Inhibiting influenza virus transmission using a broadly acting neuraminidase that targets host sialic acids in the upper respiratory tract

ABSTRACT The ongoing transmission of influenza A viruses (IAV) for the past century continues to be a burden to humans. IAV binds terminal sialic acids (SA) of sugar molecules present within the upper respiratory tract (URT) in order to successfully infect hosts. The two most common SA structures that are important for IAV infection are those with α2,3- and α2,6-linkages. While mice were once considered to be an unsuitable system for studying IAV transmission due to their lack of α2,6-SA in the trachea, we have successfully demonstrated that IAV transmission in infant mice is remarkably efficient. This finding led us to re-evaluate the SA composition of the URT of mice using in situ immunofluorescence and examine its in vivo contribution to transmission for the first time. We demonstrate that mice express both α2,3- and α2,6-SA in the URT and that the difference in expression between infants and adults contributes to the variable transmission efficiencies observed. Furthermore, selectively blocking α2,3-SA or α2,6-SA within the URT of infant mice using lectins was necessary but insufficient at inhibiting transmission, and simultaneous blockade of both receptors was crucial in achieving the desired inhibitory effect. By employing a broadly acting neuraminidase to indiscriminately remove both SA moieties in vivo, we effectively suppressed viral shedding and halted the transmission of different strains of influenza viruses. These results emphasize the utility of the infant mouse model for studying IAV transmission and strongly indicate that broadly targeting host SA is an effective approach that inhibits IAV contagion. IMPORTANCE Influenza virus transmission studies have historically focused on viral mutations that alter hemagglutinin binding to sialic acid (SA) receptors in vitro. However, SA binding preference does not fully account for the complexities of influenza A virus transmission in humans. Our previous findings reveal that viruses that are known to bind α2,6-SA in vitro have different transmission kinetics in vivo, suggesting that diverse SA interactions may occur during their life cycle. In this study, we examine the role of host SA on viral replication, shedding, and transmission in vivo. We highlight the critical role of SA presence during virus shedding, such that attachment to SA during virion egress is equally important as detachment from SA during virion release. These insights support the potential of broadly acting neuraminidases as therapeutic agents capable of restraining viral transmission in vivo. Our study unveils intricate virus-host interactions during shedding, highlighting the necessity to develop innovative strategies to effectively target transmission.

Therefore, there is an urgent need to understand the biology of IAV transmission and devise improved strategies to contain its spread.
Although mice are not the natural hosts for IAV, they have become a valuable tool for studying IAV biology (1)(2)(3).However, mouse-to-mouse transmission has historically been inefficient and inconsistent, dampening its appeal as an IAV transmission model (1,(4)(5)(6)(7)(8)(9).Furthermore, mice have been shown to lack α2,6 sialic acid (SA) in their trachea, which is an important entry receptor for IAV strains significant to humans.
Avian influenza viruses preferentially bind α2,3-SA, which is expressed in the intestine of birds and is a site of viral replication.Recent studies have shown that the tracheal epithelium of birds also expresses both α2,3-SA and α2,6-SA, indicating a broader distribution of influenza virus receptors in birds than previously known (10,11).In contrast, humans predominantly express α2,6-SA in the upper respiratory tract (URT), and IAV that preferentially bind α2,6-SA have demonstrated efficient human-to-human transmission.While α2,6-SA is found in parts of the lower respiratory tract (LRT) such as the trachea, bronchi, and bronchioles, it is absent in alveolar cells.In contrast, α2,3-SA is mainly present in the LRT, particularly in alveolar cells, but can also be found in specific cells of the URT (12).
Studies evaluating the composition of SA receptors in the respiratory tract of mice have been limited to LRT-derived cells and tissues.In ex vivo murine tracheal epithelial cultures (mTEC), only α2,3-SA was detected (13), while in murine lungs, both α2,3-SA and α2,6-SA were present (14,15).However, SA distribution in the murine URT (nose, pharynx, and larynx), the anatomical compartment where initial infection takes place, has never been evaluated.Furthermore, while there have been indications that α2,6-SA might not be essential for human IAV infection and replication (14), its direct role in transmission was never assessed in an in vivo model.Studies of transmission have examined viral mutations of surface viral proteins, receptor binding preferences, and their correlation with transmission in ferrets (16)(17)(18)(19).Employing a tractable model like mice to study how host factors contribute to the dynamics of transmission can provide insights into the molecular intricacies of viral-host interactions crucial for our under standing of IAV contagion.
There are key bottlenecks determining transmission in mice which can be categorized into three main stages: virus replication in the URT of the infected (index) host (stage 1), release of virus from the URT (shedding) of the infected host (stage 2), and arrival of virus in the URT of uninfected (contact) hosts and establishment of productive infection (stage 3) (Fig. 1).Successfully thwarting transmission requires interruption of any of these stages.Thus, devising vaccines and therapeutics capable of diminishing viral replication, shedding, and/or transmission to contacts is pivotal in our battle against IAV spread.
Our previous studies illustrate that, in contrast to adult mice, infant mice support efficient transmission of influenza viruses among littermates in an age-dependent and a strain-dependent fashion (9).Furthermore, co-infection with S.pneumoniae reduced the efficiency of IAV transmission through the activity of bacterial nanA, a broadly acting neuraminidase (ba-NA), and not nanB, a neuraminidase that selectively targets α2,3-SA.This implied that α2,6-SA may be present and functional in the URT of infant mice.To explore this further, we naturally directed our in vivo studies to the URT in order to determine the SA repertoire of both adult and infant mice and evaluate the direct role of SA moieties in IAV shedding and transmission.

Mice express both α2,3-SA and α2,6-SA in the URT
Successful transmission of IAV in mice was previously determined to be more efficient in younger mice than older mice (9).This finding exhibited biological plausibility in humans, given that children are widely recognized as significant contributors to IAV transmission (20,21).Hence, we sought to evaluate which stage of transmission was most efficient in infants compared to adults.To achieve this, we infected C57BL/6 J mice with A/X-31(H3N2) using a low-volume intranasal (IN) inoculum to prioritize the infection to the URT (7,22).Infants received 300 plaque-forming units (PFU) in 3 µL, and adults received 10 5 PFU in 10 µL.After infection, index mice were placed back in the cage with the dam and naïve littermates.Daily shedding samples were collected by gently dipping the nares of each pup in viral medium.Transmission was assessed in contact pups by measuring infectious titers from URT lavages at 4 days post-infection (dpi), a timepoint where >90% transmission is achieved (Fig. 2A).Interestingly, index URT titers in adults were comparable to that in infants despite being infected with 1,000-fold more virus (Fig. 2B, left).However, the overall shedding titers from days 1 to 4 were higher in infants compared to adults.This difference was particularly pronounced during the initial 2 days when peak shedding is expected (Fig. 2B, middle) (9,23).Furthermore, transmission of A/X-31 to contact pups was significantly higher in infants (100%) compared to adults (33%) (Fig. 2B, right).Collectively, this suggest that while URT titers were similar between infants and adults at 4 dpi, there were notable disparities in shedding (stage 2) and in transmission to contacts (stage 3).Given that shedding is a correlate of transmission efficiency in this model (9), we focused our studies on virus-host interactions at the epithelium surface.Specifically, we investigated the distribution of SA in the murine URT and their role in IAV shedding and transmission.
To quantify the levels of α2,3-SA and α2,6-SA in the URT of infants and adults, we performed immunoblots of URT lavages collected after sacrifice.Interestingly, naïve infant mice had higher overall sialylation compared to those of adults by immunoblot of URT lavages and by in situ IHC immunofluorescence (IF) quantification (Fig. 2D through  E).Together, these findings indicate that both infant and adult mice possess the neces sary terminal SA moieties important for IAV infection, replication, and transmission.Although the expression of sialylated receptors in the URT of adult mice is comparatively lower, viral infection and replication can still take place when a higher virus dose is administered.

Blocking both α2,3-SA and α2,6-SA is necessary to inhibit IAV transmission
To evaluate the in vivo role of URT α2,3-SA and α2,6-SA in transmission, we infected infant mice with A/X-31 as previously described.Both index and contacts were subsequently treated twice a day (BID) with unconjugated lectins IN for 4 days, with the first treatment starting at ~6 h post-infection (hpi).Virus shedding was collected daily, and intralitter transmission was assessed in contacts at 4 dpi (Fig. 3A).We observed reduction of transmission in single-or double-treated lectin groups in a dose-dependent manner, which confirms the presence of functional α2,3-SA and α2,6-SA in the URT of infant mice and underscores their critical role as host receptors involved in the transmission of A/X-31 in this model (Fig. 3B).
We next examined the impact of SA blockade at different stages of transmission.To avoid potential molecular crowding in the URT from higher lectin concentrations, we used a mid-range dose of 4 mg/kg/day.In stage 1, blocking α2,3-SA and/or α2,6-SA did not affect index URT titers (Fig. 3C, left).In stage 2, the overall shedding titers during days 1-4 were similar between mock-infected infants and those with α2,3-SA and/or α2,6-SA blockade (Fig. 3C, middle).However, during the initial 2 days when peak shedding is expected, there was a significant decrease in shedding only when both α2,3-SA and α2,6-SA were simultaneously blocked.In stage 3, selectively blocking either α2,3-SA or α2,6-SA led to a small reduction in URT titers and transmission efficiency in contact mice.However, simultaneously blocking both α2,3-SA and α2,6-SA resulted in the most significant decrease of URT titers and transmission to naïve contacts (Fig. 3C, right).Collectively, these findings suggest that the presence of URT SA in index mice is essential for effective shedding (stage 2), and their presence in contact mice is crucial for efficient virus acquisition (stage 3) during exposure.
To investigate the specific role of α2,6-SA in shedding (stage 2) and transmission to contacts (stage 3), we used mice lacking the gene st6gal1 (st6gal1-KO), which results in the absence of β-galactoside α2,6-sialyltransferase 1 (ST6GAL1) (24).This enzyme catalyzes the addition of α2,6-SA to terminal β-galactose of glycoproteins and glycolipids present in mucosal surfaces.Hence, we confirmed the absence of α2,6-SA throughout the URT of st6gal1-KO infant mice via IHC (Fig. S2) as it has been previously shown for the LRT (14).Interestingly, the shedding and transmission efficiency of A/X-31 in st6gal1-KO was comparable to that in wild-type (WT) C57BL/6 J mice (Fig. 3D).To examine if residual α2,6-SA added by a second sialyltransferase, the ST6GAL2 enzyme, might be contributing to this phenotype, we administered SNL treatment to both index and contact st6gal1-KO to provide additional α2,6-SA blockade.Nevertheless, shedding from index mice remained unaffected, and transmission to contact mice was slightly reduced from 100% to 78.6%.This confirmed that the absence of α2,6-SA alone is insufficient to prevent viral shedding and transmission, as was previously shown for viral replication (14).We next determined if dual SA blockade was necessary to inhibit shedding and transmission by complementing the α2,6-SA deficiency in the st6gal1-KO mice with additional α2,3-SA blockade in trans.We again found that when both α2,3-SA and α2,6-SA are unavailable in blotted on a nitrocellulose membrane and immunostained with biotinylated lectins to assess for sialylation of eluted glycoconjugates.Signals were normalized to background, and fold over adult mean signal was plotted.Median is denoted by a line.(E) Fluorescent signal from immunohistochemistry (in C) was measured following color unmixing using Image J. Measurements were taken from at least three random positions along the epithelial surface of each image, selected from at least two representative tissues.Background was subtracted, and fold change over the average adult values was calculated and adjusted for head size.Bars represent median ± interquartile range.All panels: Differences among two group medians were analyzed using the Mann-Whitney test.Experiments represent at least two biological replicates.Transmission rates were calculated using the Fisher's exact test.Infant mice were 4-7 days old; Adult mice were 8-10 weeks old; * (P < 0.05), ** (P < 0.01), *** (P < 0.001), **** (P < 0.0001), ns = not significant.Red brackets in statistical comparisons represent the reference group.H&E, hematoxylin and eosin stain; SNL, Sambucus nigra lectin; MAL, Maackia amurensis lectin; ECL, Erythrina crysta-galli lectin; S/N, signal-to-noise ratio.Dotted line denotes limit of detection (10 PFU/mL), where indicated; zero values were set to 1 PFU/mL.Cartoons were created with BioRender.com. the URT, the index shedding kinetics is altered and inhibited during peak viral shedding on days 1-2, and transmission to contacts is substantially reduced from 100% to 16.7% (Fig. 3D).These findings again indicate that the presence of both α2,3-SA and α2,6-SA is necessary for efficient shedding and transmission of IAV.
In vivo URT desialylation of both α2,3-SA and α2,6-SA inhibits IAV transmis sion DAS181 is a recombinant bacterial NA with enhanced bioavailability in vivo.It contains the enzymatic domain from Actinomyces viscosus NA fused to the epithelium-anchoring domain of the amphiregulin protein (25).DAS181 is a broadly acting neuraminidase promoting desialylation of both α2,3-SA and α2,6-SA and inhibiting influenza viruses (26)(27)(28)(29).This is in contrast to zanamivir, an NA inhibitor (NA-I), which targets the viral NA and prevents desialylation of host surface SA.DAS181 has been shown to desialylate murine lungs effectively, yet its effect on the URT had not been evaluated.Hence, naïve infants were either infected with A/X-31 or treated with ba-NA via IN administration for 2 days before sacrifice.URT desialylation was assessed by in situ IHC via lectin staining or by URT lavage immunoblots (Fig. 4A through C).Pups infected with A/X-31 showed detectable signals for IAV nucleoprotein (NP) in the olfactory and respiratory epithelium.As expected, both A/X31-infected and ba-NA-treated pups had increased desialylated β-galactose, indicated by a loss of ECL binding.Together, these results confirm that ba-NA is as effective as IAV infection in desialylating SA within the URT of pups.
To evaluate the therapeutic potential of in vivo desialylation on IAV transmission, we first infected infant mice with A/X-31 as previously described, then treated both index and naïve contact pups IN daily with ba-NA or NA-I for 4 days, with the first treatment starting at ~2 hpi.Virus shedding was collected daily, and intralitter transmission was assessed at 4 dpi (Fig. 5A).The inhibition of A/X-31 transmission was observed in a dose-dependent manner for both ba-NA and NA-I and exhibited different inhibitory profiles (Fig. 5B).To determine the therapeutic window of ba-NA, the inhibition of transmission was evaluated when pups were treated at delayed intervals after infection.Infant mice were infected with A/X-31, then treated with ba-NA or NA-I as previously described.The first cohort received treatment starting at 2 hpi, the second at 24 hpi, the third at 48 hpi, and the fourth at 72 hpi.Transmission was assessed in contacts at 4 dpi.Transmission inhibition by ba-NA was 100%, 86.7%, 35.8%, and 27.3% when treatment was initiated at 2, 24, 48, and 72 hpi, respectively (Fig. 5C).NA-I exhibited a similar transmission inhibition profile to ba-NA, with rates of 89.7%, 74.3%, 35.9%, and 29.5%, respectively.Next, we analyzed the effects of delayed ba-NA treatment at each transmission stage individually.In stage 1, delayed treatment with ba-NA had no significant impact on URT viral titers (Fig. 5D, left).In stage 2, viral shedding was reduced only when treatment with ba-NA treatment was initiated early after infection, specifi cally prior to 24 h (Fig. 5D, middle).In stage 3, ba-NA was effective at preventing viral transmission to contacts by >40% if administered within 48 h of exposure to the infected index (Fig. 5D, right).Collectively, these findings indicate that broad URT desialylation lowers index shedding and contact transmission in a time-dependent manner and is most effective within 48 h of index infection.
To determine which cohort had the most influence on IAV transmission, we treated index and/or contacts with ba-NA, and analyzed each transmission stage individually.Compared to mock, URT viral titers in the index were unaffected by any of the treatment combinations (Fig. 5E, left).Interestingly, overall shedding was reduced only when index pups received ba-NA treatment (Fig. 5E, middle), consistent with our prior observation that dual-receptor blockade using lectins resulted in reduced index shedding (Fig. 3C, middle).Furthermore, viral acquisition in contact pups was significantly inhibited when either the index, the contact, or both were targeted for URT desialylation, resulting in transmission efficiencies of 10.5%, 6.3%, and 0%, respectively (Fig. 5E, right).Together, these findings suggest that SA presence in the URT of index pups is important for effective viral shedding and that inhibiting IAV transmission can be achieved by reducing index shedding (stage 2) and/or viral acquisition by contacts (stage 3).
While it was anticipated that depleting α2,3-SA and α2,6-SA in the URT of contacts would prevent viral infection upon exposure, the unexpected finding was that depleting SA receptors in the URT of index pups also resulted in decreased viral shedding.Virus elution from the URT surface is not known to necessitate the presence of sialylated glycoconjugates.In contrast, it is known that the cleaving of terminal SA by viral NA promotes virus release (30).To investigate if the presence of SA and its interaction with the virus during egress is an important requirement for efficient shedding from the URT, we targeted host SA using ba-NA at different concentrations and assessed shedding and transmission of A/X31.Increasing the dose of ba-NA resulted in a corresponding decrease in overall shedding dynamics, peak shedding (days 1-2), and viral acquisition in contacts in a dose-dependent manner (Fig. 5F, left).To confirm that the dissociation of the virus from SA is also important for shedding, we inhibited the ability of viral NA to cleave SA using NA-I at various concentrations and assessed shedding and transmission.As the dose of NA-I increased, the overall shedding dynamics did not show significant changes compared to the mock group, but both peak shedding (days 1-2) and transmis sion decreased correspondingly (Fig. 5F, right).Collectively, these findings suggest that the mechanism of viral shedding from the URT is dependent on virion attachment to surface SA during egress, followed by virion detachment from SA during elution.

In vivo URT desialylation does not enhance IAV URT replication or transmis sion
Ba-NA functions by depleting SA from the surface of cells, thereby preventing important pathogen interaction.However, there exists a possibility that URT desialylation might lead to unintended enhancement of transmission by facilitating the release of progeny virions by complementing viral NA in the catalysis of surface SA (31)(32)(33).To address this concern, we first verified that the quantity and size of A/X-31 plaques were diminished in the presence of ba-NA using a plaque reduction assay in MDCK cells (Fig. 6A).This indicated that ba-NA directly hampers viral replication and spread, aligning with earlier observations from cell-protection assays (25).
Next, we evaluated the preventive effect of ba-NA on primary infection in vivo.Pups in a litter were pre-treated IN with PBS or ba-NA 1 day prior to A/X-31 IN challenge with 300 PFU.After sacrifice at 1 dpi, virus titers were assessed on URT lavage samples.Pretreatment with ba-NA led to a significant inhibition of URT infection following IN challenge, and in some cases, pups were completely protected from infection (Fig. 6B).An inhibitory effect on URT viral replication in the presence of ba-NA was also observed when the treatment was administered 2 hpi and URT replication assessed at 1 dpi.In both prophylaxis and therapeutic scenarios, ba-NA treatment did not enhance virus replication in the URT.
We proceeded to investigate whether ba-NA could enhance the transmission of other influenza virus subtypes that inherently exhibit a low transmission phenotype in this model (9).Index pups in a litter were first infected IN with selected viruses with 300 PFU, as previously described.Only index pups then received daily IN treatment with ba-NA for 4 days, starting at 2 hpi.Virus shedding was collected daily, and intralitter transmission was assessed at 4 dpi in mock (black circles) and ba-NA-treated (blue squares) groups at the different stages of transmission (Fig. 6C).The impact of ba-NA treatment on index URT titers (stage 1) varied among viral strains, with some displaying greater susceptibility to ba-NA than others, although none resulted in enhanced URT replication at 4 dpi (Fig. 6C-1).Index shedding titers (stage 2) were significantly reduced across all viral strains tested upon ba-NA treatment (Fig. 6C-2).Consistently, titers in untreated contact pups were also lower for all viral strains in the ba-NA index treatment group, particularly for those strains exhibiting moderate and high transmission rates in mock treatment group (Fig. 6C-3).Notably, there was no enhancement observed in any virus at any transmission stage following ba-NA treatment.Next, we investigated the impact of ba-NA treatment on the transmission of another respiratory pathogen, Streptococcus pneumoniae (Spn).Spn expresses its own NA, enabling it to remove SA from hosts and facilitate immune evasion, attachment, and invasion (34).Furthermore, during co-infection, IAV NA has been shown to enhance Spn colonization through the release of free SA which is used by the bacteria for nutrient (35,36).To assess whether the supplementation of ba-NA could increase URT Spn density, shedding, and transmission to naïve contact littermates, we employed the infant mouse model of Spn transmission (37).Index pups in a litter were first infected IN with 1,000 CFU of Spn.Only index pups then received daily IN treatment with ba-NA or PBS (mock) for 4 days.Spn shedding was collected daily by gently tapping the nares of pups onto bacterial plates 10 times.Intralitter bacterial transmission was assessed in contact littermates by measuring Spn levels in URT lavages at 4 dpi.Compared to the mock (purple circles), ba-NA (green squares) did not enhance URT Spn density, shedding, or transmission (Fig. 6C).In a separate experiment, Spn-infected index pups were co-infec ted with A/X-31 at day 3, and bacterial transmission was assessed in contact littermates at 4 dpi.Compared to mock, URT Spn density was moderately reduced in the ba-NA treatment group (P = 0.0571), which likely resulted in significant Spn shedding reduction (P = 0.0299) and Spn transmission efficiency (P = 0.0169).These findings suggest that URT desialylation by ba-NA did not result in an increase in Spn transmission at any stage, even if co-infected with IAV.In fact, the effect was inhibitory.Collectively, we demonstra ted the efficacy of ba-NA in preventing replication, shedding, and transmission of various IAV strains and Spn in our in vivo infant mouse model.Importantly, no enhancement of transmission for any pathogen was observed after therapeutic desialylation in any of the tested conditions in this model.

DISCUSSION
The interaction of IAV with host SA during viral entry and shedding is complex.The current understanding of IAV transmission efficiency in humans is heavily dependent on viral tropism studies and in vitro categorization of viruses as "avian-like" if they preferentially bind α2,3-SA or "human-like" if they preferentially bind α2,6-SA.The SA preference of IAV, however, is not sufficient to explain the strainspecific transmission phenotypes observed in humans and animal models.For example, A/Aichi/68(H3N2) and A/Victoria/75(H3N2) all have preferential binding to α2,6-SA in in vitro assays (38), yet transmission efficiencies in infant mice are 95% and 36%, respectively (9).Furthermore, the recombinant strain A/X-31(H3N2) that also has preferential binding to α2,6-SA in in vitro assays, and was confirmed to contain HA "human-like" residues 226L and 228S (GenBank: OQ925911.1),successfully transmits in WT mice and those lacking α2,6-SA (Fig. 3D) (38)(39)(40)(41)(42)(43).This implies that IAV may exploit multiple SA linkages during transmis sion, which may vary during the entry and exit phases of the life cycle.Gaining insight into the virus-SA interactions during shedding and entry is essential for our understand ing of IAV transmission biology.With the infant mouse model, we can dissect each stage of viral transit and study both viral and host factors important for transmission.In this study, we focused on understanding the contribution of the URT SA repertoire by leveraging mouse biology and demonstrating the critical role of the host's sialylated glycoconjugates in the shedding and acquisition of influenza viruses in vivo.We confirm that both α2,3-SA and α2,6-SA are present in the URT and lungs of infant and adult mice (Fig. 2C).Similar to the results in primary mTECs (13), α2,6-SA was not detected in infant or adult mice tracheas (Fig. S1).This implies that α2,6-SA expression in the trachea, the conduit to the LRT, is not essential for IAV transmission from the URT.Furthermore, IAV may be more promiscuous in their SA preference than previously thought, given that in the absence of its "preferred" SA, viruses may still utilize "non-pre ferred" moiety to drive its transmission (Fig. 3C through D).Our findings illuminate the intricacies of transmission and underscore the importance of incorporating in vitro and in vivo assays in future studies to gain a better understanding of the viral-host factors driving IAV transmission.
We also learned that the amount of SA present in the URT of infant and adult mice corresponded with the efficiency of shedding (stage 2) and transmission to contacts (stage 3) (Fig. 2).Using two independent methods, we found that adult mice, in general, had lower amounts of both α2,3-SA and α2,6-SA in the URT compared to infants, which may contribute to the lack of efficient transmission in this cohort.This implies that a threshold of SA molecules is necessary to facilitate effective shedding and ensure sustained infection in the receiving contact.We tested this hypothesis by decreasing the available SA in the URT and demonstrating inhibition of both index shedding and contact transmission using three different and independent approaches: (i) simultane ously blocking both α2,3-SA and α2,6-SA using SAspecific lectins (Fig. 3C); (ii) using an st6gal1-KO mouse, which fails to conjugate terminal α2,6-SA in all tissues, complemented with selective α2,3-SA blockade (Fig. 3D); and (iii) simultaneously removing both α2,3-SA and α2,6-SA in the URT using exogenous ba-NA (Fig. 4 and 5).Selectively inhibiting either α2,3-SA or α2,6-SA was insufficient at hindering IAV transmission.Hence, broad URT desialylation early during infection and exposure is a promising strategy to decrease the efficiency of IAV transmission.
The observation that decreasing the availability of SA in the URT led to a suppres sion of virus shedding was surprising because the presence of SA is not traditionally considered essential for virus URT release.In contrast, tissue culture studies have demonstrated that removing SA with exogenous NA promoted viral replication and elution from cells (32).To delve deeper into this discrepancy, we examined the shedding efficiency of IAV under two conditions: one where SA was absent (achieved with ba-NA) and another where the virus-SA interaction was disrupted (achieved with NA-I).Our results indicated that shedding was reduced in a dose-dependent manner in both conditions, suggesting that the production of progeny virions relies on a sequential process involving virus binding to SA to assist virion egress and formation, followed by its release from SA for final elution.This implies that SA could be a prospective host target for future antivirals.
Because of the unexpected success of ba-NA in suppressing shedding and inhibiting transmission of various IAV subtypes (Fig. 6C), we became interested in its effect on other URT pathogens that are known to cause secondary infections following IAV infection.Considering that IAV-Spn co-infection enhances bacterial colonization and pathogenesis through NA activity, we were apprehensive that ba-NA therapeutics may have inadver tent impacts on co-colonizing bacterial pathogens (36).To investigate this possibility, we evaluated the effects of ba-NA on Spn replication, shedding, and transmission in the presence or absence of co-infection with IAV.Our results indicated that ba-NA significantly reduced Spn shedding in the IAV-infected cohort.Moreover, none of the Spn cohorts, regardless of co-infection status with IAV, exhibited any enhancement of Spn replication, shedding, or transmission to contacts.These results expand upon earlier findings that demonstrated how ba-NA treatment of IAV-infected mice indirectly lowered secondary Spn pneumonia (44).This finding strengthens our confidence in the safety of employing ba-NA as a therapeutic, targeting the host to impede IAV shedding and transmission.Furthermore, a host-targeting therapeutic might have a lower risk of developing resistant variant strains compared to viral-targeting compounds.As a result, ba-NA could potentially become the first antiviral agent capable of obstructing viral shedding when directly applied to the nasal mucosa therapeutically, as well as providing protection against viral infection when used prophylactically.Ba-NA inhibitors hold the potential to be a significant advancement in antiviral therapy, offering the possibility of outwitting viral evolution.

Mice
C57BL/6J (000664, Jackson-Laboratories) and st6gal1-KO (Joseph Lau, Roswell Park Cancer Institute, NY) (24) were bred in a conventional animal facility.Pups were housed with their dam for the duration of experiments.Studies and procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (73) and the Biosafety in Microbiological and Biomedical Laboratories recommendations (45) and were approved by the Institutional Animal Care and Use Committee of NYU Langone Health (Assurance A3317-01).KAPA2G HotStart Mix (07961316001, Roche) was used in genotyping.

FIG 3
FIG 3 Therapeutic blockade of both α2,3 and α2,6-SA is required to inhibit influenza virus transmission.(A) Schematic of lectin treatment to block influenza virus transmission.At day 0, pups were infected IN with 300 PFU of A/X-31 (index) and cohoused with uninfected littermates (contact) for 4 days prior to being evaluated for transmission via URT lavage after sacrifice.Shedding was collected on days 1-4 from index pups, and infectious virus was quantified via plaque assay.IN lectin treatment was given to both index and contacts and started 6 hpi and continued twice daily until end of experiment.(B) Transmission rates of at least two independent experiments represented as percent (%) transmission (relative to mock), where 100% represents the average transmission rate of the mock-treated group [formula: y = (y/83.3)×100,where 83.3 is the transmission rate for the mock group].Error bars represent the standard error of the mean.Curves were fit using linear regression.(C) Transmission blockade by lectin inhibitor is represented by virus titers plotted at each stage of transmission.Left panel (stage 1, index URT replication), middle panel (stage 2, index shedding), right panel (stage 3, contact acquisition).(D) Transmission assay in wild-type (WT) and st6gal1-KO infants with or without additional α2,6-SA inhibitor (SNL) or α2,3-SA inhibitor (MAL) blockade.Panels C and D: Median is denoted by a line within box plots and bar graphs ± interquartile range.Simple linear regression was plotted in shedding panels.Shedding titers have a darker shade on days 1 and 2, and a lighter shade on days 3 and 4. Differences among multiple group medians were analyzed using the Kruskal-Wallis test.Experiments represent at least two biological replicates.Transmission rates were calculated using the Fisher's exact test.Infant mice were 4-7 days old; * (P < 0.05), ** (P < 0.01), *** (P < 0.001), **** (P < 0.0001), ns = not significant.Red brackets in statistical comparisons represent the reference group.Mock group represents the vehicle-treated (placebo) group.Dotted line denotes limit of detection (10 PFU/mL), where indicated; zero values were set to 1 PFU/mL.Cartoons were created with BioRender.com.

FIG 4 FIG 5
FIG 4 In vivo (Fig. 3) desialylation with broadly acting neuraminidase.(A) Immunohistochemistry of uninfected, A/X31-infected, or ba-NA-treated infant mice nasopharynx at 0.5 mg/kg/day.After sacrifice, heads were fixed with 4% paraformaldehyde, paraffin embedded, sectioned through the nasopharynx, and stained with influenza anti-NP antibody (white) or conjugated ECL (yellow).Zoomed-in images of the boxed inserts are represented by rows b and c.The scale bars correspond to a length of 100 µm in row a, and 10 µm in rows b and c.Experiments represent at least two biological replicates.Infant mice were 4-7 days old.(B) Fluorescent signal from immunohistochemistry (in A) was measured following color unmixing using Image J. Measurements were taken from at least three random positions along the epithelial surface of each image and selected from at least two representative tissues.Background was subtracted, and fold change over the average mock values was calculated.Bars represent median ± interquartile range.(C) Mock or ba-NA-treated or A/X31-infected pups were subjected toURT lavage, and samples were blotted on a nitrocellulose membrane and immunostained with biotinylated Flu-M antibody or biotinylated ECL to assess for viral infection or desialylated β-galactose in lavage fluid, respectively.Signals were normalized to background, and fold over mock mean signal was plotted.Median is denoted by a line.All panels: Differences among two group medians were analyzed using the Mann-Whitney test, and multiple comparisons were done using the Kruskal-Wallis test.Experiments represent at least two biological replicates.Infant mice were 4-7 days old; * (P < 0.05), ** (P < 0.01), *** (P < 0.001), **** (P < 0.0001).Mock group represents the vehicle-treated (placebo) group; ba-NA group represents the DAS181-treated group.AB/PAS, alcian blue and periodic acidSchiff stain; NP, Flu-NP antibody-stained group; β-gal, ECL, Erythrina crysta-galli lectin-treated group; S/N, signal-to-noise ratio.Dotted line denotes limit of detection.

FIG 5 (FIG 6
FIG5 (Continued)    for transmission via URT lavage after sacrifice.Shedding was collected on days 1-4 from index pups, and infectious virus was quantified via plaque assay.IN compound treatment was started 2 hpi and continued once daily for ba-NA (DAS181) or twice daily for NA-I (zanamivir) until the end of experiment.(B) Transmission inhibition rate in contact mice represented as percent inhibition over decreasing compound (ba-NA or NA-I) concentration.Percent inhibition was calculated using the following formula: y = 100 × [1 − (y/91.7)]for ba-NA, where 91.7 is the average transmission rate of ba-NA mock (vehicle-treated) litters, and y = 100 × [1 − (y/65)] for NA-I, where 65 is the average transmission rate of NA-I mock (vehicle-treated) litters.Mean transmission is plotted with error bars representing the standard error of the mean (shaded).(C) Transmission inhibition rate in contact mice represented as percent inhibition over increasing time to compound (inhibitor) treatment.Percent inhibition was calculated using the formulas described in (B).Mean transmission is plotted with error bars representing the standard error of the mean.(D) Transmission inhibition by ba-NA treatment time is represented by virus titers plotted at each stage of transmission.Left panel (stage 1, index URT replication), middle panel (stage 2, index shedding), right panel (stage 3, contact acquisition).(E) Transmission inhibition by ba-NA treatment of individual cohorts is represented by virus titers plotted at each stage of transmission.Left panel (stage 1, index URT replication), middle panel (stage 2, index shedding), right panel (stage 3, contact acquisition).(F) Shedding titers of A/X31-infected index pups treated with increasing compound (ba-NA or NA-I) concentration.All panels: Median is denoted by a line within box plots and bar graphs ± interquartile range.Simple linear regression was plotted in shedding panels.Shedding titers have a darker shade on days 1 and 2, and a lighter shade on days 3 and 4. Differences among multiple group medians were analyzed using the Kruskal-Wallis test.Experiments represent at least two biological replicates.Transmission rates were calculated using Fisher's exact test.Infant mice were 4-7 days old; * (P < 0.05), ** (P < 0.01), *** (P < 0.001), **** (P < 0.0001), ns = not significant.Red brackets in statistical comparisons represent the reference group.Mock group represents the vehicle-treated (placebo) group.Dotted line denotes limit of detection (10 PFU/mL), where indicated; zero values were set to 1 PFU/mL.Cartoons were created with BioRender.com.