Salivary proteomic analysis in asymptomatic and symptomatic SARS-CoV-2 infection: Innate immunity, taste perception and FABP5 proteins make the difference

Abstract

Background and aim: SARS-CoV-2 infection spawns from an asymptomatic condition to a fatal disease. Age, comorbidities, and several blood biomarkers are associated with infection outcome. We searched for biomarkers by untargeted and targeted proteomic analysis of saliva, a source of viral particles and host proteins.

Methods: Saliva samples from 19 asymptomatic and 16 symptomatic SARS-CoV-2 infected subjects, and 20 controls were analyzed by LC-MS/MS for untargeted peptidomic (flow through of 10 kDa filter) and proteomic (trypsin digestion of filter retained proteins) profiling.

Results: Peptides from 53 salivary proteins were identified. ADF was detected only in controls, while IL1RA only in infected subjects. PRPs, DSC2, FABP5, his-1, IL1RA, PRH1, STATH, SMR3B, ANXA1, MUC7, ACTN4, IGKV1-33 and TGM3 were significantly different between asymptomatic and symptomatic subjects. Retained proteins were 117, being 11 highly different between asymptomatic and symptomatic (fold change ≥2 or ≤-2). After validation by LC-MS/MS-SRM (selected reaction monitoring analysis), the most significant discriminant proteins at PCA were IL1RA, CYSTB, S100A8, S100A9, CA6, and FABP5.

Conclusions: The differentially abundant proteins involved in innate immunity (S100 proteins), taste (CA6 and cystatins), and viral binding to the host (FABP5), appear to be of interest for use as potential biomarkers and drugs targets.

Keywords: Biomarkers; CA6; FABP5; Proteomic profiling; SARS-CoV-2; Saliva.

Fig. 1

String Network Analysis of the proteins identified as significantly different between asymptomatic and symptomatic groups. Lines connecting the different nodes represent functional and/or physical interactions. Red and blue nodes represent proteins involved in the immune system and in the innate immune system, respectively.

Fig. 2

Manhattan plot reporting the enrichment analysis of GO terms associated to the proteins identified as differently abundant in asymptomatic and symptomatic groups. The top twelve most significant GO-terms are listed.

Fig. 3

Boxplots with individual data points (dots) of log₁₀ transformed peptide levels, subdivided by patients’ groups (A = Asymptomatic and S = Symptomatic).

Fig. 4

Hierarchical clustered correlation matrix of all studied peptides, including age.

Fig. 5

Principal Component Analysis (PCA) results, obtained using scaled data of all the studied proteotypic peptides and patients’ age. Panels A and B report the contribution of the 8 most important variables in explaining Dimensions 1 and 2, respectively. Panel C reports the bidimensional plot of PCA, with individual patients divided by asymptomatic (A) and symptomatic (S) and with superimposed the two ellipsoids (each including 95% of individuals); Panel D reports the contribution of the 8 most important variables in explaining Dimensions 3. Panel E shows the three-dimensional plot of PCA, with the superimposed the two ellipsoids.

Fig. 6

Hierarchical clustered heatmap analysis, performed including peptides, age and patients’ groups.