Integrated salivary proteomic profiling reveals microbial–immune interactions in recurrent aphthous stomatitis

This dataset explores the molecular mechanisms underlying recurrent aphthous stomatitis through integrated salivary proteomics and metaproteomics. Using mass spectrometry-based analysis of the cellular (pellet) saliva fraction, the study characterizes host–microbe crosstalk and immune remodeling across clinical states of disease and health. Approximately 1,700 proteins were identified, revealing persistent complement activation, dipeptidyl peptidase 4 (DPP4)–mediated chemokine signaling, and overexpression of Mammaglobin-B (SCGB2A1) as consistent hallmarks of epithelial and immune interaction. Microbial profiling by MALDI-Biotyper and qPCR confirmed the enrichment of Streptococcus pneumoniae and Clostridium species, while functional metaproteomics indicated bacterial metabolic reprogramming associated with inflammation. These results provide a comprehensive molecular map of the saliva–mucosa interface, highlighting key pathways linking microbial activity, immune modulation, and epithelial stress relevant to recurrent oral ulceration.<b>Sample Processing Protocol:</b> Unstimulated saliva was collected in the morning after overnight fasting and rinsing with water, under informed consent and ethical approval (University of Antofagasta protocol #156/2018; Maule Health Service #22-11-2018). Samples were immediately treated with protease inhibitors and centrifuged at 3,000 × g for 10 min at 4 °C to separate the cellular pellet, containing desquamated epithelial cells, microbial aggregates, and extracellular vesicles. Saliva pellets were pooled into nine groups: healthy controls (n = 3 pools, 11 samples each), ulcerative disease (n = 3 pools, 12 samples each), and remission (n = 3 pools, 12 samples each). Proteins were extracted with 8 M urea/25 mM ammonium bicarbonate, reduced with 20 mM dithiothreitol for 1 h, and alkylated with 20 mM iodoacetamide for 1 h in the dark. After eightfold dilution, proteins were digested overnight with sequencing-grade trypsin (1:50 w/w) at 37 °C. Peptides were acidified with formic acid (10%), purified with Sep-Pak C18 (Waters), and dried in a rotary concentrator. A total of 200 ng of peptides was analyzed on a nanoElute UHPLC (Bruker Daltonics) coupled to a timsTOF Pro mass spectrometer using a 25 cm × 75 µm C18 column (IonOpticks). Separation was achieved with a 90-min gradient (2–35% acetonitrile, 0.1% formic acid). Data were acquired in PASEF mode (10 cycles, m/z 100–1700). Microbial analysis was performed by MALDI-Biotyper and verified by qPCR amplification of the Streptococcus pneumoniae lytA gene. Functional metaproteomics was performed using Unipept (UniProt reference proteomes).<b>Data Processing Protocol: </b>Raw data were analyzed using MSFragger v3.5 via FragPipe v18.0. Searches were conducted against the UniProt human and bacterial proteomes with a 1% FDR threshold. Trypsin specificity was applied (max. 2 missed cleavages), with fixed carbamidomethylation of cysteine and variable modifications for methionine oxidation, N-terminal acetylation, and deamidation. Precursor and fragment tolerances were ±20 ppm. Protein intensities were median-normalized and log₂-transformed. Differentially abundant proteins were defined by |log₂ FC| &gt; 1 and adjusted p &lt; 0.05 (Benjamini–Hochberg). Functional enrichment was performed with Reactome, integrating human and bacterial pathways. Metaproteomic and keratinocyte secretome data were processed with Proteome Discoverer (Thermo Scientific) and annotated with Reactome. Quality control involved randomized sample runs and pooled QC analyses every 10 injections.