Drak2 Contributes to West Nile Virus Entry into The Brain and Lethal Encephalitis

Abstract

Drak2, a member of the death-associated protein family of serine/threonine kinases, is specifically expressed in T and B cells. In the absence of Drak2, mice are resistant to experimental autoimmune encephalomyelitis due to a decrease in the number of cells infiltrating the central nervous system (CNS). In the present study, we investigated the role of Drak2 in West Nile virus (WNV)-induced encephalitis and found that Drak2^−/− mice were also more resistant to lethal WNV infection than wild-type mice. Although, Drak2^−/− mice had an increase in the number of IFN-γ-producing T cells in the spleen after infection, viral levels in the peripheral tissues were not significantly different between these two groups of mice. In contrast, there was a reduced viral load in the brains of Drak2^−/− mice, which was accompanied by a decrease in the number of Drak2^−/− CD4⁺ and CD8⁺ T cells in the brain following WNV infection. Moreover, we detected viral antigens in T cells isolated from the spleen or brain of WNV infected mice. These results suggest that following a systemic infection, WNV might cross the blood brain barrier and enter the CNS by being carried by infected infiltrating T cells.

Introduction

West Nile virus (WNV), a vector-borne neurotropic flavivirus, has caused annual outbreaks of viral encephalitis in North America since its arrival here in 1999 (1, 2). Following initial virus replication at a site peripheral to the central nervous system (CNS), WNV can enter the brain either via a hematogenous pathway or by retrograde axonal transport from a peripheral site (3–5). Severe neurological disease (including encephalitis) and death have been observed in more than 30% of the confirmed WNV cases with a higher frequency in the elderly and immunocompromised (1, 2). Currently, no vaccines are available for humans, and treatment is mostly nonspecific and supportive.

The murine model is an effective in vivo experimental model to investigate WNV pathogenesis and host immune response to human infections (6–8). Work from animal models has demonstrated that type 1 interferons (IFNs), γδ T cells and humoral immunity are critical in controlling dissemination of WNV (9–15). In addition, CD4⁺ (16) and CD8⁺ αβ T-cells (17, 18) contribute to host survival following WNV infection. WNV can gain access to the CNS after a brief viremia in the periphery, a process called neuroinvasion, that may turn a mild viral infection into severe lethal encephalitis (6, 19, 20). Recent studies have shown that a WNV replication induces a proinflammatory response that modulates the blood brain barrier (BBB) permeability, which in turn may enable viral entry into the brain and induce lethal encephalitis (5, 21). Nevertheless, the underlying mechanism(s) in which WNV crosses the BBB and enters the CNS are not yet clearly understood. Once inside the brain, WNV-induced CNS disease might be caused by neuronal degeneration, a direct result of viral infection, and/or by bystander damage from the immune response to the pathogen, including lymphocyte and microglial cell responses (18, 22–24). The contributions of direct cell damage by WNV and damage induced by the host immune response in the development of disease are still under investigation.

DAP kinase-related apoptosis inducing kinase-2 (Drak2), is specifically expressed in T and B cells (25, 26). Drak2 functions to negatively regulate signals involved in T cell activation (27). T cells from Drak2^−/− mice exhibited enhanced sensitivity to T cell receptor-mediated stimulation with a reduced requirement for co-stimulation (26). However, Drak2^−/− mice were shown to be remarkably resistant to experimental autoimmune encephalomyelitis (EAE), and this resistance was due in part, to a reduction of infiltrating cells into the CNS (26). In the present study, we investigated the role of Drak2 in WNV induced encephalitis.

Materials and Methods

Mice

Drak2^−/− mice were bred on the C57BL/6 (B6) background and had been fully backcrossed (26). Littermates were used as controls. Experiments were performed with 6- to 14-wk-old animals. Groups were age- and sex-matched for each experiment and were housed under identical conditions. All animal experiments were approved by the Animal Care and Use Committee at Colorado State University.

Infection in mice

We inoculated mice intraperitoneally (i.p.) with 800 plaque-forming units (PFU) (a dose close to LD₁₀₀) of WNV isolate 2741 in 200 μl of phosphate-buffered saline (PBS) with 5% gelatin. For intracranial (i.c.) infection, we anesthetized mice and infected them with 160 PFU of WNV isolate 2741 in 20 μl PBS with 5% gelatin. Infected mice were monitored twice daily for morbidity, including lethargy, anorexia and ataxia.

Quantitative PCR (Q-PCR) for determining viral load, T cell levels and cytokine production

Blood, spleen, and brain tissues from control and experimental mice were collected at indicated days post-infection. H36.12j cells (a mouse macrophage cell line) were obtained from the American Type Culture Collection (Manassas, VA) and incubated on 12-well plates (1 × 10⁵ cells/well) for 24h at 37 °C and then infected with WNV (multiplicity of infection = 1). Cells were harvested at day 1 post-infection. Viral RNA was extracted using RNeasy extraction (Qiagen, Valencia, CA). The extracted RNA was eluted in a total volume of 50 sl of RNase-free water. Two hundred fifty nanograms of each extracted RNA sample was used to synthesize cDNA using the ProSTAR First-strand RT-PCR kit (Stratagene, Cedar Creek, TX). The sequences of the primer-probe sets for WNV envelope gene (WNVE) and PCR reaction conditions were described previously (5, 28). Probes contained a 5′ reporter, FAM, and a 3′ quencher, TAMRA (Applied Biosystems, Foster City, CA). The assay was performed on an iCycler (Bio-Rad, Hercules, CA). The sequences of the primer sets for CD4, CD8, and CXCL10 genes were described in previous publications (10, 29). Q-PCR analysis for these genes was performed using RT² Real-Time^™ SYBR Green/Fluorescein PCR master mix (Superarray, Frederick, MD). To normalize the samples, the same amount of cDNA was used in a Q-PCR for β-actin. The ratio of the amount of amplified gene compared with the amount of β-actin cDNA represented the relative levels in each sample.

ELISA

Microtiter plates were coated with recombinant WNV-E protein expressed in Drosophila melanogaster S2 cells (30) overnight at 4°C at 100 ng/well in coating buffer [0.015 M Na₂CO₃, 0.03 M NaHCO₃, and 0.003 M NaN₃ (pH 9.6)]. Sera from infected mice were diluted from 1/40 or 1/100 in PBS with 2% BSA, added to the duplicate wells, and incubated for 1 h at room temperature. Plates were washed three times with PBS-Tween (PBST). Alkaline phosphatase-conjugated goat anti-mouse IgG or IgM (Sigma-Aldrich, St. Louis, MO) at a dilution of 1/1000 in PBS-T was added for 1h at room temperature. After washing three times with PBS-T, color was developed with p-nitrophenyl phosphate (Sigma-Aldrich) for 10 min and intensity determined at an absorbance of 405 nm using a spectrophotometer.

Isolation of brain leukocytes

Brain leukocytes were isolated at day 11 post-infection based on a previously described method (31). Prior to harvest, extensive cardiac perfusion was done using PBS to deplete intravascular leukocytes. Brains were collected and homogenized. The cell homogenates were centrifuged and resuspended in 7 ml PBS with 2% fetal bovine serum mixed with 3 ml of 90% percoll (Sigma-Aldrich) in PBS. The suspension was next underlaid with 1 ml of 70% Percoll in RPMI and centrifuged at 800 × g for 20 min at 22°C. Leukocytes at the interface were harvested and counted.

T cell purification from spleen and brain

Single cell suspensions of T cells were made from spleens of WNV-infected mice or brain leukocytes by a positive selection method, using anti-CD90 magnetic beads (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s instructions.

In vitro T cell infection and virus titration

Purified splenic T cells were incubated on 96-well plates (2 × 10⁵ cells/well) for overnight at 37 °C and infected with WNV (multiplicity of infection = 0.5) for 1h. Cells were washed after infection and incubated at 37 °C. Supernatant was harvested at days 2, 3 and 5 post-infection. Virus titers were determined by plaque assay. Briefly, Vero cells were seeded in 6-well plates in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Sigma, St. Louis, MO) 24h before infection. Serial dilutions of supernatant from infected T cell culture were added and incubated for 1h. Subsequently, DMEM containing 1% low-melting-point agarose was added and the plates were incubated for 4 days. A second overlay of 2.5ml 1% agarose-medium containing 0.01% neutral red was added to visualize plaques. Virus concentrations were determined as PFU/ml.

Flow cytometry

Freshly isolated splenocytes were stained with antibodies specific for CD4, CD8 (BD Biosciences, San Diego, CA) and CXCR3 (R&D, Minneapolis, MN). Isolated brain leukocytes were stained with antibodies for cell surface markers, including CD3, CD4, CD8 and CD45 (BD Biosciences). After staining, cells were fixed with 0.5% paraformaldehyde in PBS and examined using a Coulter XL flow cytometer (Beckman Coulter, Fullerton, CA). Dead cells were excluded on the basis of forward and side light scatter. Data were analyzed using FCS express 2 (De Novo Software Ontario, Canada).

Intracellular cytokine staining

To measure cytokine production, splenocytes from WNV-infected mice were isolated and then stimulated at 2.5 × 10⁶ cells/ml with 50 ng/ml PMA (Sigma-Aldrich) and 500 ng/ml ionomycin (Sigma-Aldrich) for 5 h at 37°C. Golgi-plug (BD Biosciences) was added during the final 3.5h. The cells were then harvested, stained with antibodies (BD Biosciences) for TCRαβ and CD8α and fixed in 2% paraformaldehyde. The cells were then permeabilized with 0.5% saponin before adding Phycoerythrin (PE)-conjugated anti-IFN-γ (clone XMG 1.2) or control PE-conjugated rat IgG₁ (BD Biosciences). Cells were analyzed using a Coulter XL instrument as described above.

Cytometric bead array (CBA)

At days 2 and 4 post-infection, sera were collected from mice and measured for IFN-γ production using a mouse Th1/Th2 cytokine kit (BD Biosciences, which includes assays for IL-2, IL-4, IL-5, IFN-γ, and TNF-α) using a FACSArray analyzer (BD Biosciences).

Immunofluorescence microscopy

Mice were transcardially perfused with PBS, and brains were placed in 4% paraformaldehyde in PBS overnight at 4 °C, and, in the case of cryoimmunohistochemistry, cryoprotected with 10% (w/v) followed by 20% (w/v) and then 30% (w/v) sucrose in PBS for 24h each before embedding in Optimal Cutting Temperature compound. Subsequently, specimens were processed and histological slides were prepared for staining with various antibodies. For paraffin sections, we processed antigen retrieval at 90°C for 30 min in 10% (v/v) target retrieval solution (Dakocytomation, Denmark) or by treatment with 88% formic acid for 8 min. The endogenous peroxidase activity was quenched in 0.3% H₂O₂ for 30 min at room temperature. We performed the staining using the Vectastain Elite ABC kit coupled to the 3′-3 diaminobenzadine substrate (Vector Laboratories, Burlingame, CA). Rat antibody specific for mouse CD45 (clone YW 62.3, Serotec, Raleigh, NC), or mouse ascitic fluid WNV antibodies were used. Sections were digitized with Kodak scientific imaging software (Eastman Kodak, Rochester, NY). For splenic and brain T cell staining, 80,000 cells were fixed by submerging them in acetone at −20°C for 30 min. Cells were stained and antigens detected with PE-Cy5 conjugated antibodies to CD8 or CD4 and the flavivirus E protein-specific monoclonal antibody 4G2 ((32), 1:150 dilution) for 1.5 h at 37°C followed by biotinylated anti-mouse IgG (1:200 dilution) for 1.5 h and streptavidin-fluorescein isothiocyanate (GE Healthcare Bio-Sciences Corp., Piscataway, NJ) for 30 min at 37°C.

Statistical analysis

Survival curve comparisons were performed using Prism software (GraphPad Software, San Diego, CA) statistical analysis, which uses the log rank test (equivalent to the Mantel-Haenszel test). Values of p for viral burden, cytokine production, antibody titer and T cell number experiments were calculated with a non-paired Student’s t test or Mann-Whitney test.

Results

Drak2^−/− mice are more resistant to WNV infection

Drak2^−/− mice were previously shown to be more resistant to EAE than wild-type mice due, in part to a decrease in the number of cells infiltrating the CNS (26). WNV infection also induces the migration of lymphocytes into the CNS (10, 18). In order to examine if Drak2 is involved in WNV-induced encephalitis, we challenged Drak2^−/− mice and wild-type mice intraperitoneally i.p. with 800 p.f.u. of WNV and monitored the mice twice daily for survival. Drak2^−/− mice were more resistant (47% survival) to lethal WNV infection than wild-type controls (16.7% survival) (Fig. 1, P < 0.05). Throughout infection, the kinetics and magnitude of viremia in Drak2^−/− mice were not significantly different from that of wild-type controls (Fig. 2A, P > 0.05). Although viral load was slightly reduced in spleens of Drak2^−/− mice at day 2 (Fig. 2B, P < 0.05), no difference was noted between these two groups at the later stage of viral infection (days 4 and 6, P > 0.05). Nevertheless, there was a nearly sixty-fold reduction in viral load in the brains of Drak2^−/− mice at day 6 (peak of infection in the brain) and seven to eightfold reduction at day 9 following WNV challenge (Fig. 2C, P < 0.05). These data suggest that Drak2 plays a role in the lethality and affects the virus levels in the CNS following WNV challenge.

Figure 1.

Drak2^−/− mice are more resistant to lethal WNV infection. Wild-type and Drak2^−/− mice were infected with a dose close to an LD₁₀₀ of WNV and monitored twice daily for mortality. Data shown are pooled from three independent experiments. P = 0.02 < 0.05 for wild-type (n = 18) vs. Drak2^−/− mice (n = 17).

Figure 2.

Viral load analysis in wild-type and Drak2^−/− mice. Viral load was determined in blood (A), spleen (B), and brain (C) of wild-type and Drak2^−/− mice at the indicated days following WNV infection using Q-PCR. The y-axis depicts the ratio of the amplified WNV-E cDNA to β-actin cDNA of each sample (unitless ratio ± SEM). 5–8 samples were collected at each condition. P < 0.05 compared to wild-type mice. Data shown were representative of two independent experiments.

Humoral responses remain normal in Drak2^−/− mice upon WNV infection

Drak2 is specifically expressed in T and B cells (26). B cell-mediated humoral immune responses are critical for the host defense against disseminated infection by WNV (14, 33). Therefore, we next measured WNV-specific IgM and IgG levels in the sera from Drak2^−/− and wild-type mice following primary WNV infection. At days 5 and 9 post-infection, we did not detect significant differences between the two groups in regard to either the IgM or IgG response (Figs. 3A & 3B, P > 0.05). These data suggest that the absence of Drak2 does not influence the development of antibody following WNV infection.

Figure 3.

Humoral responses remain normal in Drak2^−/− mice upon WNV infection. Sera were collected from wild-type or Drak2^−/− mice at the indicated days during WNV infection. The development of specific IgM (A) or IgG (B) antibodies to WNV was determined after incubating sera with absorbed purified r-WNV-E protein. Data shown are representative of three similar experiments and are the average of four mice in each experiment per time point, performed in duplicate.

There are more IFN-γ-producing splenocytes from Drak2^−/− mice after ex vivo stimulation

Cellular immunity plays an important role in host survival during WNV infection (8, 13). To understand the role of Drak2 in regulating peripheral T cell responses during WNV infection, we analyzed IFN-γ production of splenic CD4⁺ and CD8⁺ T cells from WNV infected mice using an ex vivo intracellular cytokine assay. As shown in Fig. 4A, at day 5 post-infection, more splenic CD4⁺ T cells in Drak2^−/− mice produced IFN-γ upon ex vivo stimulation with PMA and ionomycin than those of wild-type mice, while there was not a difference in the number of IFN-γ-producing CD8⁺ T cells at this time. However, at day 9 post-infection, there was a two-fold increase in the number of CD4⁺ and CD8⁺ T cells that produced IFN-γ in Drak2^−/− mice compared to wild-type mice (Fig. 4A). These results suggest that Drak2^−/− T cells respond to WNV infection; in fact, there is an increase in the number of cells that produce IFN-γ upon ex vivo stimulation compared to wild-type mice during WNV infection. Surprisingly, we noted no significant differences in production of IFN-γ (Fig. 4B, P < 0.05) or other Th1/Th2 cytokines (data not shown here) in sera of these two groups of mice at days 2 and 4 post-infection.

Figure 4.

IFN-γ production in wild-type and Drak2^−/− mice following WNV infection. A, Splenocytes were isolated from WNV-infected wild-type and Drak2^−/− mice (4 samples per condition) at days 5 and 9 post-infection and were cultured ex vivo with PMA plus ionomycin, as described in Materials and Methods, and stained for IFN-γ, CD4 or CD8. Total number of IFNγ⁺CD4⁺ (left panel) and IFNγ⁺CD8⁺ (right panel) is shown. Data presented are representative of two similar experiments. P < 0.05 compared to wild-type mice. B, Sera were collected from wild-type or Drak2^−/− mice at the indicated days during WNV infection. IFN-γ production was measured using mouse Th1/Th2 cytokine kit. Data shown are representative of two independent experiments and are the average of four mice in each experiment per time point.

There are fewer Drak2^−/− T cells in the brain following WNV infection

In the EAE model, Drak2^−/− mice exhibited a reduced number of infiltrating cells in the spinal cord (26), suggesting a role for Drak2 in T cell infiltration into the CNS. A hallmark of WNV encephalitis is the accumulation of inflammatory infiltrates, which reportedly varies between brain regions and consists predominantly of lymphocytes and macrophages (10). Thus, to study T cell infiltration into the CNS, we first measured CD4⁺ and CD8⁺ T cell levels in the brain by Q-PCR analysis. In the brains of Drak2^−/− mice, we observed more than a two-fold reduction of CD4⁺ T cells (Fig 5A, left panel) at day 9 post-infection and at least a two-fold reduction of CD8⁺ T cells (Fig. 5A, right panel) at days 6 and 9 post-infection, compared to wild-type mice. To verify these results, we isolated brain leukocytes at day 11 post-infection. As shown in Fig. 5B, the percentage of CD3⁺CD4⁺ T cells in the brains of Drak2^−/− mice was lower than in those of wild-type mice (2.8% vs. 7.3%, Drak2^−/− vs. wild-type). The percentage of CD3⁺CD8⁺ among Drak2^−/− mice brain leukocytes was also reduced (7.0% vs. 19%, Drak2^−/−vs. wild-type). The total number of CD4⁺ and CD8⁺ T cells in the brains of Drak2^−/− mice was also more than two-fold lower compared to wild-type mice (Fig. 5D, P < 0.05). Consistent with these results, the percentage of CD45⁺ CD3⁺ leukocytes in Drak2^−/− mouse brains was lower than in those of wild-type mice (Fig. 5C). Interestingly, no differences were noted when we compared the number of CD45⁺CD3⁻ cells in the brains of Drak2^−/− mice and wild-type mice (Fig. 5D). These results further suggest that Drak2 is important for T cell infiltration into the CNS upon WNV infection.

Figure 5.

There are fewer Drak2^−/− T cells in the CNS following systemic WNV infection. A, CD4 (left panel) and CD8 levels (right panel) in WNV infected mice brains were measured by Q-PCR at indicated days. The y-axis depicts the ratio of the amplified CD4 or CD8-cDNA to β-actin cDNA of each sample (unitless ratio ± SEM). P < 0.05 compared to wild-type mice. Data shown are one representative of three independent experiments. Four samples were collected from each condition. (B) & (D), Numeric reduction of CD4⁺ and CD8⁺ T cells in the brains of wild-type and Drak2^−/− mice at day 11 post-infection. (C) & (D), Number of CD45⁺CD3⁻ cells in the brains of Drak2^−/− mice and wild-type mice was not different. At least three samples were collected at each condition. Experiments were repeated four times.

T cells are potential carriers for WNV entry into the CNS

Earlier Q-PCR analyses revealed a significantly reduced viral load in Drak2^−/− mice brains at days 6 and 9 following WNV challenge (Fig. 2C). At day 13 post-infection, immunofluorescence double-staining of wild-type mouse brains (olfactory bulbs, cerebellum and brain stem) showed either numerous CD45⁺ leukocytes (Fig. 6A) or CD4⁺ T cells (Fig. 6B) in the vicinity of WNV-positive cells, but in the same regions of Drak2^−/− mouse brains the staining of both WNV antigen and T cells was less obvious. These data indicate that reduced T cell migration into the CNS is correlated with decreased viral levels in the brains. To test whether non-T cells in the CNS contribute to difference of susceptibility to WNV infection between wild-type and Drak2^−/−mice, we next infected mice i.c. with 160PFU WNV isolate 2741. As shown in Fig. 6C, viral load in the brains of Drak2^−/− mice was not significantly different from that of wild-type controls around the peak of infection (day 5) (P > 0.05). Therefore, it is possible that WNV enters the CNS through infiltrating T cells. To test this possibility, we determined whether peripheral T cells were permissive to WNV infection. Following i.p. challenge, splenic T cells were isolated at day 6 post-infection. The purity of these cells was close to 90% as analyzed by flow cytometry (Fig. 7A). Q-PCR analysis for viral load of the purified T cells revealed a similar level of infection as the day 1 WNV infected H36.12j cells (Fig. 7B). Immunofluorescence staining (Fig. 7C) also demonstrated WNV antigens in these cells. Further, we isolated brain-infiltrating T cells from wild-type mice at day 10 post-infection and stained for WNV antigen and CD4 or CD8 T cell markers. Immunofluorescence double-staining shows CD4⁺WNV⁺ and CD8⁺WNV⁺ populations in these purified brain T cells (Fig 7D). Plaque assay showed WNV replicated productively in purified splenic T cells from both groups of mice at day 2 and 3 post-infection, but decreased at day 5, suggesting Drak2^−/− T cells supporting a short-term productive virus replication as efficiently as wild-type T cells (Fig. 7E). Overall, these results indicate that infected T cells might carry WNV as they migrate into the brains, thereby serving as one of the pathways for WNV entry into the CNS.

Figure 6.

T cell migration into the CNS is correlated with enhanced viral load in the brains following systemic WNV infection. A & B, Wild-type (n = 4) and Drak2^−/− mice (n = 4) mice were infected with WNV, brains were isolated at day 13 post-infection, and immunostained with antibodies against CD45 (A, red signal) or CD4 (B, red signal) and WNV antigen (green signal) to reveal infiltrating leukocytes or T cells in WNV infected brain regions (olfactory bulbs, cerebellum, and brain stem sections). DAPI was used as a nuclear counterstain (blue signal), and merged images are shown. C, Viral load in wild-type and Drak2^−/− mice. Brains were harvested at day 5 following i.c. infection. P = 0.84 > 0.05 for wild-type (n = 12) vs. Drak2^−/− mice (n = 12).

Figure 7.

T cells are permissive to WNV infection. A, Splenocytes from WNV infected mice were harvested at day 6 post-infection. Splenic T cells were purified with CD90 microbeads, followed by positive selection with an autoMACS Separator (Miltenyi Biotec). These T cells were stained with antibody to CD3 and analyzed by flow cytometer. Open area represents purified T cells stained with anti-CD3; gray area, unstained purified T cells. B, WNV infection in purified splenic T cells of infected mice as measured by Q-PCR. Negative and positive controls represent splenic T cells isolated from naïve mice and WNV infected H36.12j cells (MOI = 1, day 1 post-infection) respectively. The y-axis depicts the ratio of the amplified WNV-E cDNA to β-actin cDNA of each sample (unitless ratio ± SEM). C, Immunofluorescence photomicrographs of purified splenic T cells stained for WNV antigen with various magnifications. D, Immunofluorescence photomicrographs of purified brain T cells isolated at day 10 post-infection. T cells were isolated by positive selection and double-stained for WNV antigen and CD4 (top panel) or CD8 (bottom panel). Arrows point to CD4⁺WNV⁺ and CD8⁺WNV⁺ populations. E, In vitro WNV infection in purified splenic T cells isolated from wild-type (n = 4) and Drak2^−/− mice (n = 4) as measured by plaque assay.

Drak2 affects chemokine expression during WNV infection

Interactions between chemokines and their receptors is an important step in the control of leukocyte migration into sites of inflammation. Neurons are the primary target for WNV in the CNS (23). Infected neurons secrete chemokines, such as CXCL10, which recruits effector T cells via the chemokine receptor CXCR3 (31) (10). To investigate whether Drak2 plays a role in chemokine signaling, we analyzed CXCL10 levels in the brain following WNV infection. At day 4 post-infection, there was a two-fold reduction of CXCL10 mRNA in the brain of Drak2^−/−mice as measured by Q-PCR. This difference was more dramatic at day 6 post-infection (Fig 8A). In addition, we found that the percentage of splenic CD8⁺ T cells expressing CXCR3 was slightly lower in Drak2^−/− mice than wild-type mice at day 5 post-infection; whereas the percentage of CD4⁺ T cells expressing CXCR3 did not differ in mice of these two groups (Fig. 8B left panel). Furthermore, the total numbers of CD4⁺ T cells (20.7 ± 1.2 × 10⁶ vs. 13.9 ± 0.5 × 10⁶, wild-type vs. Drak2^−/−, P < 0.05) and CD8⁺ T cells (15.6 ± 0.8 × 10⁶ vs. 9.8 ± 0.8 × 10⁶, wild-type vs. Drak2^−/−, P < 0.05) were higher in wild-type mice than those of Drak2^−/− mice. The total numbers of CD4⁺ and CD8⁺ spleen T cells expressing CXCR3 were also higher in wild-type mice than in Drak2^−/− mice (Fig. 8B right panel).

Figure 8.

Drak2 affects chemokine expression during WNV infection. A, CXCL10 levels in the brains of infected mice at indicated time points as measured by Q-PCR. The y-axis depicts the ratio of the amplified CXCL10 cDNA to β-actin cDNA of each sample (unitless ratio ± SEM). P < 0.05 compared to wild-type mice. Data shown are representative of three independent experiments. n = 4 for each group. B, Peripheral CXCR3 expression was reduced on peripheral T cell surface in Drak2^−/− mice as compared to wild-type mice. Splenocytes were stained with CD4, CXCR3 or CD8. Percent (left panel) and total number (right panel) of CD4⁺CXCR3⁺ or CD8⁺CXCR3⁺ are shown. In each experiment, three to four mice per group were analyzed. Similar experiments were repeated twice. *A value of P < 0.05 compared to wild-type mice.

Discussion

The mechanism(s) by which WNV evades the immune system in susceptible organisms and enters the CNS is not completely understood. In this study, we have investigated the role of Drak2 in WNV-induced encephalitis. We demonstrated that Drak2^−/− mice were more resistant to lethal WNV infection than wild-type controls. In the peripheral tissues, virus infection was not significantly different between wild-type and Drak2^−/− mice. A defect in Drak2 signaling did not affect the development of antibodies following WNV infection. Although more Drak2^−/− T cells isolated from WNV-infected mice produced IFN-γ upon ex vivo stimulation, no significant difference in IFN-γ production was found in the sera of mice in these two groups at days 2 and 4 post-infection. These data indicate that Drak2 is not involved in the control of virus dissemination in the peripheral tissues. Nevertheless, we noted a drastically reduced viral load in the brains of Drak2^−/− mice, compared to that in wild-type controls. Interestingly, there was also a reduction in CD4⁺ and CD8⁺ T cell infiltrates in the brains of Drak2^−/− mice. Together, these results indicate a correlation between the number of T cells in the brain and virus dissemination in this organ. Moreover, we demonstrated that upon WNV challenge, T cells from either a peripheral organ or from brain infiltrates contained WNV antigens. Overall, our data suggest that Drak2 is important for maintaining CD4⁺ and CD8⁺ T cells in the CNS following a systemic WNV challenge, and may also be involved in WNV entry into the CNS.

There are several possible explanations for the reduction of T cells in the brain of Drak2^−/− mice. One possibility is that Drak2 is required for T cell migration into the CNS. However, a recent report by Ramos et al. (34) revealed that Drak2^−/− T cells migrate to the brain following mouse hepatitis virus (MHV) infection. However, in this experiment, the virus was injected directly into the brain, and the migration of T cells into the CNS may be mediated by different mechanisms when physical trauma is induced. Nevertheless, when Drak2^−/− mice were infected i.p. with lymphocytic choriomeningitis virus (LCMV), a similar number of T cells were present in the brain compared to wildtype mice early after infection. However, by thirteen days post-infection, there were fewer T cells in the brain of Drak2^−/− mice, and of the T cells that were present, a higher percent were undergoing apoptosis (McGargill MA, and Hedrick SM. et. al. Manuscript submitted). Likewise, during EAE, there were similar numbers of Drak2^−/− and wild-type T cells early after the induction of disease, however the Drak2^−/− T cells were more sensitive to apoptosis, and therefore did not accumulate and induce disease similar to the wild-type T cells. These data suggest that T cells can migrate to the CNS in the absence of Drak2, however, the cells are more sensitive to death, and therefore do not remain in this organ. Although, we did detect a difference in chemokine expression in the brain, and a slight reduction in chemokine receptor on T cells, this may be secondary to the fact that the T cells do not remain in the brain in the absence of Drak2. It is possible that after systemic WNV infection, T cells carry virus into the brain, which results in the infection of neurons and upregulation of CXCL10, and in turn, amplifies the recruitment of T cells to the CNS. However, in Drak2^−/− mice, T cells do not remain in the CNS, which may limit the opportunity of virus to spread to the neurons, and therefore there is not an upregulation of chemokines. It is interesting to speculate whether there are fewer T cells in the CNS of Drak2^−/− mice due to a decrease in viral antigen in the brain, or whether there is a decrease in viral antigen in the CNS as a result of the reduction of T cells in this organ. Regardless, there is a decrease in WNV-induced lethality in the absence of Drak2, which is likely due to both the decrease of T cells and viral load in the CNS.

Recent studies have shown that WNV-induced proinflammatory responses could modulate BBB permeability, which might enable viral entry into and infection of the brain, leading to induction of lethal encephalitis (5, 21). As a result, mice which are deficient in proinflammatory cytokine signaling, including macrophage migration inhibitory factor and tumor necrosis factor-alpha receptor 1, have a reduced viral load in the brain and are subsequently more resistant to lethal WNV infection (5, 21). Following BBB compromise, WNV may enter the CNS via one or more pathways. The virus may invade the CNS by crossing the BBB directly (20, 35) or by being carried into the brain by infected infiltrating myeloid cells (36, 37) and inducing encephalitis. In this study, we have shown that the absence of Drak2 attenuates lethality from WNV encephalitis by 30%, and this is accompanied by a decrease of T cells in the CNS.

The role of T cells in protecting the host against WNV infection has been the subject of recent investigations. Both αβ and γδ T cells have been shown to contribute to host survival during WNV infection (8, 13). Among αβ T cells, CD4⁺ T cells respond vigorously in the periphery (16), whereas CD8⁺ T cell responses have been observed in both the spleen and brain following WNV infection (38). CD4⁺ T cells are known to provide help for antibody responses and to sustain WNV-specific CD8⁺ T cell responses in the CNS enabling viral clearance (39). CD8⁺ T cells have important functions in clearing infection from peripheral tissues and CNS, and in preventing viral persistence (40, 41). Further, FasL effector or perforin-dependent mechanisms (17, 42) are critical for CD8⁺ T cells to clear WNV infection in the CNS. CD40-CD40L interactions have also been demonstrated to be important for T cell trafficking into the CNS and for protection of the host from a low dose WNV challenge (43).

Immune responses are also involved in viral pathogenesis. For example, one study has shown that CD8⁺ T cells contribute to immunopathology upon high-dose WNV challenge (18). Nevertheless, little is known about the role of T cell-mediated pathology in WNV-related brain damage. Our data now provide the first evidence that Drak2-mediated T cell survival in the CNS plays an important role in WNV pathogenesis. These results may prompt the development of new pharmacotherapeutic treatment using Drak2 signaling inhibitors, alone or in combination with currently existing antiviral agents, in treating encephalitis caused by WNV and perhaps, other viral encephalitides.