4’-Fluorouridine mitigates lethal infection with pandemic human and highly pathogenic avian influenza viruses

Influenza outbreaks are associated with substantial morbidity, mortality and economic burden. Next generation antivirals are needed to treat seasonal infections and prepare against zoonotic spillover of avian influenza viruses with pandemic potential. Having previously identified oral efficacy of the nucleoside analog 4’-Fluorouridine (4’-FlU, EIDD-2749) against SARS-CoV-2 and respiratory syncytial virus (RSV), we explored activity of the compound against seasonal and highly pathogenic influenza (HPAI) viruses in cell culture, human airway epithelium (HAE) models, and/or two animal models, ferrets and mice, that assess IAV transmission and lethal viral pneumonia, respectively. 4’-FlU inhibited a panel of relevant influenza A and B viruses with nanomolar to sub-micromolar potency in HAE cells. In vitro polymerase assays revealed immediate chain termination of IAV polymerase after 4’-FlU incorporation, in contrast to delayed chain termination of SARS-CoV-2 and RSV polymerase. Once-daily oral treatment of ferrets with 2 mg/kg 4’-FlU initiated 12 hours after infection rapidly stopped virus shedding and prevented transmission to untreated sentinels. Treatment of mice infected with a lethal inoculum of pandemic A/CA/07/2009 (H1N1)pdm09 (pdmCa09) with 4’-FlU alleviated pneumonia. Three doses mediated complete survival when treatment was initiated up to 60 hours after infection, indicating a broad time window for effective intervention. Therapeutic oral 4’-FlU ensured survival of animals infected with HPAI A/VN/12/2003 (H5N1) and of immunocompromised mice infected with pdmCa09. Recoverees were protected against homologous reinfection. This study defines the mechanistic foundation for high sensitivity of influenza viruses to 4’-FlU and supports 4’-FlU as developmental candidate for the treatment of seasonal and pandemic influenza.

Introduction Influenza A viruses (IAV) are zoonotic negative sense, segmented RNA viruses with high pandemic potential [1]. Wild aquatic birds are a major natural reservoir for avian influenza viruses, creating high opportunity for spread into the poultry industry and spillover of HPAI viruses into the human population with potentially catastrophic consequences [2,3]. The 2021-2022 H5N1 HPAID epidemic specifically is the largest on record, having reached an unprecedented geographical expansion and requiring the culling of 48 million birds so far [4]. Approximately 14 IAV pandemics have occurred in the past 500 years [5], six of which were concentrated in the last 120 years alone. Most deadly of these outbreaks was the 1918 "Spanish Flu" pandemic, which was caused by an H1N1 subtype IAV of unusual virulence that accounted for approximately 500 million infected people, equating to nearly a third of the world population at the time, and 50 million death [6]. Case fatality rates were particularly high in young children,  year old adults, and the elderly above 65 years of age [6]. Subsequent pandemics in the 20 th century included the Asian influenza of 1957 and the Hong Kong influenza of 1968, which were caused by IAV of H2N2 and H3N2 subtypes, respectively [7], and the swine origin H1N1 outbreak of 2009 [8]. Although considered less pathogenic than several of the earlier pandemic strains, pdm09 IAV is still estimated to have caused 151,000 to 575,400 deaths worldwide and major economic damage [9].
Influenza viruses cause rapid-onset disease with high fever, cough, body aches, rhinorrhea, and nasal congestion [10]. Advance to severe disease is marked by virus invasion of the small airways and viral pneumonia, which can lead to sepsis due to a hyperinflammatory systemic response [11]. Other serious complications include myocarditis, encephalitis, myositis, and multi-organ failure [12,13]. At greatest risk of life-threatening influenza are children less than 2 years of age, older adults, pregnant women, and patients with pre-existing conditions such as asthma, diabetes, or chronic heart disease [14]. Vaccine prophylaxis is available, but the efficacy of the seasonal tetravalent influenza vaccine is moderate, ranging from 40 to 60% even under optimal conditions in interpandemic years [15]. Effectiveness of the vaccine can be substantially lower in older adults, who are at greatest risk of progressing to severe disease, or when vaccine components and circulating strains are poorly matched [15]. Multiple avenues are being pursued to develop a broad-spectrum influenza vaccine, but none have yet achieved regulatory approval [16].
To improve disease management especially in vulnerable patient groups, rapidly respond to endemic outbreaks of seasonal influenza when the vaccine is not optimally matched, and to prepare for the pandemic threat originating from zoonotic transmission of HPAI viruses, effective therapeutics are urgently needed. Three classes of antiviral drugs have received approval by the FDA for treatment of influenza thus far. The M2 channel blocking adamantanes (amantadine and rimantadine), the neuraminidase inhibitors (oral oseltamivir, inhaled zanamivir, intravenous peramivir), and, most recently, the PA endonuclease inhibitor baloxavir marboxil [17]. Of these, use of the adamantanes has been discontinued due to widespread pre-existing resistance in human influenza viruses and natural IAV reservoirs [18]. Pre-existing signature resistance mutations against the neuraminidase inhibitors were furthermore detected in 2009 pandemic viruses [19], giving rise to the concern that future effectiveness of this inhibitor class may be questionable [20]. Although baloxavir marboxil is the newest influenza drug, resistance has developed rapidly and was even observed in human patients during the early clinical trials, resulting in rebound of virus replication [21][22][23]. Novel therapeutic options are therefore needed to broaden the available anti-influenza virus drug arsenal.
In search of suitable broad-spectrum developmental candidates, we explored in this study antiviral potency, in vivo efficacy, and mechanism of action of 4'-FlU against the influenza virus indication. Recently, we reported that 4'-FlU is a highly potent inhibitor of SARS-CoV-2 and pathogens of the mononegavirus order [24,25], establishing oral efficacy in animal models of SARS-CoV-2 and RSV infection. Having demonstrated that this nucleoside analog functions as an immediate chain terminator of influenza virus polymerase, we assessed oral efficacy in two different animal models, mice and ferrets, against pandemic human and HPAI viruses. Influenza treatment paradigms developed in these models outline an unusually broad time window for successful therapeutic interference with a lethal viral challenge. These results support the development of 4'-FlU as a therapeutic candidate for the treatment of influenza with the potential to improve preparedness against zoonotic spillover of avian IAVs with pandemic potential into the human population.

Results
To explore whether the indication spectrum of 4'-FlU ( Fig 1A) extends to the orthomyxoviruses, we examined activity of the compound against seasonal and pandemic IAVs representing H1N1 and H3N2 subtypes and influenza B virus (IBV) isolates representing the Yamagata and Victoria lineages in cell culture ( Fig 1B). In parallel, we determined activity of 4'-FlU against a recombinant pandemic A/California/07/2009 (H1N1) (pdmCa09), A/swine/Spain/ 53207/2004 (H1N1), B/Memphis/20/1996, and B/Malaysia/2506/2004 on non-differentiated primary human airway epithelium (HAE) cells (Fig 1C), explored 4'-FlU potency against a recombinant A/WSN/33 (H1N1)-based reporter virus [26] on a panel of non-differentiated primary HAEs representing 10 distinct donors (S1A Fig), and generated an airway epithelium model through differentiation of HAEs at air-liquid interface to test inhibition of pdmCa09 through basolateral 4'-FlU in a disease-relevant ex vivo tissue model (S1B Fig). Most viral isolates tested were efficiently inhibited by the compound, consistently returning 50% and 90% inhibitory concentrations (EC 50 and EC 90 , respectively) in the nanomolar range with steep Hill slopes independent of host cell type. The least sensitive were A/swine/Spain/53207/2004 (H1N1), B/Memphis/20/1996, and B/Malaysia/2506/2004, but 4'-FlU inhibitory concentrations remained in the low-micromolar range. Recapitulating our previous experience with 4'-FlU performance on respiratory epithelium cultures [24], potency of 4'-FlU against pdmCa09 was greater in the differentiated HAE model (EC 90 0.07 μM) than on cell lines (EC 90 0.7 μM) or undifferentiated primary cells (EC 90 3.33 μM). Combined with a 50% cytotoxic concentration (CC 50 ) of 4'-FlU of 468 μM [24] in primary HAEs from the same donor as used in the differentiated HAE model, we calculated a selectivity index (SI = CC 50 /EC 50 ) of 4'-FlU against pdmCa09 of 15,600 in this disease-relevant human tissue model. Transepithelial electrical resistance (TEER) between basolateral and apical chamber of the air-liquid interface cultures was unchanged throughout the study (S1C Fig), confirming our previous observation that 4'-FlU does not disrupt epithelium integrity [24] and, in the case of vehicle-treated wells, that the epithelium maintains barrier function despite productive IAV replication [27].

4'-FlU is an immediate chain terminator of influenza virus polymerase
Through biochemical RdRP assays using purified recombinant RSV and SARS-CoV-2 polymerase complexes and synthetic template RNA, we have demonstrated that incorporation of 4'-FlU triphosphate (4'-FlU-TP) into nascent chain RNA triggers delayed termination of the

4'-Fluorouridine mitigates lethal influenza virus infection
RNA-dependent RNA polymerases from a negative sense (RSV) and a positive sense (SARS--CoV-2) RNA virus. Whereas variations in stalling efficiencies were observed between sequences and polymerase, position i+3 marked the predominant stalling site [24]. Following expression of IAV RdRP PA, PB1, and PB2 subunits in insect cells, we subjected affinity chromatography purified, reconstituted IAV polymerase to the equivalent biochemical RdRP assay in the presence of increasing concentrations of 4'-FlU-TP or UTP (Figs 1D, 1E and S2A to S2E). 4'-FlU-TP was incorporated instead of UTP when the polymerase reached the first adenosine in the template RNA as we had noted before in RSV and SARS-CoV-2 RdRP assays [24], but then triggered immediate chain termination of IAV RdRP. Densitometric quantitation of relative amplicon intensities to compare incorporation rates of the analog relative to UTP revealed an approximately 4-fold higher affinity of IAV polymerase for endogenous UTP than for 4'-FlU-TP ( Fig 1F).
These results demonstrate high sensitivity of influenza virus polymerases to 4'-FlU and identify sequence-independent immediate chain termination at position i as a predominant mechanism of influenza virus RdRP inhibition by the compound.

Pharmacokinetic (PK) properties of 4'-FlU in mouse plasma and tissues
We have previously demonstrated oral bioavailability of 4'-FlU in ferrets and efficient intracellular anabolism to 4'-FlU-TP with prolonged tissue exposure levels [24]. To assess cross-species conservation of PK profiles between IAV efficacy models used in this study, ferrets and mice, we subjected the compound to a single-dose oral PK study in mice, measuring 4'-FlU plasma concentrations and corresponding 5'-triphosphate levels in selected soft tissues after oral dosing at 1.5 mg/kg body weight (Fig 2A and 2B). Free 4'-FlU reached a maximal plasma exposure of approximately 1 μM 90 minutes after oral administration, far exceeding the cell culture EC 90 concentration against the influenza virus test panel. Plasma levels were sustained high over a 12-hour period after dosing. Corresponding 4'-FlU-TP tissue exposure levels 12 hours after dosing were approximately 1 nmol/g tissue in all organs but brain and kidney, underscoring suitability of 4'-FlU for in vivo efficacy testing against IAV.

Late-onset treatment with 4'-FlU mediates complete survival of IAVinfected animals
For mouse efficacy studies, animals were infected intranasally with a lethal inoculum amount of pdmCa09, which replicates efficiently in mice without a requirement for species adaptation [28]. Infection results in severe viral pneumonia with major clinical signs such as weight loss and hypothermia. To establish the lowest efficacious dose, we examined 3 4'-FlU dose levels in a survival study, each administered by oral gavage on three consecutive days in a once daily (q. d.) regimen starting at the time of infection ( Fig 2C). All animals treated at the 2 mg/kg dose level survived, whereas all mice of the vehicle-treated and lowest dose (0.08 mg/kg) groups succumbed to infection within 5 days of inoculation ( Fig 2D)  Having established 2 mg/kg 4'-FlU administered q.d. as a low efficacious dose, we next determined the latest time for initiation of effective treatment at this dose level ( Fig 2E). All vehicle-treated animals again succumbed to infection within 5 days of inoculation ( Fig 2F). In contrast, treatment start at 2.5 days after infection, which corresponds to 0.5 days after the onset of major clinical signs (S3C and S3D Fig

PLOS PATHOGENS
4'-Fluorouridine mitigates lethal influenza virus infection treatment was corroborated by a significant reduction in lung virus load compared to vehicletreated controls at 4.5 days after infection, determined in a parallel set of equally treated animals ( Fig 2G). Within 7 days, virus became undetectable in lungs of animals that were treated starting 24 hours after infection with oral 4'-FlU (2 mg/kg q.d.; 3 doses total), indicating that viral replication had ceased ( Fig 2H). No rebound of virus replication was observed within a 9-day period after infection (S3E and S3F Fig). These results define a broad therapeutic time window for effective treatment of IAV infection with 4'-FlU.
To determine the minimal number of oral 4'-FlU doses (2 mg/kg q.d.) for complete survival, we again initiated treatment 24 hours after infection, in this study comparing the effect of a single vs double or triple doses (Fig 2I). Lung virus load was assessed 4.5 days after infection and survival monitored in a parallel set of equally treated animals. The number of doses administered correlated with antiviral effect size, but even a single dose of 4'-FlU resulted in a statistically significant reduction in lung virus load of approximately one order of magnitude ( Fig 2J). Reduced viral burden translated to survival of some animals in the single and double dose groups. However, three doses were required for full therapeutic benefit ( Fig 2K)

4'-FlU effect on IAV transmission in the ferret model
Since mice, in contrast to ferrets, do not shed IAVs, we next explored anti-IAV efficacy of 4'-FlU in the ferret model. Guided by the dosing regimen established in mice, we treated animals infected intranasally with pdmCa09 with either 2 mg/kg (standard dose) or 0.5 mg/kg (low dose), administered orally starting 24 hours after infection ( Fig 3A). Animals received one (2 mg/kg level) or two doses (either 2 or 0.5 mg/kg level) total. Fever is the predominant clinical sign associated with pdmCa09 IAV infection in ferrets. Monitoring body temperature continuously telemetrically, we observed significant reduction in all treated animals within 24 hours of the first dose, whereas vehicle-treated animals showed a prolonged fever plateau (S4A Fig). Viral titers in nasal lavages were statistically significantly reduced compared to those of vehicle animals, but effect size was lower in animals of the low dose than the standard dose group, and virus rebounded 2.5 days after infection in the single standard dose group ( Fig 3B). In contrast, two standard doses 4'-FlU given q.d. were sterilizing within 2 days of treatment initiation. Virus load in nasal turbinates assessed 4 days after infection corroborated the lavage titers, revealing a dose-dependent significant antiviral effect of two doses, whereas a single dose had no significant effect on virus titer in turbinates at the time of tissue harvest ( Fig 3C). Examination of lower respiratory tract tissues, trachea, and bronchoalveolar lavage fluid (BALF) revealed, however, that a single standard dose of 4'-FlU administered at the time of onset of clinical signs completely suppressed the development of viral pneumonia ( Fig 3D).
To determine the effect of 4'-FlU on virus spread, we treated in a direct-contact transmission study source ferrets infected with pdmCa09 with the compound as before, starting 12 or 24 hours after infection ( Fig 3E). Co-housing of source ferrets with untreated sentinels in a 1:1 ratio was initiated 2.5 days after study start. pdmCa09 rapidly spread from vehicle-treated source ferrets to their direct contacts, which first became virus-positive in nasal lavages only 12 hours after the beginning of co-housing ( Fig 3F) and had progressed to viral pneumonia with high turbinate, lung and trachea virus load by study end (Fig 3G and 3H) and clinical signs (S4B to S4E Fig). Treatment of source animals with 4'-FlU initiated 12 or 24 hours after infection in each case resulted in exponential decline in shed virus load, but virus was undetectable in lavages of the 12-hour treatment group by the time of co-housing whereas approximately 10 4 TCID 50 units/ml lavage remained detectable in lavages of the 24-hour treatment group at that time. We did not detect any virus transmission to untreated sentinels of the

4'-Fluorouridine mitigates lethal influenza virus infection
12-hour group. However, low-grade transmission to direct contacts of the 24-hour treatment group occurred, characterized by delayed onset of virus shedding compared to sentinels of vehicle-treated source animals and a reduced peak shed virus load ( Fig 3H). Shedding from all sentinels that became virus-positive ceased approximately 12 days after the beginning of co- housing. These results confirm cross-species anti-IAV efficacy of 4'-FlU, reveal that a single oral dose is sufficient to prevent advance to severe influenza, and demonstrate a shortened window in which a treated case remains fully infectious.

Mitigation of IAV pathogenesis and lung histopathology by 4'-FlU
To better understand the impact of 4'-FlU on mitigating immunopathogenesis, the major driver of clinical signs associated with influenza virus infection, we determined inflammatory cytokine profiles for a 10-day period after infection of mice with pdmCa09 using a 6-plex Th17 cytokine assay ( Fig 4A). Treated animals received the 3-dose standard 4'-FlU regimen as before, started 24 hours after infection. Lung tissue for histopathology was extracted from parallel sets of equally treated animals 5 or 21 days after infection. Lung titers in vehicle-treated animals demonstrated fast-onset viral pneumonia, approaching a burden of 10 6

4'-FlU control of influenza in immunocompromised hosts and of HPAI infection
In the clinic, the risk of progression to severe disease and poor overall outcome is greatly increased when immunocompetence of the patient is impaired or after zoonotic spillover of HPAIs into the human population [29,30]. To first assess the potential of 4'-FlU to pharmacologically manage influenza in high-risk hosts, we infected interferon α/β receptor knockout (IFNAR1 KO) and V(D)J recombination activation gene RAG-1 KO mice with pdmCa09. Whereas the former cannot mount an effective type I IFN innate response [31,32], the latter are compromised in their cell-mediated adaptive immunity due to a lack of mature T and B cells [33]. Oral treatment with 4'-FlU was initiated 24 hours after infection at standard and a high-dose (10 mg/kg) level, then continued q.d. for a total of 7 doses of each group (Fig 5A). Independent of 4'-FlU dose level, all treated animals of the RAG-1 KO and IFNAR1 KO groups survived, whereas all animals in the vehicle-treated groups developed severe clinical signs (S10A to S10D Fig) and succumbed to infection within 4 days (Fig 5B). Reduction in lung virus load of treated vs vehicle animals after three q.d. 4'-FlU doses (day 4 after infection) was dose level-dependent, but even at the standard dose, this reduction exceeded two orders of magnitude in either knockout mouse strain ( Fig 5C).
Next, we explored 4'-FlU efficacy against HPAIs, initially testing potency against polymerases derived from A/CA/07/2009 (H1N1), A/VN/12/2003 (H5N1), and A/Anhui/1/2013 (H7N9) in minireplicon dose-response assays (Fig 5D). The compound dose-dependently inhibited activity of either HPAI-derived polymerase complex, reaching active concentrations resembling that observed against pdmCa09 RdRP. Selecting A/VN/12/2003 (H5N1), which causes lethal encephalitis in mice [34], for a proof-of-concept efficacy study of 4'-FlU against zoonotic HPAIs, we started treatment at the 5 mg/kg dose level 12 or 24 hours after infection and continued q.d. thereafter (Fig 5E). This elevated dose was chosen to compensate for lower brain tissue exposure of 4'-FlU (Fig 2A) when treating infection by a CNS-invasive IAV. None of the animals of either 4'-FlU treatment group developed clinical signs even when treatment was initiated 24 hours after infection (Fig 5F), and all treated animals survived (Fig 5G). In contrast, 19 out of 20 vehicle-treated animals succumbed to A/VN/12/2003 (H5N1) with a median survival time of 9 days after infection, developing clinical signs on day 7 after infection followed by rapid further deterioration (Fig 5F). Vehicle-treated animals experienced high lung virus loads three days after infection ( Fig 5H). Treatment reduced HPAI burden by several orders of magnitude to limit of detection. These results confirm uncompromised efficacy of 4'-FlU in high-risk immunocompromised hosts and against highly pathogenic zoonotic viruses with high pandemic concern.

Anti-IAV immune status of 4'-FlU-treated recoveries
Highly efficacious antivirals may suppress host immune activation, especially when treatment is started early in the course of infection, which may leave the host susceptible to subsequent re-infection and may potentially even result in exacerbated disease [35][36][37]. To explore the effect of 4'-FlU on subsequent homotypic IAV rechallenge of recoveries, we infected animals with pdmCa09 and started standard 3-dose 4'-FlU treatment at three time points, therapeutically at 24 and 48 hours after infection and prophylactically at 12 hours before infection ( Fig  6A). For this study, mice were inoculated with aerosolized virus in a temperature and airflowcontrolled aerosol chamber using an Aeroneb nebulizer (S11A and S11B Fig) to better mimic natural infection. Prior to study start, we performed pilot experiments to determine suitable exposure time (S12A to S12C Fig), infectious dose (S13A and S13B Fig) and proof-of-concept for 4'-FlU efficacy (S14A to S14D Fig) after infection with aerosolized virus. Mirroring the outcome of our previous efficacy studies, animals of all treatment groups survived the original infection, whereas all animals of the vehicle group reached predefined clinical endpoints 9 days after infection and had to be terminated (Fig 6B) with major clinical signs (S15A and S15B Fig). To assess the status of humoral anti-Ca09 immunity, we collected blood samples from the treated recoverees 4, 8, and 12 weeks after the original infection and determined neutralizing antibody (nAb) titers. Mice in both therapeutic treatment groups mounted a robust humoral anti-IAV response, whereas no nAbs were detectable in the prophylactically treated animals 8 weeks after the original infection (Fig 6C). Upon homotypic rechallenge with aerosolized pdmCa09 10 weeks after the original infection, however, all 4'-FlU-experienced recoverees had a significant benefit over a new set of IAV-naïve mice (Fig 6D). All animals of the therapeutic groups survived with significant boost in nAb titers relative to week 8 levels.
Remarkably, four of the five animals treated prophylactically with 4'-FlU at the time of the original infection also survived the homotypic rechallenge and mounted a robust humoral anti-Ca09 response by week 12. In contrast, all newly added IAV-inexperienced animals succumbed to infection within 9 days, confirming efficient virus delivery through aerosolization of two individual animals per study arm are shown at 10× magnification; scale bar denotes 100 μm; Br, Bronchiole; Bl, Blood vessel. f) Histopathology scores of animals treated as in (a). Lungs were extracted 5 days after infection. Scores for each animal represents a mean of individual alveolitis, bronchiolitis, vasculitis, pleuritis, perivascular cuffing (PVC), and interstitial pneumonia (IP) scores. Columns represent data medians with 95% CI; symbols show mean scores for each individual animal; 1-way ANOVA with Dunnett's post hoc test, P values and n values for each study arm are specified. https://doi.org/10.1371/journal.ppat.1011342.g004

PLOS PATHOGENS
4'-Fluorouridine mitigates lethal influenza virus infection at rechallenge. Animals of the treatment group remained disease-free during rechallenge, and clinical signs were alleviated in mice of the prophylactic group (S15C and S15D Fig). Independent of the original 4'-FlU regimen applied, we did not detect any signs of exacerbated disease upon rechallenge of treatment-experienced recoverees.

Discussion
Treatment of influenza has remained challenging, despite the current availability of two different FDA-approved drug classes, the neuraminidase and endonuclease inhibitors [38]. The main obstacles to more successful antiviral therapies are narrow therapeutic time windows [39], poor patient compliance [40], and impaired immunocompetence of older adults at greatest risk of advance to severe disease [17], defining key developmental objectives for the next generation of influenza therapeutics.
Demonstrating antiviral potency in cell culture, human airway epithelium cells, and/or efficacy in two relevant in vivo infection models, ferrets and mice, this study establishes oral efficacy of once-daily administered 4'-FlU, a broad-spectrum antiviral ribonucleoside analog [24,25], against seasonal, pandemic, and HPAI IAV strains. Designed to expand the first-line pharmacological arsenal against RNA viruses of pandemic concern, we have previously shown activity of 4'-FlU against SARS-CoV-2, pneumoviruses such as RSV, paramyxoviruses, and other members of the mononegaviruses [24]. Mechanistic characterization consistently demonstrated that incorporation of 4'-FlU-TP into the nascent RNA strand induces polymerase chain termination [24]. Unique to 4'-FlU of all chain-terminating ribonucleoside analogs analyzed [25], however, the modus of 4'-FlU-induced termination varies dependent on RdRP target. IAV RdRP stalled immediately upon incorporation, RSV polymerase termination was

4'-Fluorouridine mitigates lethal influenza virus infection
delayed at position i+3 after incorporation, and SARS-CoV-2 polymerase termination was delayed and dependent on sequence context, requiring incorporation of at least two 4'-FlU moieties in close repetition [24]. Delayed chain termination typically reflects that incorporation of the nucleoside analog alters secondary structure of the nascent strand, which interferes with polymerase processivity [25,41,42]. Accordingly, we hypothesize that distinct mechanisms of action (MOAs) against betacoronaviruses, pneumoviruses, and orthomyxoviruses are due to different capabilities of the respective polymerase complexes to accommodate 4'-FlU-specific secondary structure changes.
In clinical trials of baloxavir marboxil, viral resistance emerged rapidly, resulting in rebound of virus replication in 82.1% of treated patients [23]. Whole genome sequencing of 4'-FlU-experienced virus populations isolated from infected and treated animals showed no allele-dominant divergence from the inoculum population [24]. However, future full resistance profiling of 4'-FlU against influenza viruses will be required to better appreciate the height of the genetic barrier against viral escape from inhibition and explore the molecular basis for the slight virus strain-dependent differences in sensitivity to 4'-FlU that we observed in cell culture dose-response assays. Greater potency of 4'-FlU on differentiated HAE models compared undifferentiated HAEs and immortal cell lines especially likely reflects more efficient intracellular anabolism of the compound to the bioactive triphosphate form 4'-FlU-TP and altered viral replication kinetics in differentiated versus undifferentiated HAEs.
Very likely representing a direct consequence of the distinct MOAs, we found influenza viruses to be highly sensitive to inhibition by 4'-FlU, followed by RSV and SARS-CoV-2. Dose-finding in the mouse model identified effective levels in the low mg/kg range and three doses administered q.d. to be sufficient for achieving full therapeutic benefit, which resembled efficacy performance of baloxavir marboxil [43] and far outpaced that of the neuraminidase inhibitors [44]. Whereas therapeutic oseltamivir lacked efficacy in mice infected with pdm09 IAV isolates [45] and baloxavir marboxil ensured only partial survival when treatment was started later than 24 hours after infection [43], treatment with 4'-FlU mediated unprecedented complete survival of all treated animals when the inhibitor was first administered as late as 60 hours after infection, at the time when lung virus burden approached peak titer [46] and major clinical signs had manifested. If predictive of the treatment of human disease, this substantially widened therapeutic time window may be game-changing for influenza therapy.
Since clinical signs of influenza are largely a result of immunopathogenesis [47], we asked whether 4'-FlU may impact the quality of the antiviral immune response and potentially trigger exacerbated disease upon homotypic re-infection. Cytokine profiling showed an alleviated pro-inflammatory response in animals dosed therapeutically with 4'-FlU, but assessment of neutralizing antibody titers revealed that each of the treated animals had mounted a robust humoral response. Consistent with our previous observation of low cytotoxic and cytostatic potential of 4'-FlU [24], these results indicate that mitigated immunopathogenesis in treated animals is a consequence of reduced virus load rather than a direct immunomodulatory effect of the compound. Although essentially sterilizing, even prophylactic treatment resulted in partial protection of animals at re-infection despite the absence of neutralizing antibodies, indicating that treated animals also mounted a viable cell-mediated adaptive response [48,49].
Beyond individual patient benefit through mitigation of lethal infection and alleviation of viral pneumonia and lung immunopathogenesis, 4'-FlU statistically significant reduced upper respiratory tract virus load and viral shedding when treating 12 hours after infection, resulting in fully suppressed spread to untreated direct-contact sentinels in the highly sensitive ferret transmission model [50]. Later treatment initiation still reduced transmission efficiency, manifesting in delayed onset of virus replication and reduced upper respiratory tract peak virus load, which serves as a biomarker for risk of progression to severe viral pneumonia [51]. If

4'-Fluorouridine mitigates lethal influenza virus infection
equally applicable to the human host, pharmacological block or mitigation of transmission provides a promising path towards outbreak control through efficient interruption of community transmission chains, mitigating the health threat and economic burden [52] of endemic and pandemic influenza.
At greatest risk of life-threatening disease are the immunocompromised [53,54] and patients infected with zoonotic HPAI viruses [55,56]. When assessed in an immunocompromised mouse model, baloxavir marboxil returned a significantly prolonged time to death, but ultimately all animals succumbed to the infection [57]. In contrast, treatment with 4'-FlU mediated complete survival of all animals in two distinct immunocompromised mouse models, underscoring powerful antiviral activity that does not strictly depend on a fully immunocompetent host to clear the infection [58]. Adding efficient control of highly pathogenic zoonotic IAVs, we propose that 4'-FlU meets key efficacy requirements of a next-generation antiviral clinical candidate that strengthens pandemic preparedness and provides a muchneeded alternative option for management of seasonal and pandemic influenza viruses.

Ethics statement
All animal work was performed in compliance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and the Animal Welfare Act Code of Federal Regulations. Experiments involving mice and ferrets were approved by the Georgia State University Institutional Animal Care and Use Committee (IACUC) under protocols A20012 and A21020, respectively. In vivo experimentation with HPAI viruses was approved by the University of Georgia at Athens IACUC under protocol A2020 03-033. All experiments using infectious material were approved by the Georgia State University and the University of Georgia at Athens Institutional Biosafety Committees (IBCs) and performed in BSL-2/ABSL-2 or BSL-3/ ABSL-3 containment facilities, respectively. Experiments involving highly pathogenic avian influenza (HPAI) viruses were reviewed and approved by the IBC of the University of Georgia at Athens and were conducted in enhanced biosafety level 3 (BSL3+) containment according to guidelines for the use of select agents approved by the Centers for Disease Control and Prevention (CDC).

Study design
Cells, ferrets and mice were used as in vitro, ex vivo and in vivo models, respectively, to examine efficacy of 4'-FlU against influenza infections. Biochemical RdRP assays were added for mechanistic characterization against the influenza virus target. Viruses were administered through intranasal or aerosolized (mice only) inoculation and virus load monitored periodically in nasal lavages (ferrets only), and in respiratory tissues of ferrets and mice extracted 4 days and 3-14 days, respectively, after infection. Virus titers were determined through TCID 50 -titration.

Cells, viruses, and compound synthesis
MDCK (ATCC CCL-34) and 293T (ATCC CRL-3216) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 7.5% heat-inactivated fetal bovine serum (FBS) at 37˚C and 5% CO 2 . Human primary cells were grown in epithelial differentiation medium [24]. (Victoria) were propagated on MDCK cells using serum-free DMEM supplemented with 0.5% Trypsin. All influenza A viruses were grown at 37˚C, influenza B viruses at 32˚C. A/Viet Nam/ 1203/2004 (H5N1) and A/duck/Alberta/35/76 (H1N1) were propagated in the allantoic cavity of embryonated hen eggs at 37˚C for 24 or 48 hours, respectively, clarified, aliquoted, and stored at −80˚C. Virus stock titers were determined through TCID 50 -titration or plaque assay (A/Viet Nam/1203/2004 (H5N1)), stocks were stored in aliquots at -80˚C. Recombinant influenza viruses were recovered using an 8-plasmid system generated for A/CA/07/2009 (H1N1) that was based on original reports for laboratory-adapted IAV [59]. Reporter viruses were generated through insertion of a nanoLuc or maxGFP-encoding ORF in frame at the 5'-end of the viral HA ORF, separated by a 2A cleavage site as described for other recombinant IAVs [60]. Genetic stability of the resulting nanoLuc-HA and maxGFP-HA segments was validated through sequence confirmation after 10 consecutive viral passages on MDCK cells. 4'-FlU was synthesized from 5'-(3-chlorobenzoyloxy)-2',3'-di-O-acetyl-4'-fluorouridine and 4N ammonia in methanol as previously described [24], authenticated through elemental analysis, and stored as dry powder. Working aliquots were dissolved in DMSO or 10 mM sodium citrate with 0.5% Tween 80 for in vitro and in vivo studies, respectively.

Virus titration
Virus samples were serially diluted (10-fold starting at 1:10 initial dilution) in serum-free DMEM supplemented with 0.5% Trypsin (Gibco). Serial dilutions were added to MDCK cells seeded in 96-well plate at 1×10 4 cells per well 24 hours before infection. Infected plates were incubated for 72 hours at 37˚C with 5% CO 2 , followed by transfer of culture supernatants to suspension of chicken red blood cells and scoring of wells based on hemagglutination activity. A/Viet Nam/1203/2004 (H5N1) virus and samples were titered by plaque assay; serially diluted 10-fold in DMEM with 2% FBS, added to MDCK cells in 12-well plates and overlaid with DMEM+2% FBS with 1.2% Avicel microcrystalline cellulose. Plates were incubated for 48 hours at 37˚C with 5% CO 2 , then washed, fixed with methanol:acetone (80: 20), and counterstained with crystal violet to visualize plaques.

Dose-response assays
Active concentrations (EC 50 and EC 90 ) of 4'-FlU were calculated through 4-parameter variable slope regression modeling of dose-response assay data, derived from progeny virus yield (TCID 50 ) assays, luciferase reporter virus assays, or minigenome assays as specified. To test compound potency, 4-FlU was in all dose-response assays assessed in 3-fold serial dilutions, each tested in 3 independent repeats (biological repeats), with starting concentrations of 20 μM 4'-FlU. Control wells received vehicle (DMSO) volume equivalents. Dose-response assays on MDCK cells and 2D HAE cultures were carried out in 96-well plate format (1.5 × 10 4 cells/well) or 24-well plate format (1.5 x 10 5 cells/well), respectively. MDCK cells were infected in serum-free DMEM supplemented with 0.5% TPCK-treated trypsin (Gibco) with an MOI of 0.02 TCID 50 units/cell, HAE cultures were infected in BronchiaLife media (LifeLine) supplemented with 0.75 μg/ml TPCK-treated trypsin with an MOI of 1.0 TCID 50 units/cell. Culture supernatants were removed 24 (HAE cultures) or 36-48 hours (MDCK cells) after infection and viral titers determined through TCID 50 titration. For luciferase reporter viruses, relative luciferase activity was read in situ using a BioTek Synergy H1 multimode reader and signals normalized according to the formula % relative activity = (RLU i −R-LU min )/(RLU max −RLU min ) × 100, with RLU i specifying sample luciferase value, RLU min specifying values from equally transfected, vehicle-treated cells that did not receive the PB2

Primary human airway epithelium cultures grown at air-liquid interface
Primary HAEs (3 × 10 5 /well) at passage <3 were seeded on 6.5 mm 0.4 μm pore size polyester transwell inserts (Corning Costar Transwell) and differentiated for at least 21 days at air-liquid interface using Air-Liquid Interface Differentiation Medium (LifeLine Cell Technology). Developing HAEs were washed apically at least once per week to remove mucus. Transepithelial/transendothelial resistance (TEER) was measured every 24 hours for 3 days after infection using an EVOM volt/ohm meter coupled with STX2 electrode (World Precision Instruments). For infection, virus (5,000 TCID 50 units/transwell) was added apically. Compound was administered to the basolateral chamber, starting with a highest concentration of 0.5 μM 4'-FlU in dose-response assays, and exchanged once daily. Each concentration was tested in three independent replicates (biological repeats). Apically shed virus titers were determined through TCID 50 titration on MDCK cells as described above.

Minigenome assay
293T cells were seeded and transfected with 1.0 μg of luciferase-encoding reporter plasmid [26] and 0.5 μg each of PB1, PB2, NP and PA expression plasmids derived from A/California/ 07/2009 (H1N1), A/Viet Nam/1203/2004 (H5N1), and A/Anhui/1/2013 (H7N9), respectively. Three hours post transfection, 4'-FlU or vehicle (DMSO) volume equivalents were added in 3-fold serial dilution series starting at 20 μM 4'-FlU concentration, followed by incubation for 30 hours at 37˚C and determination of luciferase activity using a BioTek Synergy H1 multimode plate reader. Relative activities were calculated according to the formula % relative activity = (RLU i −RLU min )/(RLU max −RLU min ) × 100, with RLU i specifying sample luciferase value, RLU min specifying values from equally transfected, vehicle-treated cells that did not receive the PB2 subunit encoding plasmid, and RLU max specifying values from fully transfected, vehicletreated cells.

In vitro RdRP primer extension assay
In vitro primer extension assays were carried out essentially as described [24,61]. Briefly, RdRP assay reactions were set up in 20 mM Tris-HCl pH 7.4, 10% glycerol, 1 mM TCEP, 3 mM MnCl2, 2 μM RNA template, 100 μM RNA primer, and nucleotides and 1 μCi of α 32 P-labelled ATP (Perkin-Elmer) as specified in the figure legends, followed by incubation at 30˚C for 10 minutes. Polymerase complexes were added to the reaction and incubation continued for 1 hour at 30˚C. Reactions were stopped with 1 volume of deionized formamide containing 25 mM ethylenediaminetetraacetic acid (EDTA), followed by a denaturation step at 95˚C. Amplicons were fractionated on 20% polyacrylamide gels with 7M urea, using a

Pharmacokinetic property and tissue distribution profiling in mice
Male and female C57BL/6J mice were rested for 3 days prior to study start. Then, animals were gavaged with 1.5 mg/kg bodyweight 4'-FlU in 10 mM sodium citrate with 0.5% Tween 80, final gavage volume 200 μl. Blood samples (150 μl) were taken at 90 minutes, 3 hours and 12 hours after dosing, cleared (2,000 rpm, 5 minutes, 4˚C), and extracted plasma stored at -80˚C. Organ samples (lung, brain, spleen, heart, kidney and liver) were harvested 12 hours after dosing and snap-frozen in liquid nitrogen. Animal tissues were homogenized with 70% acetonitrile in water that included internal standards. Animal plasma and tissue concentrations of 4'FlU and 4'-FlU-TP were measured by a qualified LC/MS/MS method in MRM mode on a QTRAP 5500 (Sciex, Santa Clara, CA, USA) instrument.

4'-FlU anti-IAV efficacy studies in mice
Mice were inoculated intranasally with IAV stocks as described, followed by treatment with 4'-FlU at specified dose levels and starting time points through oral gavage. Unless specified otherwise, treatment was continued q.d. Each individual study contained animals receiving equal

4'-Fluorouridine mitigates lethal influenza virus infection
volumes of vehicle through oral gavage. At study end point, lung tissue was harvested for virus titration or histopathology assessment. For tissue fixation, lungs were perfused using 10% neutral-buffered formalin [10% formalin, NaH 2 PO 4 (4 g/liter), and Na 2 HPO 4 (6.5 g/liter)], dissected, and fixed for 24 hours. Formalin-fixed lungs were transferred to 70% EtOH after 72 hours, embedded in paraffin, sectioned (5-μm thickness), and stained with hematoxylin and eosin. Slides were evaluated by a licensed pathologist. A pathology score range of 0 to 3 was used to evaluate tissue damage. For analysis of proinflammatory cytokine levels through Bioplex, lungs of euthanized mice were lavaged with 1 ml sterile PBS to extract BALFs, and cleared BALF analyzed against the Standard Th17-panel (Bio-Rad) according to the manufacturer's instructions.

Efficacy studies in ferrets
Female ferrets (6-10 months of age) were purchased from Triple F Farms. Upon arrival, ferrets were rested for 1 week, then randomly assigned to study groups and housed in groups of three animals under ABSL-2 conditions. In some studies, animals received a temperature sensor for continued telemetric measurement of body temperature. Otherwise, body temperature was determined rectally once daily. Dexmedetomidine/ketamine anesthetized animals were inoculated intranasally with 1×10 5 TCID 50 units of A/CA/07/2009 (H1N1) in a volume of 500 μl per nare. Nasal lavages were performed twice daily using 1 ml of PBS containing 2× antibioticsantimycotics (Gibco). Treatment with 4'-FlU through oral gavage followed a q.d. regimen. Ferrets were terminated four days post infection and BALF and organ (lungs, turbinates and tracheas) samples extracted for virus titrations.

Ferret transmission studies
Female ferrets (3-8 months of age) were anesthetized with dexmedetomidine/ketamine and infected intranasally with 1×10 5 TCID 50 units of A/CA/07/2009 (H1N1) as before. Nasal lavages were performed every 12 hours, bodyweight and temperature determined once daily. Treatment of source animals was initiated 12 or 24 hours after infection at a dose of 2mg/kg bodyweight, and continued q.d. for three doses total. Starting day 2.5 after infection, uninfected and untreated contact animals were added to the infected and treated source ferrets to allow direct-contact transmission, and co-housing continued until termination of the source ferrets on study day 4. Sentinels were subjected to once daily nasal lavages and monitored until study day 10 or, when infectious particles were detected in lavages of contacts of 4'-FlU-treated source animals, study day 14.

Virus titration from tissue samples
Organs were weighed and homogenized in 300 μl PBS using a beat blaster, set to 3 cycles of 30 seconds each at 4˚C, separated by 1-minute rest periods. Homogenates were cleared (10 minutes at 20,000 × g and 4˚C), and cleared supernatants stored at -80˚C until virus titration. Viral titers were expressed as TCID 50

Infection of mice with aerosolized A/CA/07/2009 (H1N1)
An Aeroneb (Kent Scientific) system was used to infect mice (female Balb/c, 6-8 weeks of age) with aerosolized IAV. Conscious animals were transferred to an aerosol chamber placed on a

4'-Fluorouridine mitigates lethal influenza virus infection
heat pad adjusted to 37˚C to prevent condensation, connected to the nebulizer unit (Aeroneb with palladium mesh) and air pump with airflow restrictor. For condition-finding experiments, the nebulizer chamber was loaded with virus inoculum diluted to 1×10 3 −1×10 5 TCID 50 A/CA/07/2009 (H1N1) per ml in sterile PBS, airflow adjusted to 1 l/minute, and animals continuously exposed for 3, 10, or 30 minutes. For treatment studies, conditions were adjusted to 1×10 5 TCID 50 /ml in sterile PBS, 10 minutes exposure time, and 1 l/minute airflow.

Re-infection of 4'-FlU-experienced mice with A/CA/07/2009 (H1N1)
Mice infected with aerosolized A/CA/07/2009 (H1N1) as above were treated q.d. with 4'-FlU or vehicle as specified, and monitored for 10 weeks or until humane endpoints were reached. Blood was collected 4 and 8 weeks after infection and A/CA/07/2009 (H1N1)-neutralizing antibody titers determined. Treated recoverees were reinfected with homotypic, aerosolized A/CA/07/2009 (H1N1) 10 weeks after the original infection along with a fresh set of naïve animals, followed by monitoring for an additional 14 days and final blood collection at study end.
No treatment was administered after reinfection.

Neutralizing antibody titers in mice
Plasma was prepared from blood samples (2,000 rpm, 10 minutes, 4˚C) and stored at -80˚C until analysis. Heat-inactivated plasma was incubated with 100 TCID 50 units of A/CA/07/2009 (H1N1) for 90 minutes, followed by serial dilution and transfer to MDCK cells, and plates incubated for 3 days at 37˚C. Remaining infectivity was visualized through hemagglutination activity on chicken red blood cells.

Statistical analysis
For statistical analysis of studies consisting of only two groups, unpaired two-tailed t-tests were applied. When comparing more than two study groups, 1-way analysis of variance (ANOVA) or 2-way ANOVA with multiple comparison post hoc tests as specified were used to assess statistical difference between samples. Statistical analyses were carried out in Prism version 9.4.1 (GraphPad). The number of individual biological replicates (n values) and exact P values are shown in the figures when possible. The threshold of statistical significance (α) was set to 0.05. All quantitative source data and statistical analyses are shown in S1 and S2 data files, respectively.