γδ T Cells Facilitate Adaptive Immunity against West Nile Virus Infection in Mice¹

West Nile (WN)³ virus, a mosquito-borne flavivirus, has caused annual epidemics of viral encephalitis in North America since 1999 (1). The virus is maintained in an enzootic cycle that primarily involves mosquitoes and birds, with humans and horses as incidental hosts (2). Although human infection is usually asymptomatic, life-threatening neurological disease including encephalitis can ensue, particularly in the elderly and immunocompromised (2). Thus, understanding the factors that contribute to pathogenesis and immunity is critical for combating this emerging viral infection.

The murine model of WN virus infection mimics human disease and has been used to address questions of viral pathogenesis and immunity (3, 4, 5). After the initial infection, the virus spreads systemically and eventually invades the CNS. Mice die rapidly when encephalitis develops, usually within 1–2 wk following infection (5). Components of the innate and adaptive murine immune system have been shown to facilitate the control of WN virus infection (6). Type 1 IFNs (7, 8, 9, 10), complement (11), γδ T cells (12), and early neutralizing IgM (13) help limit viremia and dissemination into the CNS. Macrophages, B cells, and dendritic cells as APCs play an important role in T cell activation and proliferation during WN virus infection (14, 15). Cellular immunity, including both CD4⁺ (15) and CD8⁺ αβ T cells, participate in the recovery of the host from WN virus infection (16, 17, 18). Although secondary immune responses provide long-lasting protection against WN virus, the mechanisms that link innate and adaptive immunity upon infection are not fully understood.

We recently demonstrated that γδ T cells provide a rapid response following WN virus infection in mice, limiting the viral load and invasion of the CNS within the first few days following infection, thereby protecting the host from lethal encephalitis (12). γδ T cells comprise a minority of the CD3⁺ T cells in lymphoid tissue, yet they are well represented in the peripheral blood and are abundant at epithelial and mucosal sites (19). Like αβ T cells, γδ T cells bear a TCR and can be polarized to produce Th1- or Th2-type cytokines (20, 21, 22), although they are more limited in TCR diversity (19). γδ T cells can rapidly produce cytokines in response to microbial Ags (23) and have unique features, including a lack of MHC restriction and the capacity to react with Ag without the requirement for undergoing conventional Ag processing (24, 25, 26). These features suggest a role for γδ T cells in pathogen control early after infection. Increased numbers of γδ T cells in the peripheral blood and/or localization to sites of infection have been documented in both human (27, 28, 29) and murine viral infections (30, 31).

γδ T cells are thought to bridge the innate and adaptive immune responses (32, 33). For example, human Vδ2⁺ T cells have been recently shown to act as professional APCs and induce proliferation and differentiation of naive αβ T cells, forming a unique link between innate and adaptive immune responses (34). In primates, γδ T cells have characteristics of adaptive immunity upon mycobacteria infection, including the generation of a memory-type response (35). In this study, we have explored the possibility of whether γδ T cells are also involved in the development of protective adaptive immune responses during murine WN virus infection.

C57BL/6 (B6) and γδ T cell-deficient (TCRδ^−/−) mice were purchased from The Jackson Laboratory and were maintained under specific pathogen-free conditions at the animal facilities at Yale University, School of Medicine. TCRδ^−/− mice were bred on the B6 background and had been fully backcrossed. Experiments for primary and secondary infection were performed in 6- to 10-wk-old and 11- to 15-wk-old mice, respectively. Groups were age and sex matched for each experiment and were housed under identical conditions. All animal experiments were approved by the Institutional Animal Care and Use Committee at Yale.

WN virus isolate 2741 was initially cultivated by J. Anderson (Connecticut Agricultural Experiment Station, New Haven, CT) (36). For primary infection, mice were inoculated i.p. with 1/10 of an LD₅₀ of WN virus CT 2741. For secondary infection, all mice were inoculated i.p. with an LD₁₀₀ of WN virus isolate 2741. A total of 10² PFU corresponds to the LD₅₀, and 10³ PFU represents the LD₁₀₀ for WN virus isolate CT 2741. For some experiments, mice were inoculated with a nonlethal dose (100 PFU) of WN virus isolate bird 1153, a naturally attenuated strain of WN virus (37). Infected mice were monitored twice daily for morbidity, including lethargy, anorexia, and difficulty in walking.

At days 2 and 6 after WN virus challenge, RNA was extracted from the blood of B6 mice and TCRδ^−/− mice using RNeasy extraction (Qiagen). The extracted RNA was eluted in a total volume of 60 μl 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). A total of 40 ng of cDNA was then used for real-time PCR. The sequences of the primer-probe sets for WN virus were described earlier (38). The probe contained a 5′ reporter, FAM, and a 3′ quencher, TAMRA (Applied Biosystems). The reaction mixture contained a total volume of 50 μl, including each primer pair at a concentration of 1 μM and a probe at a concentration of 0.2 μM. The assay was performed on an iCycler (Bio-Rad). The thermal cycling consisted of 95°C for 3.5 min and 48 cycles of 95°C for 30 s and 60°C for 1 min. To prepare the DNA standard for real-time PCR, the 1.5-kb WN virus envelope protein Ag (WNV-E) gene region was cloned into pBADTOPO, as described earlier (5). To normalize the samples, the same amount of cDNA was used in the β-actin Q-PCR. The quantity of each sample was determined using the standard curve of each Q-PCR. The ratio of the amount of amplified WNV-E DNA compared with the amount of β-actin DNA represented the relative infection level of each sample.

Before secondary infection, B6 mice surviving primary viral challenge were depleted of γδ T cells, as described previously (39). Briefly, mice were i.p. injected with 250 μg of hamster anti-TCRγδ mAb (UC-7; American Type Culture Collection) on three consecutive days. Depletion was confirmed via flow cytometry by staining with Abs specific for TCRαβ (clone H57-597) and TCRγδ (clone GL3) (BD Pharmingen).

Microtiter plates were coated with rWNV-E expressed in Drosophila melanogaster S2 cells (40) 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 to 1/1000 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-T. Alkaline phosphatase-conjugated goat anti-mouse IgG or IgM (Sigma-Aldrich) at a dilution of 1/1000 in PBS-T was added for 1 h at room temperature. After washing three times with PBS-T, color was developed with p-nitrophenyl phosphate for 10 min and read at an absorbance of 405 nm using a spectrophotometer.

Six-week-old B6 mice were intradermally injected with 150 μl of antiserum (diluted 1/3 in PBS) pooled either from day 28 infected wild-type or TCRδ^−/− mice, or naive B6 mice. The animals were challenged with either one LD₁₀₀ or 100 × LD₁₀₀ of WN virus isolate 2741 24 h after the serum transfer. Infected mice were monitored twice daily, as described above.

Freshly isolated splenocytes were stained with Abs specific for TCRαβ and TCRγδ, or CD8α (clone 53-6.7) and CD44 (clone IM7) (BD Pharmingen). After staining, the cells were fixed in PBS with 2% paraformaldehyde and examined using a FACSCalibur flow cytometer (BD Biosciences). Dead cells were excluded on the basis of forward and side light scatter. Data were analyzed using CellQuest software.

To measure cytokine production, splenocytes from WN virus-infected mice were isolated and were stimulated at 3 × 10⁶ cells/tube with 50 ng/ml PMA (Sigma-Aldrich) and 500 ng/ml ionomycin (Sigma-Aldrich) for 4 h at 37°C. Golgi-plug (BD Pharmingen) was added during the final 2 h. Murine endothelial cells were isolated from brains of B6 mice and immortalized, as previously described (41). These cells were seeded on 12-well plates (1 × 10⁵ cells/well), incubated for 24 h at 37°C, and infected with WN virus. A total of 8 × 10⁵ splenocytes from infected wild-type or TCRδ^−/− mice was incubated with 3 × 10⁵ to 4 × 10⁵ of WN virus-infected endothelial cells. All stimulations were performed for 5 h at 37°C in the presence of Golgi-plug. The cells were then harvested, stained with Abs (BD Pharmingen) for TCRαβ and CD8α, and fixed in 2% paraformaldehyde. The cells were then permeabilized with 0.5% saponin before adding PE-conjugated anti-IFN-γ mAb (clone XMG 1.2) or control PE-conjugated rat IgG1 (BD Pharmingen). Cells were examined using a FACSCalibur flow cytometer, as described above.

Single-cell suspensions of CD8⁺ T cells were made from spleens of WN virus-infected wild-type or TCRδ^−/− mice by positive selection method using anti-CD8α magnetic beads (Miltenyi Biotec), as described previously (42). A total of 8 × 10⁶ cells was injected i.v. into naive 6-wk-old B6 mice 24 h before infection with one LD₁₀₀ of WN virus 2741. Following challenge, infected mice were monitored twice daily, as described above.

The in vitro analysis of CTL activity was performed, as described (43). Splenocytes were isolated from naive B6 mice and were depleted of CD8⁺ T cells by negative selection MACS with anti-CD8α magnetic beads (Miltenyi Biotec) (42). These CD8-negative splenocytes were used as target cells. Briefly, they were pulsed with purified rWNV-E at 1 μg/10⁶ cells at 37°C for 30 min. The pulsed cells were then washed and labeled with the fluorescent dye CFSE (Molecular Probes) by incubation of 5 × 10⁶ cells/ml in PBS containing 50 μM CFSE for 10 min at 37°C, followed by one wash in 5 vol of ice-cold PBS and two washes in DMEM (Invitrogen Life Technologies). CD8-depleted splenocytes not pulsed with rWNV-E were labeled with 10-fold dilutions of CFSE. These two cell populations were then mixed equally and incubated with different ratios of purified CD8⁺ T cells from wild-type and TCRδ^−/− mice at 37°C for 5 h. Cells were harvested and stained with anti-CD8α mAb. After staining, cells were fixed with 2% paraformaldehyde and examined by flow cytometry. Dead cells were excluded on the basis of forward and side light scatter. Cells were gated on the CD8-negative population and analyzed using CellQuest software. Percent specific lysis was determined using the following calculation: (1 − (mean CFSE^high percentage of test sample × mean CFSE^low percentage of test sample)/(mean CFSE^high percentage of spontaneous release × mean CFSE^low percentage of spontaneous release)) × 100.

Survival curve comparisons were performed using Prism software (GraphPad) statistical analysis, which uses the log rank test (equivalent to the Mantel-Haenszel test). Values of p for viral burden, Ab titer, and memory T cell number experiments were calculated with a nonpaired Student’s t test.

We previously demonstrated that γδ T cells produced IFN-γ within the first few days following WN virus infection in mice, reduced viral dissemination, and partially protected mice from lethality (12). In the current work, to determine whether γδ T cells also play a role in promotion of adaptive immunity, we more carefully examined the subset of mice that survived the primary infection with WN virus. Wild-type mice and TCRδ^−/− mice were first infected with 1/10 of the LD₅₀ of WN virus strain CT 2741 and monitored twice daily for morbidity. The low dose of virus was used to maximize the number of mice that survived following infection, particularly in the immunodeficient animals. As expected based upon our earlier work (12), only 15–20% of the TCRδ^−/− mice survived viral challenge, while ∼80–85% of the wild-type mice survived. We have not been able to detect WN virus in surviving mice from either group, when examining selected mice by PCR after 21 days (data not shown), suggesting that the virus does not persist in these animals.

We then assessed whether the TCRδ^−/− and wild-type mice that were able to withstand primary infection with a low dose of WN virus were equally susceptible to challenge with a secondary infection by WN virus. For these studies, a higher dose of WN virus was used for the secondary challenge studies to generate a group of animals that readily succumbed so that the influence on protection could be more clearly discerned. Surviving mice were therefore challenged with an LD₁₀₀ of WN virus at day 30 after primary infection. Age- and sex-matched naive mice (both wild-type and TCRδ^−/− mice) were infected with the same dose of WN virus and used as experimental controls in the secondary challenge experiments. Wild-type mice that survived following primary infection were much more resistant to the secondary viral challenge (p < 0.01) than the age- and sex-matched naive wild-type mice, demonstrating development of a protective memory response (Fig. 1 a). Likewise, TCRδ^−/− mice surviving primary challenge were more resistant to WN virus infection than the age- and sex-matched naive TCRδ^−/− mice (p < 0.01), suggesting the development of protective memory responses in the absence of γδ T cells. As expected (12), the naive TCRδ^−/− mice were more susceptible to WN virus infection than naive wild-type mice (p < 0.01).

Surprisingly, the difference in susceptibility to WN virus infection was also sustained between the two experimental groups during secondary infection. The TCRδ^−/− mice that survived primary infection with WN virus were more susceptible to secondary infection with WN virus than the wild-type mice that had endured the initial infection with virus. We therefore analyzed the viral burden in the blood from both groups of mice during secondary challenge. At day 2 after reinfection, viremia was significantly higher in TCRδ^−/− mice than the wild-type mice (Fig. 1,b). Moreover, at day 6, viral levels in the blood of wild-type mice were no longer detectable, while the viral burden remained elevated in TCRδ^−/− mice (Fig. 1 b). We next extended these observations by performing additional primary challenge studies using WN virus bird 1153 strain, which is a naturally attenuated WN virus that lost its neuroinvasiveness (37). Infection with the attenuated WN virus enabled us to generate a larger group of animals that survive the primary infection, and thereby perform the secondary challenges expeditiously. In addition, challenge with the attenuated strain in mice produces a nonlethal infection in the majority of the wild-type animals, which more closely resembles human cases (37). As above, mice lacking γδ T cells that survived primary infection with the attenuated isolate of WN virus were more susceptible (75% survival) to secondary infection with WN virus than the wild-type mice (100% survival) following inoculation of similar doses of WN virus. Together, these data suggest that γδ T cells contribute to the protective immune response upon recurrent exposure to WN virus. Although protective immunity induced by primary infection does appear to develop in the absence of γδ T cells, it is incomplete.

The observation that TCRδ^−/− mice had a higher level of viremia and lethality than wild-type mice upon secondary challenge with WN virus suggested that γδ T cells could be contributing directly to the antiviral response, in the absence of an effect upon the adaptive immune response, or alternatively, they might contribute to the development of protection mediated by the adaptive immune response. To test the former possibility, we depleted γδ T cells from wild-type mice that survived primary infection with WN virus, and then subjected these animals to a secondary challenge. To remove γδ T cells, mice were treated with an i.p. injection of 250 μg of hamster anti-TCRγδ mAb daily for 3 days, starting at day 28 after the primary WN virus infection. The efficiency of depletion was close to 90% at 24 h after treatment as analyzed by flow cytometry, as the percentage of splenic TCRγδ⁺TCRαβ⁻ T cells dropped from 2.22% in untreated mice to 0.15% in treated mice (Fig. 2,a). At 5 days posttreatment, 60% of γδ T cells remained absent (data not shown). Twenty-four hours after the depletion of γδ T cells, mice were challenged with one LD₁₀₀ dose of WN virus. Mice treated with PBS were used as controls. Wild-type mice that had survived a primary infection with WN virus CT 2741 and then were depleted of γδ T cells were as resistant to secondary challenge with WN virus CT 2741 as control mice (Fig. 2 b). These studies demonstrate that depletion of γδ T cells before secondary challenge does not alter the course of disease and suggests that γδ T cells do not act as memory T cells per se during secondary infection with WN virus.

γδ T cells were previously reported to be involved in regulating either Ab production or Ab isotype switching (44, 45). Moreover, B cell-mediated humoral immune responses are critical for host defenses against disseminated infection by WN virus (13, 46, 47). To determine whether γδ T cells influence the development of Abs to WN virus during infection, we measured WN virus-specific IgM (Fig. 3,a, left panel) and IgG levels (Fig. 3 a, right panel) in the sera from TCRδ^−/− and wild-type mice that survived primary infection. Ab responses in both groups of mice were similar at 4 days, an early interval postinfection, and at later time points (days 21 and 28, p > 0.05, compared with sera in wild-type mice).

To determine whether neutralizing Abs to WN virus developed in both the wild-type and TCRδ^−/− mice, sera from both groups at day 28 postinfection were used for passive immunization. Naive 6-wk-old B6 mice were administered sera from either group of animals 24 h before challenge with one LD₁₀₀ of WN virus CT 2741. Sera from both the wild-type and TCRδ^−/−-infected animals protected mice from lethal infection (Fig. 3,b, left panel; p > 0.05, compared with naive mice that received wild-type sera), while mice given sera from uninfected mice all succumbed to infection (p < 0.05, compared with mice transferred with noninfected sera). In a separate experiment, mice were also challenged with a higher dose of WN virus (100 LD₁₀₀) to determine whether this quantity of virus challenge could influence the outcome of the study. Sera from both groups partially, but equally (25% survival vs 25% survival; Fig. 3 b, right panel; p > 0.05, compared with naive mice that received wild-type sera) protected naive mice from infection when the viral dose was increased, suggesting Ab production was not significantly different between these two groups during primary infection.

Finally, we measured WN virus-specific IgM and IgG during secondary challenge. At days 2 and 6 after reinfection, we did not detect significant differences between the two groups in either the IgM or IgG response (Fig. 3,c, left and right panels; p > 0.05, compared with wild-type mice). However, at day 21 after reinfection, IgM response of wild-type mice was significantly higher than that in TCRδ^−/− mice (Fig. 3 c, left panel; p < 0.05, compared with wild-type mice). Overall, these data suggest that γδ T cells do not influence the development of Ab responses during primary and at the early stages of secondary infection with WN virus.

CD8⁺ effector T cells have been shown to have important functions in clearing WN virus infection from tissues and preventing viral persistence (10, 17, 18, 48). To understand the role of γδ T cells in regulation of adaptive immunity, we next focused on CD8⁺ T cells. We adoptively transferred CD8⁺ T cells (90–95% purity; Fig. 4,a) from wild-type and TCRδ^−/− mice at day 30 after WN virus infection to naive recipients 24 h before challenge with an LD₁₀₀ of WN virus. Control groups, including mice that did not receive cells via adoptive transfer and mice given CD8⁺ T cells from naive mice, succumbed to infection within 2 wk. In contrast, mice that received CD8⁺ T cells from previously infected wild-type mice had a significantly higher resistance to lethal infection than the groups of control mice (Fig. 4,b; 50% survival; p < 0.05 compared with control groups). Furthermore, mice administered CD8⁺ T cells from previously infected TCRδ^−/− mice were much more susceptible to lethal infection than mice provided with CD8⁺ T cells from the infected wild-type mice, and succumbed to virus challenge at the same rate as control mice (Fig. 4 b; 12.5% survival vs 0%; p > 0.05 compared with control mice). These studies clearly indicate that CD8⁺ T cell responses were significantly impaired in the absence of γδ T cells during the course of WN virus infection.

To further understand the role of γδ T cells in regulating CD8⁺ T cell responses, we assessed the function of these cells following primary infection with WN virus. Splenocytes from both wild-type and TCRδ^−/− mice were isolated 30 days after primary infection and examined for the percentage of CD8 memory T cells. Wild-type mice had 36% of CD44^highCD8⁺ memory T cells compared with 30% in TCRδ^−/− mice (Fig. 5,a). The total number of CD44^highCD8⁺ memory T cells is 2.78 ± 0.07 × 10⁶ in wild-type mice, compared with 2.5 ± 0.1 × 10⁶ in TCRδ^−/− mice (Fig. 5,b; p < 0.05 compared with wild-type mice). The functional capacity of the total CD8⁺ T cell response in these mice was determined using a CTL assay. At day 30 following infection, purified CD8⁺ T cells from both wild-type and TCRδ^−/− mice were incubated with various ratios of target cells labeled with rWNV-E. Flavivirus Ags that elicit CD8 T responses have been examined (49, 50, 51, 52, 53), and the E protein has been demonstrated to be a target (51, 54, 55). Moreover, WN virus E protein has been demonstrated in our previous studies as an antigenic target. Therefore, we focused on examining CD8 T cell responses to the E protein, recognizing that other proteins may elicit responses as well (5). The CTL activity of CD8⁺ T cells in TCRδ^−/− mice was significantly reduced compared with wild-type animals (Fig. 5,c). Finally, the functional CD8⁺ T cell response at day 30 postinfection was also determined by using an ex vivo intracellular cytokine staining assay for IFN-γ after treatment with PMA and ionomycin. A total of 3.2% of the splenic CD8⁺ T cells in wild-type mice produced IFN-γ compared with 2.7% of CD8⁺ T cells from TCRδ^−/− mice (Fig. 5,d, top panel). Moreover, incubation with WN virus-infected murine endothelial cells revealed that 18% of splenic CD8 T cells in wild-type mice produced IFN-γ in comparison with 12% of CD8 T cells from TCRδ^−/− mice (Fig. 5,e). These data were consistent with an earlier time point during primary infection (day 12), in which 1.6% of splenic CD8⁺ T cells produced IFN-γ compared with 0.9% of CD8⁺ T cells from TCRδ^−/− mice (Fig. 5 d, bottom panel). Overall, these studies suggest that γδ T cells facilitate the expansion and effector function of CD8⁺ memory T cells during WN virus infection.

γδ T cells have features that are characteristic of innate immune cells: they respond and expand quickly upon exposure to a pathogen, produce inflammatory cytokines, and are not Ag specific (19, 56). Our previous work has shown a role for γδ T cells in the early response during WN virus infection (12). γδ T cells expanded significantly early during WN virus infection, produced IFN-γ, and enhanced perforin expression by splenic T cells, thereby enabling the host to clear the virus.

In the current work, we have further demonstrated an important role for γδ T cells in the development of a protective CD8⁺ T cell response against WN virus. We initially found that TCRδ^−/− mice were more susceptible than wild-type mice to secondary challenge with WN virus, suggesting possible involvement of γδ T cells in adaptive immunity. γδ T cells were not directly acting as memory T cells in these studies, because depleting these cells in wild-type mice that survived primary infection with WN virus had no effect on host susceptibility to secondary viral challenge. Moreover, γδ T cells did not contribute to any alteration in WN virus-specific Ab production, because the TCRδ^−/− and wild-type mice that survived primary infection with WN virus had similar levels of WN virus-specific IgG and IgM, and passive immunization with sera from both groups of mice could protect or partially protect naive mice from lethal WN virus infection. CD8⁺ memory T cell responses were significantly impaired in TCRδ^−/− mice that survived primary infection. Both the total number of CD8⁺ memory T cells and the functional capacity (both CTL activity and ex vivo cytokine production) of the cells were reduced in TCRδ^−/− mice that survived primary infection, compared with wild-type mice. Moreover, adoptive transfer of CD8⁺ αβ T cells from wild-type surviving mice into naive mice helped to enhance host resistance to lethal WN virus infection, while those transferred from surviving TCRδ^−/− mice did not have a similar effect. These data demonstrate that γδ T cells are required for effective CD8⁺ T cell memory responses during WN virus infection.

Depletion of γδ T cells before secondary challenge did not affect susceptibility to WN virus in wild-type mice, suggesting that γδ T cells do not act as memory T cells per se during secondary challenge with WN virus. Although the mechanisms through which γδ T cells regulate CD8⁺ T cell memory responses are still under investigation, these depletion studies also exclude the possibility that γδ T cells are involved in the maintenance of CD8⁺ memory T cells. For an acute infection such as WN encephalitis, the CD8⁺ T cell response has three characteristic phases: a period of initial activation and expansion, a contraction or death phase, and the establishment and maintenance of memory (57). We found a significant decrease in the CD8⁺ T cell response in TCRδ^−/− mice 12 days postinfection. WN virus is usually cleared from peripheral organs such as the spleen as early as day 8 (4). These data suggest that the impairment of CD8⁺ T cell response led by the deficiency of γδ T cells already exists during/before the contraction stage, possibly at the early T cell priming stage through cross talk with APCs, such as dendritic cells (58, 59, 60).

Increasing evidence from both human and primate studies has suggested that γδ T cells are involved in the genesis of adaptive immunity. Upon microbial activation, human Vδ2⁺ T cells display some of the principal characteristics of professional APCs: they efficiently process and display Ags and provide costimulatory signals sufficient for strong induction of naive αβ T cell proliferation and differentiation (34). Moreover, the killer cell lectin-like receptor G1 is expressed in a significant proportion of human Vγ9/Vδ2 T cells, which also have the phenotype for effector memory T cells (lack of CD27, CD45RA, CD62L, and CCR7 (61). In primates studies, the recall expansion of Vγ2/Vδ2⁺ T cells developed during Mycobacterium tuberculosis infection of bacillus Calmette-Guerin-vaccinated macaques (35). Although the systemic Th1 response was impaired in γδ T cell-depleted mice following an intravaginal infection with HSV type 2, it was not clear whether this contributed directly to the higher susceptibility and lethal viral infection (62). Our findings in this study show that γδ T cells play an important role in regulating CD8 T cell memory responses in the murine model. Several groups have recently demonstrated that CD8⁺ effector T cells are critical in clearing WN virus infection from tissues and preventing viral persistence (10, 17, 18, 48). However, the development of CD8 T cell memory response to WN virus infection is not fully understood. Our studies have provided important insights into understanding the immune response to WN virus infection, and they will aid identification of the linkage between γδ T cells and adaptive immunity during WN virus infection. Overall, γδ Τ cells not only provide early defense against WN virus infection, but also contribute to the memory response, suggesting a potential role in facilitating future vaccine development.

We thank Debbie Beck and Ping Zhu for technical assistance. We also thank Drs. Susan Kaech, Tim Quan, and Jin-Young Choi for helpful discussions.

The authors have no financial conflict of interest.