Antiviral effect of 9-(1,3-dihydroxy-2-propoxymethyl)guanine against cytomegalovirus infection in a guinea pig model

Cytomegalovirus antivirals and development of improved animal models

INTRODUCTION

Cytomegalovirus (CMV) is a ubiquitous pathogen that establishes a lifelong asymptomatic infection in healthy individuals. Infection of immunesuppressed individuals causes serious illness. Transplant and AIDS patients are highly susceptible to CMV leading to life-threatening end-organ disease. Another vulnerable population is the developing fetus in utero, where congenital infection can result in surviving newborns with long-term developmental problems. There is no vaccine licensed for CMV and current antivirals suffer from complications associated with prolonged treatment. These include drug toxicity and emergence of resistant strains. There is an obvious need for new antivirals. Candidate intervention strategies are tested in controlled preclinical animal models but species specificity of human CMV precludes the direct study of the virus in an animal model.

AREAS COVERED

This review explores the current status of CMV antivirals and development of new drugs. This includes the use of animal models and the development of new improved models such as humanized animal CMV and bioluminescent imaging of virus in animals in real time.

EXPERT OPINION

Various new CMV antivirals are in development, some with greater spectrum of activity against other viruses. Although the greatest need is in the setting of transplant patients, there remains an unmet need for a safe antiviral strategy against congenital CMV. This is especially important as an effective CMV vaccine remains an elusive goal. In this regard, greater emphasis should be placed on suitable preclinical animal models and greater collaboration between industry and academia.

Expression of the human cytomegalovirus UL97 gene in a chimeric guinea pig cytomegalovirus (GPCMV) results in viable virus with increased susceptibility to ganciclovir and maribavir

In lieu of a licensed vaccine, antivirals are being considered as an intervention to prevent congenital human cytomegalovirus (HCMV) infection. Ideally, antiviral therapies should undergo pre-clinical evaluation in an animal model prior to human use. Guinea pig cytomegalovirus (GPCMV) is the only small animal model for congenital CMV. However, GPCMV is not susceptible to the most commonly used HCMV antiviral, ganciclovir (GCV), rendering in vivo study of this agent problematic in the guinea pig model. Human cytomegalovirus (HCMV) susceptibility to GCV is linked to the UL97 gene. We hypothesized that GPCMV susceptibility to GCV could be improved by inserting the HCMV (Towne) UL97 gene into the GPCMV genome in place of the homolog, GP97. A chimeric GPCMV (GPCMV::UL97) expressed UL97 protein, and replicated efficiently in cell culture, with kinetics similar to wild-type GPCMV. In contrast, deletion of GP97 resulted in a virus (GPCMVdGP97) that grew poorly in culture. GPCMV::UL97 had substantially improved susceptibility to the inhibitory effects of GCV in comparison to wild-type GPCMV. Additionally, GPCMV::UL97 exhibited improved susceptibility to another antiviral undergoing clinical trials, maribavir (MBV; benzimidazole riboside 1263W94), which also acts through UL97.

Pivotal role of animal models in the development of new therapies for cytomegalovirus infections

Since human cytomegalovirus (CMV) is extremely species specific and does not replicate in experimental animal tissues, animal models for the evaluation of antiviral agents for these infections have utilized surrogate animal viruses including murine CMV, rat CMV and guinea pig CMV. Murine CMV and rat CMV infections in normal and immunocompromised animals provide models of disseminated infection and are ideal for screening of new agents. While guinea pig CMV infection in immunocompromised animals also provides a model for disseminated disease, the model for congenital CMV is unique among all the experimental models. While these models have played a major role in the development of ganciclovir, foscarnet and cidofovir, they do not provide information directly related to human CMV, nor are they useful for evaluation of agents that are active only against human CMV. The SCID-hu mouse models in which human tissue is infected with human CMV has been very useful in the development of new antiviral agents such as maribavir and cyclopropavir. Collectively these experimental CMV infections provide a variety of models representing various aspects of CMV infection in humans that are highly predictive for antiviral efficacy in humans.

The UL97 protein kinase of human cytomegalovirus and homologues in other herpesviruses: impact on virus and host

The human herpesviruses, herpes simplex virus 1 (HSV-1), HSV-2, varicella zoster virus (VZV), Epstein-Barr virus (EBV), human cytomegalovirus (HCMV), human herpesvirus 6A (HHV-6A), HHV-6B, HHV-7 and HHV-8, establish persistent infections with possible recurrence during immunosuppression. HCMV replication is inhibited by the nucleoside analogue ganciclovir (GCV), the compound of choice for the treatment of HCMV diseases and preemptive treatment of infections. The viral UL97 protein (pUL97) which shares homologies with protein kinases and bacterial phosphotransferases is able to monophosphorylate GCV. Homologues of pUL97 are found in HSV (UL13), VZV (ORF47), EBV (BGLF4), HHV-6 (U69), HHV-8 (ORF36) as well as in murine CMV (M97) or rat CMV (R97). Several indolocarbazoles have been reported to be specific inhibitors of pUL97. The protein is important for efficient replication of the virus. Autophosphorylation of pUL97 was observed using different experimental systems. Most recently, it has been shown that pUL97 interacts with the DNA polymerase processivity factor pUL44. Indolocarbazole protein kinase inhibitors are promising lead compounds for the development of more specific inhibitors of HCMV.

Cyclic HPMPC is safe and effective against systemic guinea pig cytomegalovirus infection in immune compromised animals

Cidofovir (HPMPC) is licensed for the treatment of cytomegalovirus (CMV) retinitis in patients with AIDS but its use is limited by nephrotoxicity. We evaluated the safety and efficacy of 1-[((s)-2-hydroxy-2-oxo-1,4,2-dioxaphosphorinan-5-yl)methyl]cytosi ne dihydrate (CHPMPC) the cyclic congener of cidofovir. Treatment was well tolerated both in normal guinea pigs and in animals immune compromised with cyclophosphamide. Further, blood chemistry analysis showed no adverse effects of CHPMPC treatment on kidney or liver function. In efficacy studies in immune compromised guinea pigs challenged with a virulent salivary gland passaged guinea pig CMV, CHPMPC treatment significantly reduced mortality resulting from disseminated virus infection. Quantitative culture showed that treatment also significantly reduced virus replication in the liver and spleen, but not the lungs of infected animals. The efficacy of CHPMPC combined with its improved safety profile appear to make it an attractive alternative to cidofovir for the treatment of herpesvirus infections. Further evaluation is warranted.

Sequence and transcriptional analysis of the guinea pig cytomegalovirus UL97 homolog

Guinea pig cytomegalovirus (GPCMV) displays a similar pathogenesis to human cytomegalovirus (HCMV), and the guinea pig has been used as a model system for testing anti-CMV therapies. However, not all agents active against HCMV share antiviral activity against GPCMV. For example, GPCMV appears resistant to the nucleoside analog, ganciclovir. The molecular basis for this discrepancy in antiviral susceptibility is unknown because to date there has been little analysis of the GPCMV genome. For HCMV, the antiviral effect of ganciclovir depends upon phosphorylation of the drug to its active form. This effect is mediated by the viral UL97 gene product. In order to begin to explore the molecular basis of the resistance of GPCMV to ganciclovir, experiments were undertaken to test whether the GPCMV genome encoded a homolog of the HCMV UL97 gene. Based on the prediction of co-linearity of UL97 homologs within the respective viral genomes, the EcoR I S and F fragments of the GPCMV genome were cloned and partially sequenced. A 1815 base pair open reading frame (ORF) capable of encoding a 604 amino acid (aa) protein was identified spanning portions of the EcoR I S and adjacent EcoR I F genome fragments. Computer-assisted matrix analyses revealed identity between this ORF and the HCMV UL97 gene. ORFs upstream of the GPCMV UL97 gene were identified which shared homology with the HCMV UL95 and 96 genes. Northern blot analyses identified a UL97-specific mRNA of 3.9 kb which was expressed at "early" times post-infection. RNA transcripts of 6.0 and 4.6 kb were identified which corresponded to the UL95 and UL96 homolog coding sequences, respectively. Comparison of the GPCMV UL97 sequence to that of other herpesvirus homologs as well as that of ganciclovir-resistant clinical isolates of HCMV identified nonconservative aa substitutions in two domains involved in catalysis and substrate recognition.

Prophylactic treatment of cytomegalovirus infection with traditional herbs

Hot water extracts of four traditional herbs, Geum japonicum, Syzygium aromaticum, Terminalia chebula and Rhus javanica, which have been shown to have anti-herpes simplex virus (HSV) activity in vivo, were examined for anti-cytomegalovirus (CMV) activity in vitro and in vivo in this study. They inhibited replication of human CMV and murine CMV (MCMV) in vitro. These anti-CMV activities in vivo were examined in an MCMV infection model using immunosuppressed mice. Mice were subcutaneously treated with various doses of cyclosporine, and immunosuppression and MCMV infection were monitored by suppression of antibody production and virus yield in the lung, respectively. Each herbal extract was orally administered to mice treated with 50 mg/kg of cyclosporine from a day before intraperitoneal infection, and the efficacy of herbs was evaluated by the reduction in the virus yield in the lung. Among them Geum japonicum, Syzygium aromaticum, and Terminalia chebula significantly suppressed MCMV yields in lungs of treated mice compared with water treatment. Efficacy of oral treatment with 750 mg/kg per day of Geum japonicum extract was similar to that of the intraperitoneal administration of 2 mg/kg per day of ganciclovir in increasing the body weight of infected mice and reducing the virus yield in the lungs. These herbs may be beneficial for the prophylaxis of CMV diseases in immunocompromised patients.

Combined treatment with 2'-nor-cGMP and ganciclovir against cytomegalovirus infection in a guinea pig model

The combination 2'-nor-cGMP/DHPG at fixed ratios 1:5, 1:10 and 1:20 showed synergistic antiviral effects against GPCMV replication in vitro with CI value < 1. In vivo, a fixed ratio of 1:10 at three different dosage levels of 1.25/12.5 mg, 2.5/25 mg and 5/50 mg/kg/day 2'-nor-cGMP/DHPG combination showed only additive results when compared with each drug alone. However, synergistic antiviral effects were obtained when infected guinea pigs were treated with 2'-nor-cGMP/DHPG combination 2.5/10 mg/kg/day (1:4). A significantly lower GPCMV infectivity titer was noted in the salivary gland, lung and spleen of infected guinea pigs treated with the combination of 2'-nor-cGMP/DHPG 2.5/10 mg/kg/day, as compared to animals treated with a corresponding dose of each drug alone. In addition, GPCMV-infected animals treated with the latter combination showed increased body weight than when either drug was used alone. Histopathologically, each drug alone reduced the viral induced changes in the lung and spleen, but the combination therapy reduced these changes still further. Toxic changes seen in the kidney and bone marrow of infected animals treated with 2'-nor-cGMP, 2.5 mg/kg/day were not significantly increased when DHPG 10 mg/kg/day was added to the regimen. Therefore, combined treatment with 2'-nor-cGMP/DHPG in appropriate concentration is more helpful for acute cytomegalovirus infection in guinea pigs than when either drug was used alone.

Novel method for evaluating antiviral drugs against human cytomegalovirus in mice

A virus-host cell system in which human cytomegalovirus-infected human cells are entrapped in agarose plugs has been developed. This model provides an inexpensive method for the in vivo evaluation (with outbred, immunocompetent mice) of antiviral drugs against human viruses such as cytomegalovirus that replicate primarily or only in human cells.

Combined antiviral effects of paired nucleosides against guinea pig cytomegalovirus replication in vitro

Several promising antiviral nucleosides have been tested in paired combinations against guinea pig cytomegalovirus (GPCMV) replication in guinea pig embryo (GPE) cells by plaque reduction assay; these are [9-(2-hydroxy-1-3-2-dioxaphosphorinan-5-yl)oxymethyl]-guanin e P-oxide (2'nor-cGMP, compound 164), [4-amino-5-bromo-7-(2-hydroxyethoxymethyl)-pyrrolo(2,3-d)pyrimidine] (compound 102), (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine (HPMPC), 9-(1,3-dihydroxy-2-propoxymethyl)guanine (DHPG), 9-(2-hydroxyethoxymethyl)-guanine (acyclovir, ACV) and 3'-azido-3'-deoxythymidine (zidovudine, AZT). Various degrees of interactions were observed; i.e. synergistic reactions were noted in the presence of compound 164/compound 102 and compound 164/DHPG combinations at all concentrations tested. HPMPC/DHPG combinations were synergistic at relatively lower concentrations of DHPG, but became antagonistic as the concentration of DHPG increased. Combinations of compound 164/ACV and DHPG/AZT were antagonistic.

Activity of (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine (HPMPC) against guinea pig cytomegalovirus infection in cultured cells and in guinea pigs

(S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine, HPMPC, and two HPMPC-related nucleoside analogs, (S)-9-(3-hydroxy-2-phosphonylmethoxypropyl)adenine, HPMPA, and (2-phosphonylmethoxyethyl)guanine, PMEG, were evaluated for their antiviral activities against guinea pig cytomegalovirus (GPCMV) infection in guinea pig embryo (GPE) cells and human cytomegalovirus (HCMV) infection in human diploid fibroblast (MRC-5) cells. DHPG, 9-(1,3-dihydroxy-2-propoxymethyl)guanine, was used for comparison. The antiviral activity of HPMPC against GPCMV infection in vivo and its toxicity to Hartley guinea pigs were also evaluated. The 50% antiviral effective doses (ED50) of HPMPC, HPMPA, PMEG and DHPG against GPCMV infection in GPE cells were 0.22, 1.4, 0.07 and 62 microM, respectively; and against HCMV infection in MRC-5 cells, the ED50s were 0.51, 0.72, 0.01 and 17.5 microM, respectively. Their cytotoxic doses (CyD50) in GPE replicating cells were 84, 35, 1.4 and 700 microM, respectively and in MRC-5 cells were approximately 114, 31, 0.86 and 750 microM, respectively. Based on their calculated therapeutic indexes, HPMPC was the most potent and selective of the four compounds tested. In vivo, during acute infection, the spleen indexes of all infected animals that were treated with 1.25 to 5.0 mg/kg/day of HPMPC for 5 days were significantly reduced as compared with sham-treated animals. Virus infectivity titers in blood and various tissues of infected animals treated with HPMPC, 2.5 or 1.25 mg/kg/day were not significantly lower than those of the infected, sham-treated animals; with 5 mg/kg/day, infectivity titers in the blood, spleen, and salivary gland were significantly lower in HPMPC-treated than in sham-treated animals. However, HPMPC was toxic to guinea pigs especially at doses of 5 to 10 mg/kg/day. These data showed that HPMPC was highly active and selective in cultured guinea pig cells and human fibroblast cells against CMV infection but did not effectively inhibit GPCMV infection in guinea pigs at minimum toxic concentrations.

Ganciclovir. A review of its antiviral activity, pharmacokinetic properties and therapeutic efficacy in cytomegalovirus infections

Ganciclovir is a nucleoside analogue with antiviral activity in vitro against members of the herpes group and some other DNA viruses. It has demonstrated efficacy against human cytomegalovirus infections and should be considered a first-line therapy in the treatment of life- or sight-threatening cytomegalovirus infection in immunocompromised patients. Clinical efficacy varies with the underlying aetiology of immunocompromise and the site of disease, and prompt diagnosis and early treatment initiation appear to improve the response. In patients with cytomegalovirus pneumonia, particularly bone marrow transplant recipients, concomitant administration of cytomegalovirus immune globulin may significantly improve clinical outcome. Maintenance therapy to prevent recurrence is usually required by bone marrow transplant recipients until the recovery of adequate immune function, whereas AIDS patients may require indefinite ganciclovir maintenance therapy to prevent disease progression, as ganciclovir (like other antivirals) does not eradicate latent viral infection. Haematological effects occur relatively frequently during ganciclovir administration but are usually reversible. Ganciclovir has not been directly compared with other antiviral drugs because of the absence until recently of other effective treatments. However, comparative studies with foscarnet, particularly in cytomegalovirus retinitis, will be of considerable interest. Thus, ganciclovir represents a major advance in the therapy of severe cytomegalovirus infections in immunocompromised patients. Comparative studies, and investigation of ways of reducing toxicity (intravitreal administration; concomitant use of stimulants of haematopoiesis; use in conjunction with other antivirals with differing mechanisms of action), may further expand its eventual role.

Evaluation of new antiviral agents: II. The use of animal models

Antiviral chemotherapy has become a reality in the 1980s. Since the use of animal models in the testing of new antiviral agents is an inevitable step prior to clinical trial in human patients, it is important to understand the basic principles of using model systems. Briefly reviewed in this paper are the heterologous and homologous animal models which have been used for studies of various herpesvirus infections in humans. Discussions of the use of the guinea pig models mainly, for members of the Herpesviridae are presented in more detail. Precautions needed for the development of new animal models, and suggestions proposed for the use of animal models for testing new antiviral agents are outlined. It is hoped that new animal models will be developed in the foreseeable future for evaluating the much needed effective but less toxic antiviral agents for a variety of human viral diseases.

Cytomegalovirus infection in immunocompromised guinea pigs: a model for testing antiviral agents in vivo

An experimental model for testing antiviral agents against severe cytomegalovirus (CMV) infection in immunocompromised hosts was developed. The model consisted of cyclophosphamide (Cy) treatment of CMV-infected guinea pigs to simulate CMV infection in immunodeficient individuals. Of the 3 Cy regimens tested, a single 300 mg/kg dose administered one day after virus inoculation resulted in the most severe CMV infection considering mortality rates, mean day of death and loss of body weight. Evaluation of responses to both T and B cell mitogens suggested that the severe and lethal CMV infection resulted from the combined immunosuppressive effect of Cy and CMV. The nucleoside analog [9-(1-3-dihydroxy-2-propoxymethyl)guanine (DHPG) was used to assess the usefulness of the CMV-infected immunocompromised host model. DHPG (100 mg/kg/day for 8 days) prevented death but did not reduce virus infectivity titers in blood of Cy-treated, CMV-infected guinea pigs. This model of CMV infection in immunocompromised guinea pig is a relevant and convenient experimental tool for the assessment of candidate anti-CMV agents under well-defined experimental conditions, such as appropriate CMV inoculum and Cy regimen.

Activities of two new antiviral agents against guinea pig lymphotropic herpesvirus infection in vitro

Two new antiviral agents, compound 164, also known as 2'-nor-cGMP (9-[(2-hydroxy-1,3,2-dioxaphosphorinan-5-yl)oxymethyl]-guani ne P-oxide), and compound 102 [4-amino-5-bromo-7-(2-hydroxyethoxymethyl)-pyrrolo(2,3-d)pyrimidine], together with acyclovir for comparison, were evaluated for activities against the guinea pig lymphotropic herpesvirus infection in vitro by plaque reduction and virus yield reduction assays in guinea pig embryo cells. The two new compounds were demonstrated to be more potent against guinea pig lymphotropic herpesvirus infections than acyclovir. Compound 164 was the most potent of the three; drug concentrations required to reduce the number of plaques by 50% were 2, 35.5, and 144.5 microM for compounds 164, 102, and acyclovir, respectively. The two new compounds were cytostatic but not cytotoxic to guinea pig embryo cells in cultures. Attempts were made to investigate the inhibition of viral replication by these compounds, and the influence of test conditions on antiviral evaluations is discussed.

Effect of 2'-nor-cyclic GMP against guinea pig cytomegalovirus infection

Cyclic phosphate derivative of DHPG, 2'-nor-cGMP [9-[(2-hydroxy-1,3,2-dioxaphosphorinan-5-yl)oxymethyl]-guani ne phosphate-oxide] was evaluated for activity against guinea pig cytomegalovirus (GPCMV) infection in cultured guinea pig embryo cells and in guinea pigs. By virus yield reduction and plaque reduction assays, 2'-nor-cGMP was demonstrated to be 15- to 20-fold more potent against GPCMV infection than its parental drug DHPG. The selectivity index of 2-nor-cGMP was 110, which was 10-fold higher than that of DHPG. In cultured cells, 2'-nor-cGMP attained maximal antiviral activity when added to the cells within 12 h postinfection. In the studies on GPCMV infection in guinea pigs, 2'-nor-cGMP administered subcutaneously once daily (5 mg/kg per day) for 8 days, starting 24 after virus inoculation, significantly suppressed GPCMV infectivity titers in the blood, spleen, lung, and salivary gland during acute infection (10 days postinfection) as compared with sham-treated infected animals. A greater reduction of GPCMV infectivity titers in the salivary gland was noted during chronic infection (i.e., 24 days postinfection). Clinically, splenomegaly and peripheral lymphocytosis were significantly modified as compared with the sham-treated animals (P less than 0.05). The drug, administered at this dosage, was reasonably tolerated by the guinea pigs and showed clinical benefit.

Evaluation of new antiviral agents: I. In vitro perspectives

A variety of different methods for the evaluation of antiviral agents in cell culture systems are briefly reviewed. It has been repeatedly noted that many test conditions such as the cell culture system, virus strain, virus challenge dose, virus input multiplicity of infection, and time of harvesting, etc., can substantially affect or even alter the test results, thus making comparative studies and unambiguous evaluations very difficult. Attempts are made to discuss previous test methods together with our recent studies with the aim to simplify test procedures and assay methods. Suggestions are proposed for in vitro evaluation of new antiviral agents. It is hoped that this review will alarm investigators to the problems of assaying new antiviral agents. If the suggestions made in this review can be followed, the screening of the enormous number of promising antiviral compounds may be made more efficiently in the near future.

Cytomegalovirus infection in orthotopic liver transplantation

We retrospectively studied 175 orthotopic liver transplants in 151 patients. Of the 151 patients, 59 (39.1%) were diagnosed as having cytomegalovirus (CMV) infection. The rate of infection in patients treated for rejection was 48.8% as compared to 26.2% in patients without rejection (P less than 0.01). Antirejection therapy was associated with culture-positive cases in 33 out of 43 patients as compared to 9 out of 16 patients who had CMV antibody titer elevations. Patients were treated with gancyclovir if they had simultaneous positive cultures from multiple sites and were seriously ill. Eighteen of the 19 patients thus treated had side effects, one of which was serious (bone marrow hypoplasia). Cultures became negative in 15 out of 17 (88%) of the surviving patients. Patient survival was similar to our overall survival rate of 87%.

Ganciclovir prophylaxis for cochlear pathophysiology during experimental guinea pig cytomegalovirus labyrinthitis

The effectiveness of the antiviral agent ganciclovir (9-[1,3-dihydroxy-2-propoxymethyl]guanine) against guinea pig cytomegalovirus was tested in vitro in guinea pig embryonic fibroblasts and in vivo in an experimental guinea pig cytomegalovirus labyrinthitis model. In vitro, ganciclovir completely prevented guinea pig cytomegalovirus infection of guinea pig embryonic fibroblasts at concentrations above 32.6 micrograms/ml. In vivo, antibody-negative animals had an average 17-dB elevation in their auditory nerve compound action potential thresholds (P less than 0.01, t test) and showed signs bilaterally of guinea pig cytomegalovirus labyrinthitis 8 days after intrathecal inoculation of virus. Ganciclovir administration starting 1 day before inoculation prevented the development of both cochlear histopathologic change and hearing loss. Guinea pig cytomegalovirus meningitis was observed in both the drug-treated and untreated groups. High-pressure liquid chromatography confirmed the presence of ganciclovir in the serum, perilymph, and cerebrospinal fluid of the drug recipients. Prophylactic ganciclovir thus can protect the cochlea from the histopathologic changes and hearing loss normally associated with experimental guinea pig cytomegalovirus labyrinthitis.

Efficacy of S26308 against guinea pig cytomegalovirus infection

Prophylactic use of antiviral agents against cytomegalovirus (CMV) is particularly indicated for the immunocompromised host because morbidity and mortality due to CMV occur most frequently following immunosuppression. We have evaluated the new Riker compound S26308 for its therapeutic and prophylactic antiviral activity against CMV in guinea pigs. The efficacy of the compound was assessed in vitro in guinea pig embryo cells and in vivo in both immunocompetent and immunocompromised guinea pigs. Guinea pig CMV plaque formation was reduced only in cells treated with S26308 prior to virus infection. The antiviral activity remained even when the compound was removed after virus absorption and was due to neither virus destruction nor inhibition of cell growth. The frequency of viremia was reduced in guinea pigs for which S26308 therapy was initiated 24 h prior to virus inoculation compared with sham-treated animals. This reduction in the frequency of viremia did not prevent virus spread to target tissues but did result in a reduction of the severity of CMV-induced disease in immunocompromised guinea pigs. Low levels of interferon were detected in supernatants of S26308-treated cells, and interferon was detected in the serum of guinea pigs given S26308. These results indicate that S26308 can induce interferon and reduce CMV infectivity in vivo and in vitro when used prophylactically. This antiviral activity, although modest, was accompanied by beneficial effects on CMV-induced morbidity and mortality. Prophylactic use of S26308 in combination with other therapeutic agents may be a useful strategy against CMV infections.