Influenza Antivirals and Their Role in Pandemic Preparedness

Abstract

Effective antivirals provide crucial benefits during the early phase of an influenza pandemic, when vaccines are still being developed and manufactured. Currently, two classes of viral protein–targeting drugs, neuraminidase inhibitors and polymerase inhibitors, are approved for influenza treatment and post-exposure prophylaxis. Resistance to both classes has been documented, highlighting the need to develop novel antiviral options that may include both viral and host-targeted inhibitors. Such efforts will form the basis of management of seasonal influenza infections and of strategic planning for future influenza pandemics. This review focuses on the two classes of approved antivirals, their drawbacks, and ongoing work to characterize novel agents or combination therapy approaches to address these shortcomings. The importance of these topics in the ongoing process of influenza pandemic planning is also discussed.

1.0. The influenza antiviral landscape

1.1. Overview

Pandemic influenza viruses with novel antigenicity and the capacity for significant public health and socio-economic impacts have emerged periodically throughout history. Future influenza pandemics are difficult to predict with regard to viral strain, time of onset, or geographic origin (Monto and Fukuda, 2020). Effective antiviral drugs targeting conserved viral proteins may serve as first-line countermeasures against emerging pandemic influenza strains. Properly implemented, they can reduce viral loads, may limit transmission, and help manage severe infections and thereby reduce morbidity and mortality until effective and antigenically matched vaccines are available.

1.2. Influenza viruses with pandemic potential

Origination of influenza pandemics is influenced by the complex ecology, broad host species range, and segmented genome of influenza viruses. Influenza A virus subtypes, whose nomenclature is taken from numbering of antigenically distinct surface glycoproteins hemagglutinin (HA) and neuraminidase (NA), are maintained in avian reservoirs including aquatic bird species (Monto, 2013; Webster et al., 1997). From this reservoir, the viruses may transmit to other migratory birds, domestic poultry, swine, and a variety of other mammals. The eight segments of influenza A virus genomes can lead to re-assortment of RNA segments in co-infected cells, broadening the potential for generation of novel viruses. Occasionally, zoonotic transmission of novel viruses into humans with no prior immunity has led to pandemic scenarios. The most recent pandemic, caused by the “triple reassortant” A(H1N1)pdm09 subtype, was composed of gene segments originally derived from avian, swine, and human viruses (Smith et al., 2009). Sporadically, avian influenza viruses can be directly transmitted into humans and cause severe disease outcomes (Wang, 2013). Subtypes A(H5Nx), A(H7N9), A(H9N2), A(H10Nx) are among the candidates exhibiting these characteristics and elevating their pandemic potential (Sutton, 2018; Taubenberger et al., 2019). As a component of pandemic preparedness, avian and swine influenza viruses are routinely evaluated using risk assessment tools developed by the World Health Organization (WHO, Tool for influenza pandemic risk assessment, TIPRA) (Global Influenza Programme WEP, 2020) and the Centers for Disease Control and Prevention (CDC, Influenza risk assessment tool, IRAT) (Centers for Disease Control and Prevention, 2020). By evaluating virus phenotypic properties (receptor binding specificity, transmissibility in animals, etc.), attributes of the human population (disease severity, population immunity, etc.), and viral ecology and epidemiology in non-human hosts (species of origin, diversity of species naturally susceptible and sustaining the virus, and geographic distribution of these species and viruses with consideration of their proximity to humans), each virus is scored on its likelihood of emergence and the likely impact if it causes a pandemic. Antiviral susceptibility is another important component of the risk assessment algorithm that predicts this likely impact.

1.3. Use of antiviral drugs against seasonal influenza

Historically, four classes of antivirals have been approved for treating seasonal or pandemic influenza. The first of these, the adamantanes (amantadine, rimantadine), were identified in the 1970s. They act primarily by inhibiting the M2 ion channel protein responsible for pH control of the endosome, which is critical for uncoating and release of viral genomes into the infected cell. Adamantanes are no longer recommended for clinical use, as circulating seasonal viruses are naturally resistant (Hurt, 2014). Favipiravir, a nucleoside analogue that interferes with the viral RNA-dependent RNA polymerase, received conditional approval in Japan in 2014 for treating emerging influenza viruses that are resistant to the available antivirals, but it is not approved for routine treatment of seasonal influenza. This leaves treatment of seasonal influenza to the NA inhibitors (NAIs) approved in 1999, and the cap-dependent endonuclease inhibitor (CENI) baloxavir marboxil (BXM), which was approved in 2018. Because of their lower cost and longer clinical use and market availability, NAIs are currently used more widely than BXM. These two drug classes will be the basis for antiviral support during a pandemic in the near future.

2.0. Neuraminidase inhibitors

2.1. Overview

NAIs target the NA glycoprotein, the most abundant surface glycoprotein after HA, by inhibiting the enzymatic activity necessary for budding of newly emerged virions from host cells and preventing virus spread (Gubareva and Mohan, 2022). Four NAIs (oseltamivir, zanamivir, peramivir, and laninamivir) are available for treating influenza (Figure) and should be administered within the first 48 hours of symptom onset. Although structurally similar, they differ with respect to treatment regimens, administration routes, and resistance profiles (Tables 1, 2). The most widely used NAI is oseltamivir (Tamiflu), for which the therapeutic regimen in adults is a 75-mg dose, twice daily, for 5 days. Because zanamivir (Relenza) has low (<5%) bioavailability, it was approved as an inhaled drug administered as a 5-mg dose, twice daily, for 5 days. In 2019, intravenous (IV) zanamivir (Dectova) received European Union authorization for use under exceptional circumstances (e.g., in cases resistant to other anti-influenza drugs). Efforts to increase NAI bioavailability and half-life led to the development of peramivir (Rapivab/Rapiacta/PeramiFlu), administered by IV infusion as a single 600-mg dose (McLaughlin et al., 2015), and laninamivir (Inavir), a long-lasting NAI that showed therapeutic efficacy after a single nasally inhaled dose (Sugaya and Ohashi, 2010). Peramivir and laninamivir are approved for use in a limited number of countries globally, while oseltamivir and zanamivir have geographically wider approvals.

Influenza antiviral drugs in clinical-stage active development that have received post-marketing approval in the United States. Additional drugs marked as having received conditional approval (Inavir—Japan only) or limited approval (Arbidol—Russia and China only; Inavir—Japan only). The compounds are indicated by name with the company sponsoring their development; color-coded boxes indicate the route of administration.

CENI—cap-dependent endonuclease inhibitor targeting the PA protein; NAI—neuraminidase inhibitor targeting the NA protein; mAb—monoclonal antibody targeting the HA protein; MEK inhibitor—mitogen-activated protein kinase; TLR—toll-like receptor; NSAID—non-steroidal anti-inflammatory drug; MOA—monoamine oxidase inhibitor.

Table 1:

Antiviral Drugs for Treating Influenza Virus Infection

Inhibitor	Trade name	Drug class	Viral protein target	Mechanism of action	Clinically relevant resistance (influenza A viruses)
Amantadine Rimantadine	Amantadine: Symmetrel Rimantadine: Flumadine	Adamantane	M2 (influenza A only)	Interferes with virion and endosomal acidification. Inhibits downstream HA conformation change, endosomal fusion, and release of viral genomes into the cytoplasm	M2 protein: S31N (predominant), L26F, V27A, A30T, G34E
Oseltamivir phosphate	Tamiflu	Neuraminidase inhibitor	NA (influenza A and B)	Blocks NA enzymatic cleavage of host cell sialic acid receptors. Inhibits progeny virus budding	NA protein: E119D/G/I/V, I222R, S246N, H274Y, R292K, N294S
Zanamivir	Relenza				NA protein: E119D/G, I222R, S246G/R, R292K, E119G+H274Y
Peramivir	Rapivab/Rapiacta/PeramiFlu				NA protein: S246G/R, H274Y, R292K, G147R+H274Y
Laninamivir	Inavir			A longer-lasting formulation of this class of drug	NA protein: G147R+H274Y
Favipiravir	Avigan	Nucleoside analogue	PB1 (influenza A and B)	Preferentially incorporated by PB1 into viral RNA. Leads to chain elongation termination and/or lethal mutagenesis	N/A
Baloxavir marboxil	Xofluza	Cap-dependent endonuclease inhibitor	PA (influenza A and B)	Blocks PA endonuclease activity necessary to cleave PB2-bound, capped host mRNAs. Halts viral mRNA transcription	PA protein: I38T (predominant), I38M/F/L/N/S, E23G/K, A37T, E199G

N2 NA numbering.

N/A, not applicable.

Table 2:

Influenza Antiviral Administration and Potential Pandemic Application

Inhibitor	Indicationa	Route of administration	Therapeutic dose & regimen (adult)a	Potential pandemic applications WHO guidelines for patients at risk for severe influenza illness (seasonal, pandemic, or zoonotic virus)b	Summary of expert opinionsc–h
Amantadine Rimantadine	Currently not recommended for treatment or prophylaxis	Oral	Amantadine:100 mg, twice daily, 7 days Rimantadine: 100 mg, twice daily, 7 days	Not reviewed in current GDG report	Unlikely pandemic use. Resistance is prevalent in seasonal viruses and is readily acquired post treatment. Some zoonotic viruses (swine, avian) remain susceptible, so reevaluation may be justified if the pandemic virus emerges from these reservoirs or in a study of combination therapy.
Oseltamivir phosphate	Prophylaxis (≥ 1 year of age) Treatment (≥ 2 weeks of age)	Oral (also as oral suspension)	75 mg, twice daily, 5 days	Oral oseltamivir is recommended. Treatment should be initiated early after symptoms onset. Oseltamivir may limit hospitalizations and mortality. Treatment should be discontinued if a sensitive diagnostic test indicates no evidence of infection.	Likely to continue to be stockpiled and to be the major or primary antiviral agent in a pandemic because of oral administration, availability of generics, long shelf-life, and two decades of clinical application. Resistance risk is low, but not absent. Early commencement of treatment (within 48 h of symptom onset) is critical.
Zanamivir	Prophylaxis (≥ 5 years of age) Treatment (≥ 7 years of age)	Inhaled Intravenous (emergency use authorization)	5 mg inhaled, twice daily, 5 days	Inhaled zanamivir is not recommended. Insufficient evidence to suggest inhaled zanamivir will limit hospitalizations and mortality.	Unlikely to be widely used or stockpiled because of route of administration. Risk of resistance may be lower than for oseltamivir or peramivir.
Peramivir	Treatment (> 6 months)	Intravenous	600 mg IV infusion (≥ 15 min)	Intravenous peramivir is not recommended. Data is absent or uncertain if peramivir will limit hospitalizations and mortality.	Unlikely to be widely used or stockpiled because of route of administration and limited approval. May be limited to hospitalized settings, which could be beneficial depending on the disease severity of the circulating virus. Insufficient data exists to understand efficacy in patients with severe influenza.
Laninamivir	Prophylaxis Treatment	Inhaled	20 mg, 2 doses inhaled consecutively	Inhaled laninamivir is not recommended. No evidence to suggest laninamivir will limit hospitalizations and mortality. Recommendation does not apply to a situation involving oseltamivir-resistant virus.	Unlikely to be used in a pandemic. Approval limited to Japan for seasonal virus prophylaxis.
Favipiravir	Treatment (adults only, for novel or re-emerging influenza only)	Oral	1600 mg, twice daily, 1 day 600 mg, twice daily, 4 days	Not reviewed in current GDG report	May be used in a pandemic, especially under conditions in which a virus is resistant to NAIs, but currently approved only in Japan. Preclinical teratogenicity will exclude certain risk groups.
Baloxavir marboxil	Prophylaxis (≥ 5 years of age, post-exposure criteria) Treatment (> 5 years of age)	Oral (also as oral suspension)	20–80 kg body weight: 40 mg, once ≥80 kg body weight: 80 mg, once	Guidelines have not been updated due to lack of pertinent trials when the review was initially conducted.	Next most likely to be stockpiled and/or widely used during a pandemic because of single-dose oral administration, non-inferiority to NAIs with greater viral load reductions, and potential for post-exposure prophylaxis. Cost per treatment and shorter shelf-life may affect public health decision making. Resistance variants may occur a higher rates than NAIs.

^aIndication, route, and dose as defined in prescribing information from the company that originally developed the drug and/or USFDA.gov. Guidelines are given for adults and may differ for other age groups.

^bWorld Health Organization. Guidelines for the clinical management of severe illness from influenza virus infections. PMID: 35353465

^cConsensus recommendations and discussion from expert reviews

^dUkeyki et al. Influenza. Lancet. 2020. 400 (10353). pp.693-706. (PMID: 36030813)

^eDuwe et al. Prophylaxis and treatment of influenza: options, antiviral susceptibility, and existing recommendations. GMS Infectious Diseases. 2021. 9 (Doc02). (PMID: 34113534)

^fHurt et al. Antiviral therapy for the next influenza pandemic. Tropical Medicine and Infectious Diseases. 2019. 4(2). (PMID: 31003518)

^gBeard et al. Treatment of influenza with neuraminidase inhibitors. Current Opinion in Infectious Diseases. 31(6). pp.514-519. (PMID: 30320639)

^hHayden & Shindo. Influenza virus polymerase inhibitors in clinical development. Current Opinion in Infectious Diseases. 2019. 32(2). pp.176-186. (PMID: 30724789)

2.2. Challenges confronting NAI usage and emergence of resistance

Various NAIs display different therapeutic benefits and efficacy dependent upon the subtype/lineage of influenza virus infecting the patient. Therapeutic benefits of oseltamivir for seasonal uncomplicated influenza caused by A(H1N1), A(H3N2) and B viruses include reductions of approximately 24 hours in the time to alleviation of symptoms (TTAS), resumption of usual activities, decrease in fever duration, along with decreases in illness severity, ancillary medication use, complication occurrence, frequency of antibiotic prescriptions for lower respiratory tract complications, and hospitalization rates (Hayden et al., 1999; Kaiser et al., 2003; Treanor et al., 2000). Meta-analysis of nine randomized controlled trials (RCTs) demonstrated accelerated clinical symptom alleviation in oseltamivir-treated vs. placebo-treated groups and a reduced risk of lower respiratory tract complications and hospital admission (Dobson et al., 2015). Inhaled zanamivir application is limited by the necessity to use a breath-activated plastic device, with additional care required in patients with asthma, chronic respiratory disease, immune deficiencies, in pregnant and lactating women, in elderly patients. However, IV zanamivir provides reliable drug delivery in critically ill patients and can be an important option for treatment of serious illnesses (Beigel and Hayden, 2021). However, uncertainties remain as to the effectiveness of NAIs (Jefferson et al., 2009; Jefferson et al., 2006). A meta-analysis of 20 RCTs of oseltamivir and zanamivir (as prophylaxis in four trials, as treatment in 12, and as post-exposure prophylaxis in four) in otherwise-healthy adults exposed to naturally occurring influenza revealed only modest effectiveness against influenza symptoms (Jefferson et al., 2009) and concluded that independent RCTs were needed to resolve these uncertainties for seasonal influenza infections.

NAIs were deployed in response to the 2009 swine-origin A(H1N1)pdm09 pandemic virus. A meta-analysis showed that outpatient or community-based treatment of patients with confirmed or suspected A(H1N1)pdm09 infection significantly reduced the risk of hospitalization (Venkatesan et al., 2017). NAI treatment (irrespective of timing) also significantly reduced the mortality risk of hospitalized patients, although early treatment (within 48 hours of symptom onset) was associated with a greater reduction in mortality (Muthuri et al., 2014). NAI use during influenza pandemics varies by country. As summarized elsewhere (Berera and Zambon, 2013), the 2009 pandemic showed that the strategy of using NAIs to contain virus spread was ineffective in the United States and United Kingdom, and that early detection of cases with rapid diagnostic tests and universal antiviral treatment in Japan provided better protection, especially for the high-risk population (e.g., pregnant women).

Though limited in number, NAI therapy has been applied to humans infected with influenza viruses with pandemic potential. Highly pathogenic avian influenza (HPAI) A(H5N1) viruses have occurred sporadically since first reported in Hong Kong in 1997 with >50% human mortality rates (Global Influenza Programme WEP, 2022). The vast majority of HPAI A(H5N1) were found to be fully susceptible to the NAIs, with 50% inhibitory concentrations (IC₅₀s) similar to those of commonly circulating human A(H1N1) viruses (Govorkova et al., 2013; Takashita et al., 2022). Multivariate modelling showed a 49% mortality reduction in A(H5N1) virus-infected patients when oseltamivir was started within 6–8 days after symptom onset, and this treatment appears to benefit all age groups (Adisasmito et al., 2010). Experience with oseltamivir in human A(H5N1) influenza suggests that increasing the dose could be more effective (Gambotto et al., 2008; Kandun et al., 2006). Human infections with avian A(H7N9) viruses frequently cause severe illness characterized by pneumonia that rapidly develops into acute respiratory distress syndrome, multiple organ dysfunction, and shock (Gao et al., 2013). No association was found between NAI treatment and survival of A(H7N9) virus-infected patients (Cheng et al., 2021).

The clinical outcome of antiviral therapy can be compromised by the emergence of influenza viruses with reduced susceptibility (resistance) to NAIs. Resistant profiles differ for the various NAIs and NA subtypes. Common treatment-emergent NA substitutions among A(H1N1)pdm09 viruses include NA-E119D/G (N2 numbering), NA-I222R, NA-S246N, and NA-H274Y (clinically relevant) (Baz et al., 2009; L’Huillier et al., 2015; Yates et al., 2016). With viruses carrying NA-H274Y, oseltamivir and peramivir exhibit elevated IC₅₀s, whereas zanamivir and laninamivir have IC₅₀s similar to wild-type virus. Viruses with NA-S246N possess reduced or highly reduced susceptibility to oseltamivir, zanamivir, peramivir, and laninamivir. The HPAI A(H5N1) virus with NA-H274Y also exhibits resistance to oseltamivir and peramivir but remains susceptible to zanamivir (Le et al., 2005). For the A(H3N2) subtype, reduced oseltamivir susceptibility was reported for viruses with NA-E119V, NA-R292K, and NA-N294S, but only NA-R292K affected zanamivir susceptibility (Marty et al., 2017; Okomo-Adhiambo et al., 2010; Takashita et al., 2020). Combined phenotypic and NA sequence–based analysis revealed that the global frequency of viruses displaying reduced or highly reduced inhibition by NAIs remained low at 0.5% for 2018–2019 and 0.6% for 2019–2020 (Govorkova et al., 2022).

2.3. Novel development of NAIs

Currently two novel NAIs and NA blocker are in phase 1 and 2 development (Figure). The NAI HNC042 (ZBD Pharmaceutical, China) shows potent inhibitory activity against wild-type and oseltamivir-resistant NA-H274Y viruses in pre-clinical studies. In a phase 1 trial in the United States, IV HNC042 showed good safety and tolerability profiles and linear pharmacokinetics (NCT04603989). IV HNC042 is intended for preventing and treating influenza, especially in severely ill patients who require hospitalization. Pre-clinical evaluation of AV5080 (ChemRar, Russia, Viriom, USA) revealed activity against a broad range of influenza viruses, including oseltamivir-resistant strains. Oral AV5080 demonstrated good physicochemical and metabolic properties and a favourable pharmacokinetic profile in animals (Ivachtchenko et al., 2014). Multi-centre RCT demonstrated safety and good tolerability of the AV5080 in patients with influenza, no adverse events were observed (NCT05095545). CD388 (Cidara Therapeutics Inc., USA) is a long-acting NA blocker whose structure enables two distinct and complementary mechanisms by stably coupling the NA of influenza virus to an effector domain consisting of a proprietary variant of a human antibody fragment (Fc). This fundamentally new approach to protection from and treatment of viral infections is designed to deliver universal protection throughout an influenza season. By targeting a highly conserved region on the virus, CD388 can potentially protect individuals from all influenza strains, including seasonal and pandemic influenza A, influenza B, and major clinically characterized, drug-resistant viruses (NCT05523089).

3.0. Polymerase inhibitors

3.1. Overview

The influenza virus RNA polymerase is critical for viral gene transcription and genome replication, making it a prime target for directed drug design. The heteromeric complex comprises three proteins, each with distinct functions.

The polymerase acidic protein (PA), with its CEN activity, cleaves away host mRNA caps that can be used for viral gene transcription. CEN-inhibiting compounds including those with dioxobutanoic acid and 2,6-diketopiperazine moieties were first identified in the mid-1990’s (Hastings et al., 1996; Tomassini et al., 1994; Tomassini et al., 1996), but the availability of the PA protein endonuclease domain crystal structure in 2009 has since spurred a rapid increase in drug discovery towards this target (Yuan et al., 2009). However, to date, only one CENI, baloxavir marboxil (Xofluza, BXM), has been approved (2018, Japan, USA; Europe, 2021). BXM is a relatively young drug with no generic equivalent, while oseltamivir remains standard-of-care in many countries. For these reasons, it is difficult to state the demand for BXM vs. NAIs (Dufrasne, 2021). In Japan, where the drug was developed and initially approved, usage is likely higher than other countries, with an estimated 40% of prescribed antivirals in the 2018–2019 season being BXM (Takeuchi and Kawakami, 2021). Compared with oseltamivir, BXM has demonstrated similar TTAS and greater viral load reductions in healthy adults vs. oseltamivir (Hayden et al., 2018) and similar TTAS vs. oseltamivir in high-risk adults (Ison et al., 2020). It has also been used for post-exposure prophylaxis for household contacts of infected individuals (Ikematsu et al., 2020), and a retrospective study showed possible reductions in hospitalizations with BXM vs. NAIs (Komeda et al., 2021). A significant benefit of BXM over NAIs is its single weight-adjusted (40 or 80 mg) oral dose regimen (Tables 1, 2). However, phase 2 and 3 trials identified PA-I38X and PA-E23G/K substitutions (present in 2.9%-9.7% of treated individuals) that reduce BXM efficacy and could lead to virus rebound (Uehara et al., 2020). Additional PA endonuclease domain substitutions have been sporadically implicated in BXM reduced susceptibility (Govorkova et al., 2022; Ison et al., 2021), and in vitro studies identified non-endonuclease domain substitutions that occur in combination with PA-I38T during CENI drug passage (Jones et al., 2018). Therefore, it is important to consider that additional, undescribed PA gene substitutions may ultimately affect the efficacy of BXM and the CENI class of drugs. Despite the low barrier to resistance, BXM is highly effective and could be leveraged in a pandemic, considering its pan-influenza efficacy and it’s in vitro/in vivo inhibition of multiple zoonotic H5, H7, and H9 subtypes (Takashita et al., 2022; Taniguchi et al., 2022).

The polymerase basic protein 1 (PB1) is an RNA-dependent RNA polymerase that can use PA-stolen caps as primers to transcribe viral mRNA. Later during the viral cycle, PB1 replicates the viral genome. Favipiravir (Avigan, T-705) is a prodrug whose active metabolite functions as a purine nucleoside analogue targeting various RNA viruses (Furuta et al., 2017) including influenza (A, B, and C). PB1 preferentially incorporates the drug into the growing RNA strand, causing chain termination (Sangawa et al., 2013; Wang et al., 2021) and/or lethal mutagenesis (Baranovich et al., 2013). Favipiravir demonstrated reduced TTAS vs. placebo in one phase 3 trial but failed to find a statistically significant difference in TTAS in the second confirmatory study (Hayden and Shindo, 2019). Favipiravir’s complex pharmacokinetics (Hayden and Shindo, 2019) necessitate a high loading dose (1600 mg, twice daily, 1 day; 600 mg, twice daily, 4 days) that is not fully defined for acute/severe influenza (Wang et al., 2020b). Although PB1-K229R has been identified as a marker for reduced susceptibility to favipiravir in vitro (Goldhill et al., 2018; Goldhill et al., 2021), neither this nor other PB1 substitutions that reduce drug susceptibility have been found in clinical studies (Shiraki and Daikoku, 2020; Takashita et al., 2016), suggesting a high barrier to resistance. Favipiravir-associated teratogenicity was observed in animal models, thus limiting eligible treatment populations, and excluding women with reproductive potential (Hayden and Shindo, 2019; Nagata et al., 2015). Like BXM, favipiravir has in vitro and/or in vivo activity against zoonotic A(H5Nx), A(H7N9), and A(H9N2) viruses (Marathe et al., 2016; Takashita et al., 2022).

The polymerase basic protein 2 (PB2) binds host mRNA caps via its cap-binding domain (CBD) (residues 321–482) to facilitate PA cap-cleavage activity. Pimodivir (VX-787) targets the PB2 CBD, halting the initiation of cap-snatching. In phase 2a studies of healthy volunteers experimentally infected with influenza, individuals treated with pimodivir vs. placebo had less viral shedding and milder clinical symptoms (Trevejo et al., 2018), and pimodivir reduced viral RNA loads in a phase 2b study vs. placebo (Finberg et al., 2019). Pimodivir demonstrated less-than-ideal pharmacokinetics (Clark et al., 2014; Zhao et al., 2020) in a phase 2a study and the PB2 substitutions (M431I, M431L/R/V, S324C, K376R) reduced pimodivir susceptibility 57-fold (Trevejo et al., 2018). Furthermore, the dissimilarity of the CBDs of influenza A and B viruses means that pimodivir works only against influenza A infection. In 2020, Janssen halted ongoing phase 3 studies (NCT03376321, NCT03381196) because interim analyses suggested that pimodivir would provide no additional benefit beyond the standard-of-care that included oseltamivir treatment for some participants (Janssen, 2020). However, pimodivir demonstrates in vitro and in vivo activity against multiple influenza subtypes, including zoonotic A(H5N1) virus (Byrn et al., 2015), suggesting that this class of drugs could be beneficial in a pandemic, pending future study or drug refinement.

3.2. Development of novel polymerase inhibitors

The drawbacks associated with BXM, favipiravir, and pimodivir highlight the need to identify additional polymerase-targeted molecules that broadly inhibit influenza A and B viruses. Preclinical studies of various PB2-targeted small molecules are ongoing (Chen et al., 2022), with at least one (CC-42344) being in phase 1 (NCT05202379) and another (ZSP1273) being in phase 3 studies (NCT04024137) (Figure). Nucleoside inhibitor alternatives to favipiravir, such as molnupiravir, may have some benefit in treating influenza because of their efficacy against various RNA viruses (Toots et al., 2019), but they may also have teratogenic side effects similar to those of favipiravir. Development of nucleoside-based and non–nucleoside-based PB1 inhibitors is proceeding preclinically (Hou et al., 2022), but none are currently in clinical testing. Multiple patents and in vitro studies have focused on CENI compounds or variations on the BXM scaffold (Stevaert and Naesens, 2016). Although some showed promising in vitro efficacy and pharmacodynamics in mice (Ivashchenko et al., 2021; Ivashchenko et al., 2020), candidates such as AL-749 did not progress past phase 1 testing. The CENIs TG-1000 and GP681 recently completed phase 2 trials (NCT04706468, NCT04736758) and recruiting for a phase 1 study of ZX-7101A is underway (NCT05217732). Finally, various small molecules and peptides that disrupt important protein–protein interactions among the polymerase complex components, including critical contacts between PA and PB1 or between PB1 and PB2, are in preclinical development (Massari et al., 2021).

4.0. Additional Treatment Options

4.1. Non-NA or polymerase targeted inhibitors

Small-molecule inhibitors targeting the NA or polymerase protein are the most commonly approved anti-influenza antivirals worldwide, but treatments targeting different viral proteins may be useful in pandemic settings. Arbidol (Umifenovir) is a non-nucleoside inhibitor with multiple proposed mechanisms of action (Blaising et al., 2014), and it likely targets the influenza (A and B) virus HA, which mediates receptor binding and entry/fusion. The drug is approved for influenza prophylaxis and treatment in Russia and China (since 2006), but there have been no recent clinical studies or approvals. Additionally, there are insufficient data concerning its impact on zoonotic virus infections (Ahmed and Abouzid, 2021; Fediakina et al., 2005).

HA-targeted monoclonal antibodies (mAbs) are an additional non-small molecule treatment option. Preclinical development of NA-targeted mAbs is ongoing (Gubareva and Mohan, 2022), but the mAbs currently being tested clinically target either the HA globular head to disrupt interaction between the protein and the sialic acid host receptor or the HA stalk to inhibit fusion/entry. mAbs would ideally neutralize both group 1 and group 2 HAs and show neutralizing activity against zoonotic viruses. Clinical results have been mixed, with some candidates, such as CR8020 and CR6261, not meeting phase 2 primary endpoints or studies being halted (Han et al., 2021). MHAA4549A failed to demonstrate clinical efficacy in a phase 2 study (NCT02623322) (Lim et al., 2022) but reduced virus shedding in a human challenge trial, although this effect was inconsistent over the tested doses (NCT01980966) (McBride et al., 2017). The VIS410 mAb reduced late-infection viral loads in a phase 2 trial vs. placebo (NCT02989194) (Hershberger et al., 2019), and it reduced viral loads in a phase 2 human challenge trial vs. placebo (NCT02468115) (Sloan et al., 2020). Recruitment is anticipated for a phase 2 trial of pre-exposure prophylaxis with VIR-2482 (NCT05567783) (Figure). Overall, the mixed efficacy results with current mAbs, along with costs, present barriers to their widespread use, but they may be considered for use during a pandemic. Currently, there are no approved mAbs or immune plasma–related treatments for influenza.

4.2. Combination therapy of antiviral agents

Additive or synergistic effects among classes of influenza antivirals could enhance treatment and reduce the emergence of drug resistance during a pandemic. Early studies of NAI combinations were inconclusive or failed to show superiority over single NAI use for seasonal or zoonotic influenza infections (Carrat et al., 2012; Duval et al., 2010; Escuret et al., 2012). Oseltamivir + peramivir combination therapy was not superior to oseltamivir monotherapy for adults with A(H7N9) infection (Zhang et al., 2016), and the persistently high fatality rate warrants more evidence-based research to understand whether NAIs reduce mortality in these patients and to guide clinical practices. The oseltamivir + rimantadine + ribavirin triple combination significantly reduced viral loads on day 3 vs. NAI monotherapy but did not provide significant benefits with respect to clinical outcomes (eg. symptom duration) in a phase 2 study (Beigel et al., 2017). The introduction of polymerase-targeted antivirals and mAbs has increased the candidates available for combination therapy. In a phase 3 study of NAI + baloxavir vs. NAI alone, the combination therapy did not provide superior clinical outcomes in hospitalized patients, and a dual NAI–BXM drug-resistant virus (NA-H274Y + PA-I38T) was isolated from two immunocompromised patients in the former group (Kumar et al., 2022). In a study of 40 patients, favipiravir + oseltamivir led to improved day 14 clinical outcomes in critically ill patients and a higher proportion of samples with undetectable virus at day 10 (Wang et al., 2020a). Phase 2 studies of pimodivir + oseltamivir vs. pimodivir monotherapy for uncomplicated influenza demonstrated faster viral clearance with the combination and a lower percentage of pimodivir-associated PB2 mutations (Finberg et al., 2019). Another phase 2 study to measure pharmacokinetics in hospitalized patients given pimodivir + oseltamivir vs. NAI alone reported shorter TTAS and lower incidence of influenza-related complications in the combination group (O’Neil et al., 2020). Combinations of mAbs + oseltamivir have so far yielded no superior results. In a phase 3 study of critically ill patients, MHAA4549A + oseltamivir did not reduce the time to discontinuation of supplemental oxygen (the primary endpoint), reduce viral loads, or lead to clinical improvement when compared with oseltamivir monotherapy (Lim et al., 2020). MEDI8852 + oseltamivir did not yield better outcomes than MEDI18852 monotherapy in outpatients with uncomplicated influenza A infection (Ali et al., 2018). Currently, combination therapy has demonstrated limited clinical benefit. Additional clinical trials are needed, with considerations of adjusting the dosing and timing of administration, target population (high risk vs. uncomplicated infections), and incorporating new combinations including CENIs.

4.3. Host-targeted immunomodulators

Severe influenza infections are characterized by elevated pro-inflammatory cytokines (Hui et al., 2018), and host-targeted immunomodulators may provide adjunct treatment benefits. Although no immunomodulator agents are currently recommended as adjuncts for influenza, several have been used in combination with antivirals and have demonstrated clinical benefit in RCTs. Macrolide antibiotics have been used in combination with oseltamivir in hospitalized patients. Azithromycin + oseltamivir mediated declines in plasma concentrations of IL-6, IL-17, and CXCL9 that were significantly faster than those in patients receiving oseltamivir monotherapy; however, viral load reduction and mortality were not significantly different (Lee et al., 2017). Interestingly, the azithromycin + oseltamivir combination did not significantly reduce pro-inflammatory cytokines or chemokines in ambulatory patients with influenza beyond what was observed with oseltamivir monotherapy (Kakeya et al., 2014). In an RCT of 217 hospitalized adults infected with A(H3N2) seasonal influenza, patients receiving clarithromycin + oseltamivir + naproxen (a non-specific cyclooxygenase 2 (COX) inhibitor with anti-NP activity) had reduced viral titers and reduced 30-day mortality when compared with patients receiving oseltamivir alone (Lee et al., 2021). In hospitalized children, this combination significantly reduced viral loads and time to defervescence when compared with oseltamivir alone (Lee et al., 2021).

The COX-2 inhibitor celecoxib may be another anti-inflammatory option, demonstrating promising preclinical efficacy against lethal A(H5N1) challenge in mice when co-administered with mesalazine (an anti-inflammatory drug approved for ulcerative colitis) and zanamivir (Zheng et al., 2008). Preliminary results of an RCT (NCT02108366) suggested that 28-day mortality was lower with celecoxib–oseltamivir combination therapy than with oseltamivir monotherapy (Koszalka et al., 2022). Another non-selective COX inhibitor, flufenamic acid, in combination with clarithromycin + zanamivir, reduced mortality in mice lethally challenged with a mouse-adapted A(H1N1)pdm09 virus (Lee et al., 2018). An RCT comparing the efficacy of flufenamic acid + clarithromycin + oseltamivir vs. oseltamivir monotherapy is ongoing (NCT03238612).

Although immunomodulators are not currently recommended for severe influenza, RCTs conducted during the COVID-19 pandemic have established the clinical benefits of corticosteroids, IL-6 receptor blockers, and a Janus kinase (JAK) inhibitor for managing COVID-19 in hospitalized patients according to disease severity. The data may provide useful comparisons with respect to treatments for other severe respiratory viral infections. Corticosteroids may reduce mortality in patients with severe community-acquired pneumonia (Stern et al., 2017) or severe COVID-19 (Recovery Collaborative Group et al., 2021; van Paassen et al., 2020); however, there is insufficient evidence to support using corticosteroids as adjunctive influenza therapy, and a meta-analysis of observational studies suggested an increased risk of mortality with corticosteroid use for severe influenza infection (Lansbury et al., 2020). Factors including study design (e.g., the lack of RCTs of corticosteroids in severe influenza), time (e.g., duration and time of initiation of administration), corticosteroid dosing, and differences between influenza and SARS-CoV-2 viruses in terms of pathogenesis and pro-inflammatory cytokine induction require further investigation (Winkler et al., 2022). Tocilizumab and sarilumab are mAbs targeting the IL-6 receptor that are authorized for treating rheumatoid arthritis. They reduce the mortality of hospitalized patients with COVID-19 and received emergency use authorization (EUA) for treating critical COVID-19 patients (WHO Rapid Evidence Appraisal for COVID-19 Therapies Working Group et al., 2021). Tocilizumab acted synergistically with corticosteroids to reduce the mortality of severely or critically ill patients with COVID-19 (Albuquerque et al., 2022). No clinical evidence supports using IL-6 receptor antagonists to treat severe influenza, but an observational study found that tocilizumab treatment reduced inflammation (reducing fever and the levels of C-reactive proteins) in juvenile patients with idiopathic arthritis who received tocilizumab after influenza infection (Kawada et al., 2013). Similarly, the JAK inhibitor baricitinib is also approved for rheumatoid arthritis treatment and severe or critical COVID-19 (Agarwal et al., 2020) and could be a further avenue of study for influenza treatment. Additional RCTs are needed to understand if there is benefit of tocilizumab, sarilumab, or baricitinib in influenza infection.

5.0. Drug stockpiling for a future pandemic

In the 2009 influenza pandemic, treatment was limited to the NAIs, as neither favipiravir nor BXM were clinically available. Adding BXM to antiviral stockpiles for future influenza pandemics would enable targeting of different viral proteins with distinct mechanisms of action. If the only NAI in the stockpile is oseltamivir, both drugs would be orally administered, easing use in non-hospital settings, and expediting distribution and administration within the critical treatment window of 48 hours after symptom onset.

The 2009 pandemic prompted many countries to stockpile NAIs (Berera and Zambon, 2013; Wan Po et al., 2009), mostly oseltamivir. In the United States, these stockpiles were replenished post pandemic (United States Dept. of Health and Human Services, 2017), and the United Kingdom government also expressed intent to maintain an NAI stockpile (O’Dowd, 2014). Both Japan and Taiwan have stockpiled favipiravir, in addition to oseltamivir, as part of their pandemic preparedness (Shiraki and Daikoku, 2020). Careful consideration must be given to the duration of such stockpiles. In collaboration with Roche, the United Kingdom government completed a program to extend the shelf-life of stockpiled oseltamivir (European Medicines Agency, 2009; Wan Po et al., 2009). Present FDA guidelines advise retaining Tamiflu capsules for up to 15 years (United States Food and Drug Administration, 2020), although no similar guidance on shelf-life is available for equivalent generics. The current stated shelf-life for BXM is 3–5 years, depending on the formulation (European Medicines Agency), and efforts to extend this period and updates on efficacy after longer storage would provide additional justification for stockpiling this drug.

Mathematical modelling may prove useful for justifying stockpiles. One model has suggested a benefit to distributing NAIs liberally to outpatients, with potential reductions in hospitalizations and deaths (Moss et al., 2016), and another based on the United Kingdom population suggested similar outcomes, assuming the stockpile could treat 80% of the population (Watson et al., 2016). A more recent model incorporating BXM or a CENI-type drug drew no conclusion with respect to stockpiling, but it demonstrated how factors such as stockpile size, the treated population percentage, and, importantly, the higher potential resistance rates and prices of these drugs can be incorporated to inform government health agencies (Kim et al., 2022).

Future pandemic viruses will probably be sufficiently antigenically distinct from circulating viruses or even from anticipated zoonotic candidate vaccine viruses. During the 4–6 months of lead time necessary to generate a pandemic-matched vaccine, and depending on the severity of the pandemic virus, significant health and economic costs may be incurred. Stockpiled influenza antivirals with a safety profile permitting their broad use across all populations could lower transmission, reduce healthcare resource burdens, and provide hope to a frightened population.

6.0. Additional perspectives on anti-influenza drug development

The availability of two classes of influenza antivirals with distinct mechanisms of action provides a foundation for treating seasonal influenza and for a pandemic response. NAIs, principally oseltamivir, are widely used and are considered standard-of-care; in addition, oseltamivir is available as a generic allowing for greater access globally. The approval of BXM in the United States, Japan, and elsewhere has provided an additional treatment option. The identification of viral resistance or reduced susceptibility to oseltamivir and BXM highlights the need for continued vigilance. Combination treatments with antivirals and/or host-targeted therapeutics may be additional options for responding to resistance development. However, the composition of these combinations must be carefully considered, as potential outcomes of host-targeted corticosteroids may prolong viral shedding and increase potential of resistance emergence when combined with virus-targeted drugs.

The lack of an effective influenza antiviral for treating severely ill hospitalized patients highlights the need for a new or modified approach for this population. In several clinical trials of antivirals in hospitalized patients with influenza, the antiviral alone demonstrated no clinical benefit, whereas recent experience with hospitalized patients with COVID-19 has shown that a combination of direct-targeted and host-targeted therapeutics may improve clinical outcomes.

Clinical studies have been initiated to determine whether pre-exposure prophylaxis (i.e., long-acting mAb VIR-2482, NCT05567783 and drug-Fc conjugate CD388, NCT05523089) or post-exposure BXM prophylaxis (Ikematsu et al., 2020) can prevent influenza infections. If so, this would be an option for people most at risk of developing severe influenza. It would also provide coverage during a pandemic for immunocompromised individuals and would be a useful addition to the armamentarium for pandemic response.

6.1. Clinical and preclinical data needed for future influenza antiviral drug development

Preclinical development of novel influenza antivirals should give special consideration to the following: 1) modifying existing drug scaffolds and/or developing novel molecules that contact multiple binding points of the target protein and that may help overcome the emergence of resistance to BXM and other drugs; 2) employing medicinal chemistry strategies to improve drug efficacy or to produce longer-lasting versions of existing drugs (e.g. laninamivir); 3) revisiting the M2 protein as a target to address adamantane resistance; and 4) evaluating other nucleoside analogues with similar efficacy profiles, but better safety profiles than favipiravir. Preclinical animal studies could also include delayed (>48 hours) treatment regimens and/or transmission studies in suitable animal models to expand the potential benefits of drug candidates.

Clinical testing of existing drugs may need to include higher doses of the active compound and multiple doses of BXM to address resistance concerns. A post–phase 3 trial of BXM + oseltamivir combinations in patients with severe influenza used two doses of BXM (on days 1 and 4), with an additional dose at day 7 if the patient had not met predefined improvement criteria by day 5. These adjustments may be more beneficial for higher-risk groups, including children and immunocompromised patients who may exhibit prolonged virus shedding. Recent NAI + BXM combinations have not proved superior to NAI monotherapy, casting doubt on their use in non-hospitalized patients. Despite this, triple combinations that utilize an NAI, a CENI, and a nucleoside analogue have not yet been pursued but may increase treatment efficacy and/or overcome resistance to individual drugs.

6.2. Concluding remarks

The severe public health burdens and widespread morbidity and mortality during the early months of the COVID-19 pandemic highlighted the consequences of confronting a viral pandemic with no readily available vaccine or virus-targeted antivirals. Although this is not the case for influenza, only two classes of antiviral drugs with different mechanisms of action are currently available. Although stockpiling can mitigate the health and economic impacts of a pandemic, a few mutations can render the drugs ineffective, and this is an especially important consideration given the evolutionary potential and broad host-range of influenza viruses. Efforts to develop additional treatment options for seasonal and pandemic influenza should continue to focus on new mechanisms of action, potential combination strategies, and ease of use for mass distribution. As the COVID-19 pandemic proceeded, a range of potential treatment options was developed, leading to EUAs of several treatments with various mechanisms of action. As a result, multiple large-scale RCTs have been developed to test whether various virus-targeted, host-targeted compounds, or non–small molecule inhibitors prevent death from COVID-19. Some of the drugs developed and authorized for treating COVID-19 could be applied to influenza and other emerging infectious diseases. These large-scale collaborations, including the RECOVERY Trial (NCT04381936) and WHO Solidarity Trial (ISRCTN18066414) may serve as examples that can be modified for influenza antivirals and strategies discussed here with potential to address novel virus and host-targeted drugs, alone or in combination. Initiating such investigations before the next pandemic, rather than in reaction to it, will not only benefit the next influenza pandemic response, but also help reduce morbidity and mortality associated with annual seasonal influenza.