Abstract
The COVID-19 pandemic has highlighted the urgent need for the development of broad-spectrum antivirals to enhance preparedness against future spillover of zoonotic viruses with pandemic potential into the human population. Currently, the direct-acting orally available SARS-CoV-2 inhibitors molnupiravir and paxlovid are approved for human use under emergency use authorization. A promising next-generation therapeutic candidate is the orally available ribonucleoside analog 4′-fluorouridine (4′-FlU) that had potent antiviral efficacy against different viral targets, including SARS-CoV-2 in human organoids and animal models. Although a nucleoside analog inhibitor such as molnupiravir that targets the viral RNA-dependent RNA polymerase (RdRP) complex, 4′-FlU showed a distinct mechanism of activity, delayed chain termination, compared with molnupiravir's induction of viral error catastrophe. This review will focus on some currently approved and emerging medicines developed against SARS-CoV-2, examining their potential to form a pharmacological first-line defense against zoonotic viruses with pandemic potential.
Introduction
SARS-CoV-2
Product Profile of First-Line Therapeutics
Although novel vaccines were developed in record time, the broad-spectrum nucleoside analog antiviral remdesivir was approved for treatment of SARS-CoV-2 infection, and antibody therapeutics were rapidly generated and approved for clinical use, these measures came too late to influence the course of the critical first months of the COVID-19 pandemic. To reach the overarching goal of improving preparedness against future viral threats with pandemic potential, effective first-line antivirals are urgently needed. The SARS-CoV-2 experience has highlighted several product profile criteria that must be met for this intended use:
Oral bioavailability. Pharmacological intervention in acute RNA virus infections has been challenging due to typically narrow therapeutic windows (Meganck and Baric, 2021; Plemper, 2021; Sourimant et al., 2021). Because clinical signs are largely consequences of the host antiviral immune responses, viral titers are typically in rapid decline when patients advance to complicated disease and are hospitalized (van Kampen et al., 2021), marking a transition point at which antiviral therapeutics designed to interfere with the viral life cycle and reduce viral load cease to be efficacious and treatment goals shift to immunomodulatory therapy or supportive care. Remdesivir must be delivered intravenously and is, therefore, unavailable to most patients before hospitalization (Beigel et al., 2020). Mixed remdesivir performance likely reflects the problem that the therapeutic window for a direct acting antiviral has essentially closed when the patient presents with severe disease. We propose that efficiently reaching an outpatient population before advance to severe disease is best achieved with an orally bioavailable drug that does not require trained health care professionals and sterile equipment for delivery.
Preapproved for human use. Spillover of animal viruses into the human population stands at the beginning of most viral pandemics. COVID-19 has re-emphasized that an antiviral must be approved for human use before a new pathogenic emerges to potentially stamp out a nascent pandemic. Despite unprecedented short timelines from first appearance of SARS-CoV-2 to vaccine launch and emergency approval of paxlovid and molnupiravir, the window of opportunity to regionally contain the virus and avoid a global pandemic was likely very short-lived once efficient human-to-human transmission had begun. This short timeline is incompatible with even the most accelerated approval process. Accordingly, developmental antiviral candidates must have a well-defined primary indication that provides a path to clinical testing and regulatory approval for human use in interpandemic years, or before a novel pandemic viral threat first emerges.
Broad indication spectrum. A wide antiviral target range spanning ideally viral pathogens of different families will often coincide with the request for a viable primary indication for clinical testing. Beyond ensuring a path to approval, broad-spectrum activity of first-line antivirals will be imperative for improving pandemic preparedness, since the nature of the next pandemic viral threat is unknown. Several pathogens in at least eight different viral families have been recognized for their high pandemic potential (Fernandez-Montero et al., 2020), undermining feasibility of preemptively developing and clinically advancing a very large panel of individually targeted antivirals.
High tolerability. The drug safety profile must be compatible with administration to outpatients at the earliest disease stages. This intended application will require high tolerability with a wide safety margin. Especially use at the onset of a potential pandemic, before the scope of the health threat has been fully appreciated by the public, sets narrow limits for acceptable side effects. Tolerability profiles and pharmacokinetic properties must furthermore answer to the high replication dynamics of RNA viruses causing acute diseases, which requires rapid buildup of high drug levels and sustained exposure in disease-relevant tissues.
Suitable for stockpiling. To be able to rapidly respond to a newly emerged threat, a sufficiently large number of doses must be readily deployable. This requirement is best met if the drug product is amenable to cost-effective large-scale production and is shelf-stable at ambient temperature for stockpiling.
Oral SARS-CoV-2 Therapeutics with Emergency Use Authorization
None of the currently available direct-acting antivirals used for the treatment of SARS-CoV infection meets all these requests. Of the orally bioavailable candidates, emergency use authorization has been granted by the FDA in 2021 to the nucleoside analog inhibitor molnupiravir and the viral protease inhibitor paxlovid.
Molnupiravir
Molnupiravir is the oral ester prodrug of N 4-hydroxycitidine (NHC) (Fig. 1A), which has demonstrated broad-spectrum antiviral activity against influenza viruses (Toots et al., 2019; Toots and Plemper, 2020), paramyxoviruses and pneumoviruses (Yoon et al., 2018), alphaviruses (Urakova et al., 2018), and the beta-coronaviruses (Agostini et al., 2019). Only free NHC efficiently reaches the circulation after oral administration of molnupiravir due to rapid hydrolysis of the ester during intestinal absorption (Kabinger et al., 2021). Upon cellular uptake, NHC is anabolized to the bioactive triphosphate form, NHC-TP, which is incorporated by the viral polymerase into progeny viral antigenomes and genomes in place of endogenous cytidine triphosphate (CTP).

Structures and mechanism of action of different SARS-CoV-2 inhibitors.
Accordingly, molnupiravir does not act as a chain-terminator but, rather, once incorporated it spontaneously switches between two tautomeric forms, the oxime and hydroxylamine, which base-pair as uridine triphosphate (UTP) and CTP, respectively (Kabinger et al., 2021). Owing to this spontaneous change in base-pairing, progeny genomes or antigenomes generated from a template containing incorporated NHC will harbor randomly distributed transition mutations in low allele frequency (Toots et al., 2019), which rapidly compromise integrity of the genomes and ultimately result in collapse of the viral population. This mechanism of antiviral activity follows the random lethal viral mutagenesis or error catastrophe model that was first described for ribavirin by example of picornavirus inhibition (Crotty et al., 2001).
Presumably due to the close chemical similarity of NHC to endogenous cytidine, molnupiravir is not recognized by the proofreading activity of the beta-coronavirus polymerase (Bouvet et al., 2012; Robson et al., 2020), escaping excision postincorporation. Furthermore, the genetic barrier of molnupiravir against emergence of viral resistance is unusually high, very likely reflecting that the viral RNA-dependent RNA polymerase (RdRP) complex cannot solve the sterical challenge to differentiate between NHC-TP and endogenous CTP. Despite extensive attempts to resistance profile the drug against different viral targets, no valid escape mutations have been reported so far (Urakova et al., 2018; Toots et al., 2019) and no viral resistance has been noted in the clinic (Holman et al., 2021; Painter et al., 2021).
Whereas there is demonstrated evidence that NHC-TP is also recognized by host RNA polymerases and incorporated into host RNAs when present at very drug high concentration (Toots et al., 2019), host RNAs are typically short-lived and never serve as template for further amplification. Clinical trials have demonstrated that a 5-day molnupiravir regimen of 800 mg orally twice daily is very well tolerated with no evidence of hematologic toxicities or other side effects (Fischer et al., 2021; Painter et al., 2021; Jayk Bernal et al., 2022a). Of greater concern are long-term consequences due to potential incorporation of NHC into host DNA.
However, metabolism of NHC leads to uridine and no evidence of transribosylation has been noted (Hernandez-Santiago et al., 2004). Animal studies revealed low teratogenic potential of molnupiravir (Merck, 2022) and no mutagenic activity at antiviral concentrations (Zhou et al., 2021). Although these animal toxicology data are highly encouraging, long-term follow-up clinical data are still lacking, underscoring that any therapeutic intervention decision must be based on careful evaluation of the individual risk/benefit ratio for the patient.
Molnupiravir was equally efficacious against the original SARS-CoV-2 strain and subsequently emerged variants of concern (VOC) in cultured cells and primary human airway epithelium organoids, including VOC delta and omicron (Lieber et al., 2022). Interesting, clinical data indicated that patients infected with VOC gamma experienced greater benefit from molnupiravir than those suffering from VOC delta (Jayk Bernal et al., 2022b).
Although not recapitulated by ex vivo models, a Roborovsky dwarf hamster model of acute SARS-CoV-2 induced lung injury revealed a significantly greater effect size of molnupiravir-mediated virus load reduction when animals were infected with VOC gamma compared with delta (Lieber et al., 2022). Only VOC omicron-infected dwarf hamster showed a significant biological sex-dependent difference in response to treatment. The relevance of this observation for human treatment is unknown at present, but the dwarf hamster data re-emphasize that treatment efficacy may vary dependent on VOC and must, therefore, be re-evaluated continuously as SARS-CoV-2 further evolves.
Paxlovid
Paxlovid is a combination drug of the SARS-CoV-2 3-chymotrypsin-like cysteine protease (3CLpro) inhibitor nirmatrelvir (Fig. 1B) and the antiretroviral drug ritonavir (Reina and Iglesias, 2022). 3CLpro is instrumental for beta-coronavirus replication (Chen et al., 2005), proteolytically processing the viral polyproteins to liberate individual viral polypeptides such as the RdRP subunits and viral helicase that are essential for replication and transcription (Mody et al., 2021).
Nirmatrelvir is a peptidomimetic that directly targets the active site of 3CLpro, forming a permanent covalent bond with the catalytic cysteine (Pavan et al., 2021). Similar to many peptidomimetic protease inhibitors, however, nirmatrelvir shows poor metabolic stability in vivo, rapidly becoming a target of liver oxidation through the cytochrome P450 system (Eng et al., 2022). In initial animal model proof-of-concept studies against SARS-CoV-2, nirmatrelvir monotherapy had to be administered at an unrealistic 1000 mg/kg body weight dose to compensate for its poor pharmacokinetic profile (Owen et al., 2021).
Originally developed as an inhibitor of HIV protease, ritonavir acts off-target as a strong inhibitor of cytochrome P450 key enzymes P450-3A4 (CYP3A4) and P450-2D6 (CYP2D6) (Kumar et al., 1996), reducing the rate of P450-mediated drug detoxification. This ability of ritonavir to improve drug pharmacokinetic profiles through slowing liver oxidation has resulted in its use as combination drug with, for instance, other HIV protease inhibitors and most recently nirmatrelvir (Eng et al., 2022). However, the downside of this fix of the pharmacokinetic shortcomings of nirmatrelvir is the very high potential of ritonavir for severe and sometimes fatal drug–drug interactions with dozens of approved medications, since drug levels are broadly affected when activity of the P450 system is manipulated.
Contraindications include drugs for the treatment of heart conditions, blood pressure regulators, immunosuppressants, some chemotherapeutics, and asthma and allergy medications (Talha and Dhamoon, 2022). Accordingly, paxlovid is poorly compatible with treatments for many conditions that represent major risk factors for the advance of a COVID-19 patient to severe life-threatening disease (Hammond et al., 2022).
Taken twice daily in a 5-day regimen similar to that of molnupiravir (Jayk Bernal et al., 2022b), paxlovid has shown strong efficacy in a clinical trial, reducing COVID-19-related deaths or hospitalizations by a claimed 89% in unvaccinated patients compared with the placebo control group (Mahase, 2021). In contrast to a large body of literature characterizing molnupiravir efficacy in multiple SARS-CoV-2 animal models (Cox et al., 2021b), experimental animal model data for nirmatrelvir are limited (Owen et al., 2021; Eng et al., 2022), and none have yet been released for paxlovid. Although sparse, the available evidence does not provide an obvious reason for the seemingly greater therapeutic benefit of paxlovid in clinical trials since antiviral effect size of both drugs appears to be similar in animals.
Possible explanations for the conundrum may be differences in trial design (Fischer et al., 2022) and patient groups eligible for enrollment—paxlovid could not be tested in patients taking other medications with metabolism highly dependent on cytochrome P450 CYP3A4 or known as strong inducers of CYP3A4. As more clinical data with recently emerged VOC become available, it will be important to evaluate whether the claimed efficacy difference between molnupiravir and paxlovid is real. Urgently needed are furthermore head-to-head comparisons of both drugs under experimentally controlled conditions in relevant animal models.
Next-Generation First-Line Antiviral Developmental Candidate 4′-Fluorouridine
4′-Fluorouridine (4′-FlU) (Sourimant et al., 2022) is a recently identified novel competitive ribonucleoside analog inhibitor (Fig. 1C) that is structurally distinct from known nucleoside analogs with activity against SARS-CoV-2 such as molnupiravir and remdesivir (Kokic et al., 2021). Carrying a ribose modification, the compound showed broad-spectrum activity against a wide range of negative and positive-sense RNA viruses, including pathogens of the paramyxovirus, pneumovirus, and beta-coronavirus families (Sourimant et al., 2022).
Biochemical RdRP assays with purified recombinant SARS-CoV-2 and respiratory syncytial virus (RSV) polymerase complexes identified sequence context-dependent delayed chain termination, preferentially three to four nucleotides after incorporation of bioactive 4′-FlU triphosphate (4′-FlU-TP), as the mechanism of activity. This antiviral effect resembles that described for the bioactive parent compound of remdesivir (Kokic et al., 2021) and is most likely due to altered RNA secondary structure after single or tandem incorporation of the analog that ultimately cannot be accommodated by the viral polymerase (Sourimant et al., 2022). In contrast to molnupiravir, 4′-FlU has not shown any mutagenic effect in biochemical and cell-based assays.
Unlike remdesivir, however, 4′-FlU was efficiently delivered orally, resulting in sustained tissue exposure of bioactive 4′-FlU-TP in different animal species. Dosed orally once daily, the compound potently inhibited original SARS-CoV-2, recently emerged SARS-CoV-2 VOC, and RSV replication in different animal models (Sourimant et al., 2022). Efficacy against SARS-CoV-2 and RSV was furthermore confirmed in disease-relevant primary human airway epithelium organoids, although RSV and other mononegaviruses showed greater sensitivity to 4′-FlU than SARS-CoV-2.
This difference in potency likely reflects that some incorporated 4′-FlU moieties may be recognized by the coronavirus proofreading machinery and excised, and/or that the coronavirus polymerase may have greater tolerance to accommodate altered RNA secondary structures than mononegavirus polymerases. Based on high oral bioavailability, strong metabolic stability in vivo confirmed broad-spectrum antiviral efficacy, and suitability for stock-piling; however, 4′-FlU meets the key criteria of a first-line antiviral. A unique strength compared with existing COVID-19 medicines are pharmacokinetic properties compatible with once-daily dosing, which represents a major advance in ensuring patient compliance.
Going forward, the best first-line defense against future pandemic viral threats will in our opinion realistically not come in the form of a single “magic bullet” drug, but of a panel of broad-spectrum orally available antivirals with overlapping, but not identical, indication range. These drugs must be urgently approved for human use and will likely have distinct strengths and liabilities, providing alternative options to address specific needs of individual patient subgroups.
Currently existing drugs that we consider to be part of this panel are molnupiravir, favipiravir (Ghasemnejad-Berenji and Pashapour, 2021), and—with the restriction of a limited indication spectrum covering only the beta-coronaviruses—paxlovid. Promising future contenders that should be fast-tracked are experimental antivirals such as 4′-FlU and, for instance, an orally available chemical analog of remdesivir (Cox et al., 2021a). If ultimately grown to a size of 5–10 qualified drugs, this panel could be game-changing in establishing control at an early stage of future pandemic viral threats.
Footnotes
Acknowledgments
We thank members of the Plemper Lab for discussion and A.L. Hammond for critical reading of the article.
Disclosure Statement
No competing financial interests exist.
Funding Information
This study was supported, in part, by Public Health Service grants AI071002 (to R.K.P.) and AI141222 (to R.K.P.), from the NIH/NIAID. The funders had no role in study design, data collection, and interpretation, or the decision to submit the study for publication.
