Implications of salivary metabolomics sample processing in biomarker discovery

Objective: Saliva is a practical, non-invasive biofluid to identify biomarkers of health and disease. Saliva processing techniques pose challenges with a lack of standardization, producing variable results which may impact biomarker discovery. Mucins, a major constituent of mucous, influence salivary composition, retention time of proteins identified, and biological activity. Mucinase removes mucins which reduces mucosal viscosity and may improve salivary handling overall. Therefore, this pilot study's aim was to evaluate the effects of saliva processing and treatment on the metabolomic profile after exercise-heat stress exposure. Methods: 5 Participants (n=5) completed heat tolerance testing (HTT) before (pre) and after (post) 5 days of heat acclimation (HA) 40°C, ~40% relative humidity. Saliva was collected immediately post-HTT (5.0km·hr-1, 2% grade, 120 mins) using a Salimetrics passive drool collection kit (Salimetrics, State College, PA), snap frozen, and stored at –80°C for metabolome analysis. Salivary metabolome was extracted and normalized to 10mg·mL-1 (0.1% Formic acid) and run in positive mode Waters SynaptG2-Si (mass range 80-2000, resolution mode), peaks aligned to Metlin database (adducts M+H, M+NA, M+K, 2M+Na, 2M+K) using ProgenesisIQ software, MSDial (MoNA), NIH Metabolomics Workbench (RefMet). Intraday/Interday variability: saliva from one donor was collected in the morning and afternoon of two sequential days and frozen as above. Saliva treatment performed according to Figure 1. Results: No significant differences were observed between the number of peaks or compounds detected between time point and treatment (filtered, unfiltered, mucinase, no mucinase, see Figure 1). 13,368 compound ions and 633 features were identified when matched to the Metlin database. 315/633 features were present in >50% of the samples. 6560/12844 metabolites including non-identified compounds were present in >50% of samples. Top compounds differentially abundant (p<0.05) between timepoints were minimally overlapping (<10) across treatment groups. LDA reveals global-metabolome clusters between mucinase-treated and non-mucinase treated samples. OPLS-DA analysis demonstrates cluster separation between timepoints (pre and postHA), with main metabolites contributing to the variation being peptides, with the exception of endomorphin 2 and N-Acetylneuraminic acid. Hierarchical clustering suggests the strongest variability among interday collection, followed by intraday collection and neat versus supernatant processed samples. Conclusion: Salivary metabolomics offers a practical option for biomarker discovery, however, pre-processing, software tools, and collection timing greatly influences biomarker reliability. Funding: DARPA This abstract was presented at the American Physiology Summit 2025 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.