support this argument in that RANTES expression clearly has a prominent role in both regulating leukocyte entry into the CNS and contributing to the pathogenesis of virus-induced CNS inflammation and demyelination.

An overall picture of the relationship between chemokine expression and MHV-induced CNS disease is possible based on previous studies by members of our group as well as the data presented in this paper (17). This theory holds that early following MHV infection of the CNS, there is a rapid expression of the C-X-C chemokine CRG-2/IP-10 by astrocytes (17). CRG-2/IP-10 is a potent chemoattractant for T cells and macrophages; therefore, it is likely that this chemokine serves to bring in these cells during the early stages following infection (17). Activated CD4⁺ T cells enter the brain and produce RANTES, which accelerates the severity of inflammation and demyelination by helping to attract additional T cells and macrophages. The accumulation of macrophages ultimately results in myelin destruction. The results presented in the paper strengthen this argument, and this is, to our knowledge, the first report that has clearly shown that the severity of virus-induced CNS inflammation and demyelination can be reduced by treatment with neutralizing antiserum against RANTES. In addition, our results support and extend earlier studies which have suggested that targeting chemokines may be novel therapeutic intervention strategies to treat human CNS inflammatory diseases, including MS (14–16, 31).