SJL/J mice were purchased from the Charles River Laboratories (Charles River, MA) through the National Cancer Institute (Frederick, MD). B6; 129S-Tlr3^tm1Flv /J mice (TLR3KO-B6) were purchased from Jackson Laboratories (Bar Harbor, ME). TLR3KO-B6 mice were backcrossed to SJL/J mice for 6 generations to obtain TLR3KO-SJL mice. The absence/presence of TLR3 in TLR3KO-SJL and the littermate mice (NLM) were typed based on the electrophoresis patterns of TLR3 and neomycin resistant genes. PCR products from tail genomic DNA of NLM and TLR3KO-SJL mice were determined using PCR-based genotyping analysis established by the Jackson Laboratory (Additional file 1, Figure S1). Experimental procedures that were approved by the Animal Care and Use Committee of Northwestern University in accordance with NIH animal care guidelines were used in this study.

The BeAn and GDVII strains of TMEV were propagated in BHK-21 cells grown in DMEM medium supplemented with 7.5% donor calf serum. Viral titer was determined by plaque assay on BHK cell monolayers. The cells were incubated for 4-5 days in infection-medium (DMEM supplemented with 0.1% bovine serum albumin) with TMEV at 10 MOIs and the cell lysates were cleared by centrifugation. The cleared lysates yield 3-5 × 10⁸ PFU and a pooled batch was used as a viral stock. If necessary the viral stock was diluted in DMEM before inoculation.

Approximately 30 μl of TMEV was injected into the right hemisphere of 5- to 7-week-old mice anesthetized with isofluorane. Resistant B6 and TLR3KO-B6 mice were infected with 1 × 10⁶ PFU and susceptible SJL and TLR3KO-SJL mice were infected with 2 × 10⁵ PFU TMEV. Clinical symptoms of disease were assessed weekly on the following grading scale: grade 0 = no clinical signs; grade 1 = mild waddling gait; grade 2 = moderate waddling gait and hindlimb paresis; grade 3 = severe hind limb paralysis; grade 4 = severe hind limb paralysis and loss of righting reflex; and grade 5 = death.

After cardiac perfusion with cold Hank's balanced salt solution (HBSS) (Mediatech), brain and spinal cords were removed. The tissues were homogenized in HBSS using a tissue homogenizer. A standard plaque assay was performed on BHK-21 cell monolayers 33 . Plaques in the BHK monolayer were visualized by staining with 0.1% crystal violet solution after fixing with methanol.

Mice were perfused through the left ventricle with 30 ml of sterile HBSS. Excised brains and spinal cords were forced through wire mesh and incubated at 37°C for 45 min in 250 μg/ml of collagenase type 4 (Worthington). CNS-infiltrating lymphocytes were then enriched at the bottom 1/3 of a continuous 100% Percoll (GE) gradient after centrifugation for 30 min at 27,000 × g.

CNS-infiltrating lymphocytes were isolated and Fc receptors were blocked using 100 μl of 2.4G2 hybridoma (ATCC) supernatant by incubating at 4°C for 30 minutes. The indicated antibodies were subsequently used to stain various cell types. VP3_159-166-loaded H-2K^s tetramer labeled with PE was used to assess levels of virus-specific CD8⁺ T cells in the CNS of TMEV-infected mice. Cells were analyzed using a Becton Dickinson LSRII flow cytometer.

Freshly isolated CNS-infiltrating mononuclear cells were cultured in 96-well round bottom plates in the presence of viral or control peptides and Golgi-Plug™ (BD) for 6 h at 37°C. Cells were then incubated in 100 μl of 2.4G2 hybridoma (ATCC) supernatant for 30 minutes at 4°C to block Fc receptors. Anti-CD8 (clone 53-6.7) antibody or anti-CD4 (clone L3T4) antibody was added, and cells were incubated for an additional 30 minutes at 4°C. After two washes, intracellular IFN-γ staining was performed according to the manufacturer's instructions (BD) using PE-labeled rat monoclonal anti-IFN-γ (XMG1.2) antibody. Cells were analyzed by flow cytometry.

Total RNA was isolated by TRIzol reagent (Invitrogen) and reverse transcribed to cDNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen). The cDNAs were amplified with specific primer sets using the SYBR Green Supermix (Bio-Rad) on an iCycler (Bio-Rad). The sense and antisense primer sequences used for cytokines are as follows: TMEV (VP1), (5'-TGACTAAGCAGGACTATGCCTTCC-3' and 5'-CAACGAGCCACATATGCGGATTAC-3'); IL-1β, (5'-TCATGGGATGATAACCTGCT-3' and 5'-CCCATACTTTAGGAA-GACACGGAT-3'); IFN-α, (5'-ACCTCCTCTGACCCAGGAAG -3' and 5'-GGCTCTCCAGA-CTTCTGCTC-3'); IFN-β, (5'-CCCTATGGAGATGACGGAGA-3' and 5'-CTGTCTGCTGG-TGGAGTTGA-3'); IFN-γ, (5'-ACTGGCAAAAGGATGGTGAC-3' and 5'-TGAGCTCATT-GAATGCTT GG-3'); IL-10, (5'-GCCAAGCCTTATCGGAAATGATCC-3' and 5'-AGACA-CCTTGGTCTTGGAGCTT-3'); TNF-α, (5'-CTGTGAAGGGAATGGGTGTT-3' and 5'-GGTCACTGTCCCAGCATCTT-3'); IL-6, (5'-AGTTGCCTTCTTGGGACTGA-3' and 5'-TCCACGATTTCCCAGAGAAC-3'); IL-17, (5'-GGGGATCCATGAGTCCAGGGAGAGC-3' and 5'-CCCTCGAGTTAGGCTGCCTGGCGGA-3'); CXCL10, (5'-AAGTGCTGCCGTC-ATTTTCT-3' and 5'-GTGGCAATGATCTCAACACG-3') and GAPDH, (5'-AACTTTGG-CATTGTGGAAGGGCTC-3' and 5'-TGCCTGCTTCACCACCTTCTTGAT-3'). GAPDH expression served as an internal reference for normalization. Real-time PCR was performed in triplicate.

The statistical significance of the differences between experimental groups (two-tailed p value) was analyzed with the unpaired Student's t-test using the InStat Program (GraphPAD). Comparisons of the disease courses between 2 groups were also performed using the paired t-test. Values of P < 0.05 were considered to be significant.

It has previously been reported that TLR3 plays a critical role in TMEV-induced inflammatory cytokine and chemokine responses 16 17 . To examine the role of TLR3 in the development of TMEV-induced disease, we compared the development of clinical signs and viral loads in the CNS of control B6 and TLR3-deficient B6 (TLR3KO-B6) mice following infection with the BeAn strain of TMEV (1 × 10⁶ PFU). Viral levels in the CNS of TLR3KO-B6 mice at 7 and 21 days post-infection (dpi) were significantly higher in the brain and spinal cord than the viral levels of the B6 control mice (Figure 1A). However, neither the B6 nor the TLR3KO-B6 mice developed detectable clinical signs of disease (data not shown). These results indicate that TLR3 signals are important in controlling TMEV loads in the CNS, although the increased viral levels did not lead to the development of demyelinating disease. Flow cytometric analysis of the CNS cells indicated that the level of mononuclear cells, including T cells and macrophages, that infiltrated the CNS of the TLR3KO-B6 mice were similar to those of the B6 mice at 7 dpi (Figure 1B). However, the levels of these cells in the CNS of TLR3KO-B6 mice were significantly higher than those of control B6 mice at 21 dpi. The elevated viral load in the TLR3KO mice may have caused a higher cellular infiltration into the CNS by activating higher levels of inflammatory cytokines and chemokines. To further determine the levels of virus-specific CD4⁺ and CD8⁺ T cells that infiltrated into the CNS, mononuclear cells were isolated from the CNS of TMEV-infected mice at 7 and 21 dpi, and these cells were stimulated with TMEV-specific viral epitope peptides. Subsequently, the ability of these T cells to produce IFN-γ was assessed by flow cytometry after intracellular cytokine staining (Figure 1C). The proportions of IFN-γ-producing, TMEV-specific CD4⁺ T cells and CD8⁺ T cells in the CNS were similar between TLR3KO-B6 and B6 mice.

As there were no differences in the development of TMEV BeAn-induced demyelinating disease between TLR3KO-B6 and B6 control mice, we further explored the potential differences in the susceptibility of these mice to a highly virulent GDVII strain of TMEV 34 that was administered via intraperitoneal injection (Figure 1D). After a low dose of viral infection (100 PFU), fewer than 26% of the control B6 mice developed fatal encephalitis, whereas greater than 50% of TLR3KO-B6 mice developed disease. In addition, virus-infected TLR3KO mice showed significantly higher levels of viral load in the CNS at 7 dpi compared to the infected B6 mice (Figure 1E), consistent with the differences noted in disease severity. These results indicate that TLR3-mediated signaling in resistant B6 mice plays an important role in controlling viral infection, particularly for highly virulent, encephalitic strains of TMEV.

To examine whether TLR3 plays a more prominent role in TMEV-susceptible SJL mice, we infected TLR3KO-SJL mice and normal littermates (NLM) at the 6^th generation of backcrossing to SJL/J mice with a low dose (2 × 10⁵ PFU) of TMEV BeAn. These virus-infected mice were then assessed for the progression of demyelinating disease for 80 dpi (Figure 2A). We chose the low dose of the virus to maximize the differences in disease development. Interestingly, TLR3KO-SJL mice showed exacerbated development of TMEV-induced demyelinating disease compared to the NLM. We further examined the viral loads in the CNS (brains and spinal cords) of both mouse groups at 7, 21 and 50 dpi using plaque assays (Figure 2B). The levels of infectious virus in the CNS of TLR3KO-SJL mice were significantly higher both in the brain and spinal cord compared to those for the control NLM mice. These results indicate that TLR3 signaling plays an important role in controlling viral load in the CNS and in preventing the development of TMEV-induced demyelinating disease following infection with a less virulent BeAn strain in mice of the susceptible SJL background, unlike mice of the resistant B6 background.

To compare levels of demyelination in the CNS of TMEV BeAn-infected TLR3KO-SJL and NLM SJL mice, histopathologic examinations were performed (Figure 3). First, Hematoxylin-eosin (HE) staining (Figure 3Aa and 3Ad), Kluver-Barrera's (KB) staining (Figure 3Ab and 3Ae) and immunohistochemical staining for GFAP (Figure 3Ac and 3Af) were conducted. In each experiment, mice from the NLM or TLR3KO groups were blindly selected, beforehand, for histological examination, and these mice were sacrificed at 27 dpi. The HE staining results showed that slight mononuclear cell infiltration (arrow) and mild demyelination were observed in the white matter of the spinal cord from NLM mice (Figure 3Aa and 3Ab). GFAP staining showed a lack of astrocytes in the demyelinated lesion (arrow) (Figure 3Ac). In contrast, markedly increased mononuclear cell infiltration (arrow) and extended demyelination (arrow) were observed in the white matter of the spinal cord from TLR3KO mice (Figure 3Ad and 3Ae). GFAP staining showed markedly increased number of activated astrocytes in the white matter of the spinal cords of these mice (Figure 3Af).

Next, we examined the spinal cords of TMEV-infected NLM and TLR3KO-SJL mice at days 10 and 27 post infection using immunohistochemical staining for CD3, a marker of T cells (Figure 3Ba, 3Bd, 3Bg, and 3Bj); CD45R, a marker of B cell (Figure 3Bb, 3Be, 3Bh, and 3Bk); and F4/80, a marker of macrophages (Figure 3Bc, 3Bf, 3Bi, and 3Bl). Increased T cell infiltration was observed in the white matter of the spinal cord from TLR3KO mice (Figure 3Bd and 3Bj) compared to NLM mice (Figure 3Ba and 3Bg), based on immunohistochemical staining for CD3. Few B cells were observed in the NLM (Figure 3Bb) and the TLR3KO mice at day 10 post-infection (Figure 3Be). On day 27 post-infection, B cell infiltration was increased in the white matter of the spinal cord from TLR3KO mice (Figure 3Bk) compared to NLM mice (Figure 3Bh). Similarly, macrophage infiltration was determined by staining for the F4/80 marker, and higher levels of macrophages were found in the white matter of the spinal cord from TLR3KO mice (Figure 3Bf and 3Bl) compared to NLM mice (Figure 3Bc and 3Bi) at 10 and 27 dpi.

To determine the levels of CNS-infiltrating mononuclear cells, we compared the mononuclear cells that accumulated in the CNS of NLM and TLR3K-SJL mice. The numbers of CNS-infiltrating mononuclear cells were elevated throughout the course of viral infection (days 7, 21 and 80) in TLR3KO-SJL mice compared to NLM (Figure 4A). Flow cytometric analysis of the CNS infiltrating mononuclear cells indicated that the proportions of both the CD4⁺ and CD8⁺ T cells in TLR3KO-SJL mice were significantly higher than those of the NLM group at 7 and 21 dpi (Figure 4B). The proportions of macrophages (CD11b⁺CD45^high) and neutrophils (Ly6G/6C⁺) were also higher in TLR3KO-SJL mice compared to NLM.

To understand the underlying mechanisms of exacerbated susceptibility to TMEV-induced demyelinating disease in TLR3KO-SJL mice, we compared the expression levels of TMEV RNA and cytokine genes in the CNS of virus-infected NLM and TLR3KO-SJL mice at 7 and 21 dpi (Figure 4C). The viral message level was significantly elevated at both time points in TLR3KO-SJJL mice, consistent with the higher replicating virus levels determined using plaque assays (Figure 2B). The expression of various inflammatory cytokine genes, such as type I IFNs, IL-10, TNF-α, IL-6, and IL-1 was similarly elevated in the CNS of TLR3KO-SJL mice. It was interesting to note that the expression of CXCL-10, associated with T cell infiltration to the CNS, was also highly elevated in TLR3KO-SJL mice. Because TLR3 is known to play an important role in activating the expression of these cytokine genes 16 , an increased viral load in the absence of TLR3 signaling may be sufficient to overcome the TLR3 deficiency via other receptors, such as MDA5, leading to elevated cytokine gene expression. Consequently, higher viral loads accompanied by more proinflammatory cytokines may result in elevated cellular infiltration and exacerbated development of demyelinating disease in TLR3KO-SJL mice.

To determine the levels of virus-specific T cell responses in the CNS, mononuclear cells isolated from the CNS of TMEV-infected NLM and TLR3KO-SJL mice at 7 and 21 dpi were stimulated with viral epitope peptides and assayed for the production of IFN-γ (Figure 5A-D). The proportion of TMEV-specific IFN-γ-producing CD4⁺ T cells and CD8⁺ T cells in the CNS of TLR3KO-SJL mice were consistently similar or higher than those of the NLM mice at 7 dpi. The proportion of H-2K^s-VP3_159-166-tetramer reactive CD8⁺ T cells in the CNS of TLR3KO mice was also similar to that of the NLM mice at the early stage (7 dpi) of infection, indicating the similarities in the function of virus-specific CD8⁺ T cells in both mouse groups. However, the overall numbers of virus-specific CD4⁺ and CD8⁺ T cells in the CNS were higher due to the increased cellular infiltration to the CNS in TLR3KO-SJL mice. The proportion and number of anti-viral CD4⁺ T cells became similar or lower in the mice at 21 dpi. Similarly, VP3_159-166-tetramer reactive CD8⁺ T cells were lower at 21 dpi in the TLR3KO-SJL mice, although IFN-γ-producing CD8⁺ T cells remained similar. It has previously been shown that Th17 cells are preferentially developed following TMEV infection and IL-17 promotes the pathogenesis of chronic demyelinating disease 35 . To further determine the levels of IL-17-producing T cells relative to IFN-γ-producing cells in the virus-infected mice, the overall levels of IFN-γ and IL-17 messages expressed in the CNS were assessed using real-time PCR (Figure 5E). The results confirmed the higher level of IFN-γ-producing cells observed by flow cytometry in virus-infected TLR3KO-SJL mice. In addition, the level of IL-17-producing T cells was similarly higher in TLR3KO mice compared to the control littermates. These results suggest that antiviral T cell responses are not drastically altered but rather, are elevated in TMEV-infected SJL mice in the absence of TLR3 signals.

To further examine the status of T cell activation in TLR3KO-SJL mice during early viral infection, the expression of the CD69 activation marker on T cells and MHC molecules on microglia and macrophages in the CNS of virus-infected mice were analyzed at 7 dpi by flow cytometry (Figure 5F and 5G). Levels of CD69 expression on CD4⁺ and CD8⁺ T cells were higher in the TLR3KO-SJL mice compared to NLM mice, consistent with the higher proportions of virus-specific T cells (Figure 5A and 5B). The expression levels of both MHC class I (H-2K^s) and II (I-A^s) molecules were also higher on microglia (MG) and macrophages (MP) from TMEV-infected TLR3KO mice (Figure 5G). These results suggest that early efficient T cell activation, in the absence of TLR3 signaling, may be due to the elevated expression of MHC molecules on antigen presenting cells.

To activate TLR3 signaling, susceptible SJL mice were intraperitoneally treated with poly IC, the representative TLR3 ligand, at 1 day prior to or 8 days post TMEV-infection. The progression of TMEV-induced demyelinating disease was assessed over 63 dpi. Mice treated with poly IC at 1 day prior to viral infection displayed an exacerbated development of disease, whereas mice treated with poly IC at 8 dpi resulted in a slower onset of the disease compared to virus-infected SJL mice without poly IC administration (Figure 6A). Viral message levels in the brain and spinal cord of mice pretreated with poly IC were significantly higher than those of the mice that were either untreated or treated with poly IC at 8 dpi (Figure 6B). These results indicate that the activation of TLR3 prior to viral infection leads to an increased viral load in the CNS and accelerated pathogenesis of demyelinating disease; however, such activation after viral infection does not alter the development of disease.

To further understand the immunological mechanisms of the acceleration of TMEV-induced demyelinating disease in poly IC-pretreated SJL mice, we first compared the levels of mononuclear cells accumulated in the CNS of mice at 14 and 28 dpi (Figure 6C and 6D). Flow cytometric analysis showed that the proportion of macrophages (CD11b⁺CD45^high) in the CNS of poly IC pretreatment mice was elevated, whereas the proportion in the CNS of poly IC post-treated mice remained the same as that of virus-infected mice without poly IC treatment (Figure 6C). It is interesting to note that poly IC-pretreated mice maintained the elevated macrophage level at 28 dpi, while in poly IC-post-treated mice the level decreased. Similarly, the proportion of CD4⁺ and CD8⁺ cells was higher in the CNS of the poly IC pre-treatment mice and remained higher at 28 dpi compared to the untreated virus-infected mice. However, the proportion of these T cells in the CNS of poly IC-post-treated mice was lower (Figure 6D), and these decreases appear to reflect decreased levels of viral message in the CNS (Figure 6B).

To further assess the relative levels of virus-specific T cell responses in the CNS of these poly IC treated mice, mononuclear cells isolated from the CNS of infected mice at 14 and 28 dpi were stimulated with viral epitope peptides to determine the ability of these cells to produce IFN-γ. The flow cytometry profiles of the mononuclear cells at day 28 post-infection are shown in Figure 6 (panels E and F). The proportions of both IFN-γ-producing virus-specific CD4⁺ and CD8⁺ T cells in the CNS of poly IC pre-treated mice were markedly decreased at both time points (results at day 14 post-infection not shown) compared to those of the control mice without poly IC-treatment. In contrast, the proportion of CD4⁺ and CD8⁺ T cells in the poly IC-post-treated mice were increased particularly around the onset of disease development (28 dpi). These results strongly suggest that activation of TLR3 signaling prior to viral infection hinders the induction of protective IFN-γ-producing CD4⁺ as well as CD8⁺ T cell populations. In contrast, activation of these signals after viral infection appears to improve the induction of IFN-γ-producing CD4⁺ as well as CD8⁺ T cells.

To further determine whether the decrease of IFN-γ-producing T cells in poly IC-pretreated mice (Figure 6E and 6F) reflects the inability of antigen presenting cells to stimulate T cell responses in the CNS of poly IC pre- or post-treated SJL mice, expression levels of CD69, an activation marker of CD4⁺ and CD8⁺ cells, were compared at 14 and 28 dpi (Figure 7A). Overall, the expression levels of CD69 on CD4⁺ and CD8⁺ cells were similar among the untreated control and the poly IC pre- and post-treated mice at both 14 and 28 dpi, although the expression of CD69 on CD8⁺ T cells of poly IC-post-treated mice was somewhat lower. These data suggest that the decrease in IFN-γ-producing T cell responses in poly IC-pretreated mice does not reflect the status of T cell activation in the CNS.

To verify the status of antigen presenting cells in the CNS, we examined the expression levels of MHC classes I and II, and CD40 molecules, which are associated with T cell activation, on the major antigen-presenting CD11b⁺ cells, including microglia and macrophages (Figure 7B and 7C). The expression levels of MHC class I (H-2K^s) and II (I-A^s) molecules on CD11b⁺ cells in the poly IC pretreated mice were markedly increased, whereas the levels in the poly IC post-treated mice were decreased compared to those in the untreated control virus-infected mice. These data suggested that the poor IFN-γ-producing T cell responses were not due to a deficiency in the expression of molecules associated with antigen presentation. It was also interesting to note that the levels of both T cell activation and expression of MHC and CD40 molecules appeared to correlate with viral load in the CNS.

To further explore possible mechanisms underlying the poor IFN-γ-producing T cell responses in poly IC-pretreated mice, we assessed the expression levels of PDL-1, an inhibitory molecule for both CD4⁺ and CD8⁺ T cell responses 36 , on CD11b⁺ cells in the CNS (Figure 7D). We chose PDL-1 as a candidate inhibitory molecule because this molecule is known to play a critical role in anti-viral T cell functions in virus-infected hosts, and the expression of PDL-1 is inducible by activation of TLR3 with poly IC treatment 36 37 . The expression levels of PDL-1 on CD11b⁺ cells were drastically increased in the poly IC pretreated mice at both 14 and 28 dpi, whereas the expression was markedly decreased in post-treated mice compared to those of TMEV-infected control mice without poly IC treatment. These results strongly suggest that the compromise in the immune response of mice treated with poly IC prior to virus infection was due in part to the over-expression of the inhibitory PDL-1 molecule on antigen presenting cells rather than deficiencies in the activation of T cells.

It is possible that the poor immune responses in the poly IC-pretreated mice may also have been associated with the induction of a higher level of regulatory FoxP3⁺ CD4⁺ T cells (Treg), which are known to inhibit the function of anti-viral T cell responses 38 39 40 41 42 . To examine this possibility, levels of Foxp3 expressing CD4⁺ T cells in the CNS of control mice and mice treated with poly IC were assessed at 14 and 28 dpi (Figure 7E). The level of Treg cells in the CNS of poly IC-pretreatment mice was significantly higher, particularly at the preclinical stage (14 dpi) compared to that of untreated control mice. In contrast, the Treg levels in mice treated with poly IC following viral infection were similar to the untreated control mice. These results suggest that an elevated induction of FoxP3⁺ Treg cells may also partially contribute to the low T cell response in poly IC-pretreated mice.