In this study, it was shown that the viral isolation efficiency in saliva specimens is significantly lower than that of nasal/nasopharyngeal swab specimens for COVID-19. As a result of analysis by Ct value, the viral isolation efficiency in saliva specimens was significantly lower in the groups with Ct values of 20–25 and 25–30, and no difference was observed in the group with Ct values of 30 or more. Therefore, it was found that the viral isolation efficiency differs depending on the type of specimens when the Ct value is 30 or less. In addition, previous studies have shown that the viral isolation efficiency decreases when the Ct value is 30 or more 3,6−9. In the comparison of the viral isolation efficiency for each viral strain, the viral isolation efficiencies from the saliva specimens were significantly lower between the Wuhan strain, Delta and Omicron variants of SARS-CoV-2, but no difference was observed in the alpha variant.
In this study, specimens with low Ct values were not sufficiently collected among the saliva specimens containing the Wuhan strain and Delta variant of SARS-CoV-2. Therefore, it is possible that the viral isolation efficiency of the saliva specimens was significantly lower than the real isolation efficiency. On the other hand, the number of specimens could be the same in all Ct value groups, and there was no difference in the viral isolation efficiency between the nasal/nasopharyngeal swabs and saliva specimens for the Alpha variant of SARS-CoV-2. It will be necessary to increase the number of specimens to further verify whether there is a real difference among the variants. In the comparison of the viral isolation efficiency by Ct value group for each variant, the group with the significantly lower Ct value was different. However, although there was no significant difference, the viral isolation efficiency of saliva specimens tended to be low except for the groups with Ct values of 30 or more. Combined with the result of the comparison of the Ct value groups between variants, it is inferred that the viral infectivity decreases due to some interference by saliva components in the saliva specimens with Ct values of 30 or less. In a comparison of the viral isolation efficiency among the strain or variants in the same type of specimen, the viral isolation efficiency in the Omicron variant was significantly lower than that of the other strains or variants in both nasal/nasopharyngeal swab and saliva specimens. The Omicron variant has been reported to differ from other variants in terms of its dependence on TMPRSS2 during cell entry and the cleavage efficiency of viral spike protein10. Therefore, this may also affect viral isolation efficiency.
In the isolation of SARS-CoV-2 using simulated specimens, the viral isolation efficiency from the saliva-containing specimen was significantly lower than that of the nasal/nasopharyngeal swab-containing specimens in 1.0×102 copy (2.8×10⁻¹ PFU). The saliva used in the 1.0×102 copy group was from patients 16–76 years old (median 42.5) and had a sex ratio of 7:3 (14 males and 6 females). The virus was not isolated in 15 of 20 saliva-containing specimens, and no significant difference was observed in age and sex compared to the five viral isolated specimens. All of the saliva used in this study was collected from healthy individuals who had negative SARS-CoV-2 PCR tests and had no history of vaccination against COVID-19, so the involvement of specific antibodies against the SARS-CoV-2 in saliva could be ruled out. Therefore, the decrease in infectivity of SARS-CoV-2 is considered to be due to the nature of saliva itself. By comparing the nasal/nasopharyngeal swabs and saliva samples using simulated specimens, the effect of viral interference by saliva on the clinical specimens in the previous experiment was experimentally shown.
Finally, we investigated which component of saliva was responsible for the lower isolation rate of saliva specimens by SARS-CoV-2pv. The results showed that among four candidate substances, lactoferrin and amylase inhibited SARS-CoV-2 infection the most. It has been reported that lactoferrin has an inhibitory effect on the entry and replication of SARS-CoV-2 11,12. The concentration of lactoferrin in normal saliva is considered to be about 10 µg/mL 13. Although that concentration is lower than what was used in this experiment, it is thought to have some anti-SARS-CoV-2 action in the saliva. In fact, it has been reported that human breast milk contains higher concentrations of lactoferrin than saliva, and that no infectious virus is present in breast milk even if the mother is infected with SARS-CoV-2 14. Although the same experiment was performed with the control, VSVpv, for comparison, no significant inhibition was observed, although a concentration-dependent decreasing trend was observed. Therefore, it was suggested that lactoferrin has a higher inhibitory effect on SARS-CoV-2 infection. In addition, lactoferrin was not cytotoxic at the concentrations used in this study, suggesting that it may be useful as a preventive medicine against SARS-CoV-2 infection.
Amylase also inhibited SARS-CoV-2 infection in a dose-dependent manner at concentrations of 100 U/mL or higher. However, the concentration of amylase in normal saliva is reported to be about 30 U/mL 15, which is much lower than the significant concentration used in this experiment. Therefore, it is thought that amylase has slight protective effect against in vivo infection. It also shows a similar inhibitory effect against VSVpv infection, and differs from the action of lactoferrin in that it is not specific for SARS-CoV-2. The reason for the decrease in the infectivities of both pseudotyped viruses is not well understood, but since amylase is an enzyme that degrades part of the sugar chain, it is thought that it has some effect on the viral spike protein or envelope protein for the entry. As for cathelicidin and mucin, neither SARS-CoV-2pv nor VSVpv showed an inhibitory effect for the infection. Rather, cathelicidin had an enhancing effect for the infection at high concentrations. From these experiments, it was considered that some substances contained in saliva had enhancing effects for SARS-CoV-2 infection, but that the inhibitory effect may have been largely derived from lactoferrin. In addition, although physiological concentrations of amylase alone cannot completely prevent SARS-CoV-2 infection, it is thought to synergize with lactoferrin to exert some anti-SARS-CoV-2 activity. It is assumed that these factors led to the lower isolation rates of saliva specimens compared to the nasal/nasopharyngeal swab specimens. Although this study examined four substances that are considered to be abundant in saliva, it is suspected that other substances that are specific to saliva may also be involved.
From these findings, it was speculated that infectious SARS-CoV-2 viral particles are more abundant in the nasal cavity and nasopharynx than in the saliva, even if the number of genome copies is the same as on the result of a rRT-PCR test. It was found that saliva acts in defensive manner with regard to infectious diseases, although not completely, and inhibits SARS-CoV-2 infection. In the future, it will be necessary to continue to study how these components of saliva act during SARS-CoV-2 infection.