The antiviral effects of baloxavir marboxil against influenza A virus infection in ferrets

Abstract Background Baloxavir marboxil (BXM), the oral prodrug of baloxavir acid (BXA), greatly reduces virus titers as well as influenza symptoms of uncomplicated influenza in patients. Objectives To investigate the pharmacokinetic profiles of BXA and its efficacy against influenza A virus infection in ferrets. Methods Ferrets were dosed orally with BXM (10 and 30 mg/kg twice daily for 1 day), oseltamivir phosphate (OSP) (5 mg/kg twice daily for 2 days) or vehicle to measure the antiviral effects of BXM and OSP. The pharmacokinetic parameters of BXA was determined after single oral dosing of BXM. Results The maximum plasma concentrations of BXA were observed at 1.50 and 2.00 hours with the two BXM doses, which then declined with an elimination half‐life of 6.91 and 4.44 hours, respectively. BXM at both doses remained detectable in the plasma in ferrets, which may be due to higher stability in liver microsomes. BXM (10 and 30 mg/kg twice daily) treatment at Day 1 post‐infection (p.i.) reduced virus titers by ≥3 log10 of the 50% tissue culture infective doses by Day 2, which was significantly different compared with vehicle or OSP. Body temperature drops over time were significantly greater with BXM than with vehicle or OSP. Significant reduction in virus titers was also demonstrated when BXM was administrated after symptom onset at Day 2 p.i. compared with vehicle and OSP, although body temperature changes largely overlapped between Day 2 and Day 4. Conclusions The results highlight the rapid antiviral action of BXM with post‐exposure prophylaxis or therapeutic dosing in ferrets and offer support for further research on prevention of influenza virus infection and transmission.


| BACKG ROU N D
Seasonal influenza, an acute respiratory infectious disease caused by influenza viruses, spreads easily from person to person and causes seasonal epidemics worldwide. [1][2][3][4] Although influenza vaccination is the most effective way to prevent the disease, epidemics continue to occur, indicating the need for anti-influenza drugs to control influenza virus infection. 5,6 Currently, neuraminidase inhibitors (NAIs), including oseltamivir phosphate (OSP), are the most widely used class of anti-influenza drugs. 7,8 Baloxavir marboxil (BXM), and its active form baloxavir acid (BXA), is a first-in-class, small-molecule inhibitor of cap-dependent endonuclease (CEN), 9 which has been approved for clinical use (single dose of 40 mg for patients 40 to <80 kg or 80 mg for patients ≥80 kg) in uncomplicated adults and adolescents in Japan and the United States for the treatment of influenza type A and B virus infections. The CEN in the polymerase acidic protein (PA), a subunit of influenza virus polymerase, mediates the cap-snatching process during viral mRNA biosynthesis. 10,11 BXA exhibits broad-spectrum inhibitory activity against seasonal, avian and swine influenza viruses including NAI-resistant strains in vitro and in vivo. 9,12,13 A previous phase 1 trial revealed that a single oral dose of BXM from 6 to 80 mg was rapidly metabolized to BXA, and the terminal elimination half-life of BXA ranged from 49 to 91 hours. 14 The phase 3 trial (CAPSTONE-1) showed that infectious virus titers rapidly declined within 1 day after single, weight-based oral doses of BMX (40 or 80 mg), and the average titer remained low thereafter, which were superior to both placebo and OSP. 15 The time to alleviation of symptoms was also superior compared with placebo but not significantly different from the effect of OSP treatment. Treatment-emergent amino acid substitution at amino acid position 38 in the PA (PA/I38T) was a major pathway for reduced susceptibility to BXA in the clinical study, as expected from non-clinical evidence. 15,16 Ferrets are naturally susceptible to human influenza viruses without adaptation and develop upper respiratory tract infection. 17,18 The clinical signs of infected ferrets, including fever, nasal discharge, and decreased activity, closely reflect those manifested in humans. 19 In addition, ferrets have the ability to transmit the influenza virus among each other. 20,21 Therefore, the ferret model is a useful tool for research on the evaluation of antivirals or vaccines as well as influenza virus transmission. [22][23][24][25][26][27] The aims of our study were to describe the pharmacokinetic profile and the antiviral effects of BXM against influenza A virus infection in ferrets to support our understanding of the pathogenesis of influenza.

| Animals
Female ferrets (Japan SLC Inc), confirmed to be negative for antiinfluenza antibodies, approximately 14-26 months old were used.
All ferret studies were conducted under applicable laws and guidelines and after approval from the Shionogi Animal Care and Use Committee. For stability tests, male Sprague Dawley (Crl:CD(SD)) rats and female cynomolgus monkey were obtained from Charles River Laboratories Japan, Inc. Male TOYO beagle was obtained from Oriental BioService, Inc.

| Pharmacokinetic analysis
One day prior to administration, group assignment was carried out taking body weight into account. Under isoflurane anesthesia, fasted ferrets (four per group) received BXM (10 or 30 mg/5 mL/kg) suspension orally using an intragastric tube. Blood samples were collected Dosing type was used as "Variable." The computation method used for interpolation and extrapolation of untransformed data was selected as logarithm (Log). The terminal elimination half-life for 10 and 30 mg/kg doses was extrapolated from changes of the mean plasma concentrations of BXA in four ferrets between 6 and 24 hours, with values of 6.57 and 4.08 hours, respectively.

| Stability in plasma
Approximately 30 mL of blood from Japanese male healthy volunteers was collected and centrifuged to obtain plasma. Blood samples from ferrets, rats, dogs, and monkeys were collected under anesthesia and centrifuged to obtain plasma. BXM was added to reach a

| Metabolic stability in liver microsomes
Liver microsomes were prepared from ferrets, rats, dogs, and monkeys, as previously described. 28 Human liver microsomes were ob-

| Analysis of virus titer in nasal washes and body temperature in ferrets
To monitor viral replication in nasal cavities, ferrets were lightly anesthetized with isoflurane and nasal wash fluids were obtained from each ferret using a disposable syringe filled with 5 mL of DPBS including penicillin and streptomycin on Days 1, 2, and 3 (Experiment 1), or Days 1, 2, 3, and 4 p.i. (Experiment 2). Nasal wash fluids were collected and filtered using 0.45 µm filter and then stored until use. For virus titration, serial dilutions of nasal washes were inoculated onto confluent MDCK cells in 96-well plates. After 1 hour incubation, the suspension was removed and the cells were cultured in MEM including 0.5% bovine serum albumin (Sigma-Aldrich) and 3 µg/mL trypsin.
The plates were incubated at 37°C in 5% CO 2 for 3 days. The presence of cytopathic effects was determined under a microscope, and virus titers were calculated as log 10 TCID 50 /mL. When no cytopathic effect was observed using undiluted viral solution, it was defined as an undetectable level, being lower than 0.5 log 10 TCID 50 /mL. Under anesthesia, a data logger (DS1921H-5F; Maxim Integrated Products, Inc) was implanted for subcutaneous body temperature monitoring. Penicillin and streptomycin were given to each ferret once daily for 3 days following surgery. Body temperature was analyzed using One Wire Viewer version 0.3.15.50. (Maxim Integrated Products, Inc) and expressed by calculating the average temperatures for 8-hour periods. Data recorded from 8 hours prior to the first sampling of nasal wash were defined as the basal body temperature.

| Statistical analysis
The uniformity of the mean body weight among groups was confirmed by one-way analysis of variance (ANOVA). For the comparison of the virus titers in nasal washes and change from baseline body temperature among BXM-, OSP-, and vehicle-treated groups, the fixed-sequence procedure or Dunnett's multiple-comparison method following the ANOVA test. P values of <.05 were considered to be statistically significant. Statistical analysis was performed using the statistical analysis software SAS version 9.2 (SAS Institute Inc).

| Pharmacokinetic analysis of BXA and BXM in ferrets
The plasma exposures of BXA after a single oral administration of BXM at doses of 10 and 30 mg/kg are shown in Figure 1A. The maximum plasma concentrations of BXA were reached at 1.50 and 2.00 hours at doses of BXM 10 and 30 mg/kg, respectively and then declined with a mean (standard deviation) elimination half-life (t 1/2 ) values of 6.91 (3.79) and 4.44 (0.67) hours, respectively (Table 1 and Table S2). BXA was still detectable at 24 hours (3.03 ± 2.33 and 9.91 ± 8.77 ng/mL at doses of BXM 10 and 30 mg/kg, respectively).
BXM was also detected in ferret plasma after a single oral administration of BXM at both doses ( Figure 1B and Table 1).
Due to a relatively short half-life in ferrets, the plasma concen-  Table S3).

| Stability of BXM in plasma and liver microsomes
The stability of BXM in plasma of humans and animals (except rats) was comparable ( Table 2 and Table S4). We also compared the stability of BXM with that of OSP and cefcapene pivoxil, a reference prodrug to confirm hydrolytic activity, in liver microsomes of humans, ferrets, rats, dogs, and monkeys. At 30 minutes, cefcapene pivoxil was undetectable in liver microsomes of all species. OSP was stable in liver microsomes of ferrets, rats, and dogs, and BXM was stable only in ferret liver microsomes (Table 3, and Table S5).

| Effects of oral BXM when administered at Day 1 p.i. in ferrets
The maximum virus titer in nasal washes in vehicle-treated ferrets was detected on Day 2 p.i., followed by a decline by Day 3 p.i.
( Figure 3B). BXM at doses of 10 and 30 mg/kg showed a similar reduction in virus titer to an undetectable level (ie, <0.5 log 10 TCID 50 / mL) at Day 2 p.i. OSP treatment at 5 mg/kg dose for 2 days resulted in a reduced burden of virus replication compared with vehicle. Virus titer levels were comparable by Day 3 among all groups, following a rebound in the BXM group ( Figure 3B). A rise in body temperature from baseline was observed in the vehicle-and OSP-treated groups from Day 2 to Day 3 p.i. (Figure 3C and Figure S2). The suppression of body temperature changes over time from 8 hours up to Day 3 p.i. was significantly greater with BXM at doses of 10 and 30 mg/kg than vehicle and OSP 5 mg/kg.

| Effects of oral BXM when administered at Day 2 p.i. in ferrets
In the second part of the experiments, ferrets were infected with influenza A/Kadoma/3/2006 (H1N1) strain intranasally at 5000 TCID 50 /ferret, which was a fivefold higher infectious dose increas- and BXM treatment groups, although a small but significant effect was observed with BXM compared with vehicle on Day 2 ( Figure 4C and Figure S3).

| D ISCUSS I ON
In this study, we examined the antiviral effects of BXM against influ-  Figure S4). Although high levels of BXM were detected in ferrets in contrast to humans and mice, its CEN inhibitory activity was ≥200-fold lower than that of BXA in in vitro studies, which indicates that the plasma concentration of BXA is more crucial for determining the dosing regimen of BXM. 9 Metabolic stability tests showed that BXM was stable in liver microsomes of ferrets, but not in any other species. BXM is metabolized by arylacetamide deacetylase (AADAC), which is expressed in the liver and the intestine in humans. 33,34 However, the tissue distribution of ferret AADAC is currently unknown. The species differences in AADAC expression or activity in liver may contribute to these findings.
Owing to the similarities in the pathogenesis of influenza in humans, ferrets are frequently used to describe antiviral activities of agents or to develop vaccines. 27 In this study, we have assessed TA B L E 3 Stability of baloxavir marboxil (BXM), oseltamivir, and cefcapene pivoxil in liver microsomes of a human, ferret, rat, dog, and monkey  Note: Data are expressed as the mean of duplicates. Individual values are described in Table S5. which may explain the reduced antiviral effect. In addition, the simulated BXA concentrations at 48 hours were lower than the plasma target concentration (6.85 ng/mL) (see Table S6). 36 Further studies for longer treatment periods that maintain the effective plasma con-

CO N FLI C T O F I NTE R E S T
The authors are employees of Shionogi and Co., Ltd. Osaka, Japan.

DATA AVA I L A B I L I T Y S TAT E M E N T
All relevant data are included in the manuscript, including Supporting Information. Data sharing is not applicable to this study.