|Home | About | Journals | Submit | Contact Us | Français|
A series of 4′-thionucleosides were synthesized and evaluated for activities against orthopoxviruses and herpesviruses. We reported previously that one analog, 5-iodo-4′-thio-2′-deoxyuridine (4′-thioIDU), exhibits good activity both in vitro and in vivo against two orthopoxviruses. This compound also has good activity in cell culture against many of the herpesviruses. It inhibited the replication of herpes simplex virus type 1 (HSV-1), HSV-2, and varicella-zoster virus with 50% effective concentrations (EC50s) of 0.1, 0.5, and 2 μM, respectively. It also inhibited the replication of human cytomegalovirus (HCMV) with an EC50 of 5.9 μM but did not selectively inhibit Epstein-Barr virus, human herpesvirus 6, or human herpesvirus 8. While acyclovir-resistant strains of HSV-1 and HSV-2 were comparatively resistant to 4′-thioIDU, it retained modest activity (EC50s of 4 to 12 μM) against these strains. Some ganciclovir-resistant strains of HCMV also exhibited reduced susceptibilities to the compound, which appeared to be related to the specific mutations in the DNA polymerase, consistent with the observed incorporation of the compound into viral DNA. The activity of 4′-thioIDU was also evaluated using mice infected intranasally with the MS strain of HSV-2. Although there was no decrease in final mortality rates, the mean length of survival after inoculation increased significantly (P < 0.05) for all animals receiving 4′-thioIDU. The findings from the studies presented here suggest that 4′-thioIDU is a good inhibitor of some herpesviruses, as well as orthopoxviruses, and this class of compounds warrants further study as a therapy for infections with these viruses.
Iterative medicinal chemistry studies can optimize molecules for the inhibition of viral enzymes and yield molecules that are highly effective and remarkably specific. This approach, as well as others, has led to the development of highly selective therapies for the treatment of viral infections and has been reviewed previously (5, 9, 20). While these therapeutic agents have excellent activities against the target enzymes from specific viruses, they can have reduced or poor efficacy against other closely related viruses. Thus, it is important to assess the activities of candidate molecules against an array of viruses to determine their spectra of activity. The development of compounds with activities against many viruses is challenging, yet essential, given the vast array of viral pathogens that cause human disease. A broad spectrum of activity is particularly important in the development of therapies for infections caused by emerging pathogens.
Nucleoside phosphonate antiviral agents have been approved for the treatment of infections with hepatitis B virus and human immunodeficiency virus and are under development for infections caused by hepatitis C virus and DNA viruses (6). Cidofovir (CDV) is one such compound that has activity against many viruses, and recently, alkoxyalkyl ester prodrug forms of CDV have been reported to be active at concentrations up to 3 orders of magnitude lower than the active concentrations of the parent compound (36). The compound CMX001 (hexadecyloxypropyl CDV) has excellent activity against adenoviruses, orthopoxviruses, and polyomaviruses, as well as the herpesviruses (13, 18, 32, 36). Equally important is the increased oral bioavailability conferred by the alkoxyalkyl substituent (1, 3, 17, 30), which appears to reduce drug exposure in the mouse kidney and may reduce the potential of the drug for renal toxicity (4). While the discovery of truly broad spectrum antiviral drugs will be difficult, the development of CMX001 represents a significant advance.
Previously, the efficacies of a series of nucleoside analogs against vaccinia and cowpox viruses were evaluated and a number of thymidine analogs that have good activities against these viruses were identified (8, 26). This result was intriguing since the orthopoxviruses encode thymidine kinase (TK) homologs that are closely related to the human cytosolic TK (TK1) (15). The TK homologs encoded by the orthopoxviruses and the human cytosolic TK are both type II TK homologs, which are characterized by a number of features including their homotetrameric structures and their comparatively narrow substrate specificities (2). Some herpesviruses encode distinct type I TK enzymes, which are active as dimers and have broader substrate specificities that include thymidine, cytidine, and even some purine analogs such as acyclovir (ACV) (16). Results from genetic studies with orthopoxvirus and herpesvirus TK mutants suggested that these enzymes could confer sensitivity to a few thymidine analogs, presumably through selective phosphorylation to the level of the monophosphate (26, 29). This result was confirmed subsequently by the findings of enzymatic studies, which showed that the orthopoxvirus and herpesvirus TK enzymes share the ability to catalyze the phosphorylation of certain thymidine analogs, some of which are not substrates for the cellular cytosolic TK (25). This group of analogs includes a few 2′-deoxyuridine analogs with large substituents in the 5 position (25). The carbocyclic analog N-methanocarbathymidine has also been reported previously to exhibit good activity against both the orthopoxvirus and herpesvirus families (26) and is selectively phosphorylated by each of the TK homologs encoded by these viruses (data not presented). The unexpected finding that orthopoxvirus TK homologs possess broader substrate specificities than the human cytosolic TK homolog was interesting and suggested a potential avenue for the development of therapies for infections with these viruses. Importantly, the overlapping specificities of vaccinia virus TK and herpes simplex virus (HSV) TK also suggested that it might be possible to identify additional thymidine analogs with spectra of activity that included both the orthopoxviruses and some of the human herpesviruses (HHVs). Compounds with this property would be desirable since therapies for herpesvirus infections are more viable economically than those for orthopoxvirus infections, which may help drive the continued development of these compounds. Ideally, they should also retain activity against drug-resistant strains of these viruses.
A series of 2′-deoxy-4′-thiopyrimidine nucleosides were synthesized previously and reported to be active against human cytomegalovirus (HCMV) (33). Other related analogs identified by Rahim and colleagues were also reported to have good activities against varicella-zoster virus (VZV), as well as HSV (31). One analog in this series, 5-iodo-4′-thio-2′-deoxyuridine (4′-thioIDU), also exhibited activity against HCMV. Recently, we described the antiviral activities of this and related molecules against vaccinia virus and cowpox virus (19). The most active analog was 4′-thioIDU, which inhibited viral replication in vitro at submicromolar concentrations. This compound also significantly reduced the mortality of mice infected with cowpox virus when administered orally at concentrations of 5 mg/kg of body weight and when therapy was initiated as late as 4 days after infection. The activity of the compound was shown to be largely dependent on the orthopoxvirus TK homologs, so we examined the activities of this and related compounds against HSV-2 and confirmed that 4′-thioIDU had the highest activity. The spectrum of activity of this compound also included HSV-1, VZV, and cytomegalovirus, but selective activity against the other HHVs was not observed. Immunofluorescence studies showed that the compound was specifically phosphorylated in infected cells, incorporated into viral DNA, and dependent on viral TK activity. However, in contrast to previous results that indicated good in vivo efficacy against two orthopoxviruses, the findings of the present study revealed that the compound did not significantly reduce the mortality of mice infected intranasally with HSV-2.
4′-Thiothymidine (compound 1), 1-(2-deoxy-4-thio-β-d-ribofuranosyl)-5-bromouracil (compound 2), 1-(2-deoxy-4-thio-β-d-ribofuranosyl)-5-trifluoromethyluracil (compound 3), and 1-(2-deoxy-4-thio-β-d-ribofuranosyl)-5-iodouracil (compound 4) (Fig. (Fig.1)1) were synthesized according to the procedures described previously by Secrist et al. (33) and Rahim et al. (31). 1-(4-Thio-β-d-arabinofuranosyl)-cytosine (compound 5) and 1-(4-thio-β-d-arabinofuranosyl)-flucytosine (compound 6) were prepared as described by Tiwari et al. (34, 35). 1-(2-Deoxy-4-thio-β-d-ribofuranosyl)-5-bromocytosine (compound 7) was prepared by coupling of the appropriate 4-thiosugar with 5-bromocytosine by the methodology described by Secrist et al. (33).
Primary human foreskin fibroblast (HFF) cells were prepared from newborn human foreskins, which were minced and washed repeatedly with phosphate-buffered saline (PBS) deficient in calcium and magnesium. The tissue was incubated with 0.25% trypsin for 1 h at 37°C in a CO2 incubator. The supernatant containing individual cells was filtered through sterile gauze into a flask containing growth medium consisting of minimum essential medium with Earl's salts, supplemented with 10% fetal bovine serum (FBS; HyClone, Inc., Logan, UT), l-glutamine, vancomycin, and amphotericin B (Fungizone). Cells were collected by centrifugation, seeded into 25-cm2 flasks, and incubated in a 37°C humidified CO2 incubator until they reached confluence. The cells were then expanded until passage 4, when the antibiotics in the growth medium were changed to penicillin and gentamicin. All studies with HFF cells were conducted with cells at passage numbers below 10.
Strains of HSV-1 used to assess antiviral activity included the wild-type (wt) E-377 strain, as well as drug-resistant isolates DM2.1, PAAr5, and SC16-S1 (gifts from Jack Hill, Burroughs Wellcome) and B-2006 (a gift from Pamela Chatis, Beth Israel Hospital) (7). The isolation and characteristics of SC16-S1 were reported previously (21). The wt HSV-2 strain MS, as well as resistant isolates 12247, 13077, and 11680 (gifts from Jack Hill, Burroughs Wellcome) and AG-3 (a gift from Pamela Chatis, Beth Israel Hospital) (7), was used to evaluate the activities of the compounds. The susceptibility of each of these strains to ACV was confirmed and reported recently (14), and the TK and DNA polymerase open reading frames from each of the viruses were sequenced to confirm the genotypes. Polymorphisms in the inferred amino acid sequences of the TKs from HSV-1 strains relative to that from strain 17 (accession number NP_044624) are as follows: E-377, R212K, G240E, A265T, and R281Q; DM2.1, N23S, K36E, L42P, V90E, L91M, A93E, S94A, and a stop at E95 (E95stop); SC16-S1, C6G, N23S, K36E, L42P, A265T, C336Y, and V348I; PAAr5, N23S, K36E, R89Q, and A265T; and B-2006, N23S, K36E, R89Q, I143V, Q185R, and a mutation resulting from a frameshift following nucleotide 547. TK polymorphisms relative to the TK sequence from HSV-2 strain HG-52 (accession number NP_044492.1) are as follows: MS, G39E; AG-3, G39E, N78D, Q105P, and L140F; 12247, G39E, N78D, L140F, and C337Y; 13077, G39E and R223H; and 11680, G39E, N78D, and R223H. Polymorphisms in the inferred amino acid sequences of DNA polymerases from HSV-1 strains relative to that from strain 17 (accession number NP_044632.1) are as follows: E-377, A330R, V383M, A646T, P920S, and T1208A; DM2.1, A330R, N522S, V905M, A1203T, and T1208A; SC16-S1, S33G, A330R, N522S, V905M, A1203T, and T1208A; PAAr5, S33G, A330R, A566T, R842S, P1124H, and T1208A; and B-2006, A27 deletion (A27del), S33del, G314del, A330R, V905M, and T1208A. Polymorphisms in the HSV-2 strain DNA polymerase sequences relative to that from strain HG-52 (accession number NP_044500.1) are as follows: MS, A9T, P15S, L60P, E139K, and deletion of residues 680 to 681; AG-3, E678G; 13077, A9T, P15S, L60P, and S1043N; 12247, A9T, P15S, and L60P; and 11680, G39E, N78D, and R223H.
HCMV wt strain AD169 was obtained from the American Type Culture Collection (ATCC), and the resistant strains C8914-6, 759rD100, 1117r, VR 4955r, and PFArB300 were a gift from Karen Biron (Burroughs Wellcome). The ganciclovir (GCV) susceptibilities of the strains were determined and reported recently (14). Construction of a recombinant HCMV with a K355M mutation in the UL97 open reading frame (strain RC314) was described previously (28), and this kinase null mutant is resistant to GCV (12). The UL97 and UL54 genes from each strain were sequenced to confirm the genotypes. Polymorphisms in the amino acid sequences inferred from UL97 genes relative to the sequence from AD169 (accession number YP_002608287.1) are as follows: AD169, no mutations; RC314, K355M; C8914-6, I244V and L595F; 759rD100, deletion of residues 590 to 593, A582E, L583A, and R566L; 1117r, no mutations; VR 4955r, no mutations; and PFArB300, no mutations. DNA polymerase polymorphisms relative to the sequence from AD169 (accession number YP_002608259.1) are as follows: AD169, no mutations; RC314, no mutations; C8914-6, G347D, L501F, S655L, N685S, G874R, A885T, L890F, N898D, and N1147S; 759rD100, A987G; 1117r, K513N; VR 4955r, P628L, S655L, T700A, N685S, insertion of T885, A884S, E1089G, and K1056R; and PFArB300, L724V and T946M.
The Ellen strain of VZV was obtained from the ATCC (Manassas, VA). Epstein-Barr virus (EBV)-infected Akata cells were a gift from John Sixbey (Louisiana State University, Baton Rouge). HSB-2 cells and the GS strain of human herpesvirus 6 variant A (HHV-6A), as well as BCBL1 cells infected with HHV-8, were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH.
Plaque reduction assays were performed with six-well plates containing monolayers of confluent HFF cells. Test compounds were diluted serially in minimum essential medium containing 2% FBS and pooled human immunoglobulin (Polygam S/D; Baxter Healthcare Corp. Hyland Immuno Division, Glendale, CA) to yield six concentrations of the compounds, with the highest concentration being 100 μM. Virus suspensions were diluted in growth medium, and 0.2-ml aliquots were added to each of triplicate wells to yield 20 to 30 plaques per well. The plates were rocked every 15 min for 1 h, and then compound dilutions were added to the infected monolayers and the plates were incubated for 3 days. Cell monolayers were stained with 0.1% crystal violet in 20% methanol, and plaques were enumerated using an inverted stereomicroscope. Compound concentrations that were sufficient to reduce the plaque number by 50% (50% effective concentrations [EC50s]) were interpolated from the experimental data by using a modified version of the software MacSynergy II (27; http://main.uab.edu/peds/Templates/Inner.aspx?pid=118889).
Assays used for the evaluation of activity against HCMV and VZV were similar to those used for HSV with the following modifications. Pooled human immunoglobulin was not included in the solutions containing the test compounds, and the plates were incubated for 8 and 10 days for HCMV and VZV, respectively. Cell monolayers were stained with 2 ml of a 0.017% neutral red solution in PBS rather than crystal violet, and plaques were enumerated on an inverted stereomicroscope.
Activity against the Akata strain of EBV was analyzed by methods reported previously (24). Akata cells were routinely passaged in RPMI 1640 (Mediatech, Inc., Herndon, VA) supplemented with 10% FBS and standard concentrations of l-glutamine, penicillin, and gentamicin at 37°C in a humidified 5% CO2 incubator. Compounds were diluted to yield concentrations that ranged between 100 and 0.032 μM in round-bottom 96-well plates. Latently infected Akata cells were induced to undergo a productive infection by the addition of 50 μg/ml goat anti-human immunoglobulin G antibody, and a sample of 4 × 104 cells was added to each well. Following a 72-h incubation, 100 μl of a denaturation buffer (1.2 M NaOH, 4.5 M NaCl) was added to each well and an aliquot of the DNA was transferred onto a positively charged nylon membrane. Viral DNA was quantified by hybridization using a digoxigenin-labeled probe and an antibody specific for the hapten according to the protocol of the antibody manufacturer (Roche Diagnostics, Indianapolis, IN).
Activity against the HHV-6A variant was assessed by using 96-well plates containing 2 × 104 HSB-2 cells per well in RPMI 1640 supplemented with 10% FBS, l-glutamine, penicillin, and gentamicin. Compound dilutions in growth medium were prepared and added to triplicate wells. Plates containing HSB-2 cells were infected with the GS strain of HHV-6A and incubated at 37°C for 7 days. The accumulation of viral DNA was evaluated by adding denaturation buffer to the wells, immobilizing total DNA on nylon membranes, and hybridizing with a digoxigenin-labeled DNA probe. Bound probe was detected with an antibody specific for the hapten and quantified with QuantityOne software (Bio-Rad, Hercules, CA).
BCBL1 cells were maintained in growth medium consisting of RPMI 1640 supplemented with 10% FBS, penicillin, gentamicin, and l-glutamine. Compounds were diluted in triplicate wells of a 96-well plate, with the highest dilution yielding a concentration of 100 μM. The cells were induced to undergo a lytic infection by the addition of phorbol 12-myristate 13-acetate (Promega, Madison, WI) at a final concentration of 100 ng/ml. Plates were incubated for 7 days at 37°C in a humidified CO2 incubator. Total DNA was prepared with a Wizard SV 96-well purification kit (Promega), and viral DNA was quantified by quantitative PCR using forward primer 5′-TTC CCC AGA TAC ACG ACA GAA TC-3′, reverse primer 5′-CGG AGC GCA GGC TAC CT-3′, and probe 5′-(6-carboxyfluorescein)-CCT ACG TGT TCG TCG AC-(6-carboxytetramethylrhodamine)-3′. Compound concentrations sufficient to reduce viral DNA accumulation by 50% (EC50s) were calculated by standard methods.
Neutral red uptake assays were conducted with 96-well plates containing confluent HFF cells. Compound solutions were diluted serially to yield a set of solutions with the highest compound concentration being 100 μM, and cell monolayers were incubated for 7 days. Growth medium was then aspirated, 200 μl of a 0.066% neutral red stain solution in PBS was added, and the plates were incubated for 1 h. The remaining dye was washed from the monolayers, and the dye internalized by the cells was solubilized in a solution containing 50% ethanol and 1% glacial acetic acid. The plate contents were mixed for 15 min on a rotating shaker, and the optical densities at 550 nm were determined with a microplate reader. Cytotoxicity in lymphocytes was evaluated using MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; Promega]. Assay plates for cytotoxicity determinations were prepared in the same manner as those for the antiviral assays with EBV, HHV-6A, and HHV-8, except that the cells were not induced to undergo a lytic infection or were not infected. A set of control wells that contained only growth medium was also included. Cytotoxicity was assessed at the same time that viral replication was evaluated for each virus. Briefly, 20 μl of MTS was added to each well, and the wells were incubated at 37°C for 4 h. Levels of converted formazan products were measured at 490 nm with a microplate reader. For both the MTS and the neutral red uptake assays, compound concentrations that reduced viable cell numbers by 50% (50% cytotoxic concentrations [CC50s]) were calculated by standard methods.
Cell proliferation assays were used as another measure of cytotoxicity. HFF cells were seeded into six-well plates at a concentration of 2.5 × 104 cells per well. The plates were incubated for 24 h, and the medium was removed prior to the addition of compound dilutions in growth medium. The cells were incubated for 72 h at 37°C, and then the cells were removed with trypsin-EDTA and viable cells in triplicate wells were enumerated with a Coulter Counter. The average cell number was used to calculate the concentration of a compound sufficient to reduce the cell number by 50% (the 50% inhibitory concentration).
Monolayers of HFF cells on glass coverslips were prepared by methods similar to those reported previously (28). The MS strain of HSV-2 or the strain of this virus resistant to 4′-thioIDU was used to infect cells at a multiplicity of infection of 0.2 PFU/cell, and cells were incubated for 48 h. For studies with HCMV, cells were infected with AD169 at a multiplicity of infection of 0.05 PFU/cell and incubated for 3 days. To label and detect the incorporation of 5-bromo-2′-deoxyuridine (BrdUrd) into DNA, cells were incubated with BrdUrd prior to fixation and its incorporation into DNA was detected with a monoclonal antibody specific for this analog incorporated into DNA by using a BrdUrd labeling and detection kit according to the suggested protocol of the manufacturer (Roche Diagnostics, Mannheim, Germany). These studies were also performed with 4′-thioIDU, which was substituted for BrdUrd through the addition of a 10 μM concentration of the compound for 30 min prior to fixation. The BrdUrd-specific antibody also cross-reacts with 4′-thioIDU (as well as idoxuridine [IDU]), so the incorporation of 4′-thioIDU into DNA was also detected using the same reagents and protocols. Cells infected with HSV were detected with a monoclonal antibody specific for a 155-kDa nucleocapsid protein (U.S. Biological, Swampscott, MA) and Zenon Texas Red labeling reagent (Invitrogen Molecular Probes, Eugene, OR). HCMV-infected cells were detected with a monoclonal antibody specific for pp65 (a gift from Bill Britt, University of Alabama, Birmingham). Cells were mounted with SlowFade gold reagent with DAPI (4′,6-diamidino-2-phenylindole; Invitrogen Molecular Probes).
Female BALB/c mice 3 to 4 weeks of age were manually restrained for intranasal inoculation with HSV-2 strain MS at an approximate 90% lethal dose of 1.1 × 105 PFU/animal in a 40-μl volume. Treatments were administered twice daily at 12-h intervals by oral gavage using a volume of 0.2 ml to provide 30, 10, or 3 mg/kg of 4′-thioIDU. ACV was given similarly using a 100-mg/kg dose for the positive control. Treatments were initiated 24, 48, or 72 h post-viral inoculation and administered for seven consecutive days. Mice that developed clinical signs of encephalitis were humanely euthanized. Surviving mice were monitored for 21 days.
A series of 5-substituted 4′-thiodeoxyribonucleosides were synthesized, and we reported previously that many of these compounds exhibit good activities against both vaccinia and cowpox viruses (19). A subset of the compounds also appeared to be active against HSV-1, HSV-2, VZV, and HCMV in cytopathic effect reduction assays (data not shown). The activities of all of the active compounds against HSV-2 were confirmed in plaque reduction assays (Table (Table1).1). Concurrent evaluations of cytotoxicity in a neutral red uptake assay confirmed that the compounds were minimally toxic in stationary cells, and five of the compounds had selective indices (calculated as CC50/EC50 ratios) of >100. The cytotoxicities of the compounds were comparable to those observed previously in studies that determined their activity against the orthopoxviruses (19). The compounds were evaluated further in a proliferation assay using primary HFF cells, which is a very sensitive indicator of toxicity and in our experience has been predictive of the toxicity in animal models. The most effective compounds in this series also inhibited cell proliferation, but 4′-thioIDU appeared to have the best balance between toxicity and efficacy. Subsequent studies focused on the activity of this molecule as a representative member of this series.
Additional studies with 4′-thioIDU examined its activity against other HHVs to assess its spectrum of activity against this family of viruses (Table (Table2).2). The compound inhibited HSV-1 and HSV-2 with EC50s of 0.1 and 0.45 μM, respectively, which are somewhat lower than the EC50s of 0.8 to 1 and 10 μM reported previously(31). The compound also inhibited the replication of both VZV and HCMV with EC50s of 2 and 5.9 μM, respectively, which are also similar to the values of 3.5 and 4 μM, respectively, cited in the previous report. The activity of the compound against EBV and HHV-6A, as well as HHV-8, was also evaluated. No specific antiviral activity was detected, and this result was likely related to the increased toxicity of the compound in the continually dividing lymphocytes, consistent with results from the proliferation assay (Table (Table11).
The efficacies of 4′-thioIDU against drug-resistant strains of HSV and HCMV were also evaluated to assess its potential utility in the treatment of ACV- and GCV-resistant infections, respectively. Most ACV-resistant strains of HSV were also resistant to 4′-thioIDU (Table (Table3).3). The HSV-1 strain DM2.1, which expresses a truncated TK, was more than 100-fold less susceptible to the compound than wt virus, and EC50s for other ACV-resistant strains of HSV-1 and HSV-2 were increased compared to those for wt strains. The SC16-S1 ACV-resistant strain of HSV-1 was among the least resistant to the compound, but this strain had altered TK activity and might still be capable of phosphorylating the compound to some degree (21, 22). Mutations in the DNA polymerase of a phosphonoacetic acid-resistant strain of HSV-1 did not appear to be associated with increased sensitivity to 4′-thioIDU. In summary, mutation of the TK homologs correlates with resistance to 4′-thioIDU, but we cannot exclude the possibility that mutations in the DNA polymerases might also impart some resistance to this molecule. The cross-resistance of the ACV-resistant strains suggests that this molecule would not be a good candidate to treat ACV-resistant infections.
Since HCMV does not express a functional TK, we examined another set of mutants to correlate resistance to 4′-thioIDU with mutations in the UL97 protein kinase and the DNA polymerase (Table (Table4).4). A UL97 kinase null mutant (RC314) was fully susceptible to the compound, indicating that the compound was not phosphorylated by UL97 kinase. Three isolates, including a foscarnet (PFA)-resistant isolate (PFArB300) that has mutations in the DNA polymerase but not in the UL97 kinase, were comparatively resistant to the compound. This finding suggested that mutations in the polymerase, rather than those in the UL97 kinase, were associated with resistance to the compound. These data taken together suggest that 4′-thioIDU targets the viral DNA polymerase and that point mutations in this enzyme that confer resistance to GCV or PFA may also confer resistance to 4′-thioIDU.
The MS strain of HSV-2 was passaged serially in 0.5, 2, and 5 μM 4′-thioIDU, and an isolate was obtained from the resulting stock by plaque purifying the virus three times. This isolate proved to be highly resistant to the compound, with the EC50 for the isolate being more than 100-fold greater than that for the parent strain of the virus (Table (Table5).5). It was also resistant to IDU, which was expected given the similar structures of 4′-thioIDU and IDU. The 4′-thioIDU-resistant isolate was also highly resistant to ACV, which was consistent with the reduced susceptibilities of ACV-resistant isolates to 4′-thioIDU.
The open reading frames for both the TKs and the DNA polymerases from the resistant isolate and the parent virus were sequenced to characterize the mutations associated with the observed resistance. The resistant isolate had acquired a frameshift mutation in the TK following amino acid 197, resulting in a premature stop after amino acid 262. Thus, the truncated TK in this mutant was catalytically inactive and was sufficient to confer resistance to ACV as well as IDU. The DNA polymerase was also affected and had acquired a D214N mutation. The TK frameshift mutation was introduced into the MS strain to confirm that this mutation was related to the resistance phenotype. As expected, the mutant was resistant to ACV, IDU, and 4′-thioIDU, confirming that the mutation was sufficient to confer resistance to each of the compounds (Table (Table55).
The phosphorylation of IDU by HSV TK and its incorporation into viral DNA provide an obvious model for the mechanism of action of 4′-thioIDU. The resistance study results presented above indicated that phosphorylation by the viral TK was important for efficacy. Additional studies were undertaken to confirm a role for TK and to investigate the possible incorporation of 4′-thioIDU triphosphate into viral DNA. A monoclonal antibody specific for IDU and BrdUrd that had the potential to cross-react with 4′-thioIDU was used to investigate the potential incorporation of 4′-thioIDU into DNA. Cells infected with HSV-2 were incubated briefly with BrdUrd or with 4′-thioIDU, and the monoclonal antibody was used to detect the incorporation of the molecule into DNA. No specific staining of the cells was observed in the absence of drug treatment, and this outcome confirmed the specificity of the antibody (Fig. (Fig.2A).2A). However, the addition of either BrdUrd or 4′-thioIDU resulted in specific nuclear staining in cells infected with the virus, which suggested that both molecules were incorporated into viral DNA. This study was repeated with the 4′-thioIDU-resistant strain; specific staining was absent from most infected cells, which indicated reduced incorporation of both compounds into DNA and was consistent with reduced phosphorylation of the compounds by the viral TK (Fig. (Fig.2B).2B). Specific staining was detected in small subsets of infected and uninfected cells treated with either BrdUrd or 4′-thioIDU. This result is consistent with the established phosphorylation and incorporation of BrdUrd into DNA during the S phase of the cell cycle and suggests that 4′-thioIDU may also be phosphorylated by cellular TK1 and incorporated into DNA by a cellular DNA polymerase. A similar staining pattern was also observed in cells infected with HCMV and may indicate that TK1 was involved in the phosphorylation of the drug in these cells (Fig. (Fig.2C).2C). The pattern of incorporation in infected cells was also consistent with the incorporation of the molecules into viral DNA since incorporation was localized to the DNA replication compartment. These data confirm that the mechanism of action of 4′-thioIDU is similar to that of IDU and suggest that mutations conferring resistance to the compound may occur in either TK or DNA polymerases. They suggest also that 4′-thioIDU can be incorporated into DNA in uninfected cells, indicating that a cellular kinase, such as TK1, can also phosphorylate the compound and that the triphosphate metabolite is a substrate for a cellular polymerase. This would also be consistent with the observed inhibition of cell proliferation (Table (Table11).
Previous studies with mice infected intranasally with cowpox virus showed that 4′-thioIDU is effective in reducing mortality when administered orally, even when treatment is delayed for 4 days (19). These data showed that the compound is orally bioavailable and exhibits pharmacokinetic properties sufficient to inhibit the replication of cowpox virus in mice. In the present study, a similar experiment with mice infected intranasally with the MS strain of HSV-2 was conducted (Table (Table6).6). Mortality was reduced significantly and the mean time to death was increased significantly by the ACV control when it was administered starting at up to 72 h postinoculation. However, while the experimental compound significantly increased the mean time to death when it was administered beginning at 24, 48, and 72 h postinoculation, it did not significantly reduce mortality in mice infected with HSV-2.
Initial reports describing the antiviral activities of 4′-thiodeoxyribonucleosides focused on their efficacies against HSV-1, HSV-2, HCMV, and VZV (31). A more recent study showed that these analogs had even better activities against the orthopoxviruses (19). Additional analogs were synthesized, and their activities against HSV-2 were evaluated, since HSV-2 was one of the viruses most susceptible to the activities of these analogs. Five compounds in this series had EC50s that were less than 1 μM, and subsequent assessments of cytotoxicity indicated that 4′-thioIDU had the best balance between antiviral activity and toxicity. Additional studies with this agent confirmed its activity against HSV-1, HSV-2, VZV, and HCMV, but no specific activity against EBV, HHV-6, or HHV-8 was detected. Although EBV encodes a TK and would be expected to be susceptible to the compound, the toxicity of the compound in lymphocytes may have precluded the detection of activity against this virus.
The activity of the compound against drug-resistant strains of HSV was also evaluated, and the pattern of activity was interesting. Mutations in the HSV TK gene associated with resistance to ACV also appeared to confer resistance to 4′-thioIDU, indicating that the TK enzyme was important in the mechanism of action of 4′-thioIDU. This result was confirmed using an isolate that was highly resistant to 4′-thioIDU. The TK open reading frame in this isolate had acquired a frameshift mutation that precluded the expression of an active enzyme and was similar to mutations frequently observed in this gene in clinical isolates resistant to ACV (11, 23). The mutant also exhibited reduced incorporation of the compound into viral DNA, consistent with reduced phosphorylation. This mutation was reconstructed in the susceptible strain of the virus and was shown to be sufficient to confer resistance to ACV, IDU, and 4′-thioIDU. These data are also similar to those reported previously for orthopoxviruses, which showed that the vaccinia virus TK is involved in the mechanism of action of 4′-thioIDU (19). Indeed, we have also observed the incorporation of this compound into vaccinia virus DNA (data not presented).
A mutation in the DNA polymerase was also observed in the 4′-thioIDU-resistant mutant, which had a D214N mutation in the RNase H domain. Although the mutation in the TK was shown to be sufficient to impart resistance to the compound, we cannot exclude the possibility that D214N may also contribute to the resistance phenotype. The observed incorporation of the compound into DNA indicates that the compound is a substrate of the viral DNA polymerase and that mutations in the corresponding gene would be expected to be associated with resistance.
The susceptibilities of drug-resistant HCMV isolates to 4′-thioIDU were also examined. While HCMV does not encode a TK homolog, it encodes a serine/threonine kinase (pUL97) that can phosphorylate some nucleoside analogs, including GCV. Thus, mutations in either the UL97 kinase or the DNA polymerase might be expected to be associated with resistance to 4′-thioIDU. A recombinant virus with a K355M point mutation that inhibits UL97 kinase activity proved to be fully sensitive to the compound. These data suggested that the UL97 kinase did not contribute significantly to the phosphorylation of 4′-thioIDU and that other cellular kinases were responsible for the phosphorylation of the compound. In contrast, some mutations in the DNA polymerases of GCV-resistant and PFA-resistant strains appear to be associated with reduced sensitivity to 4′-thioIDU and are consistent with its incorporation into viral DNA.
It was interesting that the efficacy of 4′-thioIDU against the ACV-resistant mutants was reduced approximately 10-fold, in contrast to the >100-fold reduction in the efficacy of ACV. This result would be expected if there was another cellular kinase that could also phosphorylate the compound to some degree, as has been reported previously for IDU (10). The fact that the observed efficacy against TK-negative strains of HSV is similar to that against wt strains of HCMV is consistent with this hypothesis, since HCMV does not express an active TK homolog. The incorporation of 4′-thioIDU into a subset of uninfected cells is also consistent with this hypothesis, and an obvious candidate kinase is TK1, which is induced during the S phase of the cell cycle.
The observed efficacy of 4′-thioIDU in vitro showed that the compound was active against HSV-2, and the compound was shown previously to reduce mortality among mice infected intranasally with cowpox virus (19). However, the compound did not significantly reduce the mortality of animals infected with HSV-2 by the same route of administration. Oral treatment with the compound did significantly increase the mean time to death, and the results appeared to confirm those of the previous study that showed the compound to be orally bioavailable. Thus, the failure of the compound to reduce mortality was likely related to the pathogenesis of HSV-2 and may also be related to the inability of the compound to penetrate the central nervous system. Additional studies will be required to establish efficacy, and the compound may prove to be effective against infections initiated by other routes.
The results presented here suggest that the mechanism of action of the compound is similar to that described for IDU in that it is phosphorylated preferentially by the TK homologs of vaccinia virus and HSV and that the triphosphate metabolite is a substrate of the DNA polymerases in both viruses. More importantly, these results confirmed that there is significant overlap in the substrate specificities of the orthopoxvirus and herpesvirus TK homologs and validate this approach toward the development of molecules that are effective against both virus families. However, 4′-thioIDU is a much more potent inhibitor of orthopoxvirus replication than of herpesvirus replication and may have its greatest potential against monkeypox or smallpox virus infections. Data from the studies presented here confirmed many of the previous findings with the orthopoxviruses and indicate that this compound warrants further investigation as a therapy for orthopoxvirus infections. There is less enthusiasm for further development of 4′-thioIDU as a potential therapy for herpesvirus infections, but other analogs may have improved activities for these viral infections.
We thank Joshua Harden for technical assistance confirming the DNA sequences of the drug-resistant isolates.
These studies were supported by Public Health Service contracts N01-AI-30049 and N01-AI-15439 and grant 1-U54-AI-057157 from the NIAID, NIH.
Published ahead of print on 21 September 2009.