The PA Endonuclease Inhibitor RO-7 Protects Mice from Lethal Challenge with Influenza A or B Viruses

ABSTRACT

Current influenza treatment relies on a single class of antiviral drugs, the neuraminidase inhibitors (NAIs), raising concern over the potential emergence of resistant variants and necessitating the development of novel drugs. In recent years, investigational inhibitors targeting the endonuclease activity of the influenza acidic polymerase (PA) protein have yielded encouraging results, although there are only limited data on their in vivo efficacy. Here, we examined the antiviral potential of the PA endonuclease inhibitor RO-7 in prophylactic and therapeutic regimens in BALB/c mice inoculated with influenza A/California/04/2009 (H1N1)pdm09 or B/Brisbane/60/2008 viruses, which represent currently circulating antigenic variants. RO-7 was administered to mice intraperitoneally twice daily at dosages of 6, 15, or 30 mg/kg/day for 5 days, starting 4 h before or 24 or 48 h after virus inoculation, and showed no adverse effects. Prophylactic administration completely protected mice from lethal infection by influenza A or B virus. The level of therapeutic protection conferred depended upon the time of treatment initiation and RO-7 dosage, resulting in 60 to 100% and 80 to 100% survival with influenza A and B viruses, respectively. RO-7 treatment significantly decreased virus titers in the lung and lessened the extent and severity of lung damage. No PA endonuclease-inhibitor resistance was observed in viruses isolated from lungs of RO-7-treated mice, and the viruses remained susceptible to the drug at nanomolar concentrations in phenotypic assays. These in vivo efficacy results further highlight the potential of RO-7 for development as antiviral therapy for influenza A and B virus infections.

INTRODUCTION

Influenza A and B viruses are the causative agents of highly contagious respiratory diseases in humans (1). Despite the availability of vaccine programs (2, 3), seasonal influenza viruses are estimated to infect 5 to 10% of the global population annually, resulting in over 225,000 hospitalizations and nearly 50,000 deaths in the United States alone (4, 5). Older adults (>65 years of age), the very young, and immunocompromised patients (6) are particularly susceptible to these viruses. However, the past two decades have witnessed the emergence from avian reservoirs of novel influenza subtypes (H5N1, H5N6, H7N9, H9N2, and H10N8) that can cause severe disease in otherwise healthy individuals (7–9). Antivirals are a critical resource for the treatment or prophylaxis of infections by seasonal or emerging influenza viruses.

Current treatment for influenza virus infections is limited to a single class of antivirals, the neuraminidase (NA) inhibitors (NAIs) (10, 11). The NAIs (oseltamivir, zanamivir, peramivir, and laninamivir) inhibit NA enzymatic activity that is necessary for virus release from host cells (12). The development of NAI resistance is of critical concern. During the 2007-2009 influenza seasons, naturally occurring oseltamivir-resistant A (H1N1) viruses with an H274Y NA substitution (N2 numbering) appeared and predominated (with approximately 90% resistance being detected) among circulating viruses of this subtype (13, 14). The incidence of NAI resistance among seasonal influenza viruses remains low (0.1 to 3%) (15, 16). However, NAI resistance-associated substitutions have been identified in the clinic in patients infected with H1N1pdm09, H3N2, and H5N1 viruses (12, 17). Influenza B viruses are less susceptible to NAIs in phenotypic enzyme-inhibition assays (12), and although no clinical studies have exclusively analyzed NAI efficacy against infection with influenza B virus alone, meta-analysis of aggregated data suggests that NAIs may be less efficacious against influenza B than against influenza A virus (12, 17, 18). Furthermore, NAIs generally have a narrow treatment window (requiring initiation of treatment within 48 h of symptom onset) and may have less desirable routes of administration (i.e., inhalation, in the case of zanamivir and laninamivir) (10, 11, 19). These limitations and the reliance upon NAI monotherapy to treat influenza infection highlight the urgent need for new therapeutic options and justify the pursuit of novel drugs, particularly those that act against different viral protein target(s).

In recent years, the availability of high-quality structural information on the influenza virus polymerase proteins has renewed interest in developing antivirals that disrupt the critical role(s) these proteins play in virus replication (20–22). The influenza polymerase complex is a heterotrimer composed of the basic polymerases 1 and 2 (PB1 and PB2) and the acidic polymerase (PA) (23). PB1 is an RNA-dependent polymerase involved in RNA strand elongation and is inhibited by nucleoside/nucleobase analogues such as ribavirin and favipiravir (T-705) (24). PB2 binds host-cell 5′ mRNA caps, which, after further modification by the PA protein, are used to initiate PB1-mediated transcription (25). Experimental inhibitors have been described that affect PB2 cap-binding activity (PB2-39 [26], BPR3P0128 [27], Cap-3/7 [28], RO and PPT28 [29]) or that disrupt the interaction of PB2 with the polymerase complex (PB1_731–757 peptide [30] and PB2_1–37 peptide [31]). One PB2 inhibitor (VX-787/JNJ872) is currently undergoing phase 2b clinical trials (12, 32–34). The PA protein contains the endonuclease activity required to cleave PB2-bound capped mRNAs and initiate viral transcription (35). The exact mechanisms of action of PA inhibitors are not precisely known, but they may include the sequestration of divalent cations in the protein's catalytic site that is necessary for enzymatic activity and/or interaction with RNA-binding residues, resulting in substrate competition (22). The influenza PA protein is highly conserved across influenza A viruses (20, 36–38). Furthermore, structural analyses indicate a putative catalytic site depression within the PA protein that is present among influenza A, B, and C viruses and includes five key active-site residues (H41, E80, D108, E119, and K134 [influenza A numbering]) (35). Therefore, compounds that affect the endonuclease activity of the PA protein have the potential to inhibit multiple influenza genera.

Several PA inhibitors have been described in the literature as being able to inhibit PA substrate cleavage and/or virus replication in the nanomolar to micromolar range in vitro (discussed in reference 39 and reviewed in reference 22). Of these, compounds AL-794 and S-033188 have advanced to clinical trials (22, 34), but no in vivo, preclinical efficacy of these PA inhibitors has been reported to date.

Hastings and colleagues demonstrated that the administration of 4-substituted 2,4-dioxobutanoic acid PA endonuclease inhibitor (L-742,001) before and after influenza A/Hong Kong/1/1968 (H3N2) virus infection of BALB/c mice resulted in a dose-dependent reduction of virus in nasal washes (40), but this study did not address virus replication in the lungs, clinical signs of disease, mortality, or a therapeutic regimen. Recently, Yuan et al. described a novel (5Z)-2-[2-(2-oxoindol-3-yl)hydrazinyl]-5-(2-oxo-1H-indol-3-ylidene)-1,3-thiazol-4-one endonuclease inhibitor (ANA-0) that could protect mice from lethal challenge with mouse-adapted influenza A/Hong Kong/415742Md/2009 (H1N1)pdm09 virus and significantly reduced lung virus titers when administered 6 h postinoculation (41). However, this study was limited to early postexposure administration. The therapeutic window, the effects of multiple dosages, and the in vivo efficacy against challenge with influenza B viruses were not explored.

Previously, we characterized RO-7 (Fig. 1A), a small-molecule, broad-spectrum inhibitor of the influenza A and B virus PA endonuclease protein (39). In vitro, RO-7 was effective at nanomolar 50% effective concentrations (EC₅₀) in MDCK cells (3.2 to 16.0 nM) and in differentiated normal human bronchial epithelial cells (3 and 30 nM). To understand the therapeutic potential of RO-7 and to address the overall lack of in vivo data regarding experimental PA endonuclease inhibitors, we examined the ability of this drug to protect mice from lethal challenge with influenza A or B virus, to reduce virus titers in the lung, and to decrease virus-induced lung pathology. In addition, we evaluated the potential for in vivo antiviral resistance to develop under different treatment regimens.

FIG 1.

Safety profile of RO-7 treatment in mice. (A) Chemical structure of RO-7. (B) BALB/c mice (n = 5/group) received sterile PBS (control, i.p.), RO-7 (30 mg/kg/day, i.p.), or OSE (20 mg/kg/day, orally) twice daily for 5 days. Body weights were monitored over 18 days. The blue-shaded area indicates the duration of treatment.

RESULTS

RO-7 safety profile.

We initially tested whether administering RO-7 to mice in the absence of influenza virus infection caused any adverse effects. Mice receiving RO-7 showed no weight loss (Fig. 1B) or changes in clinical signs or behavior (data not shown) during the observation periods, similar to the mice receiving the clinically available drug oseltamivir phosphate (OSE) or phosphate-buffered saline (PBS) alone. These results suggest a favorable safety profile for RO-7 in this experimental system.

An RO-7 prophylactic regimen protects mice from lethal challenge with influenza A or B virus.

To determine the efficacy of RO-7 in a preexposure prophylaxis regimen, mice were inoculated with influenza A or B virus, and RO-7 was administered beginning 4 h before virus inoculation (Fig. 2A). Treatment with OSE was conducted for comparison purposes, since its efficacy against influenza virus infection in the mouse model is well established (42, 43). The PBS-treated (control) mice inoculated with A/California/04/2009 (H1N1)pdm09 virus exhibited progressive weight loss and succumbed to infection at 7 to 10 days postinoculation (dpi) (Fig. 2B). Treatment with all dosages of RO-7 resulted in 100% survival of mice, and the changes in body weight loss were dose dependent. Mice treated with RO-7 at 6 mg/kg/day lost 15 to 17% of their initial body weight, whereas mice treated with RO-7 at 15 mg/kg/day lost no more than 4% of their initial body weight (Fig. 2B). The pattern of return to initial body weight was also dose dependent; mice treated with 6 mg/kg/day of RO-7 regained their initial weight by 18 dpi compared to 12 dpi for mice treated with 15 mg/kg/day. Mice treated with 30 mg/kg/day of RO-7 lost no body weight during the study (Fig. 2B).

FIG 2.

RO-7 prophylaxis protects mice from lethal challenge with influenza A or B virus. Female 6- to 8-week-old BALB/c mice (n = 5/group) were lightly anesthetized with isoflurane and inoculated intranasally with 5 MLD₅₀ of A/California/04/2009 (H1N1)pdm09 or B/Brisbane/60/2008 virus. (A) Mice were treated with RO-7 (6, 15, or 30 mg/kg/day, i.p.) or OSE (20 mg/kg/day, orally) at 4 h before inoculation (−4), 8 hpi (+8), and twice daily for 4 days (black arrow) after virus inoculation (red arrow). Control virus-inoculated mice received sterile PBS (i.p.) on the same schedule. Body weight (B and D) and survival (C and E) were monitored for 18 days. The blue-shaded areas indicate the duration of treatment, and the dotted line indicates the endpoint for mortality (loss of 25% of initial weight).

Similar to the influenza A virus challenge, control mice inoculated with B/Brisbane/60/2008 virus lost more than 25% of their initial body weight (Fig. 2D) and succumbed to infection between 7 and 9 dpi, whereas RO-7 administration (all dosages) resulted in 100% survival of mice (Fig. 2E). Protection from weight loss was dose dependent; mice treated with 6 mg/kg/day of RO-7 lost 10% of their initial weight, whereas those treated with 15 or 30 mg/kg/day lost no more than 4% of their initial weight. Mice treated with 6 mg/kg/day of RO-7 regained their initial body weight at a similar rate to mice inoculated with influenza A virus, returning to their initial weight at approximately 18 dpi, whereas mice treated with the two higher dosages had minimal weight loss (Fig. 2D).

Overall, RO-7 administration at a dosage of 15 or 30 mg/kg/day provided better protection against challenge with A/California/04/2009 (H1N1)pdm09 (8 dpi, P ≤ 0.0001) or B/Brisbane/60/2008 (7 dpi, P ≤ 0.0001) viruses than did OSE prophylaxis. These treatment dosages also resulted in the mice recovering 100% of their initial weight in a shorter time, compared to OSE-treated animals, when challenged with influenza A (≤12 days with RO-7 versus 16 days with OSE) or B (negligible loss with RO-7 versus 18 days with OSE) viruses. Treatment of mice with a low dosage of RO-7 (6 mg/kg/day) yielded results for survival and morbidity that were similar to the results with OSE prophylaxis (20 mg/kg/day). Thus, RO-7 prophylaxis prevents morbidity and mortality in mice inoculated with a lethal dose of influenza A or B virus and confers protection that is similar to or more effective than that provided by OSE.

An RO-7 therapeutic regimen protects mice from lethal challenge with influenza A or B virus.

We next determined whether RO-7 exhibited therapeutic activity when administered to mice after virus inoculation. Mice were inoculated with influenza A or B virus and treated with RO-7 beginning at +24 or +48 h postinoculation (hpi) (Fig. 3A). As previously stated, control animals inoculated with A/California/04/2009 (H1N1)pdm09 virus lost more than 25% of their initial weight and succumbed to infection at 7 to 10 dpi (Fig. 3). Unlike prophylactic administration of RO-7, the therapeutic drug regimen displayed no dose-dependent effect on the morbidity of mice inoculated with A/California/04/2009 (H1N1)pdm09 virus. When RO-7 administration was delayed until +24 hpi, all drug dosages demonstrated the same efficacy, with 80% of mice surviving lethal virus challenge (Fig. 3B and C). When RO-7 administration was delayed until +48 hpi, the weight loss patterns were similar to those observed in the +24 hpi delayed-treatment groups, but survival ranged from 60 to 80%, depending on the dosage (Fig. 3D and E).

FIG 3.

RO-7 therapy protects mice from lethal challenge with influenza A or B virus. Female 6- to 8-week-old BALB/c mice (n = 5/group) were lightly anesthetized with isoflurane and inoculated intranasally with 5 MLD₅₀ of either A/California/04/2009 (H1N1)pdm09 or B/Brisbane/60/2008 virus. (A) Mice were treated twice daily with RO-7 (6, 15, or 30 mg/kg/day, i.p.) or OSE (20 mg/kg/day, orally) for 5 days, starting at 24 or 48 hpi (+24 and +48, black arrows) after virus inoculation (red arrow). Control virus-inoculated mice received sterile PBS (i.p.) on the same schedule. Body weights (B, D, F, and H) and survival (C, E, G, and I) were monitored for 18 days. The blue-shaded areas indicate the duration of treatment, and the dotted line indicates the endpoint for mortality (loss of 25% of initial weight).

Therapeutic administration of RO-7 to mice inoculated with B/Brisbane/60/2008 virus exhibited a dose-dependent effect on body weight loss, similar to the results observed with the prophylactic regimen. When treatment was delayed until +24 hpi, mice receiving 6 mg/kg/day of RO-7 lost 18% of their initial weight, whereas mice receiving dosages of 15 or 30 mg/kg/day lost 5 and 1% of their initial weights, respectively (Fig. 3F). The two higher RO-7 dosages resulted in 100% survival, compared to 80% survival for mice treated with 6 mg/kg/day (Fig. 3G). This dose-dependent effect on weight loss was diminished when treatment was delayed until +48 hpi, resulting in all drug-treated mice losing 10 to 17% of their initial weight (Fig. 3H). A clear dose response was observed in relation to survival: 80% of mice treated with 6 or 15 mg/kg/day of RO-7 survived lethal virus challenge compared to 100% of animals treated with 30 mg/kg/day (Fig. 3I).

The therapeutic RO-7 regimen demonstrated some advantages over OSE treatment in mice, particularly in those animals inoculated with B/Brisbane/60/2008 virus. Mice inoculated with that virus and treated with RO-7 at 15 or 30 mg/kg/day had less peak weight loss compared to OSE-treated animals (P ≤ 0.05) when treatment was initiated +24 hpi (Fig. 3F), although no survival benefits were observed. When drug therapy was initiated at +48 hpi, RO-7 treatment resulted in the survival of 80 to 100% of mice compared to the 40% survival observed with OSE treatment (Fig. 3I).

Overall, these results indicate that a therapeutic RO-7 regimen provides survival benefits for mice lethally challenged with influenza A or B virus. The beneficial effect over OSE was more pronounced when RO-7 treatment was initiated at +48 hpi and in mice infected with influenza B virus.

RO-7 decreases histologic changes and lung damage during lethal infection with influenza A or B virus.

In both RO-7–treated and untreated mice, pulmonary lesions that developed after influenza infection were localized to the virus-infected areas of the lung parenchyma and were characterized by thickened septa accompanied by both interstitial and alveolar inflammatory cells. Pulmonary edema, hemorrhage, and hyaline membrane (HM) formation were restricted to mouse lungs exhibiting more severe damage to the alveolar-capillary membrane, which resulted in the leakage of plasma proteins and red blood cells into the alveoli. In these studies, the widespread HM formation present in untreated control mice infected with influenza A or B virus (Fig. 4 and 5A and B) correlated with the high mortality observed in these animals (Fig. 2 and 3). In contrast, HMs were essentially absent in mice that received prophylactic RO-7 (Fig. 4 and 5A and C) and were markedly reduced in mice treated with RO-7 postexposure (Fig. 4 and 5A, D, and E). Scattered protein/serum aggregates suggestive of milder damage to the alveolar-capillary membrane were present in some animals when RO-7 treatment was initiated +24 hpi, and HMs were present in only some animals when RO-7 treatment was initiated +48 hpi. These reductions in HM formation correlated with the improved survival of RO-7-treated mice.

FIG 4.

RO-7 decreases lung damage associated with influenza A virus challenge. Female 6- to 8-week-old BALB/c mice (n = 3/group) were inoculated with A/California/04/2009 (H1N1)pdm09 as described in the previous figures. (A) Lungs sampled at 8 dpi were subjected to quantitative scoring of HM formation. (B to H) Representative hematoxylin and eosin-stained sections of control (PBS-treated) lungs (B) and lungs treated with RO-7 (30 mg/kg/day) (C to E) or OSE (20 mg/kg/day) (F to H), with treatment being initiated at the indicated time points (−4, +24, and +48 hpi). HM formations are indicated by black arrows. The sections are displayed at ×40 magnification.

FIG 5.

RO-7 decreases lung damage associated with influenza B virus challenge. Female 6- to 8-week-old BALB/c mice (n = 3/group) were inoculated with B/Brisbane/60/2008 as described for the previous figures. (A) Lungs sampled at 8 dpi were subjected to quantitative scoring of HM formation. (B to H) Representative hematoxylin and eosin-stained sections of control (PBS-treated) lungs (B) and lungs treated with RO-7 (30 mg/kg/day) (C to E) or OSE (20 mg/kg/day) (F to H), with treatment being initiated at the indicated time points (−4, +24, and +48 hpi). HMs are indicated by black arrows. The sections are displayed at ×40 magnification.

RO-7 administration decreases virus titers in the lungs of mice inoculated with influenza A or B virus.

To characterize the effect of RO-7 administration on the kinetics of virus replication at the site of infection, we measured virus titers in lungs of mice at 3, 6, or 9 dpi. Prophylactic administration of RO-7 (all dosages) significantly decreased titers of A/California/04/2009 (H1N1)pdm09 virus compared to those in untreated control animals, at all-time points (P < 0.05) (Table 1). Similarly, RO-7 prophylaxis significantly decreased lung titers of B/Brisbane/60/2008 virus, compared to those in untreated controls, at 3 and/or 6 dpi (P < 0.05) (Table 2).

TABLE 1.

Effect of RO-7 treatment on influenza A/California/04/2009 (H1N1)pdm09 virus replication in the lungs of BALB/c mice

Treatment regimen			Lung titer (mean log10 TCID50/ml ± SEM)a
Group	Dose (mg/kg/day)	Treatment initiation (hpi)	3 dpi	6 dpi	9 dpi
Control	0	−4	8.0 ± 0.3 (3/3)	7.2 ± 0.2 (3/3)	6.4 ± 0.1 (3/3)
RO-7	6	−4	2.3 ± 0.0 (1/3)*§	5.3 ± 0.3 (3/3)*§	1.9 ± 0.0 (1/3)*
		+24	6.6 ± 0.1 (3/3)*§	6.1 ± 0.2 (3/3)*	<(0/3)*
		+48	6.8 ± 0.2 (3/3)*§	5.7 ± 0.1 (3/3)*§	<(0/3)*
RO-7	15	−4	5.5 ± 0.5 (3/3)*	5.3 ± 0.3 (3/3)*§	<(0/3)*
		+24	6.9 ± 0.2 (3/3)*§	5.8 ± 0.3 (3/3)*	<(0/3)*
		+48	7.6 ± 0.2 (3/3)	6.4 ± 0.5 (3/3)	<(0/3)*
RO-7	30	−4	5.2 ± 0.5 (3/3)*§	5.2 ± 0.4 (3/3)*§	1.3 ± 0.0 (1/3)*
		+24	6.8 ± 0.2 (3/3)*§	5.8 ± 0.3 (3/3)*	<(0/3)*
		+48	6.8 ± 0.3 (3/3)*§	6.1 ± 0.2 (3/3)*§	<(0/3)*
OSE	20	−4	7.2 ± 0.5 (3/3)	7.4 ± 0.1 (3/3)	2.6 ± 1.2 (2/3)*
		+24	7.8 ± 0.2 (3/3)	7.0 ± 0.4 (3/3)	<(0/3)*
		+48	7.7 ± 0.1 (3/3)	7.7 ± 0.3 (3/3)	1.8 ± 0.0 (1/3)*

^aVirus titers in whole lungs from virus-inoculated BALB/c mice (n = 3/group) were determined at the indicated time points; a “<” symbol indicates a titer below the limit of detection (0.75 log₁₀ TCID₅₀/ml). Values in parentheses indicate the numbers of animals with detectable virus in the lung/total number of animals. *, P ≤ 0.05 compared to control virus-inoculated untreated mice; §, P ≤ 0.05 compared to virus-inoculated OSE-treated mice at each time point after virus inoculation (as determined by unpaired t test).

TABLE 2.

Effect of RO-7 treatment on influenza B/Brisbane/60/2008 virus replication in the lungs of BALB/c mice

Treatment regimen			Lung titer (mean log10 TCID50/ml ± SEM)a
Group	Dose (mg/kg/day)	Treatment initiation (hpi)	3 dpi	6 dpi	9 dpi
Control	0	−4	4.9 ± 0.3 (3/3)	3.7 ± 0.1 (3/3)	<(0/3)
RO-7	6	−4	3.3 ± 0.3 (3/3)*	2.7 ± 0.1 (3/3)*	<(0/3)
		+24	4.3 ± 0.3 (3/3)	3.3 ± 0.5 (3/3)	<(0/3)
		+48	4.6 ± 0.1 (3/3)	3.4 ± 0.5 (3/3)	<(0/3)
RO-7	15	−4	2.9 ± 0.3 (3/3)*§	2.4 ± 0.9 (2/3)	<(0/3)
		+24	3.2 ± 0.7 (3/3)	2.4 ± 0.9 (2/3)	<(0/3)
		+48	3.7 ± 0.6 (3/3)	2.8 ± 0.3 (3/3)*	<(0/3)
RO-7	30	−4	<(0/3)*§	<(0/3)*§	<(0/3)
		+24	3.9 ± 0.3 (3/3)	1.5 ± 0.0 (1/3)	<(0/3)
		+48	3.8 ± 0.3 (3/3)*§	<(0/3)*§	<(0/3)
OSE	20	−4	4.1 ± 0.3 (3/3)	2.8 ± 0.2 (3/3)*	<(0/3)
		+24	4.3 ± 0.3 (3/3)	3.5 ± 0.1 (3/3)	<(0/3)
		+48	4.7 ± 0.1 (3/3)	3.3 ± 0.3 (3/3)	<(0/3)

^aVirus titers in whole lungs from virus-inoculated BALB/c mice (n = 3/group) were determined at the indicated time points; a “<” symbol indicates a titer below the limit of detection (0.75 log₁₀ TCID₅₀/ml). Values in parentheses indicate the numbers of animals with detectable virus in the lung/total numbers of animals. *, P ≤ 0.05 compared to control virus-inoculated untreated mice; §, P ≤ 0.05 compared to virus-inoculated OSE-treated mice at each time point after virus inoculation (as determined by unpaired t test).

The therapeutic RO-7 regimen also decreased virus replication in the lungs of mice inoculated with influenza A virus. RO-7 treatment at 6 and 30 mg/kg/day, and in most cases at 15 mg/kg/day, significantly decreased the lung titers of A/California/04/2009 (H1N1)pdm09 virus, compared to those in control mice, at all time points (P < 0.05) (Table 1). The therapeutic RO-7 regimen only slightly decreased titers in B/Brisbane/60/2008 virus-inoculated animals. Statistically lower titers were noted only in the +48 hpi delayed-treatment groups at 3 or 6 dpi for the two highest RO-7 dosages (Table 2). However, the total virus burden in the lungs of mice inoculated with B/Brisbane/60/2008 virus was >3-fold lower than that in the lungs of mice inoculated with A/California/04/2009 (H1N1)pdm09 virus, and this likely affected the statistical comparisons between RO-7-treated and -untreated animals inoculated with different viruses.

By comparison, prophylactic administration of OSE had little effect on A/California/04/2009 (H1N1)pdm09 virus titers in the lungs of mice until 9 dpi (Table 1). Mice receiving RO-7 prophylaxis had lower titers of A/California//04/2009 (H1N1)pdm09 virus, compared to those in OSE-treated animals, at 3 and/or 6 dpi (P < 0.05) (Table 1). The highest RO-7 dosage (30 mg/kg/day) also resulted in B/Brisbane/60/2008 virus titers lower than those in OSE-treated animals at 3 and 6 dpi (P < 0.05) (Table 2). The RO-7 therapeutic regimen (all dosages) also significantly decreased virus titers (P < 0.05), compared to those in OSE-treated animals, particularly at 3 and 6 dpi (Table 1).

Since RO-7 is demonstrated to be an endonuclease inhibitor, we examined the ability of treatment to affect levels of viral mRNA (vmRNA) in the lungs of A/California/04/2009 (H1N1)pdm09 virus-inoculated mice at 6 dpi. A trend toward lower coefficient of variation values (relative amounts of vmRNA target gene to host β-actin) (44) and fold reductions compared to control (PBS-treated) was observed with early (−4) initiation of treatment at 6 and 30 mg/kg/day, and also with delayed treatment in the 30-mg/kg/day groups. No significant differences were observed with OSE treatment or in the 15-mg/kg/day groups, although a high variability was present among the samples obtained from individual mice in the latter (see Fig. S1 in the supplemental material). However, we note that quantitation of viral RNA species is not a clear indicator of either infectious virus or shedding of transmissible virus (45).

Together, these data show the ability of both prophylactic and therapeutic RO-7 regimens to restrict virus replication and spread within lung tissues. This effect may contribute to the enhanced recovery of these animals that is observed during virus infection. Of note, the reduction in influenza A virus titers by RO-7 was often more pronounced than that observed in OSE-treated animals, especially at early time points.

RO-7 treatment does not lead to the emergence of resistant influenza A or B virus in vivo.

Antiviral resistance can develop during clinical treatment with existing influenza antivirals (NAIs) (13–16), and it has been documented in clinical trials with experimental polymerase inhibitors, such as VX-787/JNJ872 (22). To determine whether RO-7 treatment of mice inoculated with influenza virus leads to the development of PA mutations that give rise to antiviral resistance, we performed genotypic analysis of the endonuclease domain of the PA genes from influenza viruses isolated from lung homogenates obtained at 6 dpi from all experimental groups (−4, +24, and +48 hpi). Molecular markers for PA inhibitor resistance are not well defined; therefore, we screened PA sequences for the presence of substitutions previously reported to reduce susceptibility to other PA inhibitors. These include I79L, F105S, and E119D, which were identified from mutagenic analyses, and T20A, which was identified by passaging of A/Puerto Rico/8/1934 (H1N1) virus in MDCK cells under drug pressure (46). We also checked for the presence of substitutions at the key catalytic-site residues that are conserved among influenza A and B viruses (H41, E80, D108, E119, and K134) (35).The stock mouse-adapted A/California/04/2009 (H1N1)pdm09 virus and all viruses isolated from lung homogenates contained the T20A PA substitution that increases resistance to the PA inhibitor L-742,001 by 3-fold (46), but this substitution is also found in the wild-type A/California/04/2009 (H1N1)pdm09 virus (GenBank accession no. ACP41104). No other molecular markers of resistance were identified in the input virus inoculum or in any lung homogenate isolates for the A/California/04/2009 (H1N1)pdm09 (Table 3) or B/Brisbane/60/2008 virus (data not shown).

TABLE 3.

Genotypic and phenotypic analysis of influenza A viruses isolated from lung homogenates of BALB/c mice treated with RO-7

Treatment group	Mouse	Treatment initiation (hpi)	Genotypic analysis, PA protein endonuclease domaina								Phenotypic analysis, mean EC50 (nM)b
			20c,e	41d,e	79b,e	80d	105b,e	108d	119d,e	134d,e
Consensus A/California/04/2009 (H1N1)pdm09	NAf	NA	T/A	H	I	E	F	N	E	K	NA
Virus inoculum, mouse-adapted A/California/04/2009 (H1N1)pdm09	NA	NA	T/A	H	I	E	F	N	E	K	28.0
Control	1	−4	A	H	I	E	F	N	E	K	4.2
	2	−4	A	H	I	E	F	N	E	K	4.2
	3	−4	A	H	I	E	F	N	E	K	28.9
Avg											12.4
RO-7 (15 mg/kg/day)	1	−4	A	H	I	E	F	N	E	K	6.9
	2	−4	A	H	I	E	F	N	E	K	1.9
	3	−4	A	H	I	E	F	N	E	K	2.7
Avg											3.8
RO-7 (30 mg/kg/day)	1	−4	A	H	I	E	F	N	E	K	13.7
	2	−4	A	H	I	E	F	N	E	K	6.5
	3	−4	A	H	I	E	F	N	E	K	6.4
Avg											8.9

^aSequence analysis of the PA protein was performed with influenza A/California/04/2009 (H1N1)pdm09 viruses isolated from lung homogenates of BALB/c mice treated with RO-7 (15 or 30 mg/kg/day) or control (PBS). Amino acid numbering is based on the influenza A PA protein (35).

^bThe susceptibility to RO-7 of influenza/California/04/2009 (H1N1)pdm09 viruses isolated from lung homogenates (6 dpi) of BALB/c mice either drug (15 and 30 mg/kg/day)- or control (PBS)-treated was determined by a cell viability assay. MDCK cells were infected with viruses at an MOI of 0.008 log₁₀ TCID₅₀/ml (n = ≥6 wells/drug concentration/virus), and the cell viability was determined at 72 hpi by a CellTiter-Glo luminescent cell viability assay. Values represent means from two independent experiments performed in triplicate.

^cPA amino acid substitutions were identified previously by serial passage of A/Puerto Rico/8/1934 (H1N1) virus in MDCK cells under the pressure of L-742,001 at 2 to 32 μg/ml and shown to possess reduced drug susceptibility (46).

^dResidues forming the catalytic domain of the PA protein (35).

^ePA amino acid substitution was identified previously by directed mutational analysis and shown to possess reduced L-742,001 drug susceptibility (53).

^fNA, not applicable.

We next examined the phenotypic susceptibility to RO-7 of selected virus-containing lung-homogenate samples in a cell-viability assay (39). We tested whether A/California/04/2009 (H1N1)pdm09 viruses present in the lung homogenates of mice that received antiviral prophylaxis at 15 or 30 mg/kg/day retained phenotypic susceptibility to RO-7. RO-7 inhibited the replication of the A/California/04/2009 (H1N1)pdm09 virus in the nanomolar range (EC₅₀ = 28 nM), and tested viruses from the lung homogenates were also inhibited by RO-7 at similar or lower EC₅₀ concentrations (EC₅₀ = 3.8 to 12.4 nM) (Table 3). A degree of variation was seen among EC₅₀s obtained with the lung-homogenate samples from the individual mice (in some cases severalfold lower than the input virus). It should be noted that these lung samples are representative of a quasispecies which may contain virus populations with slight variability in fitness and growth kinetics. Testing the phenotypic susceptibility of the virus-containing lung-homogenate is more representative of susceptibility of the in vivo, infectious virus population than if a more homogenous (i.e., tissue culture-grown virus stock) were used. Taken together, these data suggest that RO-7 treatment of mice does not readily induce the emergence of phenotypic or genotypic resistant variants of influenza A or genotypic resistant variants of influenza B viruses.

RO-7 treatment does not prevent the development of anti-HA antibody responses in mice inoculated with influenza A or B virus.

To evaluate the impact of RO-7 treatment on the induction of humoral immunity to the influenza A and B viruses used in the study, serum hemagglutination inhibition (HI) titers were measured at 21 dpi. An HI titer of at least 1:40 is often used as a correlate of protection from influenza virus infection in humans (47). All surviving mice in this study had HI titers of at least 1:40 to the challenge virus (Fig. 6). The prophylaxis regimen elicited an RO-7 dose-dependent decrease in HI titers (Fig. 6A and D), probably due to a sustained decrease in the amount of viral antigen present with early initiation of treatment. A similar, but less prominent, decrease was noted with one therapeutic regimen (+24 hpi) during influenza B virus challenge (Fig. 6E). When treatment was delayed until +24 hpi with influenza A virus challenge or until +48 hpi with either virus challenge, we observed no differences in the HI titers (Fig. 6B, C, and F). Thus, treatment with RO-7 does not prevent the development of humoral immunity, but the magnitude of the response is dependent upon the dosage administered.

FIG 6.

RO-7 does not prevent the development of anti-HA antibody responses in mice inoculated with influenza A or B virus. At 21 days after virus inoculation, the HI titers were determined from the surviving mice inoculated with mouse-adapted A/California/04/2009 (H1N1)pdm09 (A to C) or B/Brisbane/60/2008 viruses (D to F). The HI titers are expressed as reciprocal values, e.g., 40 versus 1:40.

DISCUSSION

Small-molecule inhibitors of the influenza virus PA protein are an attractive option for developing new virus protein-targeted drugs for a number of reasons: (i) the PA endonuclease activity is essential for influenza virus replication; (ii) critical residues in the PA protein enzymatic site are conserved among influenza A and B viruses, suggesting potential for broad-spectrum activity; and (iii) a high barrier to emergence of resistant variants is expected. In the present study, we evaluated the antiviral potency of the PA endonuclease inhibitor RO-7 (39) against lethal infections with influenza A or B virus in mice, providing in vivo data that is important to the drug development process. We used BALB/c mice, which are the most widely used and accepted small-animal model for testing anti-influenza drug efficacy (48, 49). Furthermore, we used two challenge viruses [mouse-adapted A/California/04/2009 (H1N1)pdm09 and B/Brisbane/60/2008] that are representative of currently circulating human influenza viruses and contemporary vaccine composition (50).

We demonstrated that prophylactic RO-7 administration completely prevented the mortality associated with influenza A or B virus infection in mice. Initiating RO-7 treatment at 24 or 48 hpi (the therapeutic regimens) improved the survival of mice inoculated with a lethal dose of influenza A or B virus. The advantages of RO-7 therapy were dose dependent, and higher dosages (15 or 30 mg/kg/day) were more effective at reducing pulmonary pathology, particularly HM formation, and improving the survival of mice. Therapeutic RO-7 regimens had similar effectiveness to OSE regimens at reducing the morbidity of mice challenged with influenza A/California/04/2009 (H1N1)pdm09 virus, but they offered advantages over treatment with the NAI OSE, particularly in terms of lowering viral titers and increased survival outcomes of mice infected with influenza B virus. To our knowledge, this is the first in vivo study to demonstrate the efficacy of an endonuclease inhibitor against a clinically relevant influenza B virus challenge.

There is an overall low amino acid sequence similarity (approximately 36 to 38%) between the PA proteins of influenza A and B viruses (51, 52). However, all five of the key catalytic-site residues necessary for enzymatic activity (H41, E80, D108, E119, and K134) are conserved (35). It is therefore not surprising that RO-7, along with other PA endonuclease inhibitors, showed activity against both influenza A and B viruses in vitro (39) and, as we have now demonstrated, in vivo. The broad-spectrum activity of RO-7 and other endonuclease inhibitors against influenza A and B viruses also makes them more advantageous than the current cap-binding PB2 polymerase inhibitors, which are capable of inhibiting infections by influenza A viruses only (33, 34).

The emergence of drug-resistant variants is a major concern with drugs that target specific virus proteins. To date, the emergence of PA inhibitor resistance has not been examined in vivo. Two groups examined the emergence of PA endonuclease inhibitor-resistant viruses during serial passages in vitro. Nakazawa et al. reported the emergence of a T20A PA substitution known to increase EC₅₀s 3-fold for the first-generation PA inhibitor L-742,001, but Song and coworkers were unable to identify any L-742,001 resistance by this method (46, 53). Both Song and Stevaert and their respective colleagues used a mutational library or directed mutagenesis to identify potential resistance markers. These groups identified residues within critical metal-binding motifs necessary for enzymatic activity (including H41A, G81T/F, and E119D) that increased the 50% inhibitory concentrations of L-742,001 in the minireplicon assay and the susceptibility to the drug in MDCK cells (53, 54). However, additional studies are required to determine how these substitutions could affect influenza B virus fitness, taking into consideration the conserved nature of the PA endonuclease domain among influenza A and B genera (36). We did not detect any previously reported molecular markers of resistance in the PA endonuclease domain of the influenza A and B viruses isolated from the lungs of RO-7-treated mice at 6 dpi, nor did we detect any differences in the overall sequence identity between the RO-7-treated groups and the untreated animals. In phenotypic assays, viruses present in lung homogenates of mice treated with RO-7 remained susceptible to RO-7 in the nanomolar range. Thus, if resistance mutations did arise during RO-7 treatment of mice, the variants did not predominate to a detectable degree. The limitation of our study is its use of the Sanger platform method for identifying resistance markers. Minor populations of resistant variants among viral quasispecies within samples may not be identified using our approach (albeit unlikely to be at significant frequency or be fit enough given the phenotypic results). Therefore, additional studies may be required using a next-generation sequencing approach.

To fully realize the potential of RO-7 as an antiviral drug, additional parameters not included in this study should be addressed. Utilization of an increased number of animals per experimental group will result in more robust statistics. Although we often observed dose-dependent responses with increasing RO-7 dosages, the optimal dosage may be higher than those used in our experiments. Assessments of cytotoxicity and pharmacokinetics in an expanded dosing study will aid in determining a selective index for RO-7 that can be used to inform future in vivo experiments. Furthermore, alternative routes of administration and extending the treatment window (particularly for influenza B virus infection) are options that may increase RO-7 efficacy but have yet to be explored. Finally, it will be important to verify our in vitro data (39) demonstrating RO-7 efficacy in the face of common NAI resistance-associated substitutions among influenza A and B viruses in the mouse model.

The current literature and ongoing clinical trials suggest a clear path for the use of influenza polymerase protein inhibitors as antiviral drugs. The PB1 protein-targeted drug T-705 has been approved for limited-use influenza therapy (in pandemic scenarios) in Japan for over 2 years (12, 34), and phase III clinical trials in the United States have recently been completed (55). Phase II clinical trials are under way with the cap-binding influenza PB2 protein inhibitor VX-787/JNJ872 and include combination therapy with NAI OSE. Finally, two endonuclease inhibitors (S-033188/Roche and AL-749/Janssen) are in phase I or II trials in Japan and the United States (34, 56). Safety and pharmacologic studies have been completed, but thorough proof-of-concept data have yet to be published (34). Despite the successful entry of these two endonuclease inhibitors into clinical trials, their structure and other preclinical data have not been disclosed. The development and refinement of the NAI class of antivirals (including the identification of NAI variants with enhanced activity, increased NA protein binding, and better bioavailability, and the development of multivalent analogues and NAIs conjugated to photodynamic and anti-inflammatory compounds) continues, despite the clinical availability of these drugs for two decades (57, 58). Similarly, it is important to continue the study of other PA endonuclease inhibitors, including RO-7, with a view to future clinical trials.

Although the broad-spectrum activity of RO-7 and other PA endonuclease inhibitors against influenza A and B viruses gives them an advantage over NAIs and cap-binding PB2 inhibitors, combination therapy with one or more members of these latter classes of inhibitor may provide enhanced protection overall. In vitro studies have demonstrated synergistic effects between NAIs and the cap-binding inhibitors VX-787/JNJ872 (33) and ANA-0 (41). To date, such an effect has not been demonstrated with RO-7 or other endonuclease inhibitors. Examination of drug interactions in vitro, along with the in vivo efficacy of RO-7 against OSE-resistant viruses and combination therapy (with NAIs or cap-binding inhibitors), is a logical step in the further development of this drug and has implications for reducing the emergence of resistant variants, extending the window of treatment efficacy, and providing additional opportunities to treat severe influenza virus infections.

In conclusion, RO-7 is an additional candidate among the existing PA endonuclease inhibitors with demonstrated in vivo activity against both influenza A and B viruses in the mouse model. The ability of RO-7 to decrease morbidity and mortality in virus-infected BALB/c mice, as well as reduce virus replication in the lung and pulmonary pathology, without inducing detectable antiviral resistance suggests it has potential for further development for clinical applications.

MATERIALS AND METHODS

Viruses.

Mouse-adapted influenza A/California/04/2009 (H1N1)pdm09 virus (59) was obtained from the St. Jude Children's Research Hospital influenza virus repository and propagated for 48 h at 37°C in Madin-Darby canine kidney (MDCK) cells in serum-free modified Eagle medium (MEM) containing TPCK (tolylsulfonyl phenylalanyl chloromethyl ketone)-treated trypsin (1 μg/ml; Worthington, Lakewood, NJ). Influenza B/Brisbane/60/2008 virus was provided by the Influenza Division at the Centers for Disease Control and Prevention (Atlanta, GA) and was grown for 72 h at 33°C in the allantoic cavity of 10-day-old embryonated chicken eggs (Marshall Durbin, Birmingham, AL).

Cells.

MDCK cells (American Type Culture Collection, Manassas, VA) were maintained in MEM (Cellgro; Corning, Manassas, VA) supplemented with 10% fetal calf serum (HyClone, Logan, UT).

Antiviral drugs.

RO-7 and oseltamivir phosphate (OSE) were provided by Hoffmann-La Roche (Basel, Switzerland). For in vitro studies, RO-7 was prepared as a 10 mM stock in dimethyl sulfoxide (DMSO) and was soluble when diluted in cell culture medium. For animal experiments, RO-7 was dissolved in vehicle containing 3% DMSO in 1× PBS (Cellgro). OSE was dissolved in distilled water. All drugs were stored at 4°C for the duration of the experiment pending use.

Animals and ethics statement.

Six- to eight-week-old female BALB/c mice (weight, 18 to 20 g; Jackson Laboratories, Bar Harbor, ME) were used for all experiments. The animal experiments were approved by the St. Jude Children's Research Hospital Animal Care and Use Committee in compliance with the National Institutes of Health and the Animal Welfare Act.

Drug safety.

BALB/c mice were treated with RO-7 (200 μl, administered intraperitoneally [i.p.], which provides 65% bioavailability [unpublished data]), OSE (100 μl, administered orally), or PBS (200 μl, administered i.p.) twice daily for 5 days. Animals were monitored for morbidity (weight loss) and mortality for 18 dpi.

Animal infections and drug treatments.

Mice (n = 5/group) were lightly anesthetized with isoflurane and inoculated intranasally with five 50% mouse lethal doses (MLD₅₀) of A/California/04/2009 (H1N1)pdm09 (1.1 × 10³ 50% tissue culture infectious doses [TCID₅₀]) or B/Brisbane/60/2008 (2.5 × 10⁵ TCID₅₀) virus. Mice in prophylaxis groups received RO-7 (6, 15, or 30 mg/kg/day) or OSE (20 mg/kg/day) 4 h before virus inoculation (−4 hpi), 8 h after virus inoculation (+8 hpi), and twice daily (BID) for 4 days. Mice in therapeutic groups received the same drug dosages BID for 5 days, beginning at 24 or 48 h after virus inoculation (+24 hpi and +48 hpi, respectively). Morbidity, clinical signs of disease (ruffled fur, hunched posture, and/or lethargy), and mortality were monitored for 18 dpi. The weight gain or loss was calculated for each mouse as a percentage of its weight before virus inoculation. Animals that showed signs of severe disease and/or experienced 25% weight loss were sacrificed.

Virus isolation.

At 3, 6, and 9 dpi, mice (n = 3/group) were sacrificed, and their lungs were homogenized in 1 ml of sterile PBS with penicillin-streptomycin/amphotericin B (Sigma-Aldrich, St. Louis, MO) and titrated on MDCK cells. The TCID₅₀/ml values were calculated by the method of Reed and Muench (60).

Viral mRNA (vmRNA) analysis.

Total RNA was isolated from mouse lung homogenates (RNeasy; Qiagen, Valencia, CA) and was used (200 ng) to synthesize cDNAs complementary to NA gene vmRNA or host β-actin with gene-specific primers as described previously (39, 61). Quantitative PCR (qPCR) with Power SYBR green PCR master mix (Invitrogen) was performed using the cDNA templates. The coefficient of variation (44) of vmRNA to β-actin was calculated from the cycle threshold values (ΔC_T) as follows: 2^{−(vmRNA ΔCT − β-actin ΔCT)}. The fold vmRNA reduction was calculated as a comparison to the control group (virus infected, no drug treatment).

Sequencing.

Total RNA was isolated from virus-containing mouse lung homogenates obtained from different experimental groups (in which treatment was initiated at −4, +24, or +48 hpi) at 6 dpi by using an RNeasy kit (Qiagen). The first 1.7 kb (or less) of the PA gene of influenza A or the first 1.2 kb (or less) of the PA gene of influenza B virus was amplified by a one-step reverse transcription-PCR (RT-PCR) assay (Qiagen) with gene-specific primers. Sequencing of the gel-extracted products was accomplished with overlapping primer pairs and the Sanger platform (by the Hartwell Center for Bioinformatics and Biotechnology at St. Jude Children's Research Hospital), with the 209-residue endonuclease domain being analyzed with SeqMan Pro (DNASTAR, Madison, WI), MEGA 6 (62), and BioEdit (v7.0.9; Ibis Biosciences, Carlsbad, CA) software. cDNA product was not amplifiable from influenza B lung homogenates from the following four experimental groups: OSE, +24 hpi (n = 1); OSE, +48 hpi (n = 1); RO-7, 30 mg/kg/day, −4 hpi (n = 3); and RO-7, 30 mg/kg/day, +48 hpi (n = 3).

Cell viability assay.

MDCK cells (1.5 × 10⁴) were plated in 96-well plates and inoculated (at a multiplicity of infection of 0.008) with influenza A virus-containing lung homogenates of mice treated with RO-7 (15 and 30 mg/kg/day). The cell viability in the presence of RO-7 (0 to 5 μM) was measured with a CellTiter-Glo luminescent assay (Promega, Madison, WI) at 72 hpi. The viability of non-virus-inoculated wells was normalized to 100% viability, and the percentage reduction was calculated for each RO-7 treatment group. The 50% effective concentrations (EC₅₀) were determined by using the log (inhibitor) versus response logistic nonlinear regression equation in Prism 6.0 software (GraphPad Software, La Jolla, CA). The data are representative of the mean values from two independent experiments.

Histology.

At 8 dpi, mice (n = 3/group) were sacrificed, and their lungs were fixed in 10% neutral buffered formalin. Paraffin-embedded tissue sections were stained with hematoxylin and eosin by the Veterinary Pathology Core at St. Jude Children's Research Hospital. The severity and extent of specific pulmonary damage, as determined by the quantitation of hyaline membrane (HM) formation, was assessed and scored in a blinded manner. The scoring of HM formation was assessed based on the following grades: 0, no lesions; 1, minimal, rare, or inconspicuous HM; 2, mild multifocal, small focal, or widely separated HM; 3, moderate, multifocal, and prominent HM; 4, marked, extensive-to-coalescing areas with HM; and 5, severe and extensive HM formation. The severity grades were converted to semiquantitative scores as follows: 0 = 0, 1 = 1, 2 = 15, 3 = 40, 4 = 80, and 5 = 100.

Serology.

Serum samples were obtained 21 days after mice were inoculated with influenza A or B virus. The sera were treated with receptor-destroying enzyme (Denka-Seiken, Chuo-ku, Tokyo), heat inactivated at 56°C for 1 h, and tested for the presence of anti-hemagglutinin (HA) antibodies by the HA inhibition assay with 0.5% turkey red blood cells (Rockland Immunochemicals), as described previously (63). The reciprocal of the last serum dilution that inhibited hemagglutination was recorded as the HI titer.

Statistical analysis.

Data were analyzed by using GraphPad Prism 6.0 software (La Jolla, CA), with individual significance being determined by unpaired t tests and/or one-way analysis of variance.