当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 病菌学杂志 > 2005年 > 第14期 > 正文
编号:11202801
Castanospermine, a Potent Inhibitor of Dengue Viru
     Departments of Medicine

    Molecular Microbiology

    Pathology and Immunology, Washington University School of Medicine, St. Louis, Missouri 63110

    Department of Microbiology, University of Pennsylvania, Philadelphia, Pennsylvania 19105

    Apath, LLC, St. Louis, Missouri 63141

    ABSTRACT

    Previous studies have suggested that -glucosidase inhibitors such as castanospermine and deoxynojirimycin inhibit dengue virus type 1 infection by disrupting the folding of the structural proteins prM and E, a step crucial to viral secretion. We extend these studies by evaluating the inhibitory activity of castanospermine against a panel of clinically important flaviviruses including all four serotypes of dengue virus, yellow fever virus, and West Nile virus. Using in vitro assays we demonstrated that infections by all serotypes of dengue virus were inhibited by castanospermine. In contrast, yellow fever virus and West Nile virus were partially and almost completely resistant to the effects of the drug, respectively. Castanospermine inhibited dengue virus infection at the level of secretion and infectivity of viral particles. Importantly, castanospermine prevented mortality in a mouse model of dengue virus infection, with doses of 10, 50, and 250 mg/kg of body weight per day being highly effective at promoting survival (P 0.0001). Correspondingly, castanospermine had no adverse or protective effect on West Nile virus mortality in an analogous mouse model. Overall, our data suggest that castanospermine has a strong antiviral effect on dengue virus infection and warrants further development as a possible treatment in humans.

    INTRODUCTION

    Dengue fever, the most prevalent arthropod-borne viral illness in humans, is caused by dengue virus (DEN). DEN is a single-stranded, positive-polarity, enveloped RNA virus that is translated in the cytoplasm as a single polyprotein and cleaved into three structural and seven nonstructural proteins. Four related serotypes of DEN exist in nature and are transmitted to humans primarily by two mosquitoes, Aedes aegypti and Aedes albopictus. DEN is a member of the Flaviviridae family and is genetically related to the viruses that cause yellow fever, hepatitis C, and the Japanese, St. Louis, and West Nile encephalitides. DEN infection results in a spectrum of disease ranging from a debilitating, self-limited illness (dengue fever) to a life-threatening syndrome (dengue hemorrhagic fever [DHF]). DEN causes disease globally with an estimated 25 to 100 million new infections per year (34). At present, no vaccine has been approved for human use and treatment is supportive.

    Flavivirus assembly takes place at the endoplasmic reticulum (ER) (7). The structural glycoproteins prM and E localize to the luminal side of the ER and form an immature particle with prM and E in a heterodimeric complex (7, 58). Furin-mediated proteolysis of prM in the trans-Golgi network (48) triggers rearrangement, homodimerization of E, and formation of the mature viral particle before release from the infected cell (1, 16). In flavivirus-infected mammalian cells, a 14-residue oligosaccharide, (Glc)3(Man)9(GlcNAc)2, is added in the ER to specific asparagine residues on the prM and E proteins. This high-mannose carbohydrate is sequentially modified in the ER by resident -glucosidases to generate N-linked glycans that lack the terminal (1,2) and both (1,3) glucose residues (19). Recent experiments suggest that trimming of N-linked carbohydrates in the ER may be required for proper assembly or secretion of DEN (8, 55).

    Castanospermine is a natural alkaloid derived from the black bean or Moreton Bay chestnut tree (Castanospermum australae). The compound is water soluble and can be readily isolated in large quantity through a rather simple purification scheme (37). In cells, castanospermine acts as an ER -glucosidase I inhibitor and reduces infection of a subset of enveloped RNA and DNA viruses in vitro (DEN type 1 [DEN-1] [8], bovine diarrhea virus [36, 54], Sindbis virus [44], cytomegalovirus [56], human immunodeficiency virus [6, 49-52], and influenza virus [15, 38]) and in vivo (herpes simplex virus [5] and Rauscher murine leukemia virus [40]). Studies of its mechanism of action suggest that castanospermine may disrupt folding of some viral proteins by preventing the removal of the terminal glucose residue on N-linked glycans. A lack of modification of the high-mannose sugars on some viral proteins may inhibit requisite interactions with the protein folding chaperones calnexin and calreticulin (33, 57). At least some flaviviruses appear sensitive to -glucosidase inhibitors, such as castanospermine and deoxynojirimycin. Both drugs inhibited DEN-1 infection in BHK-21 cells (8). Moreover, a related -glucosidase inhibitor, N-nonyl-deoxynojirimycin, inhibited infection by DEN type 2 (DEN-2) more than that by Japanese encephalitis virus in BHK cells (55).

    In this study, we further evaluated the inhibitory activity of castanospermine against flaviviruses. We demonstrate that all serotypes of DEN were inhibited by castanospermine, yet yellow fever virus (YFV) and West Nile virus (WNV) were more resistant to the effects of the drug. Castanospermine acts by blocking the production and infectivity of DEN but not WNV or virus-like particles (VLP). Administration of castanospermine prevented mortality in mice even when DEN-2 was inoculated by an intracranial route. Our data suggest that castanospermine inhibits infection of a subset of flaviviruses and may have potential in vivo as an antiviral agent against DEN in humans.

    MATERIALS AND METHODS

    Inhibitors. Castanospermine (1S,6S,7R,8R,8aR-1,6,7,8-tetrahydroxyindolizidine), isolated from the seeds of the Moreton Bay chestnut (Castanospermum australe), was kindly provided by R. Smith (Phytex Australia Pty Ltd., Peakhurst, Australia). Mycophenolic acid was purchased commercially (Sigma Chemical, St. Louis, MO).

    Virus strains, cell lines, and mice. Mouse-neuroadapted DEN-2 New Guinea C strain was kindly provided by T. Chambers (St. Louis, MO), and the prototype DEN-2 DHF strain 16681 has been previously described (41). Low-passage-number isolates (DEN-1 [Brazil], DEN-2 [N1042; Nicaragua], DEN-3 [Sri Lanka], and DEN-4 [Tahiti]) were the gift of E. Harris (Berkeley, CA). The 17D vaccine strain of YFV was obtained from T. Chambers (St. Louis, MO). The WNV strain (3000.0259) was isolated in New York in 2000 (14). For inoculation in mice, virus was diluted in Hanks balanced salt solution and 1% heat-inactivated fetal bovine serum. Inbred mouse strains C57BL/6 and A/J were obtained commercially (Jackson Laboratory, Bar Harbor, Maine) and used at 8 and 4 weeks of age, respectively. Mice were infected with WNV via a subcutaneous route and with DEN via an intracranial route. All mouse work was conducted according to both federal and Washington University ethical guidelines.

    Plaque and flow cytometric assays. Huh-7 and BHK-21 cells were infected as a subconfluent monolayer with DEN-2 strain 16681 at a range of multiplicities of infection (MOI) as indicated. Following a 2-hour infection, the monolayers were washed four times and the growth medium (Dulbecco modified Eagle medium [DMEM] with 5% fetal calf serum) was replaced with medium supplemented with a range of concentrations of castanospermine. Tissue culture supernatants were harvested 24 or 72 h later, and the titer of infectious virus was determined by plaque assay using BHK-21 cells as previously described (10). To determine the percentage of cells that expressed DEN, WNV, or YFV antigen, flow cytometric analysis was performed as described previously (10). Low-passage-number isolates representing all four DEN serotypes were used to infect BHK-21 cells at an MOI of 0.1. Infected cells were cultured in the presence of castanospermine for 48 h prior to flow cytometric analysis.

    Plaque reduction assay. A subconfluent monolayer of BHK-21 cells was infected with low-passage-number DEN isolates representing each of the four serotypes at 102 PFU per well. Cells were then overlaid with minimal essential medium containing 500 μM castanospermine, 1% low-melting-point agarose, and 2.5% fetal bovine serum. Seven days after infection, cells were fixed with 10% formalin and stained with crystal violet, and plaques were counted using a light box.

    Western blotting. BHK-21 cells (7 x 105) were infected with DEN (MOI of 0.1) in the presence or absence of castanospermine. Two days later, cells were harvested after treatment with phosphate-buffered saline (PBS) supplemented with 3 mM EDTA. After being washed once in PBS, cells were pelleted, lysed in RIPA buffer (20 mM Tris [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1% Na-deoxycholate, 0.1% sodium dodecyl sulfate [SDS], and 1x protease inhibitors [Complete-mini; Roche Pharmaceuticals, Nutley, NJ]), mixed with SDS sample buffer, boiled, and electrophoresed by 15% reducing SDS-polyacrylamide gel electrophoresis. Some of the samples were treated with endo-?-N-acetylglucosaminidase (endo H) glycosidase before electrophoresis according to the manufacturer's instructions (New England BioLabs, Beverly, MA). Transfer was performed onto Immobilon P membranes (Millipore Corporation, Bedford, MA), and blots were probed with affinity-purified rabbit polyclonal antibodies against DEN prM protein (gift of R. Levis, Food and Drug Administration) and detected with a horseradish peroxidase-conjugated goat anti-rabbit antibody (Pierce Biotechnology, Rockford, IL) and a chemiluminescent substrate (ECL Plus; Amersham Biosciences, Piscataway, NJ).

    Viral particle enzyme-linked immunosorbent assay (ELISA). To trap DEN viral particles, the 2H2 anti-prM monoclonal antibody (MAb) (20) was diluted (2 μg/ml in 0.1 M sodium carbonate buffer, pH 9.3) and adsorbed to Maxi-Sorp microtiter plates (Nalge Nunc International, Rochester, NY). Nonspecific binding sites were blocked after incubation with blocking buffer (PBS, 0.05% Tween 20, 3% bovine serum albumin, and 3% horse serum) for 1 h at 37°C. Viral supernatants from medium- or castanospermine-treated cells were clarified (14,000 x g for 10 minutes), concentrated (SW41 rotor, 100,000 x g for 2 hours at 4°C), and added to individual wells in duplicate for 2 hours at room temperature. Plates were washed four times in a biosafety hood with PBS supplemented with 0.05% Tween 20. Subsequently, biotinylated 4G2 (2 μg/ml), a MAb that recognizes a cross-reactive epitope on all flavivirus E proteins (20), was added for 1 hour at room temperature. After four additional washes in PBS with 0.05% Tween 20, horseradish peroxidase-conjugated streptavidin (2 μg/ml; Zymed Laboratories, South San Francisco, CA) was added for 1 hour at room temperature. After six final washes with PBS, signal was detected after addition of TMB substrate (DakoCytomation, Carpinteria, CA) and 0.1 N H2SO4. Plates were evaluated at 450 nm on a 96-well plate reader (Genios Pro; Tecan Instruments, Reading, United Kingdom).

    Quantitation of viral RNA. DEN and WNV viral RNAs were measured by fluorogenic reverse transcription-PCR (RT-PCR) with previously described virus-specific primers (21, 28). Briefly, supernatants were harvested from medium- or castanospermine-treated cells 24 or 48 h after DEN or WNV infection and centrifuged (14,000 x g for 10 min) to remove cellular debris. Carrier yeast tRNA (1 μg/sample) was added to an aliquot (100 μl) of supernatant, and viral RNA was recovered using the RNeasy minikit (Qiagen, Valencia, CA). Samples were processed using the TaqMan RT-PCR buffer system and an ABI 7000 sequence detection system (Applied Biosystems, Foster City, CA).

    VLP production. Pseudoinfectious VLP that incorporate the prM-E proteins of different flaviviruses were produced by complementation of a subgenomic WNV replicon and will be described in greater detail elsewhere (T. Pierson, M. Sanchez, B. Puffer, A. Ahmed, B. Geiss, L. Valentine, L. Altamura, M. Diamond, and R. Doms, unpublished data). Briefly, BHK-rep-REN cells that stably propagate a subgenomic replicon of a lineage II strain of WNV encoding Renilla luciferase were transfected with two DNA expression vectors encoding capsid (C) and the prM-E proteins of WNV or DEN (DEN-1 strain; WestPac). In each case, the prM, E, and C proteins were derived from the homologous virus. Twelve hours after transfection, the medium was replaced with a low-glucose formulation of DMEM in the presence or absence of 500 μM castanospermine. Supernatants containing infectious particles were harvested 30 hours later and clarified using 0.45-μm syringe filters. Virus particles were concentrated by ultracentrifugation for 2 hours through a 20% sucrose cushion in an SW55 rotor at 40,000 rpm (4°C). Pelleted particles were resuspended in DMEM and used to infect preplated BHK-21 target cells in 96-well plates. Infection was assayed 40 hours postinfection by measuring Renilla luciferase activity (Promega) using a Trilux 1450 MicroBeta luminometer.

    WNV and DEN subgenomic replicons. The lineage I WNV replicon plasmid pWN5'RucPur was generated from a genomic clone of the NY 1999 strain (plasmids pWN-AB1 and pWN-CG) provided by R. Kinney (Centers for Disease Control and Prevention, Fort Collins, CO). pWN5'RucPur was generated by deleting WNV nucleotides 181 to 2379 and fusing the first 31 amino acids of the C protein to a fusion protein containing Renilla luciferase (31), ubiquitin (3), and the puromycin N-acetyltransferase (PAC) gene (53). The encephalomyocarditis virus internal ribosome entry site (4) was placed downstream of the PAC stop codon, so that translation of the WNV structural proteins begins at nucleotide 2380 (methionine 794). The DEN-2 replicon plasmid pD2-hRucPac was similarly generated from a genomic clone of the 16681 strain of DEN-2 (27). pD2-hRucPac was created by deleting DEN-2 nucleotides 180 to 2342 and fusing the first 28 amino acids of the capsid protein with the Renilla luciferase, ubiquitin, and PAC genes as described above. The encephalomyocarditis virus internal ribosome entry site was placed downstream of the PAC stop codon and initiates translation of DEN nonstructural proteins at nucleotide 2343.

    DNA template for replicon RNA transcription was prepared by linearization of pWN5'RucPur or pD2-hRucPac with XbaI restriction endonuclease followed by phenol-chloroform extraction and ethanol precipitation. Replicon RNA was generated using the Amplicap T7 High Yield Message Maker kit (Epicenter Technologies, Madison, WI). Ten micrograms of T7 RNA transcripts was electroporated into 5 x 106 BHK-21 cells with a Bio-Rad Gene Pulser II with 1 pulse at 1.5 kV at 25 μF. Replicon-containing cells were stably selected with 3-μg/ml puromycin (Sigma-Aldrich, St. Louis, MO). BHK pWN5'RucPur cells or pD2-hRucPur cells were seeded into a 96-well plate at 2,000 cells/well without puromycin. One day later the medium was changed, and increasing concentrations of castanospermine were added. Cells were incubated in the presence of drug for 48 h, washed, and assayed for marker gene expression using a Renilla luciferase assay kit (Promega Corp., Madison, WI).

    Mouse infections. A/J mice (28 to 31 days old) were infected with 105 PFU of mouse-adapted DEN-2 via the intracranial route. C57BL/6 (8-week-old) mice were infected by footpad inoculation with 102 PFU of WNV. Mock infections were similarly performed except that virus was omitted. Mice were treated with a range of doses of castanospermine (25, 5, 1, or 0.2 mg/day) or vehicle by intraperitoneal injection. Mice were treated at the time of infection and daily for 10 days. Mice were monitored during treatment and for a period of 15 additional days. Mortality was recorded, and Kaplan-Meier statistics (log rank test) were used to compare the treated with the untreated groups.

    RESULTS

    Inhibition of virus infection in Huh-7 and BHK-21 cells. Previous studies suggested that castanospermine, a pharmacological inhibitor of ER -glucosidases, blocks trimming of N-linked carbohydrates and abrogates DEN-1 infection by preventing proper processing of the envelope glycoproteins (8). As a first step towards evaluating the utility of castanospermine as a broad-spectrum antiviral against DEN, we assessed its ability to inhibit the DEN-2 strain 16681, which replicates efficiently in a range of cell lines including BHK-21 and Huh-7 cells (11). Treatment of cells with castanospermine inhibited the yield of infectious virus in a dose-dependent manner (Fig. 1A, B, and C). A higher concentration of castanospermine was required to inhibit the production of infectious DEN-2 in the Huh-7 human hepatoma cell line (50% inhibitory concentration [IC50], 85.7 μM) than in BHK-21 cells (IC50, 1 μM). The IC50 of castanospermine in BHK-21 cells was relatively independent of the inoculating dose of DEN, as similar values were observed over a broad range of multiplicities of infection (Fig. 1B). As observed previously in Neuro 2a cells (8), castanospermine treatment efficiently slowed the electrophoretic mobility of DEN prM, one of the glycosylated structural proteins. This difference in size between medium- and castanospermine-treated cells was restored after incubation of both proteins with endo H glycosidase (Fig. 1D).

    Inhibitory effect of castanospermine on other DEN serotypes, WNV, and YFV. For castanospermine to be considered as a useful anti-DEN agent it should be highly active against all DEN serotypes. Treatment of BHK-21 cells with castanospermine strongly inhibited virus secretion and cell-to-cell spread of strains from all four DEN serotypes (Fig. 2A and B). Because prior studies with N-nonyl-deoxynojirimycin (55), another ER -glucosidase inhibitor, had shown that DEN-2 appeared more sensitive to inhibition than Japanese encephalitis virus, we tested the inhibitory activity of castanospermine against two other clinically relevant flaviviruses, WNV and YFV. Interestingly, comparable or even high doses (500 μM) of castanospermine had little, if any, significant inhibitory effect on WNV infection of BHK-21 cells. In contrast, an intermediate phenotype was observed with YFV. Treatment with 50 and 500 μM castanospermine resulted in a 57 and 93% reduction, respectively, in the number of cells that expressed YFV antigen, as determined by flow cytometry (Fig. 2C). A similar pattern was observed when infectious virus was measured by plaque assay of supernatants from virus-infected cells: 50 μM castanospermine reduced WNV, YFV, and DEN production by 1.7-, 6.8-, and 910-fold, respectively (data not shown). Thus, castanospermine differentially inhibited infection of related flaviviruses even within the same cell type.

    Effect of castanospermine on viral RNA replication. As a first step towards confirming its mechanism of action, we assessed the effect of castanospermine on viral replication. To separate viral replication from infectious virus production, we studied the effect of castanospermine on the replication of subgenomic replicons. BHK-21 cell lines were generated that autonomously replicate WNV and DEN subgenomic replicons. These replicons contain marker (e.g., luciferase) and nonstructural genes but lack structural genes and thus do not form viral particles. Treatment of replicon-expressing cells with MPA, a previously described inhibitor of DEN viral replication (13), reduced replicon propagation and marker gene expression by 98% (Fig. 3A). In contrast, treatment of BHK-21 cells with castanospermine (range, 15 to 500 μM) for 48 h reduced marker gene expression or propagation of WNV or DEN replicons modestly by up to 20 to 40%, respectively (Fig. 3A). Because the reduction of luciferase activity was comparable to the slight effect of -glucosidase inhibitors on overall host protein synthesis (8), there is likely little, if any, direct effect of castanospermine on viral RNA replication and translation.

    Castanospermine inhibits infectivity of DEN VLP and viruses. Previous studies have suggested that castanospermine may inhibit infection by interfering with the correct folding of DEN-1 envelope proteins (8). One group hypothesized that castanospermine attenuated DEN infection by inhibiting a required association with the molecular chaperonin calnexin (55). Based on these hypotheses, virion morphogenesis and secretion of infectious virus may be blocked by castanospermine. To test this, we produced pseudoinfectious VLP incorporating the envelope (prM-E) and capsid (C) proteins of DEN or WNV and a WNV replicon encoding a Renilla luciferase reporter gene. Pseudoinfectious VLP were produced in the presence and absence of castanospermine and used to infect BHK-21 cells. Virus entry was measured as a function of luciferase reporter gene activity. Treatment of cells producing these particles with castanospermine had little significant effect on the infectivity of WNV particles but reduced DEN particles by greater than 95% (Fig. 3B). Because castanospermine only modestly blocked replication of the viral RNA encapsidated by each particle (Fig. 3A), its primary mechanism of inhibition can be mapped directly to effects on virus envelope and membrane protein at the stage of either particle release, attachment, or entry.

    To better define the stage of infection that was primarily inhibited by castanospermine, we evaluated its effect on the viral RNA-to-infectivity ratio in supernatants from infected cells. Supernatants were harvested from DEN- or WNV-infected BHK-21 cells, centrifuged to remove cellular debris, and analyzed by viral plaque assay. In parallel, viral RNA was extracted and measured by fluorogenic RT-PCR. As expected, at 24 or 48 h after infection, castanospermine treatment had small effects on the secretion of infectious WNV or viral particles containing WNV RNA in the supernatant (Fig. 4A and data not shown). As a result, there was no significant effect on the ratio of WNV viral RNA to PFU (medium, 1.7 x 104 ± 0.4 x 104; 100 μM castanospermine, 1.8 x 104 ± 0.4 x 104; 500 μM castanospermine, 1.4 x 104 ± 0.2 x 104; P > 0.5). In contrast, at 24 and 48 h, castanospermine treatment markedly decreased the amount of DEN viral RNA and infectious virus in the supernatant (Fig. 4B and data not shown). By 48 h, castanospermine had reduced the levels of DEN viral RNA and infectious virus by 20-fold (P 0.002) and 3,000-fold (P 0.05), respectively, and thus increased the viral RNA-to-PFU ratio by 150-fold (medium, 1.0 x 102 ± 0.8 x 102; 100 μM castanospermine, 1.7 x 104 ± 0.7 x 104; 500 μM castanospermine, 1.5 x 104 ± 0.5 x 104; P 0.05).

    Based on the change in the ratio of DEN viral RNA to PFU in the supernatant of infected cells, our data suggested that castanospermine dominantly inhibited DEN by reducing the relative infectivity of secreted DEN particles. To assess this directly, we modified an antigen-capture ELISA (S. Hanna, T. Pierson, and R. Doms, unpublished results) to measure intact DEN viral particles from supernatants of infected cells. After clarification and ultracentrifugation, DEN particles were trapped by an anti-prM MAb (2H2) and detected with a biotinylated anti-E MAb (4G2). Notably, treatment with castanospermine markedly inhibited the total number of DEN particles in supernatants from infected cells, but not to the extent that infectivity was reduced (Fig. 4B). Taken together, our data suggest that castanospermine inhibits DEN infection by reducing the numbers of secreted particles and, to a greater extent, decreasing the infectivity of the secreted DEN particles.

    Effect of castanospermine in vivo against DEN and WNV. To further evaluate the protective activity of castanospermine, we assessed its ability to prevent mortality in highly lethal DEN and WNV challenge models in mice. A/J mice infected intracranially with 105 PFU of a mouse-adapted DEN-2 strain uniformly developed hind limb paralysis and succumbed to fatal central nervous system infection within 11 days of inoculation (Fig. 5A). A/J mice that were treated with castanospermine for 10 days showed marked reduction in morbidity and mortality. A/J mice treated with 0.2 mg (10 mg/kg of body weight), 1 mg (50 mg/kg), and 5 mg (250 mg/kg) per day had survival rates of 25, 90, and 85%, respectively, whereas mice treated with vehicle had a 0% survival rate (Fig. 5A, P < 0.0001 for all three doses). Of note, higher doses of castanospermine (25 mg or 1.25 g/kg) caused adverse effects including diarrhea and weight loss (data not shown). Given its efficacy in preventing lethal DEN infection in mice, we also tested its inhibitory activity, in vivo, against WNV. Based on the in vitro studies, we predicted that castanospermine would not significantly inhibit WNV-induced mortality. Moreover, because WNV infection is more severe in immunocompromised mice (12), if castanospermine had even a mildly immunosuppressive effect, we would expect increased mortality rates. Interestingly, treatment of mice with several doses of castanospermine had no effect, adverse or beneficial, on mortality after WNV infection (Fig. 5B and data not shown).

    DISCUSSION

    Despite the significant disease burden caused by various members of the genus Flavivirus, no specific antiviral therapy is currently licensed for treatment. A prior study reported that the plant alkaloid castanospermine, an ER -glucosidase inhibitor, blocked DEN-1 infection in BHK cells (8) and that this inhibition resulted from misfolding of structural glycoproteins. In this study, we extended this analysis by demonstrating the antiviral activity of castanospermine against all four serotypes of DEN. The inhibition of carbohydrate modification caused by castanospermine directly blocked the secretion and infectivity of DEN but not WNV viral particles. As a first step towards evaluating its therapeutic potential against DEN, we tested the efficacy of castanospermine in vivo. Treatment with castanospermine prevented mortality in mice after DEN infection, yet correspondingly had no effect against WNV. The lack of any effect against WNV infection in mice suggests that castanospermine does not have immunosuppressive effects, as small deficits in the innate and adaptive immune responses cause increased morbidity and mortality in this model (12).

    Castanospermine inhibited infection and viral spread of all four serotypes of DEN whereas infection by WNV was virtually unaffected by the compound. Moreover, partial inhibition of infection or VLP production was observed with YFV. Previous studies have documented that -glucosidase inhibitors reduce infection of many but not all RNA and DNA viruses (32). One hypothesis as to the selectivity of castanospermine is that the drug inhibits viruses that require carbohydrate modification and glycoprotein oligomerization as a key step in the viral life cycle (39). Surprisingly, as the structural glycoproteins, prM and E, of all flaviviruses undergo similar oligomerization, and yet clearly WNV was not susceptible to inhibition, this explanation may not apply to flaviviruses. In studies by other groups, the cell type has been shown to affect the susceptibility of vesicular stomatitis virus to castanospermine. This variation has been attributed to a Golgi apparatus-resident endomannosidase that circumvents castanospermine inhibition by cleaving an -1,3-mannose bond to release oligosaccharides (25). Interestingly, the BHK-21 and human hepatoma cell lines used in our study, which showed castanospermine-mediated inhibition of DEN, express high levels of this Golgi apparatus-resident endomannosidase (25).

    The variation in susceptibility to castanospermine among flaviviruses also could be related to a differential requirement for association with the chaperones calnexin and calreticulin. Previous studies have suggested that -glucosidase inhibitors interrupt the interaction between viral structural proteins and calnexin (18, 55). At least one of the glycans on the prM or E of DEN, but not other flaviviruses, may be essential for calnexin-mediated protein folding, oligomerization, and virion assembly (55). However, the ability to generalize from these results remains uncertain, as experiments by another group suggest that DEN structural proteins do not associate with chaperones in infected cells (8).

    The stringency of ER retention sequences in flavivirus envelope protein transmembrane domains could also explain the differential susceptibility to castanospermine. Flaviviruses that are more strongly retained in the ER would be more susceptible because they do not appreciably traffic through the Golgi network and encounter the resident "escape" mannosidase that allows carbohydrate processing and proper viral protein folding. Alternatively, a difference in transit time through the Golgi network among flaviviruses could affect exposure to "escape" mannosidases. However, at present, there are no data to support the idea that different flaviviruses have altered mechanisms for maturation and transit from the ER through the Golgi network, to the cell surface. Finally, the N-linked glycosylation of prM and E could be less critical for WNV infectivity than for that of DEN. Recent studies would suggest otherwise, as the loss of the N-linked glycosylation site on the WNV E protein was also associated with decreased viral infection in vitro (43) and in vivo (2, 45). Mutagenesis and chimerization studies among flavivirus envelope proteins will be necessary directly to test the validity of these hypotheses.

    Our complementary studies with DEN and WNV subgenomic replicons and VLP demonstrate that castanospermine has little specific effect on viral RNA replication and/or translation but a dominant effect on viral particle secretion and/or infectivity. Although recent studies have suggested that the glycosylation state of NS1, a cofactor for flavivirus RNA-dependent replication (26, 29, 30), can modulate virus production (9, 35), the carbohydrate modifications that are inhibited by castanospermine do not have a major impact on viral replication. Consistent with this, our experiments with infectious virus demonstrate that castanospermine inhibits DEN infection both at the level of secretion and at the level of infectivity of viral particles. Significant quantities of viral RNA were detected in supernatants of castanospermine-treated cells despite relatively few infectious particles, leading to a 150-fold increase in the ratio of viral RNA to infectious virus. Moreover, our DEN particle capture ELISA experiments suggest that castanospermine reduced the infectivity of the secreted particles to a greater extent than secretion. Future studies will be required to determine exactly how castanospermine decreases the infectivity of DEN viral particles, i.e., whether the changes in carbohydrate modification affect virus attachment, entry, or uncoating.

    In nature, DEN infection occurs between mosquitoes and humans. Although mice can be experimentally infected and become ill after DEN infection (17, 22, 23, 46, 47), they do not develop a pathological syndrome that resembles DHF. Nonetheless, despite being imperfect for the study of pathogenesis of DEN infection, the mouse model provides important information on the toxicity of an antiviral drug as well as its ability to inhibit viral replication and prevent viral dissemination in vivo. As an initial test of its antiviral potential in humans, castanospermine improved survival rates in mice even when DEN was inoculated intracranially. Castanospermine was well tolerated and prevented DEN-associated mortality over a broad range of doses from 10 to 250 mg/kg/day. Doses in excess of 250 mg/kg/day in mice were associated with gastrointestinal toxicity, results that are consistent with previous observations in rats (42). The treatment success with castanospermine was in spite of the relatively unfavorable pharmacokinetics, as the rate of its renal clearance in mice is up to 20 times higher than in humans (24).

    In summary, the ER -glucosidase inhibitor castanospermine is a promising antiviral agent against DEN infection, with clear evidence of in vitro and in vivo activity. Castanospermine reduced secretion and viral infectivity and improved survival rates in mice. As it is derived quite readily from trees that can be grown in tropical climates, it is intriguing to consider castanospermine as a possible affordable therapeutic agent in the regions of the globe that have the greatest DEN morbidity and yet are pressed for health care resources. Based on the preclinical data presented here, we believe that further development of castanospermine is warranted as a possible treatment for DEN infection.

    ACKNOWLEDGMENTS

    We thank A. Pekosz, K. Blight, D. Leib, L. Morrison, R. Klein, and P. Olivo and their laboratories for experimental advice. We also thank R. Smith (Phytex Australia Pty Ltd.) for the generous contribution of castanospermine, R. Kinney (Fort Collins, CO) for the WNV and DEN infectious clones, E. Harris (Berkeley, CA) for the DEN strains, and R. Levis (Bethesda, MD) for the polyclonal anti-prM antibody.

    The work was supported by the Edward Mallinckrodt, Jr., Foundation (M.S.D.); by a New Scholar Award in Global Infectious Diseases from the Ellison Foundation (M.S.D.); and by the NIH (U01 AI 538870 to M.S.D. and U54 AI57168 to R.W.D.).

    REFERENCES

    Allison, S. L., Y. J. Tao, G. O'Riordain, C. W. Mandl, S. C. Harrison, and F. X. Heinz. 2003. Two distinct size classes of immature and mature subviral particles from tick-borne encephalitis virus. J. Virol. 77:11357-11366.

    Beasley, D. W., C. T. Davis, J. Estrada-Franco, R. Navarro-Lopez, A. Campomanes-Cortes, R. B. Tesh, S. C. Weaver, and A. D. Barrett. 2004. Genome sequence and attenuating mutations in West Nile virus isolate from Mexico. Emerg. Infect. Dis. 10:2221-2224.

    Behrens, S. E., C. W. Grassmann, H. J. Thiel, G. Meyers, and N. Tautz. 1998. Characterization of an autonomous subgenomic pestivirus RNA replicon. J. Virol. 72:2364-2372.

    Blight, K. J., A. A. Kolykhalov, and C. M. Rice. 2000. Efficient initiation of HCV RNA replication in cell culture. Science 290:1972-1974.

    Bridges, C. G., S. P. Ahmed, M. S. Kang, R. J. Nash, E. A. Porter, and A. S. Tyms. 1995. The effect of oral treatment with 6-O-butanoyl castanospermine (MDL 28,574) in the murine zosteriform model of HSV-1 infection. Glycobiology 5:249-253.

    Bridges, C. G., T. M. Brennan, D. L. Taylor, M. McPherson, and A. S. Tyms. 1994. The prevention of cell adhesion and the cell-to-cell spread of HIV-1 in vitro by the alpha-glucosidase 1 inhibitor, 6-O-butanoyl castanospermine (MDL 28574). Antivir. Res. 25:169-175.

    Chambers, T. J., C. S. Hahn, R. Galler, and C. M. Rice. 1990. Flavivirus genome organization, expression, and replication. Annu. Rev. Microbiol. 44:649-688.

    Courageot, M. P., M. P. Frenkiel, C. D. Dos Santos, V. Deubel, and P. Despres. 2000. Alpha-glucosidase inhibitors reduce dengue virus production by affecting the initial steps of virion morphogenesis in the endoplasmic reticulum. J. Virol. 74:564-572.

    Crabtree, M. B., R. M. Kinney, and B. R. Miller. 2005. Deglycosylation of the NS1 protein of dengue 2 virus, strain 16681: construction and characterization of mutant viruses. Arch. Virol. 150:771-786.

    Diamond, M., T. Roberts, D. Edgil, B. Lu, J. Ernst, and E. Harris. 2000. Modulation of dengue virus infection in human cells by alpha, beta, and gamma interferons. J. Virol. 74:4957-4966.

    Diamond, M. S., D. Edgil, T. G. Roberts, B. Lu, and E. Harris. 2000. Infection of human cells by dengue virus is modulated by different cell types and viral strains. J. Virol. 74:7814-7823.

    Diamond, M. S., B. Shrestha, E. Mehlhop, E. Sitati, and M. Engle. 2003. Innate and adaptive immune responses determine protection against disseminated infection by West Nile encephalitis virus. Viral Immunol. 16:259-278.

    Diamond, M. S., M. Zachariah, and E. Harris. 2002. Mycophenolic acid inhibits dengue virus infection by preventing replication of viral RNA. Virology 304:211-221.

    Ebel, G. D., A. P. Dupuis III, K. Ngo, D. Nicholas, E. Kauffman, S. A. Jones, D. Young, J. Maffei, P. Y. Shi, K. Bernard, and L. Kramer. 2001. Partial genetic characterization of West Nile virus strains, New York State, 2000. Emerg. Infect. Dis. 7:650-653.

    Elbein, A. D., G. Legler, A. Tlusty, W. McDowell, and R. Schwarz. 1984. The effect of deoxymannojirimycin on the processing of the influenza viral glycoproteins. Arch. Biochem. Biophys. 235:579-588.

    Elshuber, S., S. L. Allison, F. X. Heinz, and C. W. Mandl. 2003. Cleavage of protein prM is necessary for infection of BHK-21 cells by tick-borne encephalitis virus. J. Gen. Virol. 84:183-191.

    Falgout, B., M. Bray, J. J. Schlesinger, and C. J. Lai. 1990. Immunization of mice with recombinant vaccinia virus expressing authentic dengue virus nonstructural protein NS1 protects against lethal dengue virus encephalitis. J. Virol. 64:4356-4363.

    Hebert, D. N., B. Foellmer, and A. Helenius. 1996. Calnexin and calreticulin promote folding, delay oligomerization and suppress degradation of influenza hemagglutinin in microsomes. EMBO. J. 15:2961-2968.

    Hebert, D. N., B. Foellmer, and A. Helenius. 1995. Glucose trimming and reglucosylation determine glycoprotein association with calnexin in the endoplasmic reticulum. Cell 81:425-433.

    Henchal, E. A., M. K. Gentry, J. M. McCown, and W. E. Brandt. 1982. Dengue virus-specific and flavivirus group determinants identified with monoclonal antibodies by indirect immunofluorescence. Am. J. Trop. Med. Hyg. 31:830-836.

    Houng, H. S., R. Chung-Ming Chen, D. W. Vaughn, and N. Kanesa-thasan. 2001. Development of a fluorogenic RT-PCR system for quantitative identification of dengue virus serotypes 1-4 using conserved and serotype-specific 3' noncoding sequences. J. Virol. Methods 95:19-32.

    Huang, K. J., S. Y. Li, S. C. Chen, H. S. Liu, Y. S. Lin, T. M. Yeh, C. C. Liu, and H. Y. Lei. 2000. Manifestation of thrombocytopenia in dengue-2-virus-infected mice. J. Gen. Virol. 81:2177-2182.

    Johnson, A. J., and J. T. Roehrig. 1999. New mouse model for dengue virus vaccine testing. J. Virol. 73:783-786.

    Kang, M. S. 1996. Uptake and metabolism of BuCast: a glycoprotein processing inhibitor and a potential anti-HIV drug. Glycobiology 6:209-216.

    Karaivanova, V. K., P. Luan, and R. G. Spiro. 1998. Processing of viral envelope glycoprotein by the endomannosidase pathway: evaluation of host cell specificity. Glycobiology 8:725-730.

    Khromykh, A. A., P. L. Sedlak, K. J. Guyatt, R. A. Hall, and E. G. Westaway. 1999. Efficient trans-complementation of the flavivirus Kunjin NS5 protein but not of the NS1 protein requires its coexpression with other components of the viral replicase. J. Virol. 73:10272-10280.

    Kinney, R. M., S. Butrapet, G. J. Chang, K. R. Tsuchiya, J. T. Roehrig, N. Bhamarapravati, and D. J. Gubler. 1997. Construction of infectious cDNA clones for dengue 2 virus: strain 16681 and its attenuated vaccine derivative, strain PDK-53. Virology 230:300-308.

    Lanciotti, R. S., A. J. Kerst, R. S. Nasci, M. S. Godsey, C. J. Mitchell, H. M. Savage, N. Komar, N. A. Panella, B. C. Allen, K. E. Volpe, B. S. Davis, and J. T. Roehrig. 2000. Rapid detection of West Nile virus from human clinical specimens, field-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. J. Clin. Microbiol. 38:4066-40671.

    Lindenbach, B. D., and C. M. Rice. 1999. Genetic interaction of flavivirus nonstructural proteins NS1 and NS4A as a determinant of replicase function. J. Virol. 73:4611-4621.

    Lindenbach, B. D., and C. M. Rice. 1997. trans-complementation of yellow fever virus NS1 reveals a role in early RNA replication. J. Virol. 71:9608-9617.

    Lo, M. K., M. Tilgner, K. A. Bernard, and P. Y. Shi. 2003. Functional analysis of mosquito-borne flavivirus conserved sequence elements within 3' untranslated region of West Nile virus by use of a reporting replicon that differentiates between viral translation and RNA replication. J. Virol. 77:10004-10014.

    McGinnes, L. W., and T. G. Morrison. 1998. Role of carbohydrate processing and calnexin binding in the folding and activity of the HN protein of Newcastle disease virus. Virus Res. 53:175-185.

    Molinari, M., and A. Helenius. 1999. Glycoproteins form mixed disulphides with oxidoreductases during folding in living cells. Nature 402:90-93.

    Monath, T. P. 1994. Dengue: the risk to developed and developing countries. Proc. Natl. Acad. Sci. USA 91:2395-2400.

    Muylaert, I. R., T. J. Chambers, R. Galler, and C. M. Rice. 1996. Mutagenesis of the N-linked glycosylation sites of the yellow fever virus NS1 protein: effects on virus replication and mouse neurovirulence. Virology 222:159-168.

    Ouzounov, S., A. Mehta, R. A. Dwek, T. M. Block, and R. Jordan. 2002. The combination of interferon alpha-2b and n-butyl deoxynojirimycin has a greater than additive antiviral effect upon production of infectious bovine viral diarrhea virus (BVDV) in vitro: implications for hepatitis C virus (HCV) therapy. Antivir. Res. 55:425-435.

    Pan, Y. T., J. Ghidoni, and A. D. Elbein. 1993. The effects of castanospermine and swainsonine on the activity and synthesis of intestinal sucrase. Arch. Biochem. Biophys. 303:134-144.

    Pan, Y. T., H. Hori, R. Saul, B. A. Sanford, R. J. Molyneux, and A. D. Elbein. 1983. Castanospermine inhibits the processing of the oligosaccharide portion of the influenza viral hemagglutinin. Biochemistry 22:3975-3984.

    Rudd, P. M., and R. A. Dwek. 1997. Glycosylation: heterogeneity and the 3D structure of proteins. Crit. Rev. Biochem. Mol. Biol. 32:1-100.

    Ruprecht, R. M., S. Mullaney, J. Andersen, and R. Bronson. 1989. In vivo analysis of castanospermine, a candidate antiretroviral agent. J. Acquir. Immune Defic. Syndr. 2:149-157.

    Russell, P. K., and A. Nisalak. 1967. Dengue virus neutralization by the plaque reduction neutralization test. J. Immunol. 99:291-294.

    Saul, R., J. J. Ghidoni, R. J. Molyneux, and A. D. Elbein. 1985. Castanospermine inhibits alpha-glucosidase activities and alters glycogen distribution in animals. Proc. Natl. Acad. Sci. USA 82:93-97.

    Scherret, J. H., J. S. Mackenzie, A. A. Khromykh, and R. A. Hall. 2001. Biological significance of glycosylation of the envelope protein of Kunjin virus. Ann. N. Y. Acad. Sci. 951:361-363.

    Schlesinger, S., A. H. Koyama, C. Malfer, S. L. Gee, and M. J. Schlesinger. 1985. The effects of inhibitors of glucosidase I on the formation of Sindbis virus. Virus Res. 2:139-149.

    Shirato, K., H. Miyoshi, A. Goto, Y. Ako, T. Ueki, H. Kariwa, and I. Takashima. 2004. Viral envelope protein glycosylation is a molecular determinant of the neuroinvasiveness of the New York strain of West Nile virus. J. Gen. Virol. 85:3637-3645.

    Shresta, S., J. L. Kyle, P. R. Beatty, and E. Harris. 2004. Early activation of natural killer and B cells in response to primary dengue virus infection in A/J. mice. Virology 319:262-273.

    Shresta, S., J. L. Kyle, H. M. Snider, M. Basavapatna, P. R. Beatty, and E. Harris. 2004. Interferon-dependent immunity is essential for resistance to primary dengue virus infection in mice, whereas T- and B-cell-dependent immunity are less critical. J. Virol. 78:2701-2710.

    Stadler, K., S. L. Allison, J. Schalich, and F. X. Heinz. 1997. Proteolytic activation of tick-borne encephalitis virus by furin. J. Virol. 71:8475-8481.

    Sunkara, P. S., M. S. Kang, T. L. Bowlin, P. S. Liu, A. S. Tyms, and A. Sjoerdsma. 1990. Inhibition of glycoprotein processing and HIV replication by castanospermine analogues. Ann. N. Y. Acad. Sci. 616:90-96.

    Sunkara, P. S., D. L. Taylor, M. S. Kang, T. L. Bowlin, P. S. Liu, A. S. Tyms, and A. Sjoerdsma. 1989. Anti-HIV activity of castanospermine analogues. Lancet i:1206.

    Taylor, D. L., M. S. Kang, T. M. Brennan, C. G. Bridges, P. S. Sunkara, and A. S. Tyms. 1994. Inhibition of alpha-glucosidase I of the glycoprotein-processing enzymes by 6-O-butanoyl castanospermine (MDL 28,574) and its consequences in human immunodeficiency virus-infected T cells. Antimicrob. Agents Chemother. 38:1780-1787.

    Taylor, D. L., P. S. Sunkara, P. S. Liu, M. S. Kang, T. L. Bowlin, and A. S. Tyms. 1991. 6-O-Butanoylcastanospermine (MDL 28,574) inhibits glycoprotein processing and the growth of HIV. AIDS 5:693-698.

    Vara, J., F. Malpartida, D. A. Hopwood, and A. Jimenez. 1985. Cloning and expression of a puromycin N-acetyl transferase gene from Streptomyces alboniger in Streptomyces lividans and Escherichia coli. Gene 33:197-206.

    Whitby, K., D. Taylor, D. Patel, P. Ahmed, and A. S. Tyms. 2004. Action of celgosivir (6-O-butanoyl castanospermine) against the pestivirus BVDV: implications for the treatment of hepatitis C. Antivir. Chem. Chemother. 15:141-151.

    Wu, S. F., C. J. Lee, C. L. Liao, R. A. Dwek, N. Zitzmann, and Y. L. Lin. 2002. Antiviral effects of an iminosugar derivative on flavivirus infections. J. Virol. 76:3596-3604.

    Yamashita, Y., K. Shimokata, S. Mizuno, T. Daikoku, T. Tsurumi, and Y. Nishiyama. 1996. Calnexin acts as a molecular chaperone during the folding of glycoprotein B of human cytomegalovirus. J. Virol. 70:2237-2246.

    Zhang, J. X., I. Braakman, K. E. Matlack, and A. Helenius. 1997. Quality control in the secretory pathway: the role of calreticulin, calnexin and BiP in the retention of glycoproteins with C-terminal truncations. Mol. Biol. Cell 8:1943-1954.

    Zhang, Y., J. Corver, P. R. Chipman, W. Zhang, S. V. Pletnev, D. Sedlak, T. S. Baker, J. H. Strauss, R. J. Kuhn, and M. G. Rossmann. 2003. Structures of immature flavivirus particles. EMBO J. 22:2604-2613.(Kevin Whitby, Theodore C.)