Antiviral activity of lambda-carrageenan against influenza viruses and severe acute respiratory syndrome coronavirus 2

Influenza virus and coronavirus, belonging to enveloped RNA viruses, are major causes of human respiratory diseases. The aim of this study was to investigate the broad spectrum antiviral activity of a naturally existing sulfated polysaccharide, lambda-carrageenan (λ-CGN), purified from marine red algae. Cell culture-based assays revealed that the macromolecule efficiently inhibited both influenza A and B viruses with EC50 values ranging from 0.3 to 1.4 μg/ml, as well as currently circulating severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) with an EC50 value of 0.9 ± 1.1 μg/ml. No toxicity to the host cells was observed at concentrations up to 300 μg/ml. Plaque titration and western blot analysis verified that λ-CGN reduced expression of viral proteins in cell lysates and suppressed progeny virus production in culture supernatants in a dose-dependent manner. This polyanionic compound exerts antiviral activity by targeting viral attachment to cell surface receptors and preventing virus entry. Moreover, its intranasal administration to mice during influenza A viral challenge not only alleviated infection-mediated reductions in body weight but also protected 60% of mice from virus-induced mortality. Thus, λ-CGN could be a promising antiviral agent for preventing infection with several respiratory viruses.

Cell culture-based antiviral assay. An antiviral assay for influenza viruses was performed as described previously 20 . Briefly, MDCK cells grown overnight in 96-well plates (3 × 10 4 cells per well) were mock-infected or infected with each viral strain at a multiplicity of infection (MOI) of 0.001 at 35 °C for 1 h. After removing unabsorbed virus, cells were treated with threefold dilutions of each compound for 3 days at the same temperature. Viability of non-infected or infected cells was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) to determine the half-maximal cytotoxic concentration (CC 50 ) and the half-maximal effective concentration (EC 50 ), respectively.
To assess anti-SARS-CoV-2 activity in an image-based system, Vero cells were grown overnight in 96-well plates (2 × 10 4 cells per well). After addition of serial threefold dilutions of compounds, cells were infected with an equal volume of SARS-CoV-2 (MOI of 0.05) at 37 °C for 2 days within a biosafety level 3 laboratory. The cells were fixed and permeabilized with chilled acetone:methanol (1:3) for probing with an anti-spike antibody (Genetex, Irvine, CA, USA) followed by Alexa Fluor 488-conjugated goat anti-mouse IgG (Invitrogen, Carlsbad, CA, USA) to determine EC 50 values. Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) to calculate the CC 50 values. The number of viral spike protein-derived or cell nuclei-derived signals detected in four spots per well was quantified from three independent samples using the Operetta high content screening system (Perkin Elmer, Waltham, MA, USA) and the built-in Harmony High-Content Imaging and Analysis software 3.5.2. For determination of 50% tissue culture infectious dose (TCID 50 ), SARS-CoV-2-infected cells were incubated in the absence or presence of antiviral compounds for 2 days. Fresh Vero cells seeded in 96-well plates were infected with serial tenfold dilutions of the culture supernatants for additional 2 days. TCID 50 s were determined by counting SARS-CoV-2 spike protein-derived green fluorescence population as well as the number of DAPI-derived nuclei distribution as mentioned above.

Western blot analysis.
For anti-influenza viral study, PR8-infected MDCK cells (MOI, 0.001) were treated with increasing concentrations of λ-CGN, pKG-03 or OSV-C at 35 °C for 1 day. For antiviral analysis against SARS-CoV-2, virus-infected Vero cells (MOI, 0.005) were treated with increasing concentrations of λ-CGN or RDV at 37 °C for 2 days. Culture lysates were harvested and loaded onto 10 or 12% SDS-PAGE gels (40 μg total protein per well) for electrotransfer. Influenza viral NP and HA proteins were detected using mouse anti-NP (Cat. No., 11675-MM03; Sino Biological, Beijing, China) and rabbit anti-HA2 (Cat. No., 86001-RM01; Sino Biological) antibodies, respectively, according to our previous report 18 . SARS-CoV-2 spike protein was probed with the mouse anti-spike antibody. Cellular β-actin used as a loading control was detected with a mouse antiβ-actin antibody (Cat. No., A1987; Sigma-Aldrich). Horseradish peroxidase (HRP)-conjugated goat anti-mouse or anti-rabbit secondary antibodies were used to detect the primary antibodies (Thermo Scientific, Waltham, MA, USA). After addition of a chemiluminescent HRP substrate (SuperSignal West Pico Chemiluminescent Substrate; Pierce, Rockford, IL, USA), images were obtained using a LAS-4000 Luminescent Image Analyzer (Fujifilm, Tokyo, Japan).
Plaque titration. Plaque reduction assay was performed as described previously, with some modifications 18 .
Briefly, MDCK cells seeded in 6-well plates were infected with PR8 at an MOI of 0.001 in the absence or presence of increasing concentrations of λ-CGN or p-KG03. On the next day, the culture supernatants were harvested and serial tenfold dilutions were used for infection of fresh MDCK cells in 48-well plates. After their incubation in overlay medium [serum-free MEM with 1.2% Avicel RC-591 (FMC Corp, Philadelphia, PA, USA) and 2 μg/ml TPCK-trypsin (Sigma-Aldrich)] at 33 °C for 3 days, the number of plaques was counted by crystal violet staining.
Time-of-addition experiments. MDCK cells were seeded in 48-well plates at a density of 1.5 × 10 5 cells per well for 1 day. In the pre-treatment assay, λ-CGN (1 μg/ml) was treated for 2 h followed by PBS-washing and infection with PR8 at an MOI of 0.001 for additional 2 h. In the co-treatment assay, the mixture of λ-CGN and PR8 were loaded onto the cells immediately or after 30-min-preincubation. In the post-treatment assay, λ-CGN was added for 2 h after infection with PR8 for 2 h. As controls, p-KG03 (1 μg/ml) or EGCG (1 μM) was treated in all sets. In each step, cells were washed with PBS to remove non-specifically bound compounds or unabsorbed virus. The cells were cultured at 33 °C for 3 days in the overlay medium for plaque titration 19 .
HA inhibition assay. Equal volumes (25 μl) of two-fold dilutions of PR8 (HA titer of the stock, 2 8 ) and increasing concentrations of λ-CGN were incubated for 20 min at room temperature in 96-well round bottom plates. In each well, 50 μl of 0.5% chicken red blood cells (RBCs) was added. After 30 min incubation, HA titers were determined by RBC agglutination 21  www.nature.com/scientificreports/ Confocal microscopy. MDCK cells were infected with PR8 virus at an MOI of 5 in the absence or presence of the sulfated polysaccharides (10 μg/ml) for 4 h at 37 °C. In parallel, the same samples were incubated for 2.5 h at 37 °C with protein synthesis inhibitor cycloheximide (10 μg/ml) (CHX; Sigma-Aldrich), of which experimental condition was optimized in our previous reports 18,22 . Viral NP was visualized using an anti-NP antibody (Cat. No., sc-80481; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and Alexa Fluor 488-conjugated goat antimouse IgG (Invitrogen), while nuclear DNA was counterstained using 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA). Images were captured under a Zeiss LSM 700 confocal microscope and data were analyzed with ZEN blue software 3.1 (Carl Zeiss, Thornwood, NY, USA).
In vivo study. Antiviral efficacy study in a mouse model was performed by modification of our previous report 18 . Briefly, female BALB/c mice (6-7 weeks old; Orient Bio Inc., Gyeonggi-do, Republic Korea) were infected with maPR8. Five units of 50% mouse lethal dose (5 MLD 50 ) of the virus were preincubated with λ-CGN for 30 min at room temperature. Mice were challenged intranasally with maPR8 alone or with maPR8 mixed with λ-CGN (1 or 5 mg/kg) in a total volume of 50 μl. The control group received OSV-P orally from days 0 For neutralization assay, Vero E6 cells were infected with each influenza A (H1N1) HA/NA-or SARS-CoV-2 spike-pseudotyped viruses. Before induction, pseudoviruses were incubated with λ-CGN for 2 h. Particularly, influenza A (H1N1) HA/NA pseudovirus was treated with 10 μg/ml of TPCK treated-trypsin for 15 min before λ-CGN treatment. After that, TPCK treated-trypsin was inactivated by 10 μg/ml Soybean solution (Sigma-Aldrich). Twenty four hours after infection, the medium was changed with fresh DMEM. Forty eight hours after transduction, BrightGlo (Promega, Madison, WI, USA) was added to the cell lysates according to the manufacturer's instructions, and relative luminescence units (RLUs) were measured by a microplate reader.
Quantitative RT-PCR. SARS-CoV-2-infected Vero cells (MOI, 0.005) were treated with increasing concentrations of λ-CGN or RDV at 37 °C. At day 2, culture supernatants were harvested for viral RNA purification using QIAamp viral RNA mini kit (Qiagen, Hilden, Germany). Relative SARS-CoV-2 RNA copies were quantified using a real-time RT-PCR kit with an N gene-specific primer set (PCLMD nCoV one step RT-PCR kit; PCL Inc., Seoul, Republic of Korea) and a CFX96 Touch real-time PCR instrument (Bio-Rad, Hercules, CA, USA). Statistical analysis. Statistical analyses were performed by unpaired, two-way ANOVA t-test according to the Dunnett's multiple comparison method using GraphPad Prism version 8.3.1. In the Kaplan-Meier survival analysis, survival statistics were calculated by Log-rank (Mantel-Cox) test. P values lower than to 0.05 were considered statistically significant. *P < 0.05; **P < 0.01; ****P < 0.0001.

Results
Anti-influenza activity of λ-CGN. To examine the antiviral activity of λ-CGN, increasing concentrations of the compound were loaded to influenza virus-infected MDCK cells. Another sulfated polysaccharide, p-KG03, of which antiviral activity has been elucidated in our previous report 19 and three different antiviral chemicals (AMT, RBV and OSV-C) were used as controls. The anti-influenza viral activity or the drug-resistance profiles of these control compounds were reproducible, indicating that the cell culture-based antiviral assay is reliable ( Table 1). The cytopathic effect (CPE) inhibition assay revealed that λ-CGN efficiently inhibited infection by both influenza A and B viruses, with EC 50 values of 0.3 to 1.4 μg/ml, with no cytotoxicity up to a maximum concentration of 300 μg/ml. Notably, the inhibitory effect was comparable with that of p-KG03. To confirm this finding, we measured changes in viral protein expression in cell lysates and infectious viral titers in culture supernatants (Fig. 1B,C, Supplementary Fig. S1). The data revealed that λ-CGN not only inhibited expression of viral proteins NP and HA in infected cells, but also suppressed production of progeny virus in a dose-dependent manner as observed in the p-KG03-treated samples. Taken together, these results suggested that λ-CGN has considerable antiviral activity against influenza A and B viruses in vitro, with selectivity index (SI) values over 263.2.

Inhibition of influenza viral entry by λ-CGN.
We wondered which step of the influenza virus life cycle is targeted by λ-CGN. Using p-KG03 and EGCG as controls for blocking virus entry, λ-CGN was treated at different time points, such as before (pre-treatment), during (co-treatment) or after (post-treatment) viral infection at 2 h intervals ( Fig. 2A). Plaque titration informed that its pre-treatment has little inhibitory effect, while www.nature.com/scientificreports/ post-treatment resulted in significant reduction but not remarkable (Fig. 2B). However, its co-treatment with the virus induced significant antiviral effect, that was more enhanced when they were pre-incubated for 30 min before treatment. All three compounds seem to target viral entry step by directly attenuating infectivity of the viral particles. The result suggested that λ-CGN could interact with a viral protein important for virus entry, possibly HA.
To evaluate this hypothesis, we examined HA inhibition assay with λ-CGN (Fig. 3A). Agglutination of chicken RBC was achieved by PR8 at the highest dilution 2 8 . However, HA titer was gradually reduced as the concentration of λ-CGN was increased, confirming that it inhibits binding of viral HA to cellular receptors. To further verify its mode-of-action, intracellular distribution of viral NP was compared in the absence or presence of the compound at 4 h p.i, a time when NP was fully localized to the nuclei for robust replication of viral RNA (Fig. 3B, Supplementary Fig. S2). The confocal microscopic images revealed that, similar to p-KG03, λ-CGN reduced the number of NP-positive nuclei when compared to the mock-treated sample. We also monitored the intracellular distribution of NP at an earlier time point (2.5 h post-infection) in the presence of CHX, a protein synthesis inhibitor that allows tracking of the input viral proteins and their localization (Fig. 3C). Under this condition, when NP was present in the cytoplasm but not reached the nucleus, λ-CGN completely blocked membrane penetration of the viral particles harboring vRNP complexes as efficiently as p-KG03. No NP accumulation on the surface of the cellular membrane strongly suggests that λ-CGN targets attachment of influenza virus to its cell surface receptors by neutralizing viral glycoprotein HA.

Protection of mice from influenza A viral infection by λ-CGN.
To investigate the antiviral activity of λ-CGN in vivo, mice were infected intranasally with maPR8 alone or with maPR8 plus λ-CGN once. As a control, maPR8-infected mice received OSV-P orally twice a day for 6 days. Antiviral activity was determined by monitoring body weight and mortality for 15 days. The results revealed that maPR8 at 5 MLD 50 caused body weight loss (Fig. 4A) and complete death within day 7 (Fig. 4B). Notably, intranasal administration of 5 mg/kg λ-CGN mitigated infection-mediated body weight loss, yielding a 60% survival rate. However, this antiviral efficacy was not observed at a lower dose (1 mg/kg). As expected, treatment with OSV-P at 10 mg/kg/day for 6 days showed remarkable therapeutic effects, ensuring the reliability of the in vivo antiviral study. Taken together, these data suggested that intranasal co-administration of λ-CGN prevents viral infection-mediated body weight loss and reduces mortality.

Anti-SARS-CoV-2 activity of λ-CGN.
To address whether λ-CGN can neutralize SARS-CoV-2 as well as influenza virus, SARS-CoV-2 spike-and influenza A (H1N1) HA/NA-pseudotyped viruses bearing a firefly luciferase reporter gene were prepared 23 . Firefly luciferase assay with their infected cell lysates revealed that λ-CGN is able to suppress entry of both SARS-CoV-2 and influenza A viral glycoproteins-derived pseudoviruses in a dose-dependent manner (Fig. 5). As a next step, we examined antiviral activity of the sulfated polysaccharide against infectious SARS-CoV-2. Vero cells infected with the virus at an MOI of 0.02 were treated with increasing concentrations of λ-CGN by using RDV as a control. On day 2, immunofluorescence microscopy with an anti-SARS-CoV-2 spike antibody revealed that viral infection was inhibited effectively by λ-CGN, without affecting cell viability (Fig. 6A). As expected, anti-SARS-CoV-2 activity was well induced in the RDV-treated cells. Quantitative analysis of antiviral dose-response and cell viability showed that λ-CGN had an EC 50 of 0.9 ± 1.1 μg/ml and a CC 50 of > 300.0 μg/ml (resulting in an S.I., > 333.3), while RDV had an EC 50 of 23.5 ± 1.2 μM and a CC 50 of > 300.0 μM (resulting in an S.I., > 12.8) (Fig. 6B). In addition, western blot and quantitative RT-PCR analyses exhibited decreases of both viral protein in cell lysate and viral RNA level in the culture supernatants by the sulfated polysaccharide (Fig. 7A,B, Supplementary Fig. S3). Decisively, reinfection of the culture supernatants into fresh Vero cells confirmed reduction of infectious viral titers in the presence of λ-CGN as observed in RDV (Fig. 7C). These results strongly demonstrate that λ-CGN is active against SARS-CoV-2.

Discussion
Sulfated polysaccharides such as heparin, dextran sulfate, and pentosan sulfate, as well as various CGNs, show antiviral or virucidal activity against diverse enveloped viruses at subtoxic concentrations [24][25][26][27][28] . These studies of the physiochemical properties and molecular structure of these compounds reveal that the antiviral efficacy is mainly due to their affinity for viral glycoproteins, resulting in blockade of viral attachment to cellular receptors. www.nature.com/scientificreports/ www.nature.com/scientificreports/  were mock-infected (black) or intranasally infected with maPR8 at 5 MLD 50 (red). As test groups, the virus was preincubated at room temperature for 30 min with λ-CGN at a lower dose (1 mg/kg/d, purple) or a higher dose (5 mg/kg/d, green), followed by intranasal administration. Control mice received OSV-P orally twice a day (10 mg/kg/d) at 8-h intervals, starting at 4 h before viral infection (blue). Body weight (A) and mortality (B) of mice were measured every day from days 0 to 14 post-infection. Data are expressed as the mean ± S.D. from five mice. Survival statistics were calculated by Log-rank (Mantel-Cox) test. **P < 0.01; n.s., non-statistically significant. The graphs were created using GraphPad Prism 8.3.1 (www.graph pad.com). www.nature.com/scientificreports/ Accordingly, it is convincing that the charge density, chain length, degree of sulfation, and detailed structural features of these macromolecules are critical for this interaction. In-depth studies of the underlying mechanisms demonstrated that the macromolecules exert anti-HIV activity by competing with polyanionic regions of host-cell-surface molecules for binding to the positively charged amino acids present in the viral enveloped glycoprotein, gp120, including the V3 loop [29][30][31] . The microbicidal activity of polystyrene sulfonate against sexually transmitted infectious diseases caused by herpes simplex virus type 2 and papillomavirus has been evaluated in vivo and in vitro 32,33 . Unfortunately, prevention of vaginal HIV transmission using topical cellulose sulfate gel failed 34 , indicating the need for development of a more potent microbicidal sulfated polysaccharide or for administration of the polymers through an alternative route, such as oral or intranasal. Regarding this issue, it is not strange to anticipate that intranasal treatment with active sulfated polysaccharides could be a promising way to prevent infection by various respiratory enveloped viruses such as influenza A and B viruses, respiratory syncytial virus, and coronaviruses. Previously, it was reported that κ-CGN with a molecular weight of 2 kDa is active against influenza A virus in vitro, with an EC 50 value of 32.1 μg/ml. In addition, ι-CGN inhibited influenza A virus infection of MDCK cells with an EC 50 value of 0.04-0.20 μg/ml; not only that, intranasal administration of ι-CGN showed therapeutic effects in an influenza A virus-infected mouse model 4,35 . Notably, a randomized double-blind study in volunteers with early symptoms of the common cold confirmed the efficacy and safety of an antiviral ι-CGN nasal spray 36 . In contrast to κ-and ι-CGNs, the antiviral activity of λ-CGN has rarely been investigated in the context of viral species that are transmitted in droplets or through the air. Therefore, we explored whether λ-CGN is able to inhibit both influenza A and B viruses and/or the emerging SARS-CoV-2. We had a great interest in λ-CGN, because this compound comprising alternating (1,3)-linked α-D-galactose-2-sulfated and (1,4)-linked β-D-galactose-2,6,-disulfated units has a higher degree sulfation with an ester sulfate content of about 32-39% and shows better solubility in cold water than the other two CGNs 37 . Accordingly, the sulfated polysaccharide was expected to have efficient and broad antiviral activity and to be easily dissolved in an aqueous solution when it is formulated for a nasal spray.
Here, we successfully observed that λ-CGN inhibits not only influenza viruses but also SARS-CoV-2 by targeting their entry process. Strikingly, its virucidal properties led to a 60% survival rate in virus-challenged mice after an exposure of infectious virus to the antiviral agent (Fig. 4). However, it remains to be investigated whether this polyanionic compound is able to protect small animals such as hACE2-expressing mice or Syrian hamsters from SARS-CoV-2 infection by blocking the viral S protein-associated entry step 38,39 . In addition, because CGNs have intrinsic anti-coagulant activity, any unwarranted side effects should be reviewed before clinical application. This is because dysfunctional or aberrant coagulation is responsible for the hyper-inflammatory responses observed in severe cases of influenza or SARS-CoV-2 infection-mediated pneumonia, and anti-coagulant signals could be over-stimulated already in the lungs of infected patients 40,41 . Nevertheless, to the best of our knowledge, this is the first report to suggest that λ-CGN potently inhibits infection by influenza B as well as influenza A viruses and emerging SARS-CoV-2. The broad spectrum antiviral activity of this compound would make it valuable especially when different respiratory viruses are circulating concurrently or when their prophylactic treatment is definitely required before diagnosis.  ) or SARS-CoV-2 spike protein (gray bars) were prepared, in which a firefly luciferase-expressing plasmid was incorporated. They were preincubated with mock or increasing concentrations of λ-CGN for 2 h at 37 °C and then infected into Vero E6 cells. On day 2, relative firefly luciferase activity (RLU) was determined by fixing the mock-treated sample at 100%. Data are expressed as the mean ± S.D. of three independent experiments. ****P < 0.0001. The graph was created using GraphPad Prism 8.3.1 (www.graph pad.com). www.nature.com/scientificreports/  On day 2, the cell culture plates were subjected to immunofluorescence assay for determination of infectious viral titer by calculating the relative ratio of spike-derived green fluorescence frequency to nuclei-derived blue-positive cell number. Data are expressed as the mean ± S.D. of three independent experiments. **P < 0.01; ****P < 0.0001; n.d., not detected. The graphs were created using GraphPad Prism 8.3.1 (www.graph pad.com).