Multi-omic Analysis of Dental Plaque in Non-diabetic and Diabetic Patients

Objectives: Oral microbial communities influence human health through complex, and yet undiscovered, microbial-microbial and microbial-host interactions. We propose that these interactions are in part mediated through metabolic intermediates and microbial proteins. Thus to explore the composition of the oral microbiome and its interaction with disease status, we have performed metabolomics, lipidomics and proteomics analysis on dental plaque collected from diabetic and non-diabetic patients with and without periodontal disease.
Methods: We recruited 99 patients from the Marshfield dental clinic (Marshfield, WI) to participate in the study. Study participants were classified as diabetic with or without periodontal disease (n=37 and 41, respectively) or non-diabetic with and without periodontal disease (n=10 and 11, respectively). We collected 3 plaque samples per participant from both buccal and palatal/lingual tooth surfaces; two plaque samples were used for mass spectrometry (MS) analysis (metabolomics, lipidomics, and proteomics analysis) and one plaque sample was reserved for later 16S/18S ribosomal RNA analysis. For the MS analysis, small molecules and lipids were extracted with 2:2:1 methanol:acetonitrile:water. Small molecules were derivatized with methoxyamine and MSTFA then separated by gas chromatography (GC) and analyzed by a Thermo Q-Exactive GC, and lipids were separated by reverse-phase liquid chromatography (LC) and analyzed by tandem MS using a Thermo Q-Exactive. In the same plaque samples, protein was precipitated from the extract then digested with endoproteinase Lys-C. The resulting peptides were analyzed by LC-MS using a Thermo Orbitrap Elite.
Results: With this multi-omic pipeline, we quantified ~250 metabolic features, included sugars, amino acids, sugar alcohols, organic acids, and fatty acids, and ~70 lipid species, including phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidylserine (PS) and triacylglycerol (TG). We also identified ~1200 microbial and host proteins per sample.
Conclusions: This rich dataset has the potential to reveal important diagnostic biomarkers and provide new insight into microbiome-host interactions.