Protein-bound sialic acid in saliva contributes directly to salivary anti-influenza virus activity

The oral cavity is an entrance for respiratory viruses, such as influenza. Recently, saliva has been shown to exert both antimicrobial and antiviral activities. Thus, saliva may be a biological factor that contributes to the prevention of influenza infection. However, the actual salivary anti-influenza A virus (IAV) activity in individuals and its determinant factors are unknown. By assessing individual variations in salivary anti-IAV activity in 92 people using an established new high-throughput system in this study, we found that the anti-IAV activity varied widely between individuals and showed a significant positive correlation with protein-bound sialic acid (BSA) level (ρ = 0.473; p < 0.001). Furthermore, the anti-IAV activity of saliva with enzymatically reduced BSA content was significantly lower. These results indicate that BSA is a direct regulator of salivary anti-IAV activity and is a determinant of individual differences. Additionally, after comparing the anti-IAV activity across the groups by age, anti-IAV activity in young people (aged 5–19 years) were lower than in adults aged 20–59 years and elderly people aged 60–79 years. Our study suggests that BSA levels in saliva may be important in preventing influenza infection.

Influenza viruses are some of the most important human pathogens. Seasonal influenza viruses (types A and B) cause epidemics worldwide, with the World Health Organization estimating hundreds of million infections and approximately 250,000-500,000 deaths per year 1,2 , leading to substantial social and economic burdens worldwide 3 . Several preventive and curative strategies exist against influenza, including vaccination and therapeutic agents, such as neuraminidase inhibitors and adamantanes 4 . However, influenza viruses evolve rapidly, which enables them to escape immunity induced by prior infections or vaccination; moreover, those viruses are transmitted efficiently from human to human via respiratory droplets. Recently, there has been substantial concern about the possibility of a pandemic of a novel influenza strain 5 , with reports of an increasing resistance of influenza A viruses to neuraminidase inhibitors and adamantanes 6 . Thus, new approaches are needed to effectively prevent and treat influenza infections.
The oral cavity is one of the entry sites for respiratory viruses. Saliva plays an essential role in maintaining the integrity of human health. It provides lubrication and protection and buffering action and clearance in the oral cavity, maintains tooth integrity, and aids taste and digestion 7 . A previous study indicated that hyposalivation may be a risk factor for acute respiratory infection 8 , suggesting that saliva may contribute to the role of innate immunity in the early stages of infection. Saliva induces both antimicrobial and antiviral activities 9 . Additionally, multiple soluble factors in saliva have potent inhibitory activity against influenza A viruses (IAVs), including sialic acid-containing mucin 5 B (MUC5B), salivary scavenger receptor cysteine-rich glycoprotein-340 (gp-340), alpha-2-microglobulin (A2M), and lectins (surfactant protein D [SP-D]) [10][11][12][13][14][15][16][17][18] . Two main mechanisms for the anti-IAV activity of these factors have been proposed. MUC5B, gp-340, and A2M acts as classic γ inhibitors, mediating the inhibition of IAV by presenting a sialic acid ligand that binds to the viral hemagglutinin (HA), thereby preventing HA attachment to target cells [12][13][14][15] . In contrast, SP-D acts as a classic β inhibitor by binding the lectin carbohydrate recognition domain (CRD) to specific oligosaccharides on the head of the viral HA [16][17][18] . It has been shown that salivary anti-IAV activity is primarily due to sialic-acid-containing proteins rather than lectin 13,19 . As the first line of defense, salivary antiviral factors are likely to play an important role in protection against respiratory infections. However, since previous studies on salivary anti-IAV activity mentioned above have only been conducted using purifying salivary proteins and small size analyses, the actual salivary anti-IAV www.nature.com/scientificreports/ activity of each individual, and its determinant factors are unknown. To solve this problem, high-throughput evaluation systems are required to evaluate multiple samples simultaneously. In this study, we constructed a new high-throughput system to evaluate salivary anti-IAV activity using neuraminidase activity derived from influenza virus as an index, and we investigated individual variations in salivary anti-IAV activity and its determinant factors.

Materials and methods
Participants Hemagglutination inhibition (HI) assay. Hemagglutinin (HA), which is present on the membrane surface of influenza viruses, binds to erythrocytes and causes erythrocyte aggregation. We evaluated the inhibitory activity of saliva on the binding of HA to erythrocytes. To remove the nonspecific inhibitory activity of salivainduced hemagglutination, 100 µL saliva samples were mixed with 200 µL of guinea pig erythrocytes and incubated at 4 °C for 2 h. The mixture was centrifuged at 2,000g for 10 min, and the supernatant (pretreated saliva sample) was used for the HI assay. The PR8 IAV strain was adjusted to 4 hemagglutinin units (HAU) per 25 µL PBS using a pre-hemagglutinin assay. In 96-well plates, a two-fold serial dilution of the pretreated saliva sample was mixed with 25 µL of 4-HAU PR8 IAV strain, followed by the addition of 50 µL of 0.7% guinea pig red blood cells and left at 4 °C for 2 h. The HI titer was calculated as the reciprocal of the highest dilution that produced complete hemagglutination inhibition.
Sialic acid assay. The concentrations of total sialic acid (TSA) and free sialic acid (FSA) in saliva were measured using a sialic acid assay kit (Bioassay Systems). The concentration of protein-bound sialic acid (BSA) was then calculated by subtracting the concentration of FSA from the concentration of TSA. www.nature.com/scientificreports/ Salivary proteome analysis. For ultrafiltration, 300 µL of saliva sample was mixed with 200 µL of 100 mM Tris-HCl buffer (pH 8.5; Nippon Gene Co., Ltd.) in an Amicon Ultra-0.5 Centrifugal Filter Unit (3 kDa; Merck Millipore Ltd), and the mixture was spun down at 14,000g for 30 min at 4 °C. The trapped saliva on the filter was filtered three times using 400 µL of Tris-HCl buffer. Finally, the filter on a new tube was spun down at 1,000g for 2 min at 4 °C, and the sample for proteome analysis was collected. The total protein concentration of each sample was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific Inc.) and the saliva sample was prepared with equal concentrations of proteins from each individual sample. The saliva sample containing 100 µg protein in Tris-HCl buffer (90 µL in total) was mixed with 5 µL of 100 mM dl-dithiothreitol (MP Biomedicals, LLC) for the reduction solution and heated at 56 °C for 45 min. The mixture was added to 5 µL of 111 mg/mL iodoacetamide (FUJIFILM Wako Pure Chemical Corporation) and incubated at 37 °C for 30 min in shade for alkylation. The mixture was then mixed with 50 µL of 0.1 µg/µL trypsin solution (Promega Corporation) and incubated at 37 °C for 16 h. The reaction solution was added to a final concentration of 0.3% (v/v) of 10% trifluoroacetic acid (FUJIFILM Wako Pure Chemical Corporation) to stop the reaction, and then subjected to LC-MS/MS analysis. All reagents were diluted with Tris-HCl buffer. Five microliters of the purified peptides were injected into a high-performance liquid chromatography system (ACQUITY UPLC: Waters) connected to a triple-quadrupole mass spectrometer (TSQ Vantage; Thermo Fisher Scientific) with an ion source of electrospray ionization in multiple reaction monitoring (MRM) mode. Chromatographic separation was performed on a Cadenza CD-C18 column (150 × 1 mm, 3 µm; Imtakt) by binary gradient elution at a flow rate of 0.05 mL/min. Eluent A was 0.1% formic acid in water, while acetonitrile containing 0.1% formic acid was used as eluent B. Ion source (ESI) parameters were optimized as follows: spray voltage, 400 V; vaporizer temperature, 100 °C; sheath gas pressure, 40 Arb; auxiliary gas pressure, 5 Arb; and capillary temperature, 250 °C. The peak areas for each target peptide were calculated using Xcalibur (Quan Browser) software (Thermo Scientific). and all data were analyzed using IBM SPSS Statistics Version 24 (IBM Japan, Ltd., Tokyo, Japan). Relationships between the two sets of data were analyzed using Spearman's rank correlation. The statistical significance of differences between groups was determined using one-way analysis of variance, followed by Tukey's ( Fig. 3) or Kruskal-Wallis (Fig. 4) multiple comparisons test. Results with p < 0.05 were considered statistically significant.

Results
Construction of high-throughput evaluation system for salivary anti-IAV activity. To simultaneously evaluate the anti-IAV activity in many saliva samples, we constructed a new high-throughput evaluation system using neuraminidase (NA) derived from IAV as an index. Figure 1A shows the schematic of the measurement flow of salivary anti-IAV activity. The measurement consisted of the following steps: (1) mixing the virus and saliva, (2) infection of MDCK cells using the IAV solution and saliva mixture, (3) collecting the supernatant of the cells after 16-18 h of infection, and (4) measurement of sialidase activities of the supernatant for multiplicated virus quantification. The NA activity value when PBS was used instead of saliva was set to 100%, and the relative value was calculated from the NA activity value of saliva samples. The value obtained by subtracting the relative NA activity value from 100 was defined as the anti-IAV activity. A high correlation (ρ = 0.916; p < 0.001) was confirmed between the NA activity value in the supernatant of infected cells and the number of infected cells stained for NP protein when MDCK cells were infected with IAV mixed with various saliva samples (Fig. 1B).
Individual differences in salivary anti-IAV activity. Using the established system, salivary anti-IAV activity against the PR8 IAV strain (H1N1) in 92 individuals was assessed. Anti-IAV activity varied widely between individuals, ranging from 2 to 97% ( Fig. 2A). Furthermore, salivary anti-IAV activity against the Memphis IAV strain (H3N2) was significantly positively correlated with anti-IAV activity against the PR8 IAV strain (ρ = 0.638; p < 0.001) (Fig. 2B). We then performed a correlation analysis of salivary anti-IAV activity against the PR8 IAV strain and the amount of various salivary components measured using proteomic analysis and some kits to explore factors associated with individual differences in salivary anti-IAV activity (Fig. 2C). The results showed the highest positive correlation between salivary anti-IAV activity and BSA levels (ρ = 0.473; p < 0.001). In addition, the concentration of several proteins, such as AGTN/gp-340, ZG16B, SGP28, and C6orf58, significantly positively correlated with anti-IAV activity.

Contribution of BSA to salivary anti-IAV activity.
To clarify whether BSA contributes directly to the anti-IAV activity of saliva, we assessed saliva samples that were heat-treated or in which BSA was enzymatically removed. The amount of BSA in the saliva was significantly reduced by α(2-3,6)-sialidase treatment and was not altered by heat treatment (Fig. 3A). While the anti-IAV activity of untreated saliva samples averaged about www.nature.com/scientificreports/ 50%, the anti-IAV activity of heat-treated saliva samples decreased to an average of 25% (Fig. 3B). In contrast, the anti-IAV activity of sialidase-treated saliva samples decreased to less than 10% on average (Fig. 3B). To further elucidate the contribution of BSA to salivary anti-IAV activity, a hemagglutination inhibition (HI) test was performed. Similar to the anti-IAV activity, large individual differences were observed in HI titers in some saliva samples (Fig. 3C). In addition, individual HI titers highly positively correlated with anti-IAV activity (Fig. 3D). Although no effect of heat treatment on HI titer was observed, sialidase treatment significantly decreased the HI titer (Fig. 3E).

Age-related differences in salivary anti-IAV activity.
We then compared salivary anti-IAV activity across a wide range of ages (5-79 years divided into three groups, young people (aged [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19], middle adults (aged 20-59 years), and elderly people (aged 60-79 years), using the high-throughput evaluation system we constructed. As a result of comparing the gender differences within each group, no difference was observed in the mean value of saliva secretion rate, anti-IAV activity, and BSA level (data not shown). Therefore, subsequent analyses did not distinguish between gender. The mean value of saliva secretion rate decreased with increasing age and was lowest in elderly people (Fig. 4A). Moderate correlation was observed for between age and salivary anti-IAV activity in individuals (rs = 0.29, p < 0.001). The mean value of salivary anti-IAV activity was similar in middle adults and elderly people, and was significantly lower in young people (Fig. 4B). The mean value of BSA concentration in saliva was significantly lower in elderly people than middle adults (Fig. 4C). In each age group, the salivary anti-IAV activity significantly positively correlated with salivary BSA level (young; ρ = 0.366, p < 0.05, middle adults; ρ = 0.589, p-value < 0.001, elderly; ρ = 608, p-value < 0.001, total; ρ = 0.511, p-value < 0.001) (Fig. 4D).

Discussion
In this study, we constructed a high-throughput system to evaluate salivary anti-IAV activity. Using this system, we found that there are large individual differences in salivary anti-IAV activity, and that the amount of BSA in saliva is important for these individual differences.
In general, the main method of measuring the viral infection titer is to count the number of infected cells by staining with antibodies against the virus and clumps of cellular degeneration associated with viral infection. While these methods are suitable for measuring the virus titer, it is difficult to evaluate multiple samples because this approach requires a great deal of time and labor because of the need for operations, such as multi-step dilutions and infected cells or colony counting. In our constructed system, the neuraminidase activity derived from the propagated influenza virus in the culture medium was measured using a multi-plate reader, which eliminates the operations described above and is suitable for the analysis of multiple samples. A significant positive correlation was observed between salivary anti-IAV activity against the PR8 (H1N1) and Memphis (H3N2) strains (Fig. 2B), indicating that individual differences in salivary anti-IAV activity are similar across influenza virus strains. It is expected that this method can be used in the evaluation of many influenza virus strains in the future. Using the constructed system, it was clarified that the individual variations in salivary anti-IAV activity most positively correlated with the amount of BSA (Fig. 2C). Moreover, the HI titers of saliva were significantly reduced by treatment with α(2-3,6)-sialidase (Fig. 3B). Some studies have reported that saliva contains soluble factors that can exert anti-influenza activity, such as sialic acid binding protein and lectins. Although glycolipids and glycopolysaccharides are also present in saliva, their concentrations are much lower than those of proteins 20,21 . Thus, their contribution as a source of sialic acid is considered to be low. It has been reported previously that www.nature.com/scientificreports/ several purified salivary proteins, such as AGTN/gp-340 and A2M, can inhibit the hemagglutination of erythrocytes by presenting a sialic acid ligand for the viral HA [12][13][14][15] . In addition, Qin et al. reported that saliva containing more α2-3/6-linked sialic acid residues has high binding properties to IAV 22 . These reports are consistent with our results, and indicate that the total amount of sialic acid possessed by these proteins determines the anti-IAV activity of saliva. In other words, the amount of BSA in the saliva is expected to be an indicator of frontline oral defenses against influenza viruses. MUC5B has potent inhibitory activity against IAV. However, our data showed that salivary MUC5B levels weakly correlated with the anti-IAV activity of saliva. White et al. reported that the sialic acid of MUC5B is more easily cleaved than that of gp-340, the levels of which highly correlated with anti-IAV activity. Our results suggest that the sialic acid of MUC5B is cleaved by sialidase in the oral cavity [23][24][25] , and the anti-IAV activity of MUC5B www.nature.com/scientificreports/ in saliva may be reduced. In considering individual variations in the anti-IAV activity of saliva, it is necessary to consider oral sialidase activity. In this study, SGP28, C6orf58, and PIP also strongly correlated with the anti-IAV activity of saliva. To date, the biological functions of SGP28 and C6orf58 have not been clarified. PIP, on the other hand, is known as a water-soluble antimicrobial protein 26 . It binds to and agglutinates bacteria, the detailed mechanism of its antimicrobial properties is not fully understood. These molecules are known to have sugar chains [27][28][29] . Although the presence of sialic acid of these molecules is not clear, they may also contribute to the anti-IAV activity of saliva via sialic acid, which is less susceptible to cleavage by sialidase. It is known that sialic acids are structurally diverse and that O-acetylated sialic acids are resistant to bacterial sialidase 30 . In the www.nature.com/scientificreports/ future, it will be necessary to clarify the presence or absence of sialic acid in these molecules, including their structural features. Interestingly, the anti-IAV activity of saliva was reduced following excessive sialidase treatment, but was not completely lost (Fig. 3A, B). Moreover, salivary anti-IAV activity also decreased following heating, which did not affect the concentration of sialic acid. Therefore, it is possible that sialic acid-independent factors are also involved in anti-IAV activity. One of the possible factors is the action of lectin-like proteins, such as ZG16B, which have shown a high correlation with anti-IAV activity 31,32 . Lectin (carbohydrate-binding proteins), including conglutinin, SP-D, MBL, and SAP, bind selectively to specific carbohydrate structures (mannose-over galactosetype sugars) located at the head of the influenza HA of susceptible strains, thereby blocking the ability of HA to bind to sialylated cell-surface receptors 16,17 . Furthermore, lectin-like proteins may also inhibit the fusion of Average plateau values were compared using one-way analysis of variance and the Kruskal-Wallis test. *p < 0.05; ***p < 0.001. www.nature.com/scientificreports/ the viral membrane with the host cell by binding to sialic acid on the surface of the host cell 18 . In our evaluation system, since host cells were exposed to a mixture of saliva and virus for 30 min, the action of lectin-like proteins in saliva that inhibit the contact between host cells and virus would also contribute to the anti-IAV activity of saliva. On the other hand, in our evaluation system, the activity of molecules that inhibit the later mechanisms of virus propagation, such as replication and release, is not reflected in the anti-IAV activity. For this reason, defensin and lactoferrin, which are known to inhibit replication of IAV 12,33-37 , showed only a weak correlation with anti-IAV activity in our study. The current thinking regarding influenza infection morbidity is that the risk factors are related to age, higher rates of infection in school-aged children relative to adults, and lower rates in the elderly. Children spend a great deal of time in communities where daily contact with other people is extensive; for example, in schools, playgroups, and daycare centers, and it is assumed that close contact favors infection 38,39 . The results of evaluating differences in salivary properties by age (Fig. 4) showed that salivary volume, which has been reported as a predictive risk factor for influenza transmission 8 , was higher in the younger age group. On the other hand, the mean value of anti-IAV activity of saliva in the young group was significantly lower than that in the middle and elderly groups. These results suggest that low salivary anti-IAV activity may also be associated with a high rate of IAV transmission in school-aged children. In the present study, although the mean value of BSA concentration in saliva of young group was not significantly different from that of the middle and elderly groups, salivary anti-IAV activity was highly correlated with sialic acid in all age group, indicating that the determinants of anti-IAV activity were not as sensitive to age. Previous studies have shown that glycoproteins capable of binding IAV are lower in the saliva of children than in adults and the elderly, supporting our results 22 . Thus, qualitative control of saliva in young people may be effective in preventing IAV infection.

Scientific
There are two major limitations to this study as follow: First, we have not been able to define the contribution of proteins such as SGP28, C6orf58, and PIP that are significantly correlated with anti-IAV activity. It is important to prepare these purified proteins from saliva and conduct experiments to determine the relationship between glycosylation levels of these proteins and anti-IAV activity. Second, we have not been able to examine the effect of salivary anti-IAV activity on respiratory illnesses. Retrospective study the relationship between salivary anti-IAV activity and history of respiratory disease and recent antibody prevalence against influenza viruses, or prospective intervention trials with techniques that provide control of salivary anti-IAV are needed to gain insight into the contribution of salivary anti-IAV activity.
Our study revealed that there are large individual differences in the anti-IAV activity of saliva, which mainly depends on the amount of BSA. BSA directly contributes to salivary anti-IAV, which may contribute to influenza morbidity in young people. In conclusion, the understanding of individual differences in salivary anti-IAV activity may provide new insights into creating effective methods to prevent influenza infection. www.nature.com/scientificreports/