INTRODUCTION

Top Abstract Introduction Materials and Methods References

The Daniel's strain of Theiler's murine encephalomyelitis virus (TMEV) is a picornavirus which upon intracerebral injection causes a persistent infection of the white matter and demyelination in susceptible mouse strains (haplotypes H-2^s, H-2^r, H-2^v, H-2^f, H-2^p, or H-2^q) but not in resistant strains (haplotypes H-2^b, H-2^d, or H-2^k) of the C57BL lineage (35, 41, 44). The demyelinating disease observed in susceptible strains of mice is an excellent model of human multiple sclerosis (6, 12, 27). While the cause of pathology in susceptible strains remains unknown, the Th1 CD4⁺ compartment has been demonstrated to mediate a delayed-type hypersensitivity inflammatory reaction with macrophages and cytokines leading to destruction of self tissue (26). However, depletion of the CD8⁺ T-cell compartments has been shown to suppress demyelination, whereas depletion of CD4⁺ T cells promotes demyelination (39). In addition, efforts to inhibit the effects of CD4 or CD8 T-cell involvement by disrupting major histocompatibility complex (MHC) class I and II expression or by knocking out CD4 and CD8 still results in demyelination and high virus titers (13, 14, 34, 37, 40). Therefore, the contribution of CD4 and CD8 T-cell subsets to demyelination in susceptible strains remains obscure.

In contrast to susceptible strains of mice, resistant mice clear TMEV from the central nervous system (CNS), and this clearance is associated with a virus-specific cytotoxic T-lymphocyte (CTL) response (43). Evidence supporting the importance of the CTL response in TMEV is indirect, but an essential role for a class I restricted response in viral clearance is demonstrated by several observations. First, the susceptible DBA/2 mice upon in vivo administration of interleukin-2 (IL-2)-secreting tumor readily clear TMEV. The viral clearance correlated with a three- to fourfold increase in virus-specific CTL activity (18). Second, perforin-deficient mice develop persistent TMEV infection (45). Perforin is an important effector molecule for both CTL and natural killer (NK) cells. Immunodepletion of NK cells in genetically resistant mice with anti-NK 1.1 antibody does not allow TMEV virus persistence, whereas a more pervasive depletion with anti-asialo GM1, which also targets some activated T cells, does (36). Third, the ₂ microglobulin (₂m) knockout mice crossed onto the resistant C57BL/6 genetic background are no longer able to clear the virus (13, 37, 40). Finally, resistance has been genetically mapped to the H-2D class I alleles, since introduction of H-2D^b and H-2D^d transgenes into susceptible strains of mice allows TMEV clearance (2, 21, 38, 41, 42). The use of MHC recombinant mice suggests that H-2D class I alleles, more effectively than H-2K class I alleles, present viral peptides to CTLs which leads to TMEV clearance, either by direct killing of infected cells or by cytokine release (21, 42).

Support for the hypothesis that the H-2D class I molecules effectively present viral peptides to CTLs during viral clearance was demonstrated by two groups simultaneously. Borson et al. (5) subjected a VP2 viral capsid fusion protein to alkali hydrolysis and used the resulting peptides to sensitize RMA-S target cells for killing by C57BL/6 CNS infiltrating lymphocytes (CNS-ILs). Viral protein 2 (VP2) amino acids 112 to 140 were identified as containing the relevant epitope. Upon synthesizing overlapping peptides from this region, VP2_121-130 (FHAGSLLVFM) was identified as the peptide that could sensitize target cells for CNS-IL killing. Independently, Dethlefs et al. (11) identified a 9-amino-acid peptide from VP2, amino acids 122 to 130, which is found in the VP2 capsid protein of both Mengo virus and TMEV. This peptide sensitized target cells for CTLs isolated from both Mengo virus- and TMEV-infected mice. While both of these studies found a relevant epitope for TMEV-specific CTL killing, the relative importance of VP2_121-130 for CTL mediated clearance of TMEV from the CNS in resistant C57BL/6 mice was not determined.

To better understand the class I-mediated T-cell response and the importance of VP2_121-130 presented in the context of D^b in a TMEV clearing inflammatory reaction, we constructed peptide-MHC tetramers. Altman et al. have described a method by which four class I heterodimers bound with a specific peptide can be assembled with fluorescent, conjugated streptavidin and used to stain T cells with T-cell receptors (TCRs) that have affinity for the peptide presented in the context of the class I molecule (1). Using this method, we assembled tetramers of D^b, VP2_121-130, and human ₂m to stain CD8⁺ T cells with TCRs specific for this epitope and then assessed their frequency in the bulk population of brain-infiltrating CD8⁺ lymphocytes. In addition, we addressed the importance of the CD4 and CD8 T-cell compartments, the H-2D^b allele, gamma interferon (IFN-), tumor necrosis factor alpha receptor I (TNF-RI), perforin, MHC class II, and CD28 in VP2_121-130-specific T-cell infiltration by using knockout mice as hosts for intracerebral viral infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods References

Virus infection and mice. Male and female C57BL/6, C57BL/6-Pfp^tm1sdz, and IFN-^/ mice were obtained from Jackson Laboratory (Bar Harbor, Maine), and the Ab^o (class II^/) mice were a gift from Chris Benoist (Strasbourg, France). All other knockout mice (CD4^/, CD8^/, TNF-RI^/, and CD28^/) were gifts from Tak Mak at Amgen (Thousand Oaks, Calif.). The FVB/D^b mice were generated with a D^b genomic transgene in Chella David's gene transfer lab at the Mayo Foundation (Rochester, Minn.). Mice ranged in age from 4 to 16 weeks. Mouse spleens, lymph nodes, and brains were harvested 7 days post-intracerebral injection with 2 × 10⁶ PFU of Daniel's strain of TMEV. Lymphocytes were then extracted from the brain through homogenization and a homogeneous Percol gradient (24). An improved step gradient containing 35% Percol layered over 70% Percol was used to isolate cleaner preparations of lymphocytes from brains (see Fig. 4). The change in isolation procedures did not alter the ratio of tetramer. Stained CD8⁺ T cells observed in C57BL/6 mice were infected with TMEV (data not shown). Erythrocytes were then removed from brain, spleen, and lymph node cell preparations through lysis with ammonium chloride, potassium carbonate, and EDTA (ACK). Handling of all animals conformed to the National Institutes of Health and Mayo Clinic institutional guidelines.

Preparation of H-2D^b tetramers. The expression vectors for H-2D^b and human ₂m were generous gifts from John Altman and Mark Davis, respectively. Proteins expressed from these reagents in bacteria were isolated and folded in gross excess of E7 or VP2_121-130 peptide as previously described (15, 25, 38). The folded monomer complexes were desalted with a PD-10 desalting column (Pharmacia, Upsala, Sweden), biotinylated with a BirA biotinylation kit (Avidity, Denver, Colo.), purified with a Mono Q cation-exchange column (Bio-Rad, Hercules, Calif.), and complexed with R-phycoerythrin-streptavidin (Molecular Probes, Eugene, Oreg.) at a 4.1:1 molar ratio. This tetramer complex was then purified by S-200 size exclusion gel filtration (Bio-Rad).

Flow cytometric analysis. Then, 10⁶ cells isolated from brain, lymph nodes, and spleen were stained with R-phycoerythrin D^b:VP2_121-130 tetramer for 1 h, adding anti-CD8 fluorescein isothiocyanate during the final 20 minutes. Samples were then washed twice with fluorescence-activated cell sorter (FACS) buffer (1% bovine serum albumin and 2% sodium azide), resuspended in cold phosphate-buffered saline and fixed in 1% paraformaldehyde. Samples were run on a Becton Dickinson FACScan instrument (Mountain View, Calif.) and are reported on a logarithmic scale (250 channels/log intensity). Gates were drawn to include CD8⁺ cells while excluding nonspecific binding of tetramer and/or anti-CD8. The data were analyzed in two ways. First, the quadrants were positioned to exclude the lymphocyte-negative signals present in normal, uninfected brain cell preparations (Fig. 1A) so that only CD8⁺ cells would be scored. This procedure counts only cells expressing higher levels of CD8 and misses some positive cells that coincide in fluorescent intensity with the negative background. Although this approach has the disadvantage of not analyzing all the cells, it has the advantage of being objectively applied to all of the samples analyzed in a given experiment. In a second approach, we drew gates around populations of fluorescent cells not present in the normal brain preparation (Fig. 1A). We defined two CD8⁺ populations in Fig. 1B: cells staining with the tetramer and cells not staining with the tetramer. The estimated ratios of tetramer-stained to nonstained cells by using either method was the same. These two strategies were applied to all the FACS data presented; both yielded comparable results.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Staining of inflammatory mononuclear cells isolated from the brains of noninfected C57BL/6 mice (A), TMEV-infected C57BL/6 mice (B), and TMEV-infected BALB/c mice (C). The circles population represents those lymphocytes that stain with both anti-CD8 and Db:VP2121-130 tetramer. The percentage given is the percentage of CD8+ lymphocytes that stain with Db:VP2121-130 tetramer. This value is corrected by subtracting the nonspecific binding of irrelevant Db:E7 tetramer. See Materials and Methods for details of the analysis.

In vitro CTL assay. A standard 4-h ⁵¹Cr release assay was done to determine CTL activity in splenocyte and CNS-IL preparations, and 10³ RMA-S target cells/well were used. RMA-S cells were labeled with ⁵¹Cr for 2 h in the presence of VP2_121-130 peptide or irrelevant K^b-restricted dEV8 peptide (49). Both peptides were synthesized in the Mayo protein core facility. Freshly isolated effector cells were incubated with labeled target cells for 4 h. The percent specific lysis of target cells was calculated from released ⁵¹Cr as follows: 100 × (experimental release  spontaneous release)/(total release in detergent  spontaneous release). The number of effector T cells added was estimated by counting the cells with lymphocyte morphology by microscopic inspection. Samples from the cell preparations were analyzed by FACS and found to contain fewer T cells than was estimated by microscopy. These effector/target ratios of brain cell preparations represent overestimates of the T cells added. This fact does not influence our conclusions, since populations containing fewer T cells exhibited cytolytic activity, while populations containing more T cells did not.

	RESULTS AND DISCUSSION

Inflammatory mononuclear cells were isolated from the whole brains of noninfected and 7-day-post-TMEV-infected C57BL/6 mice. These CNS-ILs were stained with anti-CD8 antibody and D^b:VP2_121-130 tetramer or control tetramer specific for an irrelevant papillomavirus epitope as outlined in Materials and Methods. As expected, no CNS-ILs were detected in the brains from noninfected mice and no cells had demonstrable staining with anti-CD8 antibodies or D^b:VP2_121-130 tetramer (Fig. 1A). In contrast, as determined by four separate experiments, approximately 50 to 63% of CNS-ILs from TMEV-infected mice stained double positive for both of these reagents (Fig. 1B). BALB/c (H-2^d) mice infected with TMEV did not stain with the D^b:VP2_121-130 tetramer (Fig. 1C). These observations document the specificity of the tetramer staining and demonstrate that a majority of CD8⁺ T cells in C57BL/6 mice harvested from the site of viral infection express TCR specific for the VP2_121-130 peptide presented in the context of D^b. Three-color staining revealed that the analyzed T-cell populations included approximately 30% CD4⁺/CD8 cells that are presumably T helper cells (data not shown). None of these cells bear receptors stained by the tetramer probes.

To address whether there was a systemic clonal expansion of CD8⁺ T cells specific for the D^b:VP2_121-130 epitope, lymphocytes were isolated from the brain, spleen, and cervical lymph nodes of infected C57BL/6 mice. No VP2_121-130-specific T cells were identified in the spleen or cervical lymph nodes, whereas 55% of the CD8⁺ cells in the brain stained with the D^b:VP2_121-130 tetramer (Fig. 2A to C). A control tetramer, consisting of D^b loaded with a D^b binding peptide from the human papillomavirus E7 protein (25), did not stain this population (Fig. 2F). This control tetramer, however, stains greater than 99% of D^b:E7 epitope-specific CD8⁺ T cells (data not shown). The observation that 55% of the CTLs are specific for one peptide is similar to frequencies previously reported in infections with lymphocytic choriomeningitis virus (LCMV). For example, Murali-Krishna et al., by using tetramers constructed with an immunodominant LCMV peptide presented in the context of L^d, stained up to 56% of splenic CD8⁺ T cells in BALB/c mice (30).

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Staining of lymphocytes isolated from mouse spleen, cervical lymph node, and brain at 7 days post-TMEV infection. Isolated lymphocytes were stained with either Db:VP2121-130 tetramer (A to C) or control irrelevant Db:E7 tetramer (D to F). The percentage of CD8+ T cells that stain positive or negative for Db:VP2121-130 or Db:E7 tetramer is given.

We next addressed CTL activity in the lymphocytes isolated from the spleen and brain. In agreement with the staining shown by the D^b:VP2_121-130 tetramer, RMA-S cells loaded with VP2_121-130 were killed by freshly isolated CNS-ILs but not by lymphocytes isolated from the spleen (Fig. 3). In BALB/c mice infected with the JHM strain of mouse hepatitis virus (JHMV), JHMV-specific CTL activity was also localized only to the brain and not the cervical lymph node or spleen (48). It is possible that in the case of these virus infections, the brain acquires the responsibilities of a lymph node and is the site of CTL expansion.

View larger version (25K): [in this window] [in a new window]   FIG. 3.   Db:VP2121-130 epitope-specific cytolytic activity of lymphocytes isolated from the brain and spleen at 7 days post-TMEV infection. RMA-S cells were loaded with VP2121-130 peptide or irrelevant dEV8 peptide before being subjected to killing by brain and spleen lymphocyte preparations.

The hierarchy of immunodominance among peptides in TMEV infection may be influenced by several factors. For example, peripheral immunization of C57BL/6 mice with TMEV can induce K^b-restricted splenocyte killing (21). In addition, subtle variations in TMEV itself may alter immunodominance. The TMEV protein L*, which is expressed by an alternative viral reading frame, can influence mouse susceptibility to TMEV infection and subsequent demyelinating disease (8). Recently, it has been shown that intracranial infection of C57BL/6 mice with the L*-deficient TMEV strain DAL* induces CNS-ILs capable of K^b-restricted antiviral killing (23). Whether these K^b-restricted epitopes are presented at low frequency during intracranial infection of C57BL/6 mice with Daniel strain TMEV remains to be established. Nevertheless, these results demonstrate that TMEV peptides other than VP2_121-130 are potential antigens for C57BL/6 CD8⁺ T cells. Therefore, the observed dominance of VP2_121-130 as a T-cell epitope in H-2D^b-expressing mice must be regulated in some fashion.

To address how epitope dominance might be controlled, the influence of costimulation, inflammatory cytokines, perforin, or CD4 help on D^b:VP2_121-130 epitope dominance was investigated in TMEV-infected genetic knockout mice. Class II (34)-, CD4 (31)-, CD8 (31)-, IFN- (unpublished observation)-, perforin (32)-, and CD28 (unpublished observation)-deficient mice of the C57BL/6 lineage are all susceptible to TMEV persistent infection and demyelinating disease despite expression of H-2D^b. To address the effects of these inflammatory mediators which govern resistance to the virus, we assessed the frequency of VP2_121-130-specific T cells among CNS-ILs in the brains of infected mice. CNS-ILs were isolated 7 days post-TMEV infection from mice deficient for the indicated molecules and stained with D^b:VP2_121-130 tetramer and anti-CD8 antibodies. With the exception of the CD8-deficient strain (Fig. 4E), all of the other tested mouse strains developed a major population of T cells at the site of infection that recognize VP2_121-130 presented by D^b (Fig. 4A to I). Furthermore, the frequencies of these T cells in the class II, CD28, TNF-RI, IFN-, and perforin knockout mice were not significantly different from that found in infected C57BL/6 animals. This indicates that the homing and clonal expansion of the population of VP2_121-130-specific CD8⁺ lymphocytes is not affected by the absence of IFN-, TNF-RI, class II molecules, perforin, or CD28. No staining by the D^b:VP2_121-130 tetramer was observed in TMEV-infected CD8^/ mice (Fig. 4E), demonstrating that CD8 molecules are required for the development of T cells capable of recognizing VP2_121-130 in the context of D^b. The expression of CD8 molecules could influence the development of the epitope-dominant T-cell response at the level of repertoire selection, during T-cell activation, or in the process of recruitment of T cells to the brain.

View larger version (43K): [in this window] [in a new window]   FIG. 4.   Frequency of CD8+ CNS-ILs that stain for both Db:VP2121-130 tetramer and CD8 in normal C57BL/6 mice (A), MHC-II/ mice (C), IFN-/ mice (D), CD8/ mice (E), CD28/ mice (F), CD4/ mice (G), TNF-RI/ mice (H), and perforin/ mice (I). Typical background staining of CNS-ILs from TMEV-infected C57BL/6 mice with an irrelevant Db:E7 tetramer is shown in panel B. Lymphocytes were isolated 7 days post-TMEV infection. The percentage of CD8+ cells that stained with Db:VP2121-130 tetramer is given in the righthand corner. The CD8+ population that recognizes the Db:VP2121-130 epitope is circles in panels A and B but not in panels C to I.

The presence of significant populations of D^b:VP2_121-130-specific CTL at the site of TMEV infection in susceptible mice appears paradoxical at first glance. However, it is clear that the activation of CTL is not the only factor that can determine the outcome of viral infection. Lymphokines released by T cells can have both direct antiviral function and indirect influence on antiviral immunity by upregulating MHC encoded antigen-presenting molecules and by recruiting other effector cells, such as macrophages, to the site of infection. Cells harboring actively replicating TMEV could avoid the immune response in the absence of the timely upregulation of MHC gene expression or as a consequence of an inadequate cytokine response.

Of particular interest is the observation that CD4^/ mice consistently generate a lower but still appreciable (22 to 37%) frequency of D^b:VP2_121-130-specific CD8⁺ T cells in the brain. This indicates that the focus of the CD8⁺ T-cell response on the VP2 peptide is indeed regulated but the mechanism by which CD4 influences this process is not clear. If T helper cells were important in maintaining the dominant response to the VP2 epitope, then class II knockout mice would be predicted to also display diminished epitope dominance in their responses. As shown in Fig. 4C, this is not the case. The reason for this discrepancy is not known, but it may be an indication that the functions of CD4 molecules are not completely understood. Perhaps CD4 molecules perform additional functions not involving interactions with classically defined class II molecules. For example, another stage in the development of the immune response where CD4 expression could influence a CD8-restricted T-cell response is during the differentiation of immature T cells. Perhaps the absence of CD4 expression in the knockout mice at the CD4/CD8 double-positive stage in the thymus diminishes their subsequent dominance in the hierarchy of potential responding T cells to TMEV challenge. In addition, the possibility that CD4 molecules could influence immune reactivity in ways not yet appreciated must be considered.

Why the frequency of CD8⁺ T cells specific for a single epitope is so high and how this frequency resists change despite different environmental pressures remains unknown. The inflammatory mediators studied affect the kinetics of an immune response in several ways. For example, IFN- can upregulate the expression of many factors involved in antigen presentation, including class I, class II, ₂m, TAP1, TAP2, and tapasin (4). Likewise, TNF- can modify inflammation by either working in synergy with IFN- or by activating NF-B through degradation of the inhibitory subunit IB (3, 47). Class II-restricted help could also modify an inflammatory response by providing the release of cytokines. Costimulation by B7-1 and B7-2 through CD28 is important for providing a second signal in T-cell activation, perhaps even dictating whether a Th1 or a Th2 response will occur (19). Perforin plays a central role in the effector process of T-cell-mediated cell killing. The fact that the frequency of VP2_121-130 peptide-specific T cells is essentially not different in perforin-deficient and normal C57BL/6 mice, indicates that the factors leading to epitope dominance precede the effector phase of the response. Remarkably, the mechanism responsible for focusing the immune response on the VP2_121-130 peptide is not influenced by costimulation, class II-mediated help, or the effects of IFN- or TNF-.

As part of this study, we introduced a D^b transgene into the FVB strain to determine whether a population of VP2_121-130-specific CTLs could be generated in a TMEV-infected mouse with a completely different genetic background. Highly susceptible FVB mice are H-2^q and are not expected to generate D^b-restricted CTL responses. As predicted, no D^b:VP2_121-130-specific T cells were detected in the CNS-ILs of FVB mice (Fig. 5). Introduction of a D^b transgene into the FVB strain enables these highly susceptible mice to resist TMEV-induced demyelinating disease (2). A major population (55%) of CNS-ILs isolated from TMEV-infected FVB-D^b transgenic mice stained with anti-CD8 and D^b:VP2_121-130 tetramer (Fig. 5). This frequency is similar to that observed in C57BL/6 mice (50 to 63%). This finding stresses the importance of the D^b gene in the VP2_121-130-dominated T-cell response and the minimal effects contributed by other genes differing between these distantly related mouse strains.

View larger version (25K): [in this window] [in a new window]   FIG. 5.   Epitope dominance conferred by a Db transgene. The staining patterns of CNS-ILs isolated from resistant FVB (H-2Db) transgenic mice (A) and susceptible, parental FVB (H-2q) mice (B) are shown. CNS-ILs were stained with Db:VP2121-130 tetramer and anti-CD8 antibodies and analyzed by FACS. The percentage of CD8+ T cells that are double positive is given in the upper right corner of panel A.

The concept of an immunodominant peptide eliciting a strong CTL response is not new and has been described among both human and mouse viral infections (9, 17, 28, 51). The mechanism by which this occurs is not well understood, but peptide affinity, peptide processing, and the available T-cell repertoire have all been implicated as factors in establishing immunodominance hierarchies among peptides (7, 9, 10, 16, 20, 29, 33, 46, 50, 52, 53). Here we have demonstrated that the D^b:VP2_121-130 epitope-restricted clonal dominance appears to persist in the absence of key inflammatory factors such as class II, CD4, CD28, IFN-, and TNF-RI, indicating that the clonal dominance is independent of interaction with CD8⁺ T cells and the help compartment.