Washington University School of Medicine, St Louis. MO Division of Infectious

JVI Accepts, published online ahead of print on 27 June 2012 J. Virol. doi:10.1128/jvi.00673-12 Copyright 2012, American Society for Microbiology. All Rights Reserved. 1 2 CD8 + T cells use TRAIL to restrict West Nile virus pathogenesis by controlling infection in neurons 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Bimmi Shrestha 1, Amelia K. Pinto 1, Sharone Green 4, Irene Bosch 5, and Michael. S. Diamond 1,2,3. Departments of Medicine 1, Molecular Microbiology 2, and Pathology & Immunology 3, Washington University School of Medicine, St Louis. MO 63110. Division of Infectious Diseases and Immunology 4, University of Massachusetts Medical School, Worcester, MA 01655. Massachusetts Institute of Technology 5, Cambridge, MA. Corresponding author: Michael S. Diamond, M.D., PhD, Departments of Medicine, Molecular Microbiology and Pathology & Immunology, Washington University School of Medicine, 660 South Euclid Avenue, Box 8051, St Louis. Missouri 63110. Tel: 314-362-2842, Fax: 314-362-9230, Email: diamond@borcim.wustl.edu Running title: TRAIL controls WNV infection 19 Abstract: 248 words; Figures: 9 20 1

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 ABSTRACT Previous studies in mice have demonstrated that an orchestrated sequence of innate and adaptive immune responses is required to control West Nile virus (WNV) infection in peripheral and central nervous system (CNS) tissues. TNF-related apoptosisinducing ligand (TRAIL, CD253) has been reported to inhibit infection of dengue virus, a closely related flavivirus, in cell culture. To determine the physiological function of TRAIL in the context of flavivirus infection, we compared the pathogenesis of WNV in wild type and TRAIL -/- mice. Mice lacking TRAIL showed increased vulnerability and death after subcutaneous WNV infection. Although no difference in viral burden was detected in peripheral tissues, greater viral infection was detected in the brain and spinal cord at late times after infection, and this was associated with delayed viral clearance in the few surviving TRAIL -/- mice. While priming of adaptive B and T cell responses, and trafficking of immune and antigen-specific cells to the brain was undistinguishable from normal mice, in TRAIL -/- mice, CD8 + T cells showed qualitative defects in their ability to clear WNV infection. Adoptive transfer of WNV-primed wild type but not TRAIL -/- CD8 + T cells to recipient CD8 -/- mice efficiently limited infection in the brain and spinal cord, and analogous results were obtained when wild type or TRAIL -/- CD8 + T cells were added to WNV-infected primary cortical neuron cultures ex vivo. Collectively, our results suggest that TRAIL produced by CD8 + T cells contributes to disease resolution by helping to clear WNV infection from neurons in the central nervous system. 41 2

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 INTRODUCTION West Nile Virus (WNV) is a mosquito-transmitted neurotropic flavivirus that is genetically related to other viruses causing global human disease including dengue virus (DENV), yellow fever virus, and Japanese and tick-borne encephalitis viruses. WNV cycles in nature between birds and Culex mosquitoes but can infect and cause severe neuroinvasive disease in other vertebrate animals including humans (25). While the majority of human WNV infections are asymptomatic, a subset of individuals develops a systemic febrile illness with few progressing to meningitis, encephalitis or an acute flaccid paralysis syndrome (32). Severe and symptomatic WNV infection occurs more frequently in the elderly or immunocompromised and in individuals homozygous for the CCR5Δ32 mutation (11, 18, 19, 24). Since 1999, in the United States, more than 30,000 cases of symptomatic WNV infection have been confirmed, although seroprevalence analysis estimates a much greater number of infections in the population (4). Currently there is no approved therapy or vaccine for WNV in humans. Studies in mice have helped to elucidate how the integrity of the innate and adaptive immune system is required for resistance to WNV infection. Using mice with targeted deletions in individual immune response genes, significant and protective contributions from inflammatory cytokines (e.g., type I and II interferon), chemokines, complement, B cells, and CD4 + and CD8 + T cells have been observed (reviewed in (7, 27)). In particular, receptor-ligand interactions of two members (TNF-α-TNF-α receptor and Fas-Fas ligand) of the tumor necrosis factor (TNF) superfamily of proteins are protective against WNV within the central nervous system. Antigen-specific CD8 + T cells use Fas ligand effector mechanisms to contain WNV infection in Fas-expressing 3

65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 neurons in the CNS (34) whereas TNF-α interaction with TNF-R1 protects against WNV infection by regulating migration of protective inflammatory cells into the brain during acute infection (38). TNF-related apoptosis-inducing ligand (CD253 or TRAIL) is a 281 amino acid type II transmembrane protein of the TNF superfamily. The C-terminal extracellular domain is cleaved by cell-associated proteases, resulting in a soluble TRAIL that forms a homotrimer and circulates in the bloodstream. Soluble or cell-associated human TRAIL binds to the death receptors DR4 (TRAIL-R1) and DR5 (TRAIL-R2) on tumor cells, resulting in recruitment of FADD (Fas-associated death domain), activation of caspase-8, and apoptosis of transformed cells. In addition, TRAIL also binds to decoy receptors DcR1 (TRAIL-R3) and DcR2 (TRAIL-R4) that either lack or contain a truncated cytoplasmic domain. TRAIL binding to DcR1 and DcR2 does not promote apoptosis but instead neutralizes TRAIL function (22) or promotes NF-κB activation and transcription of pro-inflammatory genes (13, 20). TRAIL binding to its receptors on non-tumor cells commonly does not trigger apoptosis. Such regulation is controlled by the binding of the endogenous inhibitor of death receptor killing, cellular FLICE-Like inhibitory protein (c- FLIP) to the intracellular domain of TRAIL receptors (23) and by expression of the X- linked inhibitor of apoptosis protein (XIAP) (26). Thus, the physiological role of TRAIL remains less well understood. TRAIL is induced as a consequence of type I and II interferon (IFN) signaling in several cell types (14) and has been reported to have antiinflammatory properties (30), which can mitigate excessive and maladaptive host immune responses in the brain (12). In mice, TRAIL binding to DR5 has been proposed to negatively regulate cellular innate immune responses (9). 4

88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 TRAIL also has been suggested to have inhibitory functions against several viruses in vivo. Administration of a blocking but non cell-depleting anti-trail antibody to mice during influenza virus infection delayed virus clearance in the lung (15), and analogously, TRAIL -/- mice showed increased disease severity after influenza infection (2, 3). In addition, type I IFN signaling induced TRAIL expression in NK cells, which facilitated control of encephalomyocarditis virus in vivo (29). Some of the antiviral effect of TRAIL may be mediated by the decoy receptors DcR1 and DcR2, as DR5 -/- mice show lower titers of murine cytomegalovirus in the spleen and higher levels of type I IFN (9). More recently, a role for TRAIL-dependent protection against flaviviruses was suggested (40). DENV infection induced TRAIL expression in blood-derived immune cells, primary myoblasts, and human endothelial cells in a type I IFN-dependent manner, and exogenous recombinant TRAIL inhibited DENV infection in human myeloid cells through an apoptosis-independent mechanism. To determine the physiological relevance of TRAIL in the context of infection by a flavivirus, we compared the virulence of WNV in wild type and TRAIL -/- mice. We found that a deficiency of TRAIL was associated with increased viral burden in the CNS, delayed clearance, and enhanced mortality despite normal priming of adaptive B and T cell immune responses and trafficking to the brain. However, CD8 + T cells showed qualitative defects in their ability to clear WNV infection. Our experiments suggest that CD8 + T cells utilize TRAIL to control WNV infection in neurons. 109 5

110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 MATERIALS AND METHODS Virus. The WNV strain (3000.0259) was isolated in New York in 2000 and passaged once in C6/36 Aedes albopictus cells to generate a stock virus as described previously (10). Mouse experiments. C57BL/6 wild-type inbred mice were commercially obtained (Jackson Laboratories, Bar Harbor, ME). The congenic, backcrossed TRAIL -/- mice were obtained from Dr. Jonathan Weiss (NIH, Bethesda, MD) and re-derived prior to initiation of infection experiments. All mice were genotyped and bred in the animal facilities of the Washington University School of Medicine under pathogen-free conditions, and experiments were performed in accordance with and approval of the Washington university animal studies guidelines. Eight to 10-week-old mice were inoculated with 10 1 and 10 2 plaque-forming units (PFU) by intracranial and subcutaneous route, respectively. Quantitation of tissue viral burden and viremia. To determine the extent of viral spread in vivo, wild type and TRAIL -/- mice were infected with WNV by subcutaneous (via footpad injection) or intracranial inoculation and euthanized at specific time points. Blood was collected by intracardiac puncture, and serum was isolated and stored aliquotted at -80ºC. After extensive tissue perfusion with PBS at 4ºC, organs were harvested, homogenized using a bead beater apparatus, and infectious virus was titered by plaque assay on BHK21-15 cells (8). For measurement of infectious virus in serum, plaque assays were performed on Vero cells (35). Measurement of WNV-specific antibodies. To determine the levels of WNVspecific antibody, IgM and IgG were measured using an enzyme-linked immunosorbent assay (ELISA) with purified WNV E protein as described previously (21). The titer was 6

133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 defined as the serum dilution yielding an optical density at 450 nm equivalent to three times above the background of the assay. The titer of neutralizing antibody was determined by using a plaque reduction neutralization assay with BHK21-15 cells (36). Plaques were counted visually, plotted and the plaque reduction neutralization titer for 50% inhibition (PRNT50) was calculated using GraphPad Prism software. Analysis of splenic CD4 + and CD8 + T cells. WNV-specific CD4 + and CD8 + T cells in the spleen were analyzed as described previously (34). Briefly, splenocytes were harvested from wild type and TRAIL -/- mice on day 8 after infection. After lysis of erythrocytes in hypotonic solution, 10 6 cells were stimulated with 0.2 µg/ml of an immunodominant D b -restricted WNV-specific NS4B peptide (SSVWNATTAI) or anti- CD3 antibody (2C11) for 6 hours at 37ºC in the presence of Golgi plug (BD Biosciences). Splenocytes were stained with anti-cd4 or anti-cd8 antibody (BD Biosciences) for 30 minutes at 4ºC. After washing in PBS supplemented with 5% goat serum, cells were fixed with 1% paraformaldehyde in PBS, permeabilized with saponin, and stained with anti-ifn-γ or anti-tnf-α antibody (BD Biosciences) for 30 minutes at 4ºC. After a final series of washes, cells were processed by flow cytometry and the percentage of CD8 + and CD4 + T cells that expressed IFN-γ or TNF-α was determined using CellQuest (BD Biosciences) and FlowJo (Treestar) software. In separate experiments, splenocytes were stained at 4ºC for 30 minutes with FITC-conjugated anti- CD8 antibody (BD Biosciences) and PE-conjugated D b -NS4B peptide (SSVWNATTAI) tetramers (prepared by the NIH tetramer facility, Atlanta GA). Uninfected mice were used as controls in all experiments. The number and percentage of CD4 + CD25 + FoxP3 + 155 regulatory T cells in the spleen of WNV-infected wild type or TRAIL -/- mice was 7

156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 measured using the regulatory T cell staining kit (ebioscience). Samples were processed by multi-color flow cytometry on an LSR II flow cytometer (Becton Dickinson) and analyzed with FlowJo software (Treestar). CNS leukocytes. Leukocytes were isolated and quantified from the brains of infected animals as described previously (36). Briefly, eight days after subcutaneous infection with 10 2 PFU of WNV, brains were harvested after extensive perfusion with PBS, dispersed into single-cell suspensions, and digested with 0.05% collagenase D, 0.1 µg/ml trypsin inhibitor TLCK (Nα-p-tosyl-L-lysine chloromethyl ketone), 10 µg/ml DNase I (all from Sigma Chemical), and 10 mm of HEPES ph 7.3 in Hanks balanced salt solution for one hour. Leukocytes were isolated by discontinuous Percoll gradient (70%37%/30%) centrifugation for 30 minutes (850 x g at 4ºC). After washing and counting, cells were stained for CD4, CD8, CD45, and CD11b using directly conjugated antibodies (BD Biosciences) for 30 minutes at 4ºC then fixed with PBS supplemented with 1% paraformaldehye. In some experiments, isolated brain leukocytes were restimulated with D b -restricted NS4B peptide or anti-cd3 antibody for 6 hours and IFNγ- and TNF-α-producing CD8 + T and CD4 + T cells were measured. Alternatively, brain leukocytes were incubated with the NS4B tetramer and anti-cd8 to detect WNV-specific CD8 + T cells. Samples were processed by multi-color flow cytometry on an LSR II flow cytometer and analyzed with FlowJo software. Primary cell isolation and infection. Primary cortical neurons were prepared from 15 days old embryos as described previously (28). Cortical neurons were seeded in 24-well poly-d-lysine- and laminin-coated plates in neurobasal medium containing B27 and L-glutamine (Invitrogen) for 24 hours. The medium was replaced and neurons were 8

179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 cultured for three additional days prior to infection. Multistep virus growth analysis was performed with primary cells at a multiplicity of infection (MOI) of 0.01. Supernatants were harvested at 6, 24, 48 and 72 hours after infection and titrated by plaque assay on BHK21-15 cells, as described previously (8). In some experiments, recombinant TRAIL (Enzo Life Sciences, 0.078 to 20 μg/ml) was added 24 hours before or one hour after WNV infection. Adoptive transfer of wild type or TRAIL -/- CD8 + T cells. CD8 + T cells from WNV-infected mice were purified and transferred as described previously (37). Splenocytes were harvested from naïve or primed wild type or TRAIL -/- mice at day 7 after WNV infection. CD8 + T cells were purified by positive selection using antibodycoated magnetic beads (Miltenyi Biotec, Auburn, CA) and transferred into CD8 -/- mice one day after WNV infection. In pilot experiments, we confirmed that equivalent numbers and percentages of wild type (4.0% + 1.6 of total lymphocytes) or TRAIL -/- (4.5% + 1.5 of total lymphocytes) primed CD8 + T cells were present in the blood of separate recipient mice two days after adoptive transfer. For virological studies, recipient mice were sacrificed at day 10 after initial infection and tissues were analyzed for viral burden by plaque assay. Addition of CD8 + T cells to WNV-infected neurons. Incubation of purified CD8 + T cells with WNV-infected neurons was performed as described previously (37). Cortical neurons were infected at an MOI of 0.001. One hour later, unbound virus was removed with four washes of warm medium. Subsequently, purified naïve or WNV-primed CD8 + T cells from wild type or TRAIL -/- mice were added at an effector to target (E:T) ratio of 50:1. Supernatants were harvested at 24 and 48 hours after infection, and WNV 9

202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 production was measured by viral plaque assay. In some experiments, recombinant TRAIL (10 μg/ml) or neutralizing anti-trail antibody or isotype control antibody rat IgG2a (BioLegend) was added at the same time as the CD8 + T cells. TRAIL, DR5, FasL and granzyme B staining. Brains were harvested from wild type or TRAIL -/- mice on day 8 after infection. After isolating leukocytes, CD4 + and CD8 + T cells were stained for expression of TRAIL using a biotin labeled anti-mouse CD253 (TRAIL) antibody (ebioscience), for FasL using an anti-mouse CD95L antibody (Biolegend), and for intracellular granzyme B using an anti-mouse granzyme B antibody (Invitrogen) for 30 minutes at 4ºC. Data collection and analysis were performed with a FACSCalibur flow cytometer and CellQuest Software (BD Biosciences). Cortical neurons were isolated as described previously (37), and infected with WNV (MOI of 0.001) for one hour. One day later, cells were fixed with 4% paraformaldehyde in PBS at 4 C for 10 minutes. After several washes in PBS and blocking in 5% normal goat serum, neurons were stained with rabbit anti-mouse DR5 (10 µg/ml anti-trail-r2, Millipore) for one hour at room temperature. After rinsing in PBS supplemented with 0.05% Triton- X-100, cells also were stained with WNV-immune rat serum (1:100 dilution) for one hour at room temperature. Fluorescence was detected after incubation with Alexa 488 (Molecular Probes) and Cy3 (Jackson Laboratories) conjugated secondary antibodies for 30 minutes at room temperature. Cells were visualized after counterstaining with ToPro-3 (Molecular Probes) using a Zeiss 510 Meta LSM confocal microscope Data analysis. All data were analyzed statistically using Prism software (GraphPad4, San Diego, CA). Kaplan-Meier survival curves were analyzed by the log rank test. Differences in viral burdens in mice were analyzed by the Mann-Whitney test. 10

225 226 Differences in viral infection in primary cell cultures and numbers of T cells were analyzed by an unpaired t test. 227 228 229 Downloaded from http://jvi.asm.org/ on October 1, 2018 by guest 11

230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 RESULTS TRAIL is required to protect mice from lethal WNV infection. Previous in vitro studies suggested that addition of high doses (e.g., 5 to 20 μg/ml) of exogenous recombinant human TRAIL to cell cultures could inhibit infection of DENV, a related flavivirus (40). To assess directly the role of TRAIL in the context of WNV infection in vivo, we compared the survival rates of 8 to 9 week-old wild type and congenic TRAIL -/- mice after subcutaneous infection with 10 2 PFU of a North American WNV isolate (New York, 2000). By day 8 after infection, all wild type and deficient mice showed clinical signs of infection, including reduced activity, weight loss, and hair ruffling. However, survival rates of the TRAIL -/- mice were markedly lower (17% compared to 75%, P < 0.0001; Fig 1A). Similarly, we observed a significant decrease in the mean time to death (11.9 + 0.3 versus 13.1 + 0.5 days, P < 0.04) in TRAIL -/- compared to wild type mice. Thus, an absence of TRAIL results in a more severe phenotype after WNV infection with poorer clinical outcome. A deficiency of TRAIL results in enhanced WNV tissue burden primarily in the central nervous system. To begin to understand how an absence of TRAIL conferred increased susceptibility to WNV infection, we measured the levels of infectious virus in tissues. Wild type and TRAIL -/- mice were infected by subcutaneous route, and the viral loads were analyzed on days 2, 4, 6, 8, and 10 after infection. (a) Peripheral tissues. We observed was no significant difference in the kinetics and magnitude of WNV infection in serum or the spleen between wild type and TRAIL -/- mice throughout the time course (Fig 1B and C, P > 0.3). Moreover, no infectious virus was detected in the liver in either wild type or TRAIL -/- mice (data not shown). In 12

253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 comparison, infectious WNV was recovered from the kidney (3 of 10 mice) at day 4 in a subset of TRAIL -/- but not from wild type mice (data not shown). The kidney usually does not support WNV infection in wild type C57BL/6 mice, whereas replication can be observed in congenic mice lacking aspects of cell-intrinsic immunity (e.g., IFNαβR -/-, IRF-7 -/-, or IRF-3 -/- mice (5, 6, 28)). Overall, TRAIL appears to have to a limited role in controlling WNV infection in peripheral organs. (b) CNS tissues. In the brain and spinal cord, similar levels of WNV (10 4 to 10 5 PFU/g) were detected in wild type and TRAIL -/- mice through day 8 after infection. However, by day 10, significantly higher levels were measured in the brain (10 4.8 versus 10 3.5 PFU/g; P = 0.01) and spinal cord (10 4.9 versus 10 3.6 PFU/g; P = 0.01) of TRAIL -/- mice (Fig 1D and 1E). These experiments suggest that TRAIL functions primarily to control WNV in CNS tissues. TRAIL signaling does not directly control WNV replication in the CNS. Based on the viral burden data after subcutaneous infection, we hypothesized that a lack of TRAIL signaling in resident cells of the CNS might contribute to higher WNV titers in the brain through direct inhibitory effects on neuronal infection. To test this, wild type and TRAIL -/- mice were infected with 10 1 PFU of WNV directly into the cerebral cortex via an intracranial route and monitored for local replication and dissemination. Viral burden in the cerebral cortex, subcortical white matter, brain stem, cerebellum, and spinal cord were measured at days 2, 4, and 6 after infection (Fig 2A-E). Somewhat surprisingly, no differences in viral burdens were observed at any time point in the different CNS regions of wild type and TRAIL -/- mice (P > 0.2). Similarly, we observed no difference in relative morbidity between wild type and TRAIL -/- mice after intracranial 13

276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 WNV infection. These results suggest that TRAIL signaling within the CNS by resident cells does not directly control WNV infection. Adaptive immune responses in TRAIL -/- mice. Since we observed elevated viral burden in CNS tissues after subcutaneous but not intracranial infection, we speculated that TRAIL might modulate adaptive immune functions (B or T cells), which are required for efficient clearance of WNV from the brain (8, 17, 35, 39). To assess this, at different time points after infection, we measured the magnitude of the adaptive immune response to WNV infection. (a) Antibody response. The kinetics and magnitude of anti-wnv IgM and IgG responses were virtually identical in wild type and TRAIL -/- mice (P > 0.9, Fig 3A and B). Similarly, there was no difference in neutralizing titer in wild type and TRAIL -/- mice at days 4 and 8 after infection (P > 0.4, Fig 3C). Thus, TRAIL was not required for the development of WNV specific antibodies, and increased viral burden in the brain and spinal cord was not due to an obvious defect in antibody production. (b) T cell response. To evaluate a possible role of TRAIL in priming of WNVspecific T cells in the periphery, splenocytes from wild type and TRAIL -/- mice were harvested at day 7 after infection and restimulated ex vivo with an immunodominant D b - restricted NS4B peptide or anti-cd3 antibody, and CD4 + and CD8 + T cells were analyzed for the production of intracellular IFN-γ and TNF-α by flow cytometry. We observed no significant differences (P > 0.2) in the percentage of IFN-γ + or TNF-α + CD8 + T cells from wild type and TRAIL -/- mice after incubation with WNV-specific peptide (Fig 4A and B). We did however detect a slight increase in the percentage (7% versus 5%, P < 0.002) but not number (P > 0.2) of NS4B-tetramer positive CD8 + T cells in 14

299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 TRAIL -/- mice (Fig 4C and D). In comparison, no difference in the percentage or absolute numbers of CD4 + T cells producing IFN-γ and TNF-α was observed between wild type mice and TRAIL -/- mice (P > 0.4) (Fig 4E and F). Finally, we also did not observe a difference (P > 0.1) in the percentage or number of CD4 + CD25 + FoxP3 + regulatory T cells in WNV-infected wild type or TRAIL -/- mice (Fig 4G, and data not shown). Overall, these experiments demonstrate that an absence of TRAIL does not affect induction of adaptive B and T cell responses after WNV infection. Trafficking of leukocytes to the brain is not impaired in TRAIL -/- mice. Although peripheral T cell responses were unchanged in TRAIL -/- mice after WNV infection, we assessed whether the enhanced viral replication in the CNS of TRAIL -/- mice might be related to blunted activation of resident microglia or migration of protective macrophages and CD8 + T cells. To assess this, leukocytes were isolated from brains of wild type and TRAIL -/- mice at day 8 after infection by gradient centrifugation and analyzed by flow cytometry. Notably, we failed to observe a difference in percentages or numbers of activated CD11b + CD45 lo microglia or macrophages (CD11b + CD45 hi ) (P > 0.4; Fig 5A and B). We did find, however, a higher percentage and number (P < 0.03, Fig 5C and D) of total CD8 + T cells in the brains of TRAIL -/- mice. This correlated with increased numbers of CD8 + T cells from TRAIL -/- mice expressing IFN-γ (P < 0.04, Fig 5E) after WNV-specific peptide restimulation ex vivo, or that stained directly with NS4B tetramers (P = 0.05, Fig 5F). In comparison, while similar percentages and numbers (P > 0.4) of CD4 + T cells and CD19 + B220 + B cells were observed, we measured greater numbers of IFN-γ + CD4 + T cells after ex vivo restimulation from TRAIL -/- compared to wild type mice (P < 0.04, Fig 5G). Overall, these data show no clear deficiency in 15

322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 recruitment or activation of leukocytes in the brains of WNV-infected TRAIL -/- mice; in general, higher levels of T cells were observed likely secondary to the increased viral burden. Adoptive transfer of wild type and TRAIL -/- CD8 + T cells. In addition to Fas-Fas ligand or perforin/granzyme cytolysis mechanisms, CD8 + T cells and NK cells can secrete TRAIL to lyse target cells (1, 3, 41). Since an absence of CD8 + T cells or their specific effector functions resulted in a failure to control WNV replication in the brain and spinal cord (34, 35, 37, 39), we hypothesized that a deficiency of TRAIL in CD8 + T cells could explain the enhanced viral replication in the CNS of TRAIL -/- mice. To evaluate this, we performed adoptive transfer studies. Primed CD8 + T cells were purified from the spleens of WNV-infected wild type or TRAIL -/- mice on day 7 by positive selection using antibody-coated magnetic beads. We adoptively transferred 10 x 10 6 CD8 + T cells into congenic CD8 -/- mice one day after WNV infection. Nine days later, we assessed the viral burden in spleen and CNS tissues of recipient CD8 -/- mice. Adoptive transfer of 10 x 10 6 naïve wild type CD8 + T cells failed to control infection, as high levels of WNV were apparent in the spleen and CNS of recipient CD8 -/- mice (Fig 6A). Transfer of 10 x 10 6 WNV-primed wild type CD8 + T cells, however, efficiently controlled infection, as no virus was detected in the target tissues. In comparison, while adoptive transfer of primed TRAIL -/- CD8 + T cells cleared virus from the spleen, they failed to control WNV in the CNS, as recipient mice (4 of 4) sustained infection (~10 4 to 10 5 PFU/g). The viral burden in the CNS of CD8 -/- mice receiving TRAIL -/- CD8 + T cells nonetheless was lower than those receiving naïve wild type cells (10 4.2 PFU/g versus 10 6.9 PFU/g, P = 0.01), confirming TRAIL-independent clearance mechanisms by CD8 + T 16

345 cells. 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 To corroborate the role of TRAIL in CD8 + T cell-mediated clearance of WNV from the CNS, a lower number of (3 x 10 6 ) of primed CD8 + T cells was adoptively transferred (Fig 6B). Nine days after transfer, WNV was not detected (< 10 1.8 PFU/g) in the spleen of recipient CD8 -/- mice that had received primed wild type cells; however, infectious WNV was recovered from mice receiving primed TRAIL -/- CD8 + T cells (~10 3.4 PFU/g). Higher levels of WNV also were recovered from the CNS of CD8 -/- mice receiving 3 x 10 6 primed CD8 + T cells from TRAIL -/- compared to wild type mice (10 3.6 to 10 6.1 PFU/g, P < 0.02). Consistent with these findings, adoptive transfer of 3 x 10 6 WNVprimed but not naïve CD8 + T cells into recipient TRAIL -/- mice resulted in an improved survival rate (71% versus 20%, P = 0.05). Overall, the adoptive transfer experiments demonstrate that a deficiency of TRAIL in CD8 + T cells negatively impacts control of WNV in mice. Delayed WNV clearance in the CNS of TRAIL -/- mice. Given our adoptive transfer study results, we hypothesized that TRAIL was required in the CNS for efficient clearance by CD8 + T cells. As prior studies showed that a deficiency of CD8 + T cell effector molecules (e.g., perforin or Fas-ligand) caused WNV persistence in the CNS for several weeks (34, 37), we evaluated the impact of a TRAIL deficiency on the kinetics of viral clearance in the brain and spinal cord. Infectious virus was measured in the few surviving wild type and TRAIL -/- mice at days 15, 21, and 28 after inoculation (Fig 6C and D). Low levels (10 2 to 10 3 PFU/g) of infectious WNV were detected on day 15 in 12.5% (1 of 8) and 25% (2 of 8) of brains and spinal cords, respectively of wild type mice. No infectious virus was recovered from the CNS at 21 days or after in wild type 17

368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 mice. In comparison, higher levels of WNV (10 2 to 10 6 PFU/g) were detected in the CNS (67%; 6 of 9, P < 0.005) of TRAIL -/- mice on day 15. Moreover, on day 21, 10 (1 of 10) to 20% (2 of 10) of TRAIL -/- mice had detectable yet low levels (10 2 to 10 3 PFU/g) of infectious virus in the brain and spinal cord, respectively. By day 28, no infectious virus was recovered from the CNS of any TRAIL -/- mice. Thus, in the surviving mice, an absence of TRAIL resulted in somewhat delayed clearance of infections WNV from the brain and spinal cord. TRAIL -/- CD8 + T cells are impaired in their ability to clear WNV from infected neurons. To demonstrate directly that CD8 + T cells require TRAIL to control WNV infection of neurons, we used a viral, clearance assay with primary cortical neurons (37). One hour after infection, bulk WNV-primed wild type or TRAIL -/- splenic CD8 + T cells were added to neurons at an effector to target ratio of 50:1. At 24 and 48 hours, the level of infectious virus in the supernatants of neuronal cultures was titrated by plaque assay. As seen previously (34), we observed reduced (2 to ~1,300-fold, P < 0.03) viral yield in supernatants from neurons incubated with wild type primed compared to naïve or no CD8 + T cells at 24 and 48 hours, respectively (Fig 6E and F). In comparison, WNVprimed TRAIL -/- CD8 + T cells less efficiently restricted infection in neurons, with no difference at 24 hours (P > 0.7) and a smaller (23-fold, P < 0.02) reduction at 48 hours. These experiments suggest that WNV primed CD8 + T cells control neuronal infection, in part through TRAIL-dependent mechanisms. As additional confirmation, we assessed surface expression of TRAIL on WNVprimed CD8 + T cells in the spleen and the brain. Wild type mice were infected with 10 2 PFU by subcutaneous inoculation, and at day 8 splenocytes and brain leukocytes were 18

391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 harvested and stained with anti-trail antibody. In the spleen, TRAIL was expressed at higher levels on CD8 + T cells compared to CD4 + T cells (mean fluorescence intensity of 56 ± 4 for CD8 + T cells and 34 ± 2 for CD4 + T cells, P < 0.006) (Fig 7A). Among total CD8 + T cells, higher levels (P < 0.0001) of TRAIL were measured on NS4B tetramer positive CD8 + T cells (Fig 7B). Analogous to data from the spleen, higher levels of TRAIL were detected on CD8 + T cells isolated from the brain compared to CD4 + T cells (mean fluorescence intensity of 40 ± 6 for CD8 + T cells and 16 ± 4 for CD4 + T cells, P < 0.02) (Fig 7C and E), and greater surface expression of TRAIL was measured on NS4B tetramer positive compared to tetramer negative CD8 + T cells (Fig 7D). These data confirm that WNV-specific CD8 + T cells express TRAIL in the spleen and brain isolated leukocytes after WNV infection. Given that CD8 + T cells in the brain expressed TRAIL, we assessed whether WNV-infected neurons expressed its cognate ligand, DR5. In the absence of infection, cortical neurons showed no apparent DR5 expression (Fig 7F, upper panels). However, within 24 hours, WNV infection resulted in efficient surface expression of DR5 (Fig 7F, lower panels). Thus, WNV infection rapidly induces DR5 expression on neurons, which could sensitize them to the inhibitory effects of TRAIL. CD8 + T cells require TRAIL to optimally control WNV infection in neurons. Our CD8 + T cell adoptive transfer and co-incubation experiments with neurons strongly suggested an inhibitory effect of TRAIL on viral replication in neurons. However, they did not provide direct insight into mechanism. We speculated that TRAIL might have a direct antiviral effect, as was seen previously with the related flavivirus, DENV (40). To address this, we added increasing concentrations of recombinant TRAIL (0.078 to 20 μg/ml) one day prior to WNV infection (MOI of 0.001) of primary cortical neurons in the 19

414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 absence of CD8 + T cells. Supernatants were collected from neurons at 24 and 48 hours after infection and virus burden was analyzed by plaque assay. Notably, we observed no reduction in WNV yield with any dose of TRAIL when added prior to infection (Fig 8A). Even when experiments were repeated with addition of recombinant TRAIL immediately after infection, no direct antiviral effect of TRAIL was observed (Fig 8B). Given these results, we suspected that the antiviral effect of TRAIL on WNV-infected neurons needed to occur in the context of CD8 + T cell-mediated clearance. Indeed, the addition of recombinant TRAIL to primed TRAIL -/- CD8 + T cells reduced WNV-infection in cortical neurons (Fig 8C; no TRAIL versus with TRAIL, 2.5-fold; P < 0.007). Reciprocally, addition of a neutralizing anti-trail antibody to wild type primed CD8 + T cells also resulted in increased WNV infection of neurons (Fig 8D: isotype control versus anti- TRAIL antibody, 3.2-fold; P = 0.05). Thus, TRAIL provides an antiviral signal to neurons but only when CD8 + T cells were present. Because of the modest effects of exogenous TRAIL and anti-trail antibodies on CD8 + T cell-mediated clearance of WNV infection in neurons, we speculated that TRAIL -/- CD8 + T cells might have additional defects in their effector functions. As prior studies had already established protective effects of cytolytic granules (perforin/granzyme) and cell-death (Fas ligand) pathways in T cell protection against WNV (34, 37), we assessed their relative expression in wild type and TRAIL -/- CD8 + T cells in the brain. Relative FasL expression was unchanged on the surface of TRAIL -/- CD4 + and CD8 + T cells in the brain (P > 0.2) (Fig 9A). This result was also observed on WNV-specific NS4B-tetramer-positive CD8 + T cells. Similarly, in the brain, granzyme B was expressed at similar levels (P > 0.2) in CD4 + and CD8 + T cells and NS4B tetramer- 20

437 438 439 positive CD8 + T cells (Fig 9B and C). Thus, the functional defect of TRAIL -/- CD8 + T cells in the context of clearance of WNV infection in the brain was not due to impaired expression of proteins associated with alternate cytolytic mechanisms. 440 441 Downloaded from http://jvi.asm.org/ on October 1, 2018 by guest 21

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 DISCUSSION TRAIL has been implicated in the control of viral infections by several different mechanisms including cytotoxic and direct antiviral effects. We initiated this study because primary human cell culture experiments showed that TRAIL was induced after DENV infection, and treatment of human myeloid cells with exogenous, recombinant TRAIL inhibited DENV infection through an apoptosis-independent mechanism (40). To address the physiological function of TRAIL in vivo in the context of infection by a flavivirus, we assessed WNV pathogenesis in wild type and TRAIL -/- mice. Mice lacking TRAIL showed relatively normal levels of WNV infection in peripheral tissues but sustained elevated viral titers in the brain and spinal cord that was associated with delayed clearance and increased mortality compared to wild type mice. No difference in WNV infection was observed in the brain and spinal cord after direct intracranial inoculation. While priming of adaptive B and T cell responses in the periphery, and trafficking of immune and antigen-specific cells to the brain was normal in TRAIL -/- mice, CD8 + T cells showed qualitative defects in their ability to clear WNV infection. Adoptive transfer studies suggested that the increased susceptibility associated with a TRAIL deficiency was due to defects in CD8 + T cell clearance mechanisms. Indeed, primed TRAIL -/- CD8 + T cells less efficiently controlled WNV infection in primary neuronal cultures ex vivo. Our experiments suggest that CD8 + T cells utilize TRAIL to control WNV pathogenesis and infection of neurons in the CNS. This conclusion is based on several experimental observations: (a) increased viral burden in the brain and spinal cord at day 10 but not day 6 or 8 after infection in TRAIL -/- mice; (b) delayed clearance of WNV 22

465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 infection in the few surviving TRAIL -/- mice; (c) less efficient clearance of infection in CD8 -/- mice after adoptive transfer of WNV-primed TRAIL -/- compared to wild type CD8 + T cells; and (d) less efficient control of WNV infection in neurons ex vivo by primed TRAIL -/- compared to wild type CD8 + T cells. This phenotype was not due to TRAILdependent effects on CD8 + T cell priming or trafficking, as equivalent percentages and numbers of polyfunctional antigen-specific CD8 + T cells were detected in WNV-infected TRAIL -/- and wild type mice. Our results are consistent with recent studies with influenza virus, which showed that a deficiency of TRAIL enhanced morbidity and viral replication through qualitatively decreased influenza-specific CD8 + T cell responses in the lung (3). Prior studies have shown that CD8 + T cells also utilize perforin and Fas ligand to clear WNV from infected neurons (34, 37). Analogous to our current studies with TRAIL - /- mice, perforin and Fas-ligand-deficient (gld) mice had elevated viral titers in the brain and spinal cord resulting in increased mortality compared to wild type mice, and in the few mice that survived infection persistent WNV infection was detected for several weeks. However, there appears to be a functional hierarchy, as a deficiency of perforin or Fas ligand resulted in less efficient clearance of WNV infection in the spleen, whereas this phenotype was absent in TRAIL -/- mice. Thus, while CD8 + T cells utilize multiple effector molecules to clear WNV from neurons in the CNS, the viral burden and survival phenotypes in TRAIL -/- mice were subordinate compared to those observed previously in perforin -/- and gld mice. Prior cell culture studies with DENV and HIV suggested that an antiviral effect of exogenous treatment of TRAIL occurs in the absence of apoptosis (33, 40). Although an analogous effect of recombinant TRAIL on WNV infection in cultured primary neurons 23

488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 was not observed, an inhibitory effect was seen in the context of a protective CD8 + T cell response. The lack of direct antiviral effect of exogenous TRAIL against WNV was consistent with some of our in vivo results: (a) we observed no difference in viral burden in peripheral tissues after subcutaneous infection; (b) we failed to detect differences in viral replication in multiple CNS regions after direct intracranial inoculation of wild type and TRAIL -/- mice. However, TRAIL did show an antiviral effect in the context of CD8 + T cell mediated clearance of WNV infection in neurons. This observation was most apparent in reconstitution experiments showing that the addition of exogenous TRAIL to TRAIL -/- CD8 + T cells enhanced the antiviral effect against WNV in neurons. What is the precise mechanism by which TRAIL facilitates optimal CD8 + T cell-mediated control of WNV infection in neurons? While studies with transformed cells in culture have shown that exogenous TRAIL sensitized respiratory syncytial virus- or human cytomegalovirusinfected cells to apoptosis (16, 31), an absence of TRAIL did not specifically affect cell death as judged by TUNEL staining of neurons in the brain or in culture (B. Shrestha and M. Diamond, unpublished data). Although further signaling analysis is warranted, our data is more consistent with a model in which TRAIL produced by CD8 + T cells provides a direct antiviral signal to neurons. Why did recombinant TRAIL alone fail to inhibit WNV infection directly in culture? We speculate that TRAIL derived by WNV-specific CD8 + T cells is necessary but not sufficient to confer the inhibitory effect, and requires additional signals through other cytolytic pathways (e.g. Fas ligand-fas and perforin/granzyme) to achieve an optimal inhibitory effect. Alternatively, activated WNV-specific effector CD8 + T cells secrete other cytokines (e.g., IFN-γ or TNF-α), which modulate expression of TRAIL ligands (e.g., DR4 and DR5) or signaling 24

511 512 components in neurons to levels that are sufficient for functional consequences. WNV infection alone can induce DR5 expression on neuronal targets in culture and in vivo. 513 In summary, our study demonstrates that CD8 + T cells require TRAIL to 514 515 516 517 518 519 520 521 522 523 524 optimally restrict WNV infection in the brain and spinal cord. Thus, CD8 + T cells utilize at least three separate effector mechanisms (TRAIL, Fas-ligand, and perforin/granzyme) to control WNV infection and injury of neurons. For highly neurovirulent viruses (e.g., New York strain of WNV), clearance mechanisms by effector CD8 + T cells appear to limit dissemination and disease and outweigh the possible pathological effects of immune targeted neuronal injury. A more complete understanding of the effector mechanisms responsible for clearance of WNV infection may inform novel vaccine strategies that optimize CD8 + T cell-mediated immunity against highly virulent viruses that target neurons in the CNS. 25

525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 FIGURE LEGENDS Figure 1. Survival and viral titer analysis of wild type and TRAIL -/- mice infected with WNV. A. Kaplan-Meier survival curves. Wild type (n = 40) and TRAIL -/- (n = 29) mice were infected via subcutaneous route with 10 2 PFU of WNV and monitored for mortality for 28 days. Survival differences were judged by the log rank test and were statistically significant (P < 0.0001). B to E. Viremia and WNV tissue burden in wild type and TRAIL -/- mice. Infectious WNV levels in (B) serum, (C) spleen, (D) brain and (E) spinal cord of wild type and TRAIL -/- mice were measured by viral plaque assay from samples harvested at the indicated time points. Data are expressed as log 10 PFU per gram of tissue or ml of serum, and reflect results for 8 to 13 mice per time point between days 2 and 10. For viral burden experiments, the horizontal bars represent the mean titers, dotted line represents the limit of sensitivity of viral detection, and asterisks indicate statistically significant differences (P < 0.05) between wild type and TRAIL -/- mice as determined by the Mann-Whitney test. Figure 2. WNV burden in the CNS after intracranial infection. Wild type and TRAIL -/- mice were infected with 10 1 PFU of WNV via an intracranial route. Different regions of brain (A, cerebellum; B, brain stem; C, cerebral cortex; D, subcortex; and E, spinal cord) were harvested at the indicated time points. Tissue homogenates were analyzed for viral burden by plaque assay. Data are shown as PFU per gram of tissue for 5 to 8 mice per time point. The dotted line represents the limit of detection. None of the differences achieved statistical significance. Figure 3. WNV-specific antibody responses in wild type and TRAIL -/- mice. Wild type and TRAIL -/- mice were infected with WNV via a subcutaneous route and serum was collected at the indicated time points. The development of WNV-specific (A) 26

549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 IgM or (B) IgG was determined by ELISA using purified WNV E protein. Data are the average of 5 mice at day 2 and 10 mice at days 4 through 8. (C) Neutralizing antibody response. Neutralizing titers were determined by a PRNT assay. All samples were serially diluted in duplicate and data are expressed as the reciprocal PRNT50, the antibody titer that reduced plaque number by 50%. Data are the average of 10 mice per time point. None of the differences achieved statistical significance (as indicated by n.s.). Figure 4. T cell responses in the spleen of wild type and TRAIL -/- mice after WNV infection. A-D. CD8 + T cells. WNV-infected splenocytes from wild type or TRAIL -/- mice were harvested at day 7 and stimulated ex vivo for six hours with (A, B) an immunodominant D b -restricted NS4B peptide. Cells were co-stained for CD8 and intracellular IFN-γ or TNF-α and processed by flow cytometry. Data is presented as the (A) percentage or (B) total number of positive cells. The (C) percentage and (D) total numbers of antigen specific CD8 + T cells in uninfected or WNV-infected wild type and TRAIL -/- mice as determined by binding to a D b -restricted NS4B tetramer. Asterisks indicate statistically significant (P < 0.05) differences between wild type and TRAIL -/- mice. E-F. CD4 + T cells. WNV-infected splenocytes from wild type or TRAIL -/- mice were harvested at day 7 and stimulated ex vivo for six hours with a stimulatory anti-cd3 antibody. Cells were co-stained for CD4 and intracellular IFN-γ or TNF-α and processed by flow cytometry. Data is presented as the (E) percentage or (F) total number of positive cells. Data from panels A-F represents from the 6 to 10 mice from three independent experiments. In panels A, C, and E, the percentage of IFN-γ, TNF-α, or NS4B tetramer positive cells represents the fraction of total gated CD8 + or CD4 + T cells. G. Regulatory T cells. Splenocytes from WNV-infected wild type (naïve or infected) and TRAIL -/- mice 27

572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 were harvested at day 7 and stained with antibodies to CD4, CD25 and FoxP3. Data is expressed as the percentage of CD4 + T cells that stained positive for CD25 and FoxP3. Data were pooled from 6 mice from 2 independent experiments. H. Flow cytometry profiles. Representative flow cytometry profiles of intracellular IFN-γ staining (upper panels) and NS4b tetramer staining (bottom panels) of splenic CD8 + T cells from wild type or TRAIL -/- mice at day 7 after WNV infection. Figure 5. Accumulation of leukocytes in the CNS of WNV-infected wild type and TRAIL -/- mice. Wild type and TRAIL -/- mice were infected with 10 2 PFU of WNV by subcutaneous route. Eight days later, brains were harvested and leukocytes were isolated by Percoll gradient centrifugation. A-B. The percentage (A) and total number (B) of microglia (CD11b + CD45 lo ) and macrophages (CD11b + CD45 hi ) are shown. C-D. The percentage (C) and total number (D) of CD8 + T cells, CD4 + T cells, and CD19 + B220 + B cells are shown. E-F. The number of WNV-specific CD8 + T cells in the brain as judged by (E) NS4B peptide restimulation and intracellular IFN-γ staining or (F) NS4B tetramer staining. G. The number of activated CD4 + T cells in the brain as judged after incubation with a stimulatory anti-cd3 antibody and intracellular IFN-γ staining. Data represent the average of two independent experiments with 6 mice per group. Asterisks indicate statistically significant (P < 0.05) differences between wild type and TRAIL -/- mice. H. Representative flow cytometry profiles of CD3 + CD8 + T cells in the brain (left) from TRAIL -/- mice eight days after WNV infection. Surface NS4b tetramer staining (middle) or intracellular IFN-γ staining after NS4B peptide restimulation ex vivo is shown (right) on brain CD8 + T cells. 28

594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 Figure 6. Role of TRAIL in CD8 + T cell mediated control of WNV infection in the brain. A-B. Viral yield in different tissues after adoptive transfer of WNV-primed CD8 + T cells. Donor CD8 + T cells (A, 10 x 10 6 ; B, 3 x 10 6 ) were purified from naïve or WNV-primed wild type and TRAIL -/- mice and transferred into recipient CD8 -/- mice one day after infection. The spleen, brain, and spinal cord of recipient CD8 -/- mice were harvested on day 10 after initial infection and WNV was titrated by plaque assay. Data represents results from 3 to 4 mice in two independent experiments. The dotted line represents the limit of sensitivity of the assay. C-D. WNV persistence in CNS tissues. Wild type and TRAIL -/- mice were infected with WNV and (C) brain and (D) spinal cord were harvested at days 15, 21, and 28 after infection, and viral yield was titrated by plaque assay. Results are from 4 to 9 mice per time point. E-F. Control of WNV infection in primary neurons ex vivo after addition of primed CD8 + T cells. Cortical neurons were infected with WNV at an MOI of 0.001. After one hour, purified naïve or WNV-primed CD8 + T cells from wild type or TRAIL -/- mice were added at an E:T ratio of 50:1. After (E) 24 or (F) 48 hours, supernatants were harvested and WNV production was measured by plaque assay. Asterisks indicate statistically significant differences in this Figure (*, P < 0.05, ** P < 0.005 and *** P< 0.0001), and n.s. indicates no significant difference. Figure 7. CD8 + T cells express TRAIL. Splenocytes and brain leukocytes from WNV-infected wild type mice were harvested at day 7 and stained for surface expression of TRAIL on CD4 and CD8 + T cells. Data were analyzed by flow cytometry and are expressed as the mean fluorescence intensity only of the positive cells. A. TRAIL expression on total CD4 + and CD8 + T cells in the spleen. B. TRAIL expression on NS4B tetramer positive and tetramer negative CD8 + T cells in the spleen. C. TRAIL expression 29

617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 on total CD4 + and CD8 + T cells in the brain. D. TRAIL expression on NS4B tetramer positive and tetramer negative CD8 + T cells in the brain. Data are pooled from four mice, and asterisks indicate differences that are statistically significant. E. Representative flow cytometry profiles after staining with anti-trail antibody on the surface of CD4 + and CD8 + T cells from the brains of mice infected with WNV. F. DR5 expression on neurons. Neurons were uninfected or infected with WNV (MOI of 0.001) for 24 hours. Cells were harvested and stained for surface DR5 and intracellular WNV antigen, counter-stained with ToPro3, and analyzed by confocal microscopy. Representative data from multiple images are shown. Figure 8. Mechanism by which TRAIL inhibits WNV infection in neurons. A. Recombinant TRAIL does not have a direct antiviral effect on cortical neurons. Primary cortical neurons were treated with increasing doses (0.078 to 20 μg/ml) of recombinant TRAIL for 24 hours prior to infection with WNV (MOI of 0.001). One hour later free virus and TRAIL were removed by washing. Supernatants were harvested 24 and 48 hours after infection and titrated by plaque assay. Results are representative of three independent experiments performed in triplicate. B. Post-exposure addition of recombinant TRAIL does not confer an antiviral effect against WNV infection of cortical neurons. Neurons were infected with WNV (MOI of 0.001). One hour later, free virus was removed by extensive washing. Subsequently, increasing doses (0.078 to 20 μg/ml) of recombinant TRAIL were added and supernatants were harvested 24 hours later and titrated by plaque assay. Results are representative of three independent experiments performed in triplicate. None of the doses of recombinant TRAIL conferred a statistically significant antiviral effect. C. Addition of recombinant TRAIL enhances the antiviral 30

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 effect of primed TRAIL -/- CD8 + T cells. Cortical neurons were infected with WNV at an MOI of 0.001. After one hour, purified WNV-primed CD8 + T cells from TRAIL -/- mice were added at an E:T ratio of 50:1. At the same time, recombinant TRAIL (10 μg/ml) was added to some of the cell cultures. Supernatants were harvested 48 hours later, and WNV production was measured by plaque assay. D. Addition of neutralizing anti-trail antibody decreases the antiviral effect of primed wild type CD8 + T cells. Cortical neurons were infected with WNV at an MOI of 0.001. After one hour, purified WNV-primed CD8 + T cells from wild type mice were added at an E:T ratio of 50:1. At the same time, a neutralizing anti-trail or isotype control antibody (10 μg/ml) was added to cells. Supernatants were harvested 48 hours later, and WNV production was measured by plaque assay. Data in panels C and D was pooled from six replicate samples generated from independent experiments. Figure 9. FasL and granzyme B expression on CD8 + T cells from the brains of WNV-infected wild type and TRAIL -/- mice. Brain leukocytes from WNV-infected wild type mice were harvested at day 8 and stained for surface expression of FasL and intracellular expression of granzyme B. Data were analyzed by flow cytometry, and all results are pooled from two independent experiments. A. FasL expression on total CD4 + and CD8 + T cells and on NS4B tetramer positive CD8 + T cells in the brain. Note, 100% of brain T cells expressed FasL on their surface. Data is expressed as the normalized mean fluorescence intensity of the FasL positive cells. B-C. Granzyme B expression on total CD4 + and CD8 + T cells and on NS4B tetramer positive CD8 + T cells in the brain. Data is expressed as the percentage of granzyme B positive cells (B) and the normalized 31

662 663 mean fluorescence intensity of the granzyme B positive cells (C). Data are pooled from nine mice, and none of the differences were statistically significant. 664 665 666 Downloaded from http://jvi.asm.org/ on October 1, 2018 by guest 32

667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 REFERENCES 1. Beraza, N., Y. Malato, L. E. Sander, M. Al-Masaoudi, J. Freimuth, D. Riethmacher, G. J. Gores, T. Roskams, C. Liedtke, and C. Trautwein. 2009. Hepatocyte-specific NEMO deletion promotes NK/NKT cell- and TRAILdependent liver damage. J Exp Med 206:1727-37. 2. Brincks, E. L., P. Gurung, R. A. Langlois, E. A. Hemann, K. L. Legge, and T. S. Griffith. 2011. The magnitude of the T cell response to a clinically significant dose of influenza virus is regulated by TRAIL. J Immunol 187:4581-8. 3. Brincks, E. L., A. Katewa, T. A. Kucaba, T. S. Griffith, and K. L. Legge. 2008. CD8 T cells utilize TRAIL to control influenza virus infection. J Immunol 181:4918-25. 4. Busch, M. P., D. J. Wright, B. Custer, L. H. Tobler, S. L. Stramer, S. H. Kleinman, H. E. Prince, C. Bianco, G. Foster, L. R. Petersen, G. Nemo, and S. A. Glynn. 2006. West Nile virus infections projected from blood donor screening data, United States, 2003. Emerg Infect Dis 12:395-402. 5. Daffis, S., M. A. Samuel, B. C. Keller, M. Gale, Jr., and M. S. Diamond. 2007. Cell-specific IRF-3 responses protect against West Nile virus infection by interferon-dependent and independent mechanisms. PLoS Pathog 3:e106. 6. Daffis, S., M. A. Samuel, M. S. Suthar, B. C. Keller, M. Gale, Jr., and M. S. Diamond. 2008. Interferon regulatory factor IRF-7 induces the antiviral alpha interferon response and protects against lethal West Nile virus infection. J Virol 82:8465-75. 7. Diamond, M. S., E. Mehlhop, T. Oliphant, and M. A. Samuel. 2009. The host immunologic response to West Nile encephalitis virus. Front Biosci 14:3024-34. 8. Diamond, M. S., B. Shrestha, A. Marri, D. Mahan, and M. Engle. 2003. B cells and antibody play critical roles in the immediate defense of disseminated infection by West Nile encephalitis virus. J Virol 77:2578-2586. 9. Diehl, G. E., H. H. Yue, K. Hsieh, A. A. Kuang, M. Ho, L. A. Morici, L. L. Lenz, D. Cado, L. W. Riley, and A. Winoto. 2004. TRAIL-R as a negative regulator of innate immune cell responses. Immunity 21:877-89. 33

697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 10. Ebel, G. D., J. Carricaburu, D. Young, K. A. Bernard, and L. D. Kramer. 2004. Genetic and phenotypic variation of West Nile virus in New York, 2002-2003. Am J Trop Med Hyg 71:493-500. 11. Glass, W. G., D. H. McDermott, J. K. Lim, S. Lekhong, S. F. Yu, W. A. Frank, J. Pape, R. C. Cheshier, and P. M. Murphy. 2006. CCR5 deficiency increases risk of symptomatic West Nile virus infection. J Exp Med 203:35-40. 12. Hoffmann, O., J. Priller, T. Prozorovski, U. Schulze-Topphoff, N. Baeva, J. D. Lunemann, O. Aktas, C. Mahrhofer, S. Stricker, F. Zipp, and J. R. Weber. 2007. TRAIL limits excessive host immune responses in bacterial meningitis. J Clin Invest 117:2004-13. 13. Hu, W. H., H. Johnson, and H. B. Shu. 1999. Tumor necrosis factor-related apoptosis-inducing ligand receptors signal NF-kappaB and JNK activation and apoptosis through distinct pathways. J Biol Chem 274:30603-10. 14. Indraccolo, S., U. Pfeffer, S. Minuzzo, G. Esposito, V. Roni, S. Mandruzzato, N. Ferrari, L. Anfosso, R. Dell'Eva, D. M. Noonan, L. Chieco-Bianchi, A. Albini, and A. Amadori. 2007. Identification of genes selectively regulated by IFNs in endothelial cells. J Immunol 178:1122-35. 15. Ishikawa, E., M. Nakazawa, M. Yoshinari, and M. Minami. 2005. Role of tumor necrosis factor-related apoptosis-inducing ligand in immune response to influenza virus infection in mice. J Virol 79:7658-63. 16. Kotelkin, A., E. A. Prikhod'ko, J. I. Cohen, P. L. Collins, and A. Bukreyev. 2003. Respiratory syncytial virus infection sensitizes cells to apoptosis mediated by tumor necrosis factor-related apoptosis-inducing ligand. J Virol 77:9156-72. 17. Lanteri, M. C., K. M. O'Brien, W. E. Purtha, M. J. Cameron, J. M. Lund, R. E. Owen, J. W. Heitman, B. Custer, D. F. Hirschkorn, L. H. Tobler, N. Kiely, H. E. Prince, L. C. Ndhlovu, D. F. Nixon, H. T. Kamel, D. J. Kelvin, M. P. Busch, A. Y. Rudensky, M. S. Diamond, and P. J. Norris. 2009. Tregs control the development of symptomatic West Nile virus infection in humans and mice. J Clin Invest 119:3266-77. 18. Lim, J. K., C. Y. Louie, C. Glaser, C. Jean, B. Johnson, H. Johnson, D. H. McDermott, and P. M. Murphy. 2008. Genetic deficiency of chemokine 34

728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 receptor CCR5 is a strong risk factor for symptomatic West Nile virus infection: a meta-analysis of 4 cohorts in the US epidemic. J Infect Dis 197:262-5. 19. Lim, J. K., D. H. McDermott, A. Lisco, G. A. Foster, D. Krysztof, D. Follmann, S. L. Stramer, and P. M. Murphy. 2010. CCR5 deficiency is a risk factor for early clinical manifestations of West Nile virus infection but not for viral transmission. J Infect Dis 201:178-85. 20. Lin, Y., A. Devin, A. Cook, M. M. Keane, M. Kelliher, S. Lipkowitz, and Z. G. Liu. 2000. The death domain kinase RIP is essential for TRAIL (Apo2L)- induced activation of IkappaB kinase and c-jun N-terminal kinase. Mol Cell Biol 20:6638-45. 21. Mehlhop, E., and M. S. Diamond. 2006. Protective immune responses against West Nile virus are primed by distinct complement activation pathways. J Exp Med 203:1371-81. 22. Merino, D., N. Lalaoui, A. Morizot, P. Schneider, E. Solary, and O. Micheau. 2006. Differential inhibition of TRAIL-mediated DR5-DISC formation by decoy receptors 1 and 2. Mol Cell Biol 26:7046-55. 23. Micheau, O. 2003. Cellular FLICE-inhibitory protein: an attractive therapeutic target? Expert Opin Ther Targets 7:559-73. 24. Murray, K., S. Baraniuk, M. Resnick, R. Arafat, C. Kilborn, K. Cain, R. Shallenberger, T. L. York, D. Martinez, J. S. Hellums, D. Hellums, M. Malkoff, N. Elgawley, W. McNeely, S. A. Khuwaja, and R. B. Tesh. 2006. Risk factors for encephalitis and death from West Nile virus infection. Epidemiol Infect 134:1325-32. 25. Petersen, L. R., A. A. Marfin, and D. J. Gubler. 2003. West Nile virus. Jama 290:524-8. 26. Rigaud, S., M. C. Fondaneche, N. Lambert, B. Pasquier, V. Mateo, P. Soulas, L. Galicier, F. Le Deist, F. Rieux-Laucat, P. Revy, A. Fischer, G. de Saint Basile, and S. Latour. 2006. XIAP deficiency in humans causes an X-linked lymphoproliferative syndrome. Nature 444:110-4. 35

757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 27. Samuel, M. A., and M. S. Diamond. 2006. Pathogenesis of West Nile virus infection: A balance between virulence, innate and adaptive immunity, and viral evasion. J Virol 80:9349-9360. 28. Samuel, M. A., and M. S. Diamond. 2005. Type I IFN protects against lethal West Nile Virus infection by restricting cellular tropism and enhancing neuronal survival. J Virol 79:13350-13361. 29. Sato, K., S. Hida, H. Takayanagi, T. Yokochi, N. Kayagaki, K. Takeda, H. Yagita, K. Okumura, N. Tanaka, T. Taniguchi, and K. Ogasawara. 2001. Antiviral response by natural killer cells through TRAIL gene induction by IFNalpha/beta. Eur J Immunol 31:3138-46. 30. Secchiero, P., F. Corallini, M. G. di Iasio, A. Gonelli, E. Barbarotto, and G. Zauli. 2005. TRAIL counteracts the proadhesive activity of inflammatory cytokines in endothelial cells by down-modulating CCL8 and CXCL10 chemokine expression and release. Blood 105:3413-9. 31. Sedger, L. M., D. M. Shows, R. A. Blanton, J. J. Peschon, R. G. Goodwin, D. Cosman, and S. R. Wiley. 1999. IFN-gamma mediates a novel antiviral activity through dynamic modulation of TRAIL and TRAIL receptor expression. J Immunol 163:920-6. 32. Sejvar, J. J., A. V. Bode, A. A. Marfin, G. L. Campbell, J. Pape, B. J. Biggerstaff, and L. R. Petersen. 2006. West Nile Virus-associated flaccid paralysis outcome. Emerg Infect Dis 12:514-6. 33. Shepard, B. D., D. De Forni, D. R. McNamara, A. Foli, S. A. Rizza, R. S. Abraham, K. Knutson, P. J. Wettstein, F. Lori, and A. D. Badley. 2008. Beneficial effect of TRAIL on HIV burden, without detectable immune consequences. PLoS One 3:e3096. 34. Shrestha, B., and M. S. Diamond. 2007. Fas Ligand interactions contribute to CD8+ T cell-mediated control of West Nile virus infection in the central nervous system. J Virol 81:11749-11757. 35. Shrestha, B., and M. S. Diamond. 2004. The role of CD8+ T cells in the control of West Nile virus infection. J Virol 78:8312-8321. 36

787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 36. Shrestha, B., T. Ng, H. J. Chu, M. Noll, and M. S. Diamond. 2008. The relative contribution of antibody and CD8+ T cells to vaccine immunity against West Nile encephalitis virus. Vaccine 26:2020-33. 37. Shrestha, B., M. A. Samuel, and M. S. Diamond. 2006. CD8+ T cells require perforin to clear West Nile virus from infected neurons. J Virol 80:119-129. 38. Shrestha, B., B. Zhang, W. E. Purtha, R. S. Klein, and M. S. Diamond. 2008. Tumor necrosis factor alpha protects against lethal West Nile virus infection by promoting trafficking of mononuclear leukocytes into the central nervous system. J Virol 82:8956-64. 39. Wang, Y., M. Lobigs, E. Lee, and A. Mullbacher. 2003. CD8+ T cells mediate recovery and immunopathology in West Nile virus encephalitis. J Virol 77:13323-34. 40. Warke, R. V., K. J. Martin, K. Giaya, S. K. Shaw, A. L. Rothman, and I. Bosch. 2008. TRAIL is a novel antiviral protein against dengue virus. J Virol 82:555-64. 41. Zamai, L., M. Ahmad, I. M. Bennett, L. Azzoni, E. S. Alnemri, and B. Perussia. 1998. Natural killer (NK) cell-mediated cytotoxicity: differential use of TRAIL and Fas ligand by immature and mature primary human NK cells. J Exp Med 188:2375-80. 37