|Home | About | Journals | Submit | Contact Us | Français|
Induction of antigen-specific CD8+ T cells bearing a high-avidity T-cell receptor (TCR) is thought to be an important factor in antiviral and antitumor immune responses. However, the relationship between TCR diversity and functional avidity of epitope-specific CD8+ T cells accumulating in the central nervous system (CNS) during viral infection is unknown. Hence, analysis of T-cell diversity at the clonal level is important to understand the fate and function of virus-specific CD8+ T cells. In this study, we examined the Vβ diversity and avidity of CD8+ T cells specific to the predominant epitope (VP2121-130) of Theiler's murine encephalomyelitis virus. We found that Vβ6+ CD8+ T cells, associated with epitope specificity, predominantly expanded in the CNS during viral infection. Further investigations of antigen-specific Vβ6+ CD8+ T cells by CDR3 spectratyping and sequencing indicated that distinct T-cell clonotypes are preferentially increased in the CNS compared to the periphery. Among the epitope-specific Vβ6+ CD8+ T cells, MGX-Jβ1.1 motif-bearing cells, which could be found at a high precursor frequency in naïve mice, were expanded in the CNS and tightly associated with gamma interferon production. These T cells displayed moderate avidity for the cognate epitope rather than the high avidity normally observed in memory/effector T cells. Therefore, our findings provide new insights into the CD8+ T-cell repertoire during immune responses to viral infection in the CNS.
Theiler's murine encephalomyelitis virus (TMEV) is a member of the Cardiovirus genus within the Picornaviridae family (43). This virus is a common enteric pathogen among wild mice but rarely causes neurological disease (57). However, when it infects susceptible mice (e.g., the SJL/J [SJL] strain) intracerebrally, it reproducibly induces a chronic immune-mediated demyelinating disease that has been studied as an infectious model of human multiple sclerosis (MS) (10, 30). In contrast, infection of resistant mice like those of the C57BL/6 (B6) strain results in strong antiviral immune responses that clear the virus effectively and prevent disease development (24, 31). Therefore, immune responses in B6 mice have been often compared to those in susceptible SJL mice to understand the nature of protective versus pathogenic immunity in these mice.
It has been shown that the major histocompatibility complex (MHC) H-2D locus is a critical genetic factor for resistance to TMEV-induced demyelinating disease (9, 49). For example, expression of the H-2Db transgene makes susceptible FVB mice resistant by inducing strong H-2Db-restricted VP2121-130-specific CD8+ T-cell responses (36). This acquired resistance is abolished when VP2121-130-specific T cells are tolerized by introducing the VP2 transgene (45). These results strongly suggest that CD8+ T cells generated in the presence of H-2Db are critical for viral clearance from the central nervous system (CNS). Since the cardinal difference between the resistant B6 and susceptible SJL strains is the quantity, not the quality, of virus-specific CD8+ T cells (23, 32), strong CD8+ T-cell responses are probably required to prevent viral persistence and the consequent development of demyelinating disease. More than threefold more virus-specific CD8+ T cells were found in the CNSs of resistant B6 mice than in those of susceptible SJL mice at the acute phase of infection. Thus, the level of virus-specific CD8+ T cells at an early phase of the immune response may be a critical factor in resistance to the disease.
Many recent investigations indicate that oligoclonal CD8+ T cells accumulate in the CNSs of MS patients (4, 38, 51). In addition, CD8+ T cells may also induce the development of experimental autoimmune encephalomyelitis (EAE) (54). Therefore, clonal expansion of certain CD8+ T cells may be associated with the pathogenesis of demyelinating diseases. However, B6 mice, which are resistant to TMEV-induced demyelinating disease, induce strong CD8+ T-cell responses to a single predominant epitope (VP2121-130), i.e., ≥70% of CNS-infiltrating CD8+ T cells (41, 42). These CD8+ T cells result in effective viral clearance yet remain at a low level in the CNS more than 120 days postinfection (dpi) without detectable pathology (42). This inconsistency led us to investigate the shape and quality of the T-cell receptor (TCR) repertoire accumulating in the CNSs of B6 mice.
The CD8+ T-cell responses induced after viral infection have previously been investigated with other animal viruses, including influenza virus, lymphocytic choriomeningitis virus (LCMV), mouse hepatitis virus (MHV), and Borna disease virus (11, 14, 35, 47, 58). Among these models, the detailed T-cell Vβ repertoire in the CNS was described only in the MHV model (46). CD8+ T-cell responses against TMEV in B6 mice are primarily against a single predominant epitope (22, 36, 41). However, virtually no study of the TCR Vβ repertoires of virus-specific CD8+ T cells has been reported. Furthermore, it is not yet known whether a particular TCR Vβ repertoire is associated with the avidity and/or function of CD8+ T cells in the CNS. Since protective versus pathogenic CD8+ T cells may correlate with their Vβ repertoire and T-cell function, these studies may help to elucidate the underlying mechanisms of protection versus pathogenesis of CD8+ T cells in the CNS.
In this study, we have addressed several important questions about the CD8+ T-cell repertoire in the CNS. First, what is the pattern of Vβ usage in TMEV-infected B6 mice? Second, are there differences in the antigen-specific CD8+ T-cell clonotypes between the CNS and periphery? Third, are the T-cell clonotypes maintained in the CNS during the viral infection? Fourth, what is the functional avidity of T cells accumulating in the CNS during this virus infection? Last, what possible factors are associated with repertoire selection and expansion in the CNS? Our results show that Vβ6+ CD8+ T cells preferentially expand in the CNS during viral infection. Further analyses of the CDR3 region of antigen-specific Vβ6+ CD8+ T cells by spectratyping and sequencing indicate that distinct T-cell clonotypes are expanded in the CNS compared to those in the periphery. T cells expressing a particular Vβ6-CDR3-Jβ1.1 sequence are preferentially retained in the CNS during the course of viral infection. Interestingly, these T cells are capable of producing gamma interferon (IFN-γ) upon stimulation and display moderate avidity for the cognate epitope. We believe that our findings will provide important information regarding the CD8+ T-cell repertoire during viral infection and that these results may help to provide a better understanding of antiviral CD8+ T-cell immunity in the CNS.
C57BL/6, SJL/J, and B6.S mice (Jackson Laboratories, Bar Harbor, ME) and 129S2/SP mice (National Cancer Institute/Charles River) were housed in the Center for Comparative Medicine, Northwestern University, Chicago, IL. Under the guidelines of the Animal Care and Use Committee, 6- to 8-week-old female mice were inoculated by intracerebral injection with 3 × 106 PFU of the BeAn strain of TMEV in 30 μl of Dulbecco's modified Eagle's medium.
All synthetic peptides purified by high-performance liquid chromatography to >95% purity were obtained from Genemed Synthesis, San Francisco, CA. These include the TMEV VP2121-130 (FHAGSLLVFM) (6, 13) and VP2M130L (FHAGSLLVFL) peptides (41).
H-2Db tetramers were generated as previously described (3). Briefly, the H-2Db and human β2-microglobulin genes were subcloned into the pET28 bacterial expression vector. BL21/DE3 was transformed, and protein expression was induced with isopropyl-β-d-thiogalactopyranoside (IPTG) for 4 h. Inclusion bodies were purified and refolded in the presence of peptides. Soluble H-2Db peptide was biotinylated with BirA at room temperature. Excess biotin was removed by ultrafiltration and then tetramerized with a streptavidin-phycoerythrin (PE) conjugate (Invitrogen, Carlsbad, CA). The H-2Db-LCMV-GP33 tetramer was purchased from Beckman Coulter (Fullerton, CA).
Mice were anesthetized with isoflurane and perfused with 30 ml of cold Hanks balanced salt solution. Brains and spinal cords were removed, minced with steel mesh, and treated with collagenase IV and DNase I (Sigma) for 45 min at 37°C. Mononuclear cells were isolated after 100% Percoll continuous gradient centrifugation at 27,000 × g for 30 min as previously described (16).
Freshly isolated CNS-infiltrating mononuclear cells were cultured in 96-well round-bottom plates in the presence of relevant peptide or phosphate-buffered saline (PBS) and Golgi-Stop for intracellular cytokine staining as previously described (50). Allophycocyanin-conjugated anti-CD8 (clone Ly2) or anti-CD4 (clone L3T4) antibody with PE-labeled rat monoclonal anti-IFN-γ (XMG1.2) antibody was used for intracellular cytokine staining. Cells were analyzed on a Becton Dickinson FACScalibur or LSRII cytometer. Live cells were gated based on light scatter properties.
The generation of mutant viruses was previously described (41). Briefly, viral RNA was transcribed in vitro from a full-length WT (pSBW) or mutant (pSBW-VP2M130L) viral plasmid using T7 RNA polymerase in the presence of RNasin (Promega, Madison, WI). Viral RNA was transfected into BHK cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Supernatant was collected 5 to 7 days posttransfection and used to infect fresh BHK cells to obtain a high-titer (~2 × 109 PFU/ml) viral stock.
IFN-γ-secreting cells were isolated using an enrichment kit (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instruction (33). Briefly, CNS and spleen cells from virus-infected mice were stimulated in vitro for 4 to 6 h in the presence of 2 μM VP2121-130. Cells were then labeled with an affinity matrix for secreted IFN-γ (catch reagent). After 45 min of incubation at 37°C, the IFN-γ bound to the catch reagent was stained with a PE-conjugated IFN-γ-specific antibody (detection reagent). IFN-γ+ T cells were isolated by magnetic cell sorting using anti-PE microbeads.
Total RNA was purified using TRIZOL reagent (Invitrogen, Carlsbad, CA), and then cDNA was generated from the RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen, Carlsbad, CA) and oligo(dT)18. Relative concentrations of cDNA were estimated on the basis of β-actin amplification levels following PCR for 25 to 30 cycles. Vβ chain usage was assessed by PCR using a sense primer specific to different Vβ genes and a common antisense Cβ primer (Table (Table11).
CDR3 length spectratyping was performed as previously described, with minor modifications (25, 44). Primary amplification of cDNAs from isolated antigen-specific T cells was performed for 35 cycles using Vβ-Cβ primer pairs at 58°C for annealing in the presence of PCR enhancer solution (GIBCO, Invitrogen Corporation, Grand Island, NY). The primary PCR product (1 μl) was subjected to a secondary PCR with a 32P-end-labeled Vβ chain primer and 12 different Jβ primers (Table (Table1).1). The secondary PCR for Vβ-Jβ combinations was carried out for 25 cycles at 60°C for annealing in the presence of 30 mM (NH4)2SO4 to avoid cross-amplification. The radioactive PCR products were separated on a 5% acrylamide sequencing gel. CDR3 lengths were then determined after exposure to X-Omat LS film (Kodak, Rochester, NY) or by using a phosphorimager (Bio-Rad Personal FX).
CNS-infiltrating mononuclear cells from mice at 8 and/or 16 dpi were restimulated in vitro with 10-fold serial dilutions of peptides starting at 1 μM. Levels of CD8+ T-cell responses were determined based on IFN-γ production by intracellular cytokine staining. Avidity thresholds were set as previously described (53): <10−3 μM, high avidity; 10−1 to 10−3 μM, moderate avidity; >10−1 μM, low avidity.
To analyze individual CDR3 spectratype bands, DNA was eluted from each gel band and reamplified with appropriate primer pairs for 30 cycles. The PCR products were cloned into the pGEM-T easy vector (Promega, Madison, WI) and sequenced (Macrogen Co., Seoul, Korea, or Northwestern University genomic core facility). CDR3 size was defined as the number of residues between the aligned C (CASS) in the Vβ gene element and the FGXG in the Jβ region minus four (25).
Primary Vβ6 PCR products from CNS mononuclear cells and splenocytes from infected mice or naïve splenocytes were used to determine their Jβ chain and CDR3 usage. Briefly, 1:500- to 1:1,000-diluted primary PCR products were reamplified by real-time PCR in SYBR green PCR mix (Bio-Rad). Amplification of the CAS conserved region to the Cβ region or only the Cβ region was used as an internal control. For Jβ chain analysis, a Vβ6 primer and 12 different Jβ chain primers were used. For CDR3 analysis, degenerate primers for CDR3 motifs corresponding to the same amino acid sequences were used (Table (Table11).
Data are presented as the mean ± the standard deviation of either two or three independent experiments or one representative result with triplicates from at least three independent experiments. Significance of differences was determined by Student's t test. P values of <0.05 were considered statistically significant.
The majority of CD8+ T cells in the CNSs of TMEV-infected B6 mice are reactive to the predominant VP2121-130 epitope. However, the clonal nature of CD8+ T cells remains unknown. To analyze the epitope-specific CD8+ T cells, the proportions of the epitope-specific cells in the CNS and periphery were analyzed at 8 days after TMEV infection using flow cytometry after staining with the H-2Db-VP2121-130 (Db-VP2) tetramer (Fig. (Fig.1A)1A) or intracellular IFN-γ (Fig. (Fig.1B).1B). Our Db-VP2 tetramer staining was specific for VP2-specific CD8+ T cells: a negative control Db-LCMV-GP33-41 tetramer did not react with CD8+ T cells from TMEV-infected mice (Fig. (Fig.1A),1A), nor did the Db-VP2 tetramer react with CNS cells from mice with myelin oligodendrocyte glycoprotein-induced EAE (data not shown). As shown previously (17, 60), the proportion of virus-specific CD8+ T cells was 7- to 10-fold higher in the CNS than in the spleen. Greater than 80% of the CNS-infiltrating CD8+ T cells and greater than 10% of the splenic CD8+ T cells from infected mice reacted to Db-VP2 tetramers, while approximately one-half of the Db-VP2 tetramer-positive cells produced IFN-γ. Therefore, the great majority of CD8+ T cells observed in the CNS after TMEV infection appear to be virus specific.
We further assessed the relative proportions of different Vβ-bearing, VP2121-130-specific CD8+ T cells in the CNSs and spleens of virus-infected mice at 8 dpi using flow cytometry after staining for Vβs, CD8, and Db-VP2. CNS-infiltrating CD8+ T cells from three individual mice showed predominant Vβ6 expression ranging from 23.4 to 41%. Splenic VP2-specific CD8+ T cells showed a similar predominance of Vβ6 usage (29 to 49.3%). Interestingly, the percentage of Vβ6-bearing VP2-specific T cells (36.7%, on average) was similar to that of Vβ8.1/2-bearing cells (32.7%, on average) in the spleen. However, a significantly reduced level (about twofold) of Vβ8.1/2-bearing cells (18.9%, on average) was found in the CNS, in contrast to Vβ6-bearing cells (36.5%, on average). These results strongly suggest that VP2-specific CD8+ T cells are differentially accumulated in the CNS in a Vβ-associate manner. However, proportions of dominant Vβ6+ cells within virus-specific CD8+ T cells remain similar in the CNS and periphery of TMEV-infected B6 mice.
Since Vβ6+ CD8+ antigen-specific T cells are predominantly found in the CNS, Vβ6+ T cells may be preferentially maintained in the CNS, perhaps in conjunction with local antigen-presenting cells (APCs). To test this possibility, the clonal nature of Vβ6+ CD8+ T cells was assessed by Vβ6-CDR3-Jβ spectratyping of isolated tetramer-positive cells and IFN-γ-producing cells at 8 dpi (Fig. (Fig.2).2). As shown in Fig. Fig.2A,2A, the Vβ6 repertoires from tetramer-binding cells in the CNS were distinct from those in the spleen. For example, among Vβ6-Jβ1.3, Vβ6-Jβ1.4, and Vβ6-Jβ1.5, the CDR3 size spectra of CNS cells were different from those of splenic cells, although Vβ6-Jβ1.1 and -Jβ1.2 were similar. Since similar CDR3 spectra may not reflect identical clonotypes, selective samples with the same-length CDR3 regions of epitope-specific CD8+ T cells from the CNS and the spleen were further sequenced (Fig. (Fig.2A,2A, parts a and b). For example, 7-amino-acid (aa) bands in the Vβ6-CDR3-Jβ1.1 spectra from the CNS and spleen were composed of different amino acids. In contrast, the Vβ6-CDR3-Jβ1.1 bands from IFN-γ-producing CD8+ T cells (Fig. (Fig.2B,2B, parts c and d) appear to selectively utilize Vβ6-MGX-TEVF in both the CNS and the spleen. Nevertheless, it appears that the CDR3 repertoire of tetramer-positive CD8+ T cells is more heterogeneous than that of IFN-γ-producing cells.
Vβ6-Jβ size determinations using PCR (Fig. (Fig.2)2) may misrepresent the T-cell repertoires due to the differential amplification by various Jβ primers. To examine this possibility, primary Vβ6-Cβ PCR products from the above-described experiments were directly sequenced after cloning into plasmids (Table (Table2).2). It is interesting that the frequency (38%) of CD8+ T cells with Vβ6-MGX-TEVF is significantly higher in the CNS than in the spleen (13%). In addition, the great majority of T cells reactive to this epitope in both the CNS and the spleen appear to utilize Vβ6-XGX-TEVF. Again, the level of VP2-specific tetramer-reactive CD8+ T cells with the XGX motif was higher in the CNS (81.4%) than in the spleen (60.8%), and this trend was even higher with VP2-specific IFN-γ-producing cells (91.1% in the CNS versus 76.9% in the spleen). Furthermore, tetramer-positive CD8+ T cells from the CNSs (74.5%) of TMEV-infected B6 mice preferentially utilized Jβ1.1 in conjunction with Vβ6, compared to those from their spleens (45.1%). However, only limited clonotypes were shared between CD8+ T cells in the CNS and spleen (e.g., CDR3 motifs paired with Jβ1.1, as indicated by asterisks). In contrast, IFN-γ-producing CD8+ T cells displayed similarly predominant Vβ6-Jβ1.1 (42%) and Jβ2.6 (42%) usages (Table (Table2).2). Interestingly, CDR3 clonotypes (indicated with double asterisks) containing MGX paired with Jβ1.1 (38% in CNS versus 27% in spleen) and MGEQY paired with Jβ2.6 (25% in CNS versus 19% in spleen) were dominant in both the CNS and the spleen. Therefore, the CDR3 region of virus-specific CD8+ T cells producing IFN-γ appears to be associated with the MGX CDR3 sequence. Since the usage of these CDR3 sequences by tetramer-positive cells is less frequent, IFN-γ-producing CD8+ T cells with this motif may represent subpopulations within the tetramer-positive cells (boxed in Table Table22).
To examine the potential expansion and/or contraction of the TCR repertoire during viral infection, CDR3-Jβ regions of CNS-infiltrating T cells were assessed by spectratyping at 4, 8, 16, and 24 dpi (Fig. (Fig.3).3). Since the number and percentage of specific cells in the CNS and spleens are too low at the early stage of virus infection, isolation of tetramer-positive cells was not feasible. Moreover, the Vβ6-positive population is the predominant VP2-specific CD8+ T-cell type (Fig. (Fig.1)1) and Vβ6+ CD4+ T cells are only a minor population in the CNSs of virus-infected mice (data not shown). Therefore, we determined the emerging Vβ6-CDR3 patterns among total Vβ6+ T-cell populations in the CNSs and spleens during viral infection to represent alterations in the magnitude of TCR CDR3 regions. As early as 4 dpi, preferential Jβ1.1 usage by Vβ6-bearing T cells in the CNS was evident when determined by real-time PCR (Fig. (Fig.3A).3A). The relative frequencies of other Jβ usages (3 mice per group) were somewhat different from the sequencing results in Table Table22 from 20 mice. In order to examine whether or not the inconsistency of other Jβ usages reflects the variation of individual mice, we further examined the Jβ usages in 4 individual mice. While the predominant Jβ1.1 usage was consistent (4/4 mice), the usages of other Jβs were not (Fig. (Fig.3B).3B). Therefore, preferential Jβ usage other than Jβ1.1 appears to be somewhat variable, depending on individual mice or experiments. The size distribution of Vβ6-CDR3-Jβ1.1 was further analyzed by spectratyping (Fig. (Fig.3C).3C). The CDR3 sizes of splenic T cells from naïve mice showed a distribution with the strongest peak at 9 aa (Fig. (Fig.3C,3C, first row). Interestingly, the predominance of the 7-aa CDR3 TCR spectrum, including Vβ6-MGX-Jβ1.1, became apparent in the CNS at 4 dpi and continued through the viral infection (Fig. (Fig.3C).3C). However, this clonotype peaked in the spleen at 8 dpi and waned at 16 dpi to levels indistinguishable from that of normal splenocytes, probably reflecting the low proportion of virus-specific CD8+ T cells, when the viral load was significantly decreased in the CNS (Fig. (Fig.3D3D).
Relative levels of CDR3 regions utilized by IFN-γ-producing VP2121-130-specific cells during viral infection were determined by real-time PCR using degenerate primers encoding deduced Vβ6-CDR3-Cβ amino acid residues containing MGN, MGF, MGE, RV, and ITPT (Fig. (Fig.4A).4A). The frequency of MGN and MGE CDR3 sequences representing dominant IFN-γ-producing virus-specific CD8+ T cells was significantly higher in the CNS at 4 dpi and continued to increase, 6,000-, 50,000-, and 8,000-fold over the level of the infrequently used ITPT sequence at 8, 16, and 24 days, respectively. The dominance of MGN and MGE was consistent in a separate experiment (data not shown). These data suggest that VP2121-130-specific CD8+ T cells with CDR3 region sequences of MGN and MGE dominantly accumulated in the CNS. It is interesting that the frequency of MGN and MGE sequences is significantly higher (P < 0.01 and P < 0.001, respectively) in naïve splenic T cells than MGF, RV, and ITPT sequences (Fig. (Fig.4B).4B). These results suggest that the dominant utilization of CDR3 regions with MGN and MGE motifs by virus-specific IFN-γ-producing CD8+ T cells represents the selection from T cells with high precursor frequency in naïve B6 mice.
To further assess the Jβ distribution within MGN and MGE Vβ6-CDR3-Cβ motifs, the corresponding primary PCR products were analyzed with individual Jβ primers (Fig. (Fig.4B,4B, parts b and d). Both MGN and MGE motifs were primarily used in conjunction with Jβ1.1 at 8 and 16 dpi, consistent with the sequencing results (Table (Table2).2). However, the use of other Jβs varied, depending on the individual mice (Fig. (Fig.3B)3B) or experiments (data not shown). These results indicate that Vβ6-MGX-Jβ1.1 is a stable dominant public repertoire and CD8+ T-cell clones bearing this CDR3 motif are preferentially accumulated in the CNS during viral infection. However, Vβ6-MGX-Jβ2.1/Jβ2.6 may be either semipublic or transient repertoires and thus CD8+ T cells bearing these CDR3s vary among different individual mice.
A high-affinity/avidity TCR is an important factor for both antiviral and antitumor CD8+ T-cell functions (1, 2). However, the relationship of functional avidity to the TCR diversity of CD8+ T cells accumulated in the CNS during viral infection is unknown. We have previously shown that the native epitope (WT, VP2121-130) and an altered peptide ligand (APL; M130L [L substituted for M at position 130 of VP2121-130]) induce similar levels of IFN-γ production at a high concentration (41). However, severalfold higher concentrations of the M130L peptide are required to induce levels of IFN-γ similar to those of the WT peptide at low concentrations, suggesting that this APL exhibits a relatively lower functional avidity (unpublished observation and Fig. Fig.5A).5A). To further explore the relationship between functional avidity and TCR diversity, we determined the ability to produce IFN-γ in response to various concentrations of the WT or M130L peptide at two different time points after viral infection. The functional avidity of CD8+ T cells against the WT peptide appears to be very high (2 × 10−6 μM) but heterogeneous among the T cells at 8 dpi (Fig. (Fig.5A).5A). However, the average avidity (8 × 10−3 μM) of the T cells was reduced to a moderate level (P < 0.001) at 16 dpi. In contrast, the avidity of CD8+ T cells against M130L APL seems to be moderate at both 8 and 16 days (4 × 10−3 and 8 × 10−3 μM, respectively, P > 0.05). The proportions of both WT and M130L tetramer-reactive CD8+ T cells increased at 16 days, while the overall numbers of WT and M130L epitope-reactive cells declined. Interestingly, the number of WT epitope-reactive cells decreased more rapidly than that of M130L-reactive cells (Fig. (Fig.5B).5B). These results suggest that virus-specific CD8+ T cells with moderate functional avidity are preferentially retained in the CNS during viral infection. Because T cells with a high-affinity TCR are known to undergo activation-induced apoptosis (37), T cells with moderate avidity may have a survival advantage under continuous viral antigenic stimulation in the CNS.
To exclude the possibility that either a low number or deficiency of APCs in the CNSs of virus-infected mice affected the measurements of the functional avidity of T cells, we further determined functional CD8+ T-cell avidity with or without additional splenic dendritic cells (DCs) from naïve B6 mice. If the number and/or function of APCs from the CNSs of virus-infected mice were compromised, the deficiencies would be overcome by the presence of excess numbers of APCs. However, the levels of IFN-γ production were similar in the presence or absence of added DCs across different concentrations of the cognate peptide (data not shown). The lack of significant deficiencies in the CNS APCs from virus-infected resistant B6 mice is consistent with our previous observation (20). Therefore, it is very unlikely that the differences in functional avidity presented in Fig. Fig.5A5A reflect deficiencies in APC function or numbers.
Since Vβ6+ CD8+ T cells reactive to M130L APL display moderate avidity and are preferentially retained in the CNS (Fig. 5A and B), we further compared the binding intensities of WT- and M130L-loaded tetramers to Vβ6- and/or Vβ8.1/2-bearing CD8+ T cells from mice at 8 dpi (Fig. (Fig.5C).5C). Interestingly, similar numbers of Vβ6+ CD8+ T cells displayed high- and low-intensity binding to the WT tetramer, whereas greater than 80% of the Vβ6+ CD8+ T cells showed low-intensity binding to the M130L tetramer (Fig. (Fig.5C,5C, left). These differences in the tetramer reactivity of Vβ6+ CD8+ T cells are also apparent in the respective mean fluorescence intensities of tetramer binding. However, such differential reactivity was not observed with subdominant Vβ8.1/2+ CD8+ T cells reactive to these tetramers (Fig. (Fig.5C,5C, right). The overall tetramer-binding intensity of Vβ8.1/2+ CD8+ T cells was much higher than that of Vβ6-bearing cells. The relationship between tetramer binding intensity and functional avidity is unclear since these are not tightly associated with each other (12, 52). Nevertheless, these tetramer binding results are consistent with the above-described functional assessment of avidity via IFN-γ production using epitope peptides (Fig. (Fig.5A),5A), supporting the idea that VP2-specific CD8+ T cells from WT virus-infected mice display intermediate avidity toward APLs.
We further compared Vβs utilized by CD8+ T cells reactive to WT- and M130L-loaded tetramers at 8 and 16 dpi by using reverse transcription (RT)-PCR (Fig. (Fig.5D).5D). Vβ6 dominance was observed in CD8+ T cells reactive to M130L at both 8 and 16 dpi, similar to WT epitope-reactive CD8+ T cells. The nucleotide sequences of CDR3-Jβ1.1 spectratyping bands equivalent to 7 aa residues of WT- and M130L-reactive cells were further analyzed to examine the potential association of a moderate-avidity TCR and the MGX motif (Fig. (Fig.5E).5E). As expected, the majority of sequences showed the MGX motif in both WT- and M130L-reactive CD8+ T cells (7/11 and 10/11, respectively) at 16 dpi, when T cells with moderate avidity are predominant in the CNS. Taken together, these results strongly suggest that virus-specific CD8+ T cells with moderate avidity display a unique CDR3 motif, are capable of producing IFN-γ, and are preferentially retained in the CNS for a prolonged time period.
It is known that the affinity of viral peptide/MHC for TCRs could affect the magnitude of antigen-specific CD8+ T-cell responses (28, 59). Since the percentage of M130L-reactive cells increased in the CNS during viral infection (Fig. (Fig.5B),5B), we examined the possibility that M130L-reactive cells expand more efficiently in vivo following infection with M130L-containing virus. At 8 dpi, we compared the magnitudes of CD8+ T cells and Vβ repertoires by using WT- and M130L-loaded tetramers (Fig. (Fig.6A).6A). Similar proportions of Vβ6+ (35 to 42%) and Vβ8.1/2 CD8+ T cells (10 to 20%) from both WT and M130L virus-infected mice recognized WT and M130L tetramers. These data strongly suggest that virus expressing the WT ligand most efficiently induces CD8+ T cells reactive not only to WT but also to M130L epitopes.
Despite the induction of a predominant Vβ6+ VP2-reactive CD8+ T-cell response by the mutant viruses, it is conceivable that these APL-bearing viruses may preferentially induce Vβ6+ CD8+ T cells with different CDR3 motifs. To examine this possibility, relative levels of MGX CDR3 in mice infected with the WT and M130L viruses were assessed at 8 dpi using real-time PCR (Fig. (Fig.6B).6B). To our surprise, M130L virus infection failed to induce a vigorous response of CNS-infiltrating T cells with MGN and MGE CDR3 (<10-fold) compared to the WT virus. However, T cells induced by the M130L virus maintained a similar hierarchy of Vβ6 TCRs with the CDR3 motifs. These results strongly suggest that moderate-avidity T cells bearing the MGX motif (Fig. (Fig.1,1, ,4,4, and and5;5; Table Table2)2) expand more efficiently in the CNS in response to WT virus expressing the unmodified ligand. Thus, the failure of APL-bearing virus to induce vigorous Vβ6+ VP2-reactive CD8+ T cells may be attributable to the inefficient induction of T cells bearing the predominant MGX motif.
We speculated that the Vβ6 dominance of VP2-specific CD8+ T cells reflects the high precursor frequencies of Vβ6-bearing VP2-specific T cells in naïve mice. To examine this possibility, we further determined the precursor frequency of Vβ6+ CD8+ T cells with the MGX CDR3 sequences in naïve mice to understand the potential mechanisms of a high frequency of VP2-reactive CD8+ T cells bearing this motif (Fig. (Fig.7A).7A). Relative frequencies of the dominant Vβ6-associated MGX motifs (MGN, MGF, and MGE) and two rare motifs (RV and ITPT) in splenic T cells from naïve B6 (H-2b), SJL (H-2s), and B6.S (H-2s in the B6 background) mice were assessed using real-time PCR (Fig. (Fig.7A).7A). Surprisingly, levels of MGX expression were similarly high in all of these naïve mice, in contrast to the low-to-undetectable levels of rare motifs. These results suggest that the precursor frequency of T cells with this motif is intrinsically high in mice, regardless of their MHC haplotypes. This observation is consistent with a very recent report indicating that T-cell repertoire diversity and response magnitude are associated with the precursor frequency of naïve antigen-specific T cells (39).
Since the levels of MGX expression are similarly high in mice with different MHC haplotypes, TMEV infection may cause nonspecific expansion of T cells with the MGX motif due to the presence of an unknown viral superantigen rather than the VP2 epitope. In order to discern whether or not expansion of the Vβ6 MGX repertoire is an epitope-specific, MHC-dependent process, B6 and SJL mice were infected with TMEV and their CNS-infiltrating T cells were analyzed at 8 dpi (Fig. (Fig.7B).7B). We reasoned that SJL mice expressing a limited Vβ repertoire (one-half the number of B6 Vβs) would better utilize Vβ6 if Vβ6 MGX dominance were not specific for VP2/H-2Db stimulation. Despite the high precursor frequency of the MGX motif in both naïve B6 and SJL mice, T cells with the MGX motif (particularly Vβ6-MGN and Vβ6-MGE) were expanded only in B6 mice bearing H-2b after virus infection (Fig. (Fig.7B).7B). Interestingly, the expansion patterns were very similar in H-2b-bearing B6 and 129 mice (Fig. (Fig.7C).7C). In particular, the preferential Jβ1.1 usages of these mice were mirror images of each other although the overall magnitude was higher in 129 mice. These results indicate that the expansion of CD8+ T cells with the Vβ6-MGX CDR3 region and Jβ1.1 restriction is both viral epitope and H-2b haplotype dependent.
In this study, we have analyzed the clonal nature of virus-specific CD8+ T cells accumulating in the CNSs of resistant B6 mice during infection with neurotropic Theiler's virus. Our results indicate that Vβ6+ CD8+ T cells are the predominant virus-specific population in the CNS. Similar preferential expansions of Vβ-restricted CD8+ T cells were reported following infection with many other viruses (8, 11, 29, 46, 48). Interestingly, the skewed Vβs are different, depending on the viruses, suggesting that the preferential expansion of CD8+ T cells reflects epitope-dependent responses. Alternatively, skewed expansion of Vβ6+ T cells after infection with TMEV may reflect the involvement of a superantigen (5, 19). However, this is unlikely since the preferential expansion of CD8+ T cells is dependent on the presence of a particular CD8+ T-cell epitope. For example, B6-P1 (virus capsid antigens) transgenic mice, which are tolerant to the CD8+ T-cell epitope (42), displayed negligible levels of CD8+ T cells including Vβ6+ populations in the CNS after TMEV infection (unpublished data). Furthermore, the precursor frequency of Vβ6-MGX CDR3 regions that are specific for the epitope is high in naïve mice regardless of the MHC haplotype, but these motif-bearing cells expanded only in response to the epitope in conjunction with H-2b (Fig. (Fig.7B).7B). Therefore, preferential use of the Vβ6 TCR may reflect the combination of a high precursor frequency of T cells with the Vβ-associated CDR3 sequences and utilization of the CDR3 region by the epitope-specific CD8+ T cells. Interestingly, however, preferential Vβ6 usage was largely maintained after infection with APL-bearing virus, suggesting that such skewing is due to the collective property of the epitope rather than a particular amino acid residue within the epitope (Fig. (Fig.66).
We have observed that CD8+ T cells accumulated in the CNSs of virus-infected resistant B6 mice displayed a skewed Vβ6 usage with distinct CDR3 sequences compared to CD8+ T cells in the spleen (Fig. (Fig.11 and and2).2). Our results differ from those found in previous studies of influenza virus or LCMV infection indicating that T-cell Vβ repertoires are similar in systemic and local sites (11, 29). However, CD8+ T cells induced in monkeys following simian immunodeficiency virus infection utilize heterogeneous Vβs with some degree of predominance in the brain (34). Similarly, infection with neurotropic MHV induced epitope-specific CD8+ T cells with a preferential Vβ in the CNS (46). Thus, virus-specific CD8+ T cells with restricted Vβs may represent a unique property of the CNS compartment. Alternatively, differences in the precursor frequencies of CD8+ T cells specific for different viruses may result in different patterns of Vβ skewing. Furthermore, it is conceivable that increases in the proportion of VP2-specific CD8+ T cells in the CNS (Fig. (Fig.1)1) may reflect infiltration and perhaps local amplification of virus-specific CD8+ T cells at the site of the reservoir of neurotropic TMEV.
CD8+ T cells producing IFN-γ have commonly been utilized to enumerate antigen-specific CD8+ T cells because the numbers of IFN-γ-producing cells and tetramer-positive populations are similar, as shown in mice infected with influenza virus and LCMV (15, 40). However, recent studies of HIV infection have demonstrated that there are distinct CD8+ T-cell populations producing IFN-γ and/or granzyme B within the tetramer-positive CD8+ T-cell population (18, 27). No attempts have previously been made to elucidate whether these functionally distinct populations represent different T-cell clonotypes or reflect different differentiation stages of the same T-cell population. Our present study strongly suggests that the IFN-γ-producing epitope-specific CD8+ T-cell subpopulation displays distinct TCR repertoires; hence, these cells express a unique Vβ6-CDR3 motif within the VP2121-130 tetramer-positive CD8+ T cells (Fig. (Fig.22 and Table Table2).2). To the best of our knowledge, this is the first example demonstrating that there are preferred TCR clonotypes of CD8+ T cells producing IFN-γ. Although the underlying mechanism of restricted TCR clonotypes associated with IFN-γ production is unknown, CD8+ T cells with certain TCR-dependent functional avidity may be able to preferentially trigger IFN-γ production or induce the differentiation of a particular type of CD8+ T cells similar to that of CD4+ T cells (7, 55, 56).
The avidity of TMEV epitope-specific CD8+ T cells in the CNS during the peak immune responses appears to be relatively high, whereas the avidity during the subacute phase following the peak responses is intermediate (Fig. (Fig.5).5). Since the expansion of CD8+ T cells with moderate avidity toward the ligand correlates with an increase in T-cell numbers with a particular TCR CDR3 motif during subacute phase of TMEV infection (Fig. (Fig.4),4), a subpopulation of T cells with restricted TCRs may be preferentially accumulated/retained in the CNS during late infection. Alternatively, prolonged antigenic stimulation during the subacute phase may lead to functional exhaustion of CD8+ T cells by upregulating the expression of inhibitory receptors. However, stimulation of T cells isolated from the CNS during late infection with a cognate epitope peptide, anti-CD3/CD28 antibodies, or phorbol myristate acetate/ionomycin (bypassing TCR signal cascade) resulted in similar levels of IFN-γ-producing CD8+ T cells. Furthermore, differences in the expression of neither PD-1 nor TIM3, which are known to be involved in CD8+ T-cell desensitization, were detected during the course of viral infection (data not shown). Therefore, CD8+ T cells with high and moderate avidities may be initially selected from the mixed TCR pool but T cells with moderate avidity are preferentially retained during the later phase in the presence of CD4+ T-cell help (26).
It is interesting that the dominance of CD8+ T cells with intermediate avidity in our study highly contrasts with the previous observation that the functional avidity of CD8+ T cells specific for LCMV is drastically increased in the periphery during the course of viral infection (52). In addition, it has recently been shown that high-affinity CD8+ T cells more likely expand and survive as memory cells rather than very-low-affinity cells, although very-low-affinity T cells specific for microbes also manage similar initial activation and proliferation (59). However, it is conceivable that CD8+ T cells with very high avidity with efficient cytotoxic function may undergo apoptosis and be removed from the CNS in order to avoid a potential pathogenic outcome. Furthermore, a recent report suggests that lower TCR avidity, resulting in shorter contact time, is more beneficial in attacking multiple target cells compared to CD8+ T cells with higher avidity (21). Therefore, T cells with moderate avidity may be able to efficiently control persistent virus without pathogenic function.
Our results indicate that intermediate-avidity CD8+ T cells specific for TMEV preferentially expand in the CNSs of virus-infected mice (Fig. (Fig.5).5). However, it is uncertain whether intermediate-avidity cells can be efficiently converted into memory CD8+ T cells in the CNS. Therefore, it would be interesting to compare the TCR repertoires of memory CD8+ T cells in the CNS and periphery to correlate them with avidity differences in future studies. In addition, susceptible mice fail to induce strong initial CD8+ T-cell responses to efficiently clear viral loads from the CNS, in contrast to resistant mice. Further studies on differences in CD8+ T cells between resistant B6 and susceptible SJL mice may also help us to understand the nature of antiviral CD8+ T-cell responses in the CNS, the site of chronic viral infection. These studies may ultimately provide the means to intervene in persistent viral infections and the pathogenesis of virally induced CNS inflammatory diseases.
We thank Ben Haley, Stacy Ryu, and Heeyoung Yang for their help in preparing the manuscript.
This work was supported by grants from the National Institutes of Health (RO1 NS28752 and RO1 NS33008) and the National Multiple Sclerosis Society (RG 4001-A6).
We have no financial conflict of interest.
Published ahead of print on 13 January 2010.