ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, June 2004, p. 2085–2090 0066-4804/04/$08.00⫹0 DOI: 10.1128/AAC.48.6.2085–2090.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Vol. 48, No. 6
␣-Galactosylceramide and Novel Synthetic Glycolipids Directly Induce the Innate Host Defense Pathway and Have Direct Activity against Hepatitis B and C Viruses Anand S. Mehta,1* Baohua Gu,1 Bertha Conyers,1 Serguey Ouzounov,1 Lijuan Wang,1 Robert M. Moriarty,2 Raymond A. Dwek,3 and Timothy M. Block1 Department of Biochemistry and Molecular Pharmacology, Thomas Jefferson University, The Jefferson Center, Doylestown, Pennsylvania 189011; Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois 606072; and Department of Biochemistry, The Glycobiology Institute, University of Oxford, Oxford, United Kingdom OX1 3QU3 Received 17 September 2003/Returned for modification 16 December 2003/Accepted 20 February 2004
␣-Galactosylceramide is a glycolipid derived from marine sponges that is currently in human clinical trials as an anticancer agent. It has also been shown to be effective in reducing the amount of hepatitis B virus (HBV) DNA detected in mice that produce HBV constitutively from a transgene. It was assumed that all of the antiviral and antitumor activities associated with ␣-galactosylceramide were mediated through the activation of NK T cells. However, we report here an additional unpredicted activity of ␣-galactosylceramide as a direct antiviral agent and inducer of the innate host defense pathway. To exploit this activity, we have developed a new class of smaller, orally available glycolipids that also induce the innate host defense pathway and have direct activity against HBV and hepatitis C virus. effect of alpha interferon) (i) that although it is antiviral, NN-DGJ does not inhibit the viral polymerase, (ii) that NN-DGJ does not inhibit HBV surface protein (HBsAg) production or other viral gene product production, and (iii) that total viral RNA levels were not influenced by these compounds (17). Recent work has shown that the mechanism of action of these compounds is similar to that of alpha interferon (8) in that they either inhibit the formation of pregenomic RNA-containing capsids or accelerate their degradation (15). Further work characterizing the structural requirements for antiviral activity highlighted a need for both a sugar head group and an unbroken alkyl tail of at least eight carbons (16). Structure activity relationship analysis led to the development of an analogue of NN-DGJ, N-9-oxadecyl-6-methylDGJ, which was as potent and efficacious as the parent molecule but with a much improved toxicity profile. Similar to the structural requirements for ␣-galactosylceramide, structural analysis of the synthetic glycolipids also highlighted the importance of a galactose head group in the antiviral activity of these compounds (16). Synthetic glycolipids with alternative head groups, such as mannose or fucose, have reduced potency compared to those with the galactose head group (16). In the case of ␣-galactosylceramide, it was assumed that all of the antiviral and antitumor activities were mediated through the activation of NK T cells (11, 12). However, we report here an additional unpredicted activity of ␣-galactosylceramide as a direct antiviral agent and inducer of the innate host defense pathway. To exploit this activity, we have developed a new class of smaller, orally available glycolipids (Fig. 1B and C) that also induce the innate host defense pathway and have direct activity against HBV and hepatitis C virus (HCV).
Hepatitis B virus (HBV) is the prototypic member of the Hepadnaviridae family of viruses that chronically infect ⬎350 million people worldwide (1). The major complication is the development of primary hepatocellular carcinoma estimated to cause ⬎500,000 deaths annually (1, 2). Although there is no cure for HBV infection, several therapeutic options now exist (9). However, the poor response rate and development of resistant mutants highlight the need for alternatives and complements to the conventional therapeutic regimens (13). ␣-Galactosylceramide, shown in Fig. 1A, is a glycolipid derived from marine sponges that is currently in human clinical trials as an anticancer agent (4). It has also been shown to be effective in reducing the amount of HBV DNA detected in mice that produce HBV constitutively from a transgene (11). These long-chain alkylated sugars bind CD1 molecules on the plasma membranes of diverse cell types and are presented to subsets of CD4⫹ CD8⫺ or CD4⫺ CD8⫺ T cells that express markers associated with NK cells and are referred to as NK T cells (18). NK T cells, when activated, secrete cytokines that have antiviral and antitumor properties and are thought to mediate important components of the non-major histocompatibility complex-dependent immune system. Previous work showed that the synthetic glycolipid N-nonyldeoxy-galactonojirimycin (NN-DGJ) could inhibit the production and secretion of HBV in vitro (16, 17). The exact molecular mechanism whereby NN-DGJ (and the chemical family it represents) induced its antiviral effect is not fully known but is analogous to the effect of inducers of the innate host defense pathway against HBV (8, 24). That is, it has been shown (similar to the
* Corresponding author. Mailing address: Department of Biochemistry, The Jefferson Center, Jefferson Medical College, 700 E. Butler Ave., Doylestown, PA 18901. Phone: (215) 489-4905. Fax: (215) 4894920. E-mail:
[email protected].
MATERIALS AND METHODS Cells and compounds. HepG2 2.2.15 cells were kindly provided by George Acs (Mt. Sinai Medical College, New York, N.Y.) and maintained in the same way as
2085
2086
MEHTA ET AL.
FIG. 1. Structures of ␣-galactosylceramide, N-9-oxadecyl-6-methylDGJ, and N-7-oxanonyl-6-methyl-DGJ. (A) ␣-Galactosylceramide is a naturally occurring glycolipid purified from marine sponges (11). (B and C) The imino sugars used in this study are composed of an imino sugar head group and an alkyl chain. The head groups found in panels B and C are imino sugar derivatives of galactose (DGJ). This is a galactose analogue in which the ring oxygen has been replaced with a nitrogen atom and the anomeric hydroxyl group of galactose is absent. N-7-oxanonyl-6-methyl-DGJ has been shown to have no activity against HBV (16). Prior structure-activity relationship analysis has indicated that an unbroken alkyl chain length of ⬎8 is required for antiviral activity (16).
HepG2 cells. The HCV subgenomic replicon cell line 9-13, a kind gift of R. Bartenschlager (14), was cultured in Dulbecco’s modified Eagle medium (Invitrogen Corp., Carlsbad, Calif.) containing 10% fetal calf serum, 1% penicillinstreptomycin, 1% nonessential amino acids, and 0.5 mg of Geneticin/ml. The cells were maintained at subconfluency prior to splitting. ␣-Galactosylceramide was the kind gift of Kirin Brewery Co., Ltd. (Tokyo, Japan). All synthetic glycolipids presented in this study were provided by Synergy Pharmaceuticals (Edison, N.J.) and United Therapeutics (Silver Spring, Md.). All compounds were dissolved in sterile double-distilled water unless otherwise indicated. Analysis of secreted viral DNA. Analysis of DNA secreted from tissue culture cells was performed by a method which would discriminate between enveloped and nonenveloped virions (23). Briefly, HepG2 2.2.15 cells were seeded at 85 to 90% confluence in T-25 flasks, and 5 days later, the appropriate drug was added at the appropriate concentrations. After 3 days, virus was concentrated from the supernatant with polyethylene glycol as described elsewhere (23). The virus was resuspended in 200 l of 10 mM TRIS (pH 7.9), 1 mM EDTA (pH 8.0), and 10 mM MgCl2. Proteinase K was added to a final concentration of 750 g/ml, and the samples were incubated for 1 h at 37°C. After 1 h, SQ1 DNase (Promega, Madison, Wis.) was added to each tube to a final concentration of 50 U/ml and incubated at 37°C for 1 h. Sodium dodecyl sulfate (SDS) was added to a final concentration of 1%, more proteinase K was added to a final concentration of 500 g/ml, and the reaction was allowed to proceed at 37°C for 4 h. DNA was purified by phenol-chloroform extraction, followed by isopropanol precipitation. The DNA was separated by electrophoresis on a 1.0% agarose gel, transferred to
ANTIMICROB. AGENTS CHEMOTHER. a nylon membrane, and probed with 32P-labeled HBV probes as described elsewhere. HBV-specific bands were subsequently identified and quantified by phosphorimager analysis (Bio-Rad, Hercules, Calif.). Toxicity analysis. Toxicity was determined by mitochondrial toxicity testing of cells that had been exposed to compounds in a manner similar to the antiviral assays except that the cells were treated in 24-well trays. Briefly, HepG2 2.2.15 cells were treated with drug as indicated in the figure legends for 3 days, and the media were removed and replaced with 100 l of a 10-mg/ml solution of tetrazolium bromide (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; Sigma Chemicals) for 1 h at 37°C. The addition of 100 to 200 l of dimethyl sulfoxide led to color development. The supernatants were removed and placed in a 96-well tray for analysis. Optical density values were read at 590 nm. Analysis of 2ⴕ,5ⴕ OAS genes and beta interferon secretion. Briefly, HepG2 cells were incubated with either ␣-galactosylceramide or the synthetic glycolipids shown in Fig. 1 for the desired length of time, and the total RNA was harvested using Tri-reagent (Gibco-BRL, Rockville, Md.) according to the manufacturer’s directions. RNA samples were further purified using the Ambion (Austin, Tex.) DNA-free kit before reverse transcriptase (RT) PCR with PCR conditions and primers exactly as reported in the literature (10). PCR was performed in the absence of RT for 50 cycles to ensure no DNA contamination. Dilution experiments were used to ensure that the PCR was within the linear range of the assay. Southern blotting–RT-PCR of RNA from HepG2 cells treated with the appropriate concentration of N-9-oxadecyl-6-methyl-DGJ or with alpha interferon (2a/2b) for 16 h was also performed to allow the quantification of induction. Briefly, limited PCR was performed as described above for 5, 10, or 15 cycles, and the PCR products were transferred to nylon membranes. Hybridization was carried out using a 1,377-bp cDNA probe from nucleotides 1 to 1377 of the published OAS-40/46 gene (accession no. X02874). OAS-40/46-specific bands were identified and quantified by phosphorimager analysis (Bio-Rad). For analysis of interferon secretion, cells were treated as described above, and the amount of beta interferon secreted into the culture medium was determined using a commercially available human beta interferon enzyme-linked immunosorbent assay kit (PBL Biomedical Laboratories, Piscataway, N.J.) according to the manufacturer’s directions. HCV replicon system and detection of HCV RNA. 9-13 cells were seeded in T25 flasks at 106 cells/flask. After allowing for adhesion of the cells, we added the appropriate concentration of interferon or N-9-oxadecyl-6-methyl-DGJ, the cells were incubated for 48 h, and the RNA was isolated using the RNAeasy kit (Qiagen, Valencia Calif.). Northern blot analysis was done to analyze the HCV replicon RNA level. Briefly, 2 g of total RNA was electrophoresed through a 1.0% agarose gel containing 2.2 M formaldehyde, transferred to a nylon membrane, and immobilized by UV cross-linking (Stratagene). Hybridization was carried out using an [␣-32P]CTP-labeled probe with random primers on a 2-kb NS5B DNA fragment in a quick-hybridization solution (Amersham Bioscience, Piscataway, N.J.) for 16 h at 65°C. The membranes were washed once in 2⫻ SSC (1⫻ SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS for 30 min at room temperature and twice in 0.1⫻ SSC–0.1% SDS for 30 min at 65°C. The radioactive signal was identified and quantified by phosphorimager analysis. Western blot analysis:. Western blot analysis was done as described elsewhere (5). A monoclonal antibody to NS5A, a kind gift of C. Liu (University of Florida, Gainesville), was used to measure the viral-protein level.
RESULTS ␣-Galactosylceramide (Fig. 1) has been reported to have antiviral activity in a transgenic animal model of HBV (11). This antiviral activity is thought to involve the activation of intrahepatic NK T cells, which inhibit HBV through the secretion of gamma interferon. However, the direct activity of this compound against HBV was not tested and remained a possibility. Thus, the direct antiviral activities of ␣-galactosylceramide and our smaller orally available glycolipids were tested in tissue culture using the stable HBV-producing cell line HepG2 2.2.15 (16, 17, 22, 23). The results, shown in Fig. 2A and B, clearly demonstrate that ␣-galactosylceramide and the smaller, orally available glycolipid N-9-oxadecyl-6-methyl-DGJ effectively reduce the amount of HBV detected in the culture medium in a dose-dependent manner. As Fig. 2A shows, the addition of ␣-galactosylceramide at concentrations of 1 to 100
VOL. 48, 2004
␣-GALACTOSYLCERAMIDE ACTIVITY AGAINST HBV AND HCV
2087
FIG. 2. ␣-Galactosylceramide and N-9-oxadecyl-6-methyl-DGJ inhibit the secretion of HBV in vitro. (A) ␣-Galactosylceramide inhibits HBV in vitro. Briefly, HepG2 2.2.15 cells, which secrete HBV, were either left untreated (Un) or treated with 103 IU of alpha interferon (IFN) (2a/2b)/ml or 0.1 to 100 nM ␣-galactosylceramide for 3 days, and the amount of HBV-specific DNA in the culture medium was determined using a method which discriminates between enveloped and nonenveloped viral particles. The HBV relaxed circular (RC) DNA, which decreases with both interferon and ␣-galactosylceramide treatment, is indicated. (B) N-9-oxadecyl-6-methyl-DGJ inhibits the secretion of HBV in vitro. HepG2 2.2.15 cells were left untreated or treated with various doses of the synthetic glycolipid N-9-oxadecyl-6-methyl-DGJ (0.8 to 25 M) for 3 days, and the level of HBV secreted into the culture medium was determined as before. 3TC lamivudine, a nucleoside analogue that inhibits the secretion of HBV, was used as a control. The HBV RC DNA, which decreases with N-9-oxadecyl-6-methyl-DGJ, interferon, and ␣-galactosylceramide treatment, is indicated. DNJ, deoxynojirimycin. (C and D) Cells treated with ␣-galactosylceramide or the synthetic glycolipid N-9-oxadecyl-6-methyl-DGJ were tested for viability using the standard mitochondrial toxicity test. (C) x axis (from left to right), untreated cells and ␣-galactosylceramide at 10, 5, 2.5, 1, 0.5, and 0.1 M; y axis, percent viability compared to untreated group. (D) x axis (from left to right), untreated cells and N-9-oxadecyl-6-methyl-DGJ at 2,000, 800, 400, 200, 100, and 10 m; y axis, percent viability compared to untreated group. The error bars indicate standard deviations.
nM to tissue culture cells inhibited the secretion of enveloped HBV with a 50% inhibitory concentration (IC50) of 0.4 ⫾ 1.1 nM. In this assay, alpha interferon is used as a control and also inhibits secretion effectively at a concentration of 103 IU/ml, consistent with other reports (8). It was noted that the potency of ␣-galactosylceramide was dependent upon the formulation of its delivery; dissolution in the lipophilic solvent provided by the supplier (intended to promote intracellular delivery) actually reduced potency (data not shown), suggesting that a surface receptor is involved. Similar to ␣-galactosylceramide, and consistent with previous reports regarding this compound class (16, 17), the synthetic glycolipid N-9-oxadecyl-6-methyl-DGJ also exerted antiviral activity at various concentrations (Fig. 2B) with an IC50 of 1 ⫾ 3.6 M. The cytotoxicity profiles of ␣-galactosylceramide and the synthetic glycolipid N-9-oxadecyl-6-methyl-DGJ were examined in parallel (and under the same conditions), as were the
antiviral profiles, and the results are shown in Fig. 2C and D. In these experiments, the cytotoxic concentration required to kill 50% of the cells with ␣-galactosylceramide is 8 M, and that for N-9-oxadecyl-6-methyl-DGJ is ⬎2,000 M. Since the IC50s for ␣-galactosylceramide and N-9-oxadecyl-6-methylDGJ were 0.4 ⫾ 1.1 nM and 1 ⫾ 3.6 M, respectively, it is safe to say that the antiviral activities of these compounds occur at concentrations well below that at which toxicity was observed, thus demonstrating selectivity for viral functions. As stated, neither ␣-galactosylceramide or N-9-oxadecyl-6methyl-DGJ had any detectable effect upon HBsAg production or secretion, core antigen production or secretion, or HBV polymerase activity at concentrations that were highly antiviral (11, 17). This indicates that the compounds were well tolerated at these concentrations and that the antiviral activity cannot be explained by a direct effect upon the synthesis of viral products
2088
MEHTA ET AL.
ANTIMICROB. AGENTS CHEMOTHER.
FIG. 3. ␣-Galactosylceramide and N-9-oxadecyl-6-methyl-DGJ induce components of the host defense pathway in tissue culture. (A) HepG2 cells were treated with the indicated doses of ␣-galactosylceramide for 16 h, and the induction of the medium (p69) and small (p40) isoforms of the 2⬘,5⬘ OAS gene were determined by RT-PCR. For each set, actin levels were also monitored as a control for loading. The negative control (⫺) in both blots was the imino sugar N-7-oxanonyl-6-methyl-DGJ, which has no antiviral activity (16) and, as shown, does not induce components of the host defense pathway. Un, untreated. IFN, interferon. (B) Southern blotting–RT-PCR of RNA from HepG2 cells treated with the indicated concentration of N-9-oxadecyl-6-methyl-DGJ or with 103 IU of alpha interferon (2a/2b)/ml for 16 h. Briefly, limited PCR was performed as described above for 5, 10, or 15 cycles, and the PCR products were transferred to nylon membranes before hybridization with a 1,377-bp cDNA probe from nucleotides 1 to 1377 of the published OAS-40/46 gene (accession no. X02874). Actin controls were performed in the same manner. The results of the 10-cycle PCR and hybridization are shown. (C) Quantification of the blot shown in panel B, indicating that N-9-oxadecyl-6methyl-DGJ induced a 20-fold induction of the 2⬘,5⬘ OAS gene at 70 M, with lower concentrations giving a dose-dependent 2- to 15-fold increase in 2⬘,5⬘ OAS gene expression.
but rather through the activation of a cellular defense mechanism. Current experimental and therapeutic antivirals against HBV target either a specific viral protein, such as the viral polymerase, or activate components of the innate host defense pathway. As both ␣-galactosylceramide and N-9-oxadecyl-6methyl-DGJ do not have a detectable impact upon HBV-specific proteins, it appeared possible that they induced components of the innate host defense pathway. Thus, the abilities of ␣-galactosylceramide and N-9-oxadecyl-6-methyl-DGJ to directly induce the innate host defense pathway in tissue culture were determined by analysis of the induction of the 2⬘,5⬘ oligoadenylate synthetase (2⬘,5⬘ OAS) genes using an RT-PCRbased methodology (10, 20). As Fig. 3A shows, alpha interferon is a potent inducer of both the medium (p69) and small (p40) 2⬘,5⬘ OAS genes. In contrast, ␣-galactosylceramide induced only the small 2⬘,5⬘ OAS gene expression over a wide dose range. Consistent with this result, the synthetic glycolipid N-9-oxadecyl-6-methyl-DGJ also induced only the small 2⬘,5⬘ OAS gene. Figure 3B shows the induction of the small 2⬘,5⬘ OAS gene utilizing a southern blotting–RT-PCR-based methodology. As the figure shows, N-9-oxadecyl-6-methyl-DGJ induced a 20-fold induction of the 2⬘,5⬘ OAS gene at 70 M, with
lower concentrations giving a dose-dependent 2- to 15-fold increase in 2⬘,5⬘ OAS gene expression (Fig. 3C). In contrast, a compound that is structurally similar to N-9-oxadecyl-6-methyl-DGJ but has limited activity against HBV (N-7-oxanonyl-6methyl-DGJ) (16) did not detectably induce 2⬘,5⬘ OAS gene expression (Fig. 3A and B). This result provides evidence for the chemical specificity of this activation, as these compounds differ in only one carbon and the localization of the oxygen in the alkyl tail (compare Fig. 1B and C). It is noted that the antiviral activity seen with this compound class correlates with the induction of the small 2⬘,5⬘ OAS gene. That is, compounds at concentrations that are antiviral induce the small 2⬘,5⬘ OAS gene (16), while compounds and doses with no antiviral activity do not (Fig. 2 and 3 and data not shown). The innate host defense pathway may also involve the induction and secretion of interferon (21). As shown in Fig. 4, beta interferon secretion was induced by both the exogenous addition of 105 U of alpha interferon (2a/2b)/ml or the addition of N-9-oxadecyl-6-methyl-DGJ. Again, the induction of components of the innate host defense pathway correlated with antiviral activity. N-7-oxanonyl-6-methyl-DGJ, which has limited antiviral activity at the doses used in this assay, does not induce any of the 2,5 OAS genes and was unable to induce the
VOL. 48, 2004
␣-GALACTOSYLCERAMIDE ACTIVITY AGAINST HBV AND HCV
FIG. 4. N-9-oxadecyl-6-methyl-DGJ induces the secretion of beta interferon in tissue culture. Briefly, HepG2 cells were either left untreated (Mock) or treated with 105 IU of alpha interferon (2a/2b)/ml (positive control), 70 M N-7-oxanonyl-6-methyl-aza-galactose (used as a negative control), or N-9-oxadecyl-6-methyl-DGJ (SP-240; 50 M), and 72 h later the amount of beta interferon in the culture medium was determined using a commercially available human beta interferon enzyme-linked immunosorbent assay kit. As shown, compounds with antiviral activity and the ability to induce the 2⬘,5⬘ OAS genes possessed the ability to induce the production and secretion of beta interferon. In contrast, the compound N-7-oxanonyl-6-methylDGJ, which has limited activity (16) against HBV and does not noticeably induce any of the 2⬘,5⬘ OAS genes, lacks the ability to induce the production and secretion of beta interferon. It was noted that the exogenous addition of alpha interferon did not result in a signal in this assay, indicating specificity for beta interferon.
2089
production of beta interferon. Consistent with other reports (8), we have found that the level of beta interferon secreted after treatment with N-9-oxadecyl-6-methyl-DGJ (250 pg/ml or 10 to 20 U/ml), while consistent with the activation of the interferon pathway in these cell lines (3), is not sufficient to inhibit HBV directly (data not shown). However, this result provides further evidence that N-9-oxadecyl-6-methyl-DGJ induces the innate host defense pathway. As N-9-oxadecyl-6-methyl-DGJ could induce the innate host defense pathway, it was of interest to determine its effect against other viruses, such as HCV. Clone 9-13 is an Huh7derived cell line that constitutively expresses the bicistronic HCV subgenomic replicon and has been shown to be sensitive to alpha interferon (14). Therefore, as replication of HCV RNA in 9-13 cells is sensitive to alpha interferon and N-9oxadecyl-6-methyl-DGJ induces an arm of the interferon pathway, it was hypothesized that N-9-oxadecyl-6-methyl-DGJ would have an antiviral effect upon HCV in this cell line. This possibility was tested by examining the amount of HCV RNA in 9-13 cells as a function of incubation in various concentrations of our lead compound, N-9-oxadecyl-6-methyl-DGJ. The results, shown in Fig. 5, demonstrate a clear, if subtle, dosedependent reduction in the NS5A protein level as a function of drug treatment. A more dramatic reduction in the steady-state level of HCV RNA is observed after treatment with various concentrations of N-9-oxadecyl-6-methyl-DGJ, with an IC50 of 1.5 ⫾ 4.0 M. Beta-actin protein and RNA were used as controls in these experiments. Thus, as predicted, N-9-oxadecyl-6-methyl-DGJ is inhibitory for HCV. DISCUSSION In this report, we make several simple but surprising points regarding the abilities of certain glycolipid molecules, such as
FIG. 5. The amounts of HCV NS5A protein and RNA are reduced in cells incubated with interferon and N-methoxynonyl-6-methyl-DGJ. 9-13 cells, which harbor the 8-kb HCV bicistronic RNA replicon, were incubated for 48 h in the absence or presence of human alpha interferon (␣-IFN) or N-9-oxadecyl-6-methyl-DGJ (SP-240) at the indicated concentrations (international units of interferon or micromolar concentrations of N-methoxynonyl-6-methyl-DGJ). (A) Total cell proteins were lysed, separated on SDS-polyacrylamide gel electrophoresis, and probed with anti-NS5A antibody. The same blot was stripped and reprobed with anti-beta-actin antibody. (B) Total RNA was isolated, resolved through agarose gels, and hybridized to radioactive probe specific for HCV NS5B. After being washed, the blot was reprobed with radioactive beta-actin-specific sequences to control for loading of the lanes. The IC50 for N-9-oxadecyl-6-methyl-DGJ is 1.5 ⫾ 4.0 M. vRNA, viral RNA.
2090
MEHTA ET AL.
␣-galactosylceramide and N-9-oxadecyl-6-methyl-DGJ, to directly (i) induce the small 2⬘,5⬘ OAS gene family, (ii) induce the secretion of beta interferon, and (iii) inhibit the production of HBV DNA and HCV RNA. Although induction of components of the innate host defense pathway is clearly observed with these compounds, the exact mechanism by which they inhibit HBV and HCV is still under investigation by our laboratory and many others. However, it does appear that it is not through the direct action of the 2⬘,5⬘ OAS gene family (7). It was noted that N-7-oxanonyl-6-methyl-DGJ, which had very limited activity in the experiments described here (Fig. 2 and 3), was shown to antagonize an ion channel activity associated with the HCV p7 protein (19). Since p7 may be essential to the life cycle of HCV, one possibility is that the mechanism of antiviral action observed for the glycolipids studied here involved inhibiting p7 function. However, since N-9-oxadecyl6-methyl-DGJ activated the small (p40) 2⬘,5⬘ OAS gene and inhibited HCV under conditions where p7 is not present, this cannot be the sole explanation for its antiviral action. More generally, how the innate host defense pathway is activated is not fully known, but the rapid induction and highdose desensitization seen with the glycolipids used here is consistent with a receptor-mediated mechanism (21). It was initially believed that the glycolipids might work through the Toll-like receptor (TLR) family. However, a key factor of TLR stimulation, NF-B activation, was not detected with these compounds (data not shown). Thus, it is possible that the glycolipids presented here activate components of the innate host defense pathway by a TLR-distinct or NF-B-independent mechanism. The analysis of a possible receptor is under way. Activation of an innate host defense pathway, as shown here, is in some respects analogous to the phenomenon observed with double-stranded RNA (6, 24). In contrast to the situation with double-stranded RNA, however, activation with N-9-oxadecyl-6-methyl-DGJ and ␣-galactosylceramide appears to induce only a subset of interferon-specific transcripts and is associated with little or no toxicity (Fig. 2 and 3). In addition, these molecules are orally available and hence represent potential orally available therapeutics. Indeed, one compound in this class is in phase 2 clinical trials for the treatment of chronic HCV infection. In conclusion, we have shown that small, orally available glycolipid mimetics, such as N-methoxynonyl-6-methyl-DGJ, can directly activate cellular defense genes (such the small 2⬘,5⬘ OAS gene) and reduce the amount of HBV and HCV replication without the recruitment of any cells other then those infected. Since the synthetic glycolipids that stimulate this response could be mimetic for pathogen glycolipids, we propose that hepatocytes have the ability themselves to autogenously recognize and react defensively to foreign pathogen molecules without assistance from any other immunological cells, and perhaps this represents a very primitive arm of the host defense system. Thus, these synthetic glycolipids represent a new class of orally available small molecules that may have therapeutic value in all cases where interferon induction is useful. ACKNOWLEDGMENTS This work was supported by the Hepatitis B Foundation of America, an appropriation from the Commonwealth of Pennsylvania, NIH
ANTIMICROB. AGENTS CHEMOTHER. grants 1R41AI/DK49924-01 and AI53884 and Synergy Pharmaceuticals, Inc. Anand Mehta is the Bruce Witte Research Scholar of the Hepatitis B Foundation. Pamela Norton is thanked for her careful reading of the manuscript. Kirin Brewery Co., Ltd., is thanked for the gift of ␣-galactosylceramide. REFERENCES 1. Block, T. M., A. S. Mehta, C. J. Fimmel, and R. Jordan. 2003. Molecular viral oncology of hepatocellular carcinoma. Oncogene 22:5093–5107. 2. El-Serag, H. B., and A. C. Mason. 1999. Rising incidence of hepatocellular carcinoma in the United States. N. Engl. J. Med. 340:745–750. 3. Fredericksen, B., G. R. Akkaraju, E. Foy, C. Wang, J. Pflugheber, Z. J. Chen, and M. Gale, Jr. 2002. Activation of the interferon-beta promoter during hepatitis C virus RNA replication. Viral Immunol. 15:29–40. 4. Giaccone, G., C. J. Punt, Y. Ando, R. Ruijter, N. Nishi, M. Peters, B. M. von Blomberg, R. J. Scheper, H. J. van der Vliet, A. J. van den Eertwegh, M. Roelvink, J. Beijnen, H. Zwierzina, and H. M. Pinedo. 2002. A phase I study of the natural killer T-cell ligand alpha-galactosylceramide (KRN7000) in patients with solid tumors. Clin. Cancer Res. 8:3702–3709. 5. Gu, B., A. T. Gates, O. Isken, S. Behrens, and R. Sarisky. 2003. Replication studies using genotype 1a subgenomic hepatitis C virus replicons. J. Virol. 77:5352–5359. 6. Guidotti, L. G., A. Morris, H. Mendez, R. Koch, R. H. Silverman, B. R. Williams, and F. V. Chisari. 2002. Interferon-regulated pathways that control hepatitis B virus replication in transgenic mice. J. Virol. 76:2617–2621. 7. Guo, J. T., Q. Zhu, and C. Seeger. 2003. Cytopathic and noncytopathic interferon responses in cells expressing hepatitis C virus subgenomic replicons. J. Virol. 77:10769–10779. 8. Hayashi, Y., and K. Koike. 1989. Interferon inhibits hepatitis B virus replication in a stable expression system of transfected viral DNA. J. Virol. 63:2936–2940. 9. Hoofnagle, J. H., and A. M. Di Bisceglie. 1997. The treatment of chronic viral hepatitis. N. Engl. J. Med. 336:347–356. 10. Hovnanian, A., D. Rebouillat, E. R. Levy, M. G. Mattei, and A. G. Hovanessian. 1999. The human 2⬘,5⬘-oligoadenylate synthetase-like gene (OASL) encoding the interferon-induced 56-kDa protein maps to chromosome 12q24.2 in the proximity of the 2⬘,5⬘-OAS locus. Genomics 56:362–373. 11. Kakimi, K., L. G. Guidotti, Y. Koezuka, and F. V. Chisari. 2000. Natural killer T cell activation inhibits hepatitis B virus replication in vivo. J. Exp. Med. 192:921–930. 12. Kikuchi, A., M. Nieda, C. Schmidt, Y. Koezuka, S. Ishihara, Y. Ishikawa, K. Tadokoro, S. Durrant, A. Boyd, T. Juji, and A. Nicol. 2001. In vitro antitumour activity of alpha-galactosylceramide-stimulated human invariant V␣24⫹NKT cells against melanoma. Br. J. Cancer 85:741–746. 13. Liaw, Y. F. 2001. Impact of YMDD mutations during lamivudine therapy in patients with chronic hepatitis B. Antivir. Chem. Chemother. 12(Suppl. 1):67–71. 14. Lohmann, V., F. Korner, J. Koch, U. Herian, L. Theilmann, and R. Bartenschlager. 1999. Replication of subgenomic hepatitis C virus RNAs in a hepatoma cell line. Science 285:110–113. 15. Lu, X., T. Tran, E. Simsek, and T. M. Block. 2003. The alkylated imino sugar N-(N-nonyl)-deoxygalactonojirimycin reduces the amount of hepatitis B virus nucleocapsid in tissue culture. J. Virol. 77:11933–11940. 16. Mehta, A., B. Conyers, D. L. J. Tyrrell, K. A. Walters, G. A. Tipples, R. A. Dwek, and T. M. Block. 2002. Structure-activity relationship of a new class of anti-hepatitis B virus agents. Antimicrob. Agents Chemother. 46:4004–4008. 17. Mehta A., B. Conyers, R. Jordan, T. Butters, R. A. Dwek, and T. M Block. 2001. Inhibition of hepatitis B virus DNA replication by imino sugars without the inhibition of the DNA polymerase: therapeutic implications. Hepatology 33:1488–1495. 18. Naidenko, O. V., Y. Koezuka, and M. Kronenberg. 2000. CD1-mediated antigen presentation of glycosphingolipids. Microbes Infect. 2:621–631. 19. Pavlovic, D., D. C. Neville, O. Argaud, B. Blumberg, R. A. Dwek, W. B. Fischer, and N. Zitzmann. 2003. The hepatitis C virus p7 protein forms an ion channel that is inhibited by long-alkyl-chain iminosugar derivatives. Proc. Natl. Acad. Sci. USA 100:6104–6108. 20. Rebouillat, D., and Ara G. Hovanessian. 1999. The human 2⬘,5⬘-oligoadenylate synthetase family: interferon-induced proteins with unique enzymatic properties. J. Interferon Cytokine Res. 19:295–309. 21. Samuels, C. 2001. Antiviral actions of interferons. Clin. Microbiol. Rev. 14:778–809. 22. Sells, M. A., M. L. Chen, and G. Acs. 1987. HepG2 cells transfected with cloned hepatitis B virus DNA. Proc. Natl. Acad. Sci. USA 84:1005–1009. 23. Wei, Y., J. E., Tavis, and D. Ganem. 1997. Relationship between viral DNA synthesis and virion envelopment in hepatitis B virus. J. Virol. 70:6455–6458. 24. Wieland, S. F., L. G. Guidotti, and F. V. Chisari. 2000. Intrahepatic induction of alpha/beta interferon eliminates viral RNA-containing capsids in hepatitis B virus transgenic mice. J. Virol. 74:4165–4173.
