Filociclovir Is a Potent In Vitro and In Vivo Inhibitor of Human Adenoviruses

Human adenovirus (HAdV) infection is common in the general population and can cause a range of clinical manifestations, among which pneumonia and keratoconjunctivitis are the most common. Although HAdV infections are mostly self-limiting, infections in immunocompromised individuals can be severe. No antiviral drug has been approved for treating adenoviruses. Filociclovir (FCV) is a nucleoside analogue which has successfully completed phase I human clinical safety studies and is now being developed for treatment of human cytomegalovirus (HCMV)-related disease in immunocompromised patients.

ABSTRACT

Human adenovirus (HAdV) infection is common in the general population and can cause a range of clinical manifestations, among which pneumonia and keratoconjunctivitis are the most common. Although HAdV infections are mostly self-limiting, infections in immunocompromised individuals can be severe. No antiviral drug has been approved for treating adenoviruses. Filociclovir (FCV) is a nucleoside analogue which has successfully completed phase I human clinical safety studies and is now being developed for treatment of human cytomegalovirus (HCMV)-related disease in immunocompromised patients. In this report, we show that FCV is a potent broad-spectrum inhibitor of HAdV types 4 to 8, with 50% effective concentrations (EC₅₀s) ranging between 1.24 and 3.6 μM and a 50% cytotoxic concentration (CC₅₀) of 100 to 150 μM in human foreskin fibroblasts (HFFs). We also show that the prophylactic oral administration of FCV (10 mg/kg of body weight) 1 day prior to virus challenge and then daily for 14 days to immunosuppressed Syrian hamsters infected intravenously with HAdV6 was sufficient to prevent morbidity and mortality. FCV also mitigated tissue damage and inhibited virus replication in the liver. The 10-mg/kg dose had similar effects even when the treatment was started on day 4 after virus challenge. Furthermore, FCV administered at the same dose after intranasal challenge with HAdV6 partially mitigated body weight loss but significantly reduced pathology and virus replication in the lung. These findings suggest that FCV could potentially be developed as a pan-adenoviral inhibitor.

INTRODUCTION

Human adenoviruses (HAdVs) are ubiquitous DNA viruses that are most commonly associated with respiratory, gastrointestinal, and eye infections in children and adults (1). Although most adenoviral infections are mild and self-limiting, disseminated forms of the disease occur in immunocompromised individuals, particularly transplant and HIV patients (2–4), resulting in high rates of fatalities. Transmission of HAdVs occurs via inhalation of infected aerosol droplets, fecal-oral transmission, and contact with contaminated fomites (5). Reactivation of latent virus after immunosuppression may also occur (6). The virus is stable in the environment, and therefore, epidemics can spread easily in closed or crowded populations, such as school children, college students, and military recruits (7, 8). More than 100 types of adenoviruses have been identified and are classified as 7 distinct species (A to G) based on immunological and genomic criteria (9). HAdVs are the most common causes of epidemic keratoconjunctivitis worldwide, accounting for up to 75% of all cases (10). Adenoviral pneumonia develops in up to 20% of newborns and infants (11, 12). HAdVs also account for >50% of pneumonia cases among unvaccinated military recruits (13). In immunocompromised individuals, dissemination and/or severe respiratory failure develops in 10 to 30% of cases, and fatality rates may exceed 50% (14, 15). A recent U.S. outbreak in New Jersey and Maryland resulted in 98 illnesses and 12 deaths (16).

No antiviral drug has been approved for treating adenoviruses. Previous reports have shown anti-adenoviral activity of a number of nucleoside analogues in cell culture and animal models, including ribavirin (RBV), ganciclovir (GCV), cidofovir (CDV), and its prodrug brincidofovir (BCV). RBV is active only against type C HAdVs and was not shown to significantly reduce viral loads or mortality in treated patients (17, 18). GCV has a modest in vitro activity against HAdVs, but its clinical efficacy remains unclear (15, 19). CDV was extensively tested in the hamster model, where it sCDV was extensively tested in thehowed moderate efficacy against HAdV5 and HAdV6 (20–22). CDV is prescribed in the clinic on an off-label basis and was recently shown to act as a competitive inhibitor of HAdV5 polymerase (23). However, controlled trials have not been completed (24–26). Furthermore, the clinical use of CDV has been associated with severe nephrotoxicity (27, 28). Brincidofovir, a lipid-linked derivative of CDV, is orally bioavailable and highly potent in vitro (29, 30). BCV has been used in the clinic on a compassionate-use basis and several studies described favorable clinical outcomes in transplant patients, although some adverse effects necessitating treatment discontinuation were also reported (31–36). In a recent randomized placebo-controlled phase II trial, the preemptive treatment of HAdV infections with BCV was shown to rapidly reduce viral loads; however, gastrointestinal (GI) toxicity prompted early termination of the study (37). Consequently, there is an urgent and unmet medical need for a safe and effective anti-HAdV drug.

Filociclovir (FCV; also known as cyclopropavir or MBX-400) is a methylenecyclopropane nucleoside analog with broad-spectrum activity against herpesviruses, including human cytomegalovirus (HCMV), varicella-zoster virus (VZV), Epstein-Barr virus (EBV), human herpesvirus 6A (HHV-6A), HHV-6B, HHV-7, and HHV-8 (38–40). It has successfully completed human phase I safety studies (38, 41, 42) and is now entering phase II human clinical efficacy studies for the treatment of HCMV-related disease in organ transplant patients. Recently, FCV was shown to be a potent inhibitor of HAdV5 in cell culture experiments (19). In this report, we demonstrate a broad-spectrum in vitro antiadenoviral activity for FCV and show that it is potent against both the systemic and respiratory forms of adenoviral disease in the Syrian hamster model (43). Since no drugs are currently approved for the treatment of adenovirus-associated diseases, we believe that FCV could potentially be a first-in-class adenovirus inhibitor.

RESULTS

FCV is a broad-spectrum HAdV inhibitor in cell culture.

To examine the spectrum of activity of FCV against adenoviral types most commonly associated with human disease, we carried out multiple dose-response cellular infection assays. Our results show that FCV is a potent inhibitor of 5 HAdV types, 4 to 8, with 50% effective concentrations (EC₅₀s) ranging from 1.24 to 3.6 μM and a high selectivity index (Fig. 1 and Table 1). This level of potency is similar to that of CDV (Table 1) and far superior to that of GCV (data not shown), which has a reported EC₅₀ of 66 μM against HAdV5 (19) and 39 μM for HAdV2 (44). The range of HAdV types tested so far suggests that FCV could potentially be developed as a pan-adenoviral inhibitor.

FIG 1.

Representative dose-response curves showing the cytotoxicity and potency of FCV and CDV against HAdV6. Human A549 cells were plated at 6 × 10³ per well 1 day prior to infection and infected with HAdV6 at an MOI of either 0.5 PFU/cell (FCV) (A) or 1 PFU/cell (CDV) (B). FCV or CDV dilutions were added to the cell plates immediately before HAdV6 infection. At 6 days postinfection, plates were stained and the results were graphed in GraphPad Prism. Values are means ± standard deviations (SD).

TABLE 1.

Spectrum of FCV activity against HAdVs^a

HAdV type	FCV		CDV EC50 (μM)
	EC50 (μM)b	CC50 (μM)
HAdV4	2.34 ± 0.16	>150	1.9
HAdV5	2.2 ± 0.4	>100	4
HAdV6	3.6 ± 0.17	>100	28.46
HAdV7	1.24 ± 0.38	>150	2.1
HAdV8	2.49 ± 0.83	>150	3.8

^aCDV was used as a positive control. EC₅₀ and CC₅₀ values were determined from the linear portion of a full dose-response curve using GraphPad Prism. HAdV4, -5, -7, and -8 were tested in HFFs at an MOI of 0.005 PFU/cell, while HAdV6 was tested in A549 cells at MOI of 0.5 PFU/cell (for FCV) and 1 PFU/cell (for CDV).

^bValues are average EC₅₀s and SD from two or more independent replicates within the same assay conditions.

FCV inhibits late stages of HAdV replication.

Immunostaining for the adenoviral DNA-binding protein (DBP) and hexon protein was used as an indicator for the stages of infection (45). DBP associates with replication centers as the infection progresses. The DBP-stained replication centers initially appear as small fluorescent dots in infected cells (each resulting from one incoming adenoviral genome), and then they expand as adenoviral DNA replication progresses. DBP staining is antinuclear and uniform during the early stage of adenoviral infection (prior to DNA replication) and punctate in the late phase of infection. Adenoviral hexon is the most abundant component of the viral capsid and is not expressed until after DNA replication takes place. Therefore, it is considered a late adenoviral protein. FCV was very effective in interfering with the late stages of the HAdV5 or HAdV6 replication cycle (Fig. 2A and B, respectively), as indicated by the lack of hexon staining. No hexon-positive cells were detected at 40 μM FCV, and few hexon-positive cells at 10 μM FCV. However, some hexon-positive cells were detected at 4 μM FCV, which showed substantially less advanced infection, as indicated by the DBP staining patterns and the size and shape of nuclei. No distinguishable differences in the potency against HAdV5 and HAdV6 were detected.

FIG 2.

Immunofluorescence of FCV-treated HAdV5 or HAdV6 infections in cultured A549 cells. Human A549 cells cultured on glass coverslips in 6-well plates were infected at 5 PFU/cell with either HAdV5 (A) or HAdV6 (B) or were mock infected. After 1 h, FCV was added to a final concentration of 40, 10, or 4 μM. A set of untreated wells were kept as controls. At 27 h postinfection, the A549 cells were fixed in paraformaldehyde (3.7% in PBS) and permeabilized with methanol. Cells were then stained for the adenovirus DNA-binding protein (DBP; green staining) and adenovirus hexon (red staining). The same field is shown for the DBP and hexon staining, and two different fields of view are shown for each condition.

Prophylactic FCV administration prevented mortality and mitigated morbidity, liver pathology, and infection induced by HAdV6.

FCV was previously shown to be orally available and effective in a murine model of cytomegalovirus infection at 10 mg/kg of body weight (46). In this study (Fig. 3A), two experiments were conducted to test the potency of oral FCV against HAdV6 over a range of doses (1 to 30 mg/kg) in the immunosuppressed Syrian hamster model (43). In both experiments, intraperitoneal CDV was used as the positive control at a dosing regimen consisting of one dose of 37 mg/kg, followed by 20-mg/kg doses 3 times weekly. In the first experiment, the 10- and 30-mg/kg doses of FCV, starting 1 day prior to infection and continuing daily, completely prevented mortality in HAdV6-infected hamsters (Fig. 3B). Initially, all animals in the infected groups lost body weight, but this effect was reversed by treatment with 10 or 30 mg/kg FCV or CDV (Fig. 3C). There was 50% mortality in the infected, untreated group; the surviving hamsters in this group recovered slowly, and by the end of the study, their body weight matched that of the hamsters in the CDV-treated group. The weight gain of hamsters in the CDV-treated group was slightly lower than that of the FCV-treated groups. At day 5 after adenovirus challenge, serum was collected from 5 animals in each group and analyzed for alanine aminotransferase (ALT) levels to assess liver damage. Four hamsters in the untreated group had high ALT levels, while none of the FCV- or CDV-treated animals had elevated serum ALT levels (Fig. 3D). Furthermore, treatment with 30 mg/kg of FCV suppressed HAdV6 replication in the liver to nondetectable levels (Fig. 3E). One animal in the 10-mg/kg FCV-treated group exhibited a very low liver virus burden, but we could not detect HAdV6 in the livers of any of the other animals in the same group. Similarly, in the CDV-treated group, 4 hamsters had undetectable or unquantifiable levels of infectious virus in the liver, and one animal had a low virus burden. Overall, the 10-mg/kg dose of FCV was potent but slightly less efficacious than the 30-mg/kg dose.

FIG 3.

Prophylactic oral administration of FCV at 30 and 10 mg/kg to HAdV6-infected hamsters (i.v.). (A) Experimental protocol. All hamsters were immunosuppressed using cyclophosphamide (CP). The hamsters were divided into 5 groups (15 each), immunosuppressed, and then injected intravenously (i.v.) with vehicle (group 1) or 2 × 10¹⁰ PFU/kg of HAdV6 (groups 2 to 5). Groups 1 and 2 received drug vehicle (once-daily oral dose). Groups 3 and 4 received FCV at doses of 10 mg/kg and 30 mg/kg (once-daily oral dose), respectively. Finally, group 5 received cidofovir intraperitoneally (one dose of 37 mg/kg, followed by 20-mg/kg doses 3 times weekly). For all groups, drug administration started 1 day before challenge and continued for the duration of the study. (B) Survival. For HAdV6 + Vehicle versus HAdV6 + FCV 10 or 30 mg/kg, P was <0.05 (log rank). (C) Mean body weight changes. Group means and standard errors of the means are shown. For HAdV6 + Vehicle versus HAdV6 + FCV 10 or 30 mg/kg, P was <0.0001 (two-way ANOVA); for HAdV6 + FCV 10 mg/kg versus Vehicle or HAdV6 + FCV 30 mg/kg, P was <0.05. (D) Serum alanine aminotransferase (ALT) levels at day 5 postchallenge. The symbols indicate values from individual animals; the horizontal bar represents the geometric mean. (E) Virus burden in the liver at day 5 postchallenge. The symbols indicate values from individual animals; the horizontal bar represents the geometric mean. NQ, not quantifiable; ND, not detectable.

In a follow-up experiment aimed at determining the lowest active dose of FCV, we tested a lower dose range in 3 treatment groups, starting with a dose of 10 mg/kg and then going down to 3 and to 1 mg/kg, to define the minimum effective dose. The 10- and 3-mg/kg dosages of FCV prevented mortality of HAdV6-infected hamsters, but the 1-mg/kg dose did not prevent mortality (Fig. 4A). Initially, all infected hamsters lost weight, but this effect was completely reversed by treatment with CDV or 10 mg/kg of FCV. The 3-mg/kg dose of FCV was partially efficacious in preventing weight loss, but the 1-mg/kg dose was ineffective (Fig. 4B). Both CDV and FCV at 10 mg/kg completely prevented HAdV6-induced liver damage, as judged by serum ALT levels (Fig. 4C). Treatment with FCV at 3 mg/kg resulted in a moderate yet significant (P < 0.05) effect on reducing ALT levels. However, FCV was ineffective in mitigating liver pathology at the 1-mg/kg dose. CDV and FCV at the 10-mg/kg dose inhibited HAdV6 replication in the liver (Fig. 4D). The 3-mg/kg dose of FCV had a moderate effect, while the 1-mg/kg dose was ineffective in inhibiting virus replication in the liver. These data suggest that the daily 10-mg/kg of FCV is the dose necessary for significant activity in this model.

FIG 4.

Prophylactic oral administration of FCV at 10, 3, and 1 mg/kg to HAdV6-infected hamsters (i.v.). The hamsters were divided into 6 groups (15 each), immunosuppressed, and then injected i.v. with vehicle (group 1) or 4 × 10¹⁰ PFU/kg of HAdV6 (groups 2 to 6). Groups 1 and 2 received drug vehicle. Group 3 to 5 received FCV at doses of 10 mg/kg, 3 mg/kg, and 1 mg/kg (once-daily oral dose), respectively. Finally, group 6 received cidofovir (one dose of 37 mg/kg, followed by 20-mg/kg doses 3 times weekly). For all groups, drug administration started 1 day before challenge and continued for the duration of the study. (A) Survival. FCV completely prevented mortality at the 3- and 10-mg/kg daily oral doses and had a moderate yet significant (P < 0.05) effect at 1 mg/kg. For HAdV6 + Vehicle versus HAdV6 + FCV at 3 mg/kg, HAdV6 + Vehicle versus HAdV6 + FCV 10 mg/kg, and HAdV6 + Vehicle versus HAdV6 + CDV, P was <0.0001 (log rank). (B) Mean body weight changes. Apart from a slight initial (from day −1 to day 5) weight loss, FCV prevented body weight loss at 10 mg/kg, was moderately efficacious at 3 mg/kg, and was not efficacious at 1 mg/kg. The symbols represent group means; the error bars indicate standard deviations. Body weight data for a given group are shown only up to time points at which unscheduled deaths occurred. (C) Serum alanine aminotransferase (ALT) levels at day 5 postchallenge. FCV mitigated the pathology caused by intravenous injection of HAdV6 at the 3- and 10-mg/kg daily doses; however, it was ineffective at 1 mg/kg. The symbols represent the data collected from individual animals; the horizontal bar indicates the group mean. The empty symbols for HAdV6 + Vehicle and HAdV6 + FCV 1 mg/kg indicate that the samples were collected from animals sacrificed ahead of schedule. *, P < 0.05; ***, P < 0.001 (one-tailed Mann-Whitney U test; HAdV6 + Vehicle and HAdV6 + FCV 1 mg/kg). (D) Virus burden in the liver at day 5 postchallenge. The symbols represent the data for individual animals; the horizontal bar indicates the group mean. The empty symbols for HAdV6 + Vehicle and HAdV6 + FCV 1 mg/kg indicate that the samples were collected from animals sacrificed ahead of schedule. NQ, not quantifiable; ND, not detectable.

Delayed FCV administration prevented mortality and mitigated morbidity, liver pathology, and infection induced by HAdV6.

Since the 10-mg/kg dose of FCV was shown to be effective prophylactically, we decided to evaluate it in a delayed-administration experiment of therapeutic efficacy. FCV was administered once daily at 10 mg/kg, starting at 1 day before (day −1) or at 1, 2, 3, or 4 days after (day +1, +2, +3, or +4) virus infection (Fig. 5A). All hamsters survived except one in the day +4 group (Fig. 5B). No body weight loss was observed for the hamsters in the day −1 and day +1 groups (Fig. 5C). Hamsters in the day +2, day +3, and day +4 groups lost weight initially and then recovered at 4, 5, and 7 days postchallenge, respectively, after which all animals gained weight similarly to uninfected hamsters (Fig. 5C). FCV completely prevented HAdV6-induced liver damage when administered starting 1 day before or 1 day after virus challenge, and it significantly mitigated it when started at day +2, +3, or +4 postchallenge (Fig. 5D). Furthermore, FCV inhibited HAdV6 replication in the liver when administration started at day +3 postchallenge or earlier (Fig. 5E). It was also very effective even when started at day +4 postchallenge; 4 of 5 animals in that group had no detectable or no quantifiable virus in their livers (Fig. 5E). Overall, the 10-mg/kg dose of FCV prevented virus replication and mitigated HAdV6-induced pathology in immunosuppressed hamsters when the treatment started at day +3 after virus challenge or earlier and provided detectable protection even when the administration started at day +4 postchallenge.

FIG 5.

Delayed oral administration of FCV at 10 mg/kg to HAdV6-infected hamsters (i.v.). (A) Experimental protocol. Since the 10-mg/kg daily dose was shown to be effective prophylactically in the previous 2 experiments, it was used in this delayed-administration experiment. The hamsters were divided into 7 groups (15 each), immunosuppressed, and then injected i.v. with vehicle or 4 × 10¹⁰ PFU of HAdV6 per kg. FCV was administered orally at 10 mg/kg once daily, starting at day −1, +1, +2, +3 or +4 relative to virus injection. An infected group that did not receive FCV and another that received vehicle only served as controls. (B) Survival. For HAdV6 + Vehicle versus HAdV6 + FCV day −1, HAdV6 + FCV day +1, HAdV6 + FCV day +2, or HAdV6 + FCV day +3, P was <0.001, and for HAdV6 + Vehicle versus HAdV6 + FCV day +4, P was <0.01 (log rank test). (C) Body weight changes. The symbols represent group means; the error bars indicate the standard errors of the means. (D) Serum alanine aminotransferase (ALT) levels at day 5 postchallenge. For HAdV6 + Vehicle versus HAdV6 + FCV day −1 or HAdV6 + FCV day +1, P was <0.01; for HAdV6 + Vehicle versus HAdV6 + FCV day +2, P was <0.01; for HAdV6 + Vehicle versus HAdV6 + FCV day +3, P was <0.01; for HAdV6 + FCV versus HAdV6 + FCV day +4, P was 0.15 (two-tailed Mann-Whitney test). (E) Virus burden in the liver at day 5 postchallenge. Symbols represent data from individual animals; the horizontal bar represents the mean. In panels D and E, the empty symbols for HAdV6 + Vehicle and HAdV6 + FCV D + 4 indicate that the samples were collected from animals sacrificed ahead of schedule. NQ, not quantifiable; ND, not detectable.

Prophylactic FCV administration reduced pathology and virus replication in the lung after intranasal challenge with HAdV6.

After showing potency in a model of adenoviral systemic infection, we further evaluated the efficacy of prophylactic FCV administration in a respiratory model of HAdV6 infection. We used the same daily dose (10 mg/kg) that was efficacious in the systemic model. As a positive control, we included a group of HAdV6-infected hamsters treated with a known effective dose (200 mg/kg, oral) of valganciclovir (VGCV) twice daily. There were no treatment-related deaths in this study. Hamsters in all the HAdV6-infected groups lost weight from the onset of the study. However, treatment with VGCV and, to a lesser but still significant (P < 0.01) effect, FCV partially counteracted this effect (Fig. 6A). Although FCV treatment tended to be less efficacious than VGCV in mitigating body weight loss, the 30-mg/kg dose of FCV reduced virus replication in the lung to a greater extent than VGCV (Fig. 6B). Also, at 7 days postchallenge, the pathology in the lungs of infected hamsters treated with FCV was significantly lower than that of untreated animals (Fig. 6C). Further, the weight of the left lung lobe of FCV-treated hamsters was also significantly less than that of the untreated animals and, by day 7 postchallenge, lower than that of the VGCV-treated hamsters (Fig. 6D). At this stage, lung weight mostly reflects the degree of hepatization of the lung.

FIG 6.

Prophylactic oral administration of FCV at 10 mg/kg to HAdV6-infected hamsters (intranasal). Hamsters were divided into 5 groups (15 each). Group 1 received vehicle; groups 2 to 5 received 1.5 × 10¹⁰ PFU of HAdV6 per kg intranasally. In addition to HAdV6, groups 3 and 4 received 10-mg/kg and 30-mg/kg oral daily doses of FCV, respectively, starting at day −1, while group 5 received 200 mg/kg VGCV orally twice daily starting 12 h before virus challenge. Drugs were administered for the duration of the study. All hamsters were observed and weighed daily. There was an intermediate time point (day 3) at which 5 hamsters were bled out and necropsied, and lung samples were collected for quantifying the infectious virus burden by TCID₅₀ assay. The remaining 10 hamsters were monitored for survival and body weight changes. All hamsters were sacrificed after 7 days. (A) Body weight changes. FCV partially mitigated HAdV6-induced morbidity. Each symbol represents the group mean; the error bars depict the standard error of the mean. For HAdV6 + Vehicle versus HAdV6 + FCV 10 mg/kg, P was <0.0001; for HAdV6 + Vehicle versus HAdV6 + FCV 30 mg/kg, P was <0.0001; for HAdV6 + Vehicle versus HAdV6 + VGCV, P was <0.0001 (two-way ANOVA). (B) Virus burden in the lung at day 3 postchallenge. FCV inhibits HAdV6 replication in the lungs of intranasally infected hamsters; symbols represent data from individual animals. (C) Gross lung pathology scores. 0, no lesions; 1, minimal; 2, moderate; 3, marked. (D) Weight of the left lung lobe. Symbols represent data from individual animals. The bars represent the group mean, and the error bars show standard errors. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (one-way ANOVA followed by Sidak's multiple-comparison test).

DISCUSSION

In this report, we provide evidence that filociclovir (FCV), a nucleoside analogue that successfully completed phase I safety studies in humans (41, 42), is a broad-spectrum inhibitor of human adenoviruses (HAdVs) in vitro and is highly potent in vivo in the Syrian hamster model. No antiviral drug has been approved for treating HAdVs, and there is an urgent and unmet medical need for a safe and effective anti-HAdV drug. In this report, we show that FCV is a potent and selective inhibitor of 5 HAdV types (types 4 to 8), with EC₅₀s ranging from 1.24 to 3.6 μM (Table 1). In addition to HAdVs, FCV is highly potent against most herpesviruses, including HCMV, VZV, EBV, HHV-6A, HHV-6B, HHV-7, and HHV-8 (38–40). FCV has been previously shown to be effective in reducing mortality and viral titers in BALB/c mice infected with murine cytomegalovirus (MCMV) and in human tissues implanted into severe combined immunodeficient (SCID) mice and infected with HCMV (46). Based on these excellent in vitro and in vivo profiles, FCV was advanced to clinical testing in humans. Phases IA and IB were conducted; no serious adverse effects were observed, and oral dosages as low as 100 mg daily achieved plasma concentrations that are sufficient to inhibit HAdVs in vitro (41, 42), suggesting that FCV is an excellent clinical candidate for phase II testing in humans.

To examine the in vivo potency of FCV against HAdV, we used the immunosuppressed Syrian hamster model, which has been used previously to evaluate several antiviral compounds (20, 22, 45, 47, 48). In this model, after intravenous injection of the virus, it replicates in most organs, with liver as a main target (43, 49). Alternatively, when the virus is instilled intranasally, it replicates in the lungs, especially in the bronchial epithelium, causing a massive infiltration by mononuclear cells, and the hamsters develop respiratory symptoms (21, 49, 50). For both the intravenous and intranasal routes, the immune system eliminates the virus within 7 to 10 days. Intravenous challenge of immunosuppressed hamsters with species C HAdV results in a disseminated disease that mimics the systemic HAdV infection seen in immunocompromised patients. Quantifiable outcomes include body weight loss, decreased survival, liver damage (elevated serum transaminase levels), and high virus burden in the liver or lung (43). In this model, FCV at a dose of 10 mg/kg administered orally 1 day before virus challenge then daily until the end of the experiment was sufficient to prevent mortality and mitigate morbidity. Further, FCV suppressed liver pathology and inhibited virus replication in the livers of infected immunosuppressed Syrian hamsters (Fig. 3 and 4). Animal efficacy at this dose (10 mg/kg) of FCV is consistent with prior studies in the murine model of cytomegalovirus (46). These promising results prompted us to examine potency in the same model, but with delayed FCV administration after virus challenge. Interestingly, similarly impressive efficacy was observed even when FCV administration started at day +4 postchallenge (Fig. 5). This suggests that FCV therapy could potentially alleviate disease symptoms and hasten recovery even if its administration is delayed after onset of disease. Compared to a known effective dose of VGCV (200 mg/kg twice daily), FCV (10 mg/kg daily) completely prevented mortality after intranasal challenge with HAdV6. FCV partially reversed body weight loss but significantly reduced viral burden in the lungs (Fig. 6). Although FCV tended to be less efficacious in mitigating body weight loss than VGCV, it performed better with regard to the more specific outcomes, namely, reducing lung pathology and suppressing virus replication. In our hands, treating respiratory HAdV infections in the immunosuppressed hamster model is notoriously difficult, and it is remarkable that thus far, FCV is one of only two antiviral compounds with any efficacy in this model.

Overall, our data demonstrate high potency for FCV against HAdVs, both in vitro and in vivo. However, the mechanism by which FCV inhibits HAdV replication remains poorly understood. Immunostaining for the adenoviral capsid hexon late protein suggests that FCV is effective in preventing progression to late stages of HAdV5 and HAdV6 infections in cell culture (Fig. 2), consistent with the hypothesis that adenoviral DNA polymerase is inhibited, as is the case with HCMV polymerase. FCV is initially phosphorylated by the HCMV UL97 protein kinase to its monophosphate derivative (FCV-MP) (39, 51). In HCMV-infected cells, FCV-MP is converted to its active triphosphate form (FCV-TP) by cellular GMP kinase (52–54). FCV-TP was shown to inhibit HCMV polymerase competitively and served as a nonobligate chain terminator of the replicating viral DNA (55). HAdVs are not very susceptible to GCV, another related nucleoside, because they are not known to encode a kinase. However, expressing HSV-1 thymidine kinase in trans in HAdV5-infected cells conferred susceptibility to GCV (56). Therefore, the demonstrated high potency of FCV against HAdVs suggests a yet-unidentified novel mechanism by which FCV is converted into its monophosphate derivative in adenovirus-infected cells. It is plausible that a novel adenoviral and/or host kinase might be involved in FCV phosphorylation. We hypothesize that the mechanism of action of FCV against HAdVs is similar to that of HCMV and that the initial phosphorylation step is carried out by an unknown adenovirus-encoded enzyme or virus-induced cellular kinase. The monophosphate form of FCV is then further phosphorylated to a triphosphate by endogenous cellular enzymes, and this triphosphate form inhibits the adenovirus-encoded DNA polymerase. In order to better elucidate the mechanism of action, further work is needed to detect and quantify FCV metabolites in adenovirus-infected cells. Other future experiments may include testing the effects of FCV-TP on adenoviral DNA synthesis using recombinant adenoviral DNA polymerase. Attempts to select for adenovirus FCV-resistant mutants are under way to possibly verify that FCV targets the adenoviral DNA polymerase, or to identify other molecular targets for FCV.

Conclusion.

Data presented here demonstrate in vitro and in vivo potency of FCV against HAdVs and support its continued clinical development.

MATERIALS AND METHODS

Cell culture and HAdV types.

For in vitro cytopathic effect (CPE) and quantitative PCR (qPCR) assays, human foreskin fibroblasts (HFFs) were prepared by methods previously described using human foreskins obtained from the University of Alabama at Birmingham tissue procurement facility with approval from the institutional review board (40). For all other assays, HFFs and human lung carcinoma (A549) cells were purchased from ATCC and cultured in a CO₂ incubator at 37°C in Dulbecco’s modified Eagle medium (DMEM) containing 2 mM glutamine, 10% fetal bovine serum, 100 IU/ml of penicillin, and 100 μg/ml of streptomycin. The human adenoviruses HAdV5 (ATCC VR-5) and HAdV6 (ATCC VR-6) were obtained from ATCC. The HAdV5 stock used in the immunofluorescence experiment (Fig. 2) was plaque purified from the original ATCC HAdV5. HAdV types 4, 7, and 8 were also obtained from ATCC.

Compound synthesis, formulation, and animals.

FCV (lot PKG-12-235) was synthesized using a previously published method by the NIH RAID pilot program (57). The powder was suspended in 0.4% carboxymethyl cellulose (Sigma) at the desired concentrations and sonicated to visual homogeneity. The drug was made up once weekly, aliquoted into daily portions, and stored at 4°C. Aliquots were allowed to equilibrate to room temperature before dosing. Male Syrian hamsters were purchased from Envigo at 60 to 80 g body weight. Approximately 100-g hamsters were dosed with 1 ml total volume of the appropriate suspension to obtain the desired doses.

Antiviral assays.

(i) CPE reduction assay. The activity of FCV against HAdV types 5, 7, and 8 was tested using the CPE reduction assay, as previously described (19). Briefly, monolayers of HFFs (5,000/well) were seeded in 384-well plates. Dilutions of FCV were prepared directly in the plates in a series of 5-fold dilutions, in duplicate wells, to yield final concentrations that ranged from 150 to 0.048 μM. Cell monolayers were infected with a multiplicity of infection (MOI) of approximately 0.005 PFU/cell. Cytopathology was determined by the addition of CellTiter-Glo reagent (Promega, Madison, WI).

(ii) qPCR assay. The activity of FCV against HAdV type 4 was evaluated utilizing qPCR to measure accumulation of viral DNA via methods that were described previously (58). Briefly, HFFs were seeded into 384-well plates and infected with HAdV4 at an MOI of approximately 0.005 PFU/cell. At 14 days following infection, total DNA was prepared and HAdV4 genome copy numbers were quantified by qPCR using the primers 5′-ACAGGACGCTTCGGAGTA-3′ and 5′-AACTTGTTCCCCAGACTGAAG-3′ and the probe 5′-CCCGCGCCACAGACACCTA-3′. The synthetic control standard sequence was GCCCCAGTGGTCTTACATGCACATCTCGGGCCAGGACGCCTCGGAGTACCTGAGCCCGGGTCTGGTGCAGTTTGCCCGCGCCACCGAGACGTACTTCAGCCTGAATAACAAGTTTAGAAACCCCACCGTGGC. Cytotoxicity was assessed in parallel with CellTiter Glo.

(iii) Neutral red survival assay. For HAdV type 6, a neutral red survival assay was performed in 96-well plates, as previously described (47). Briefly, A549 cells were plated at 6 × 10³ per well 1 day prior to infection and infected with HAdV6 at MOI of 0.5 PFU/cell. FCV was serially diluted 1:3 on separate 96-well plates with a no-drug final row. HAdV6 infections were done in 9 replicate wells for each drug concentration, and there were 3 uninfected drug control wells for each drug concentration. FCV dilutions were added to the cell plates immediately before infection with HAdV6. At 6 days postinfection (when viral CPE reached 70 to 90% for the virus-infected no-drug wells), neutral red was added for 1 h, and plates were washed 3 times with phosphate-buffered saline (PBS) to remove unattached cells. The neutral red was extracted from the remaining cells using 50% ethanol–1% glacial acetic acid. Plates were read, and the results were graphed in GraphPad Prism.

(iv) Immunofluorescence assay. An immunofluorescence assay, which detects adenovirus DNA-binding protein or hexon as indicators for the stage of infection, was conducted as previously described (45). Briefly, human A549 cells (on glass coverslips in 6-well plates) were infected at 5 PFU/cell with either HAdV5 or 6 or were mock infected. At the end of 1 h, FCV was added to final concentrations of 40, 10, and 4 μM, or no drug was added. At 27 h postinfection, the A549 cells were fixed in paraformaldehyde (3.7% in PBS) and permeabilized with methanol. Cells were stained for the adenovirus DNA-binding protein (DBP; green staining) and adenovirus hexon (red staining).

The immunosuppressed Syrian hamster model.

FCV was tested in vivo using a previously reported Syrian hamster model (43). All hamsters were immunosuppressed using cyclophosphamide (CP). CP was administered intraperitoneally as a single dose of 140 mg/kg 7 days before virus challenge and then at 100 mg/kg twice weekly for the duration of the experiment. Hamsters in the infected groups received 1.5 × 10¹⁰ to 4 × 10¹⁰ PFU of HAdV6 per kg either intravenously or intranasally. HAdV6 was used in the animal experiment because it is sequestered by tissue macrophages (mostly Kupffer cells) to a lesser extent and has a lower 90% lethal dose (LD₉₀) in hamsters than HAdV5. Further, the pathogenicity of HAdV6 is more strongly linked to virus replication in the hamster and less to the immune reaction to the input virus than that of HAdV5 (21, 49). FCV was administered orally every day for up to 16 days, as indicated in Fig. 3 to 6. The body weights and any signs of morbidity of the animals were recorded daily. Generally, throughout the experiments described in this report, test groups consisted of 15 hamsters. At day 5 (or 7 in the delayed-administration experiment) after virus challenge, 5 hamsters (designated at the start of the experiment) were sacrificed, and gross pathological observation was performed. Serum and liver samples were also collected; the virus burden in liver was determined by an assay for 50% tissue culture infective dose (TCID₅₀), and the serum was analyzed for transaminase levels. The remaining 10 hamsters were treated with FCV and then sacrificed at day 14 postchallenge. Hamsters that became moribund before day 14 were sacrificed as needed. In addition to animals judged moribund by observation, all hamsters that lost more than 20% of their original body weight were sacrificed. Thus, there were two endpoints for these experiments; the first was based on 5 hamsters from the day 3 or 5 time point (virus burden in the liver and serum transaminase levels), and the second was based on 10 hamsters (survival and body weight gain/loss).

Statistical analysis.

GraphPad Prism 4 (GraphPad Software) was used for statistical analysis. Body weight changes were compared using two-way analysis of variance (ANOVA). Overall effects for serum transaminase levels and virus burden in the liver were calculated using the Kruskal-Wallis test, and pairwise comparisons were performed using the Mann-Whitney U test or by Sidak's multiple-comparison test. A P value of ≤0.05 was considered significant.