Optimized NMR-based metabolomic protocol for salivary analysis in clinical practice
Description
Saliva remains underexplored in metabolomics compared to widely used biofluids such as blood and urine. However, its non-invasive, rapid, and cost-effective collection, as well as its suitability for self-sampling, make it particularly attractive for clinical and personalized medicine. Moreover, saliva is expected to provide complementary metabolic information to other biofluids. In this context, the present study aims to identify the most suitable sample collection and preparation protocol for maximizing relevant and informative metabolomic data from salivary NMR profiles. Four preparation protocols, selected and adapted from the literature, were systematically compared. Evaluation criteria included the number of identified metabolites, their concentrations, and the intra-day repeatability based on PCA (buckets). The most promising one is an in-house adapted method involving sequential centrifugation, freeze-drying and ultrafiltration (protocol 4). It emerges as the most robust protocol, providing the broadest metabolome coverage with 42 metabolites identified and quantified. This method was further assessed using ANOVA to determine intra- and inter-day precision. Most metabolites demonstrate excellent repeatability (coefficients of variation below 10%) and confirm the reliability of the protocol for quantitative metabolomics analysis. Overall, the optimized approach yields high-quality spectra, wide metabolic coverage and strong analytical precision, making it well suited for reproducible salivary NMR metabolomics. Beyond its methodological contribution, this work highlights saliva as a promising biofluid for diagnostics, disease monitoring and personalized medicine. With appropriate standardization of collection and preparation procedures, saliva could emerge as a biofluid of choice in clinical metabolomics, contributing to advances in preventive and precision healthcare.
Files
Steps to reproduce
Comparison of the 4 protocols Sample collection Saliva was collected under non-fasting conditions; participants waited ≥2 h after food intake and toothbrushing. After two water rinses, unstimulated saliva was collected by drooling. About 2–3 mL per participant were obtained from five individuals in sterile tubes. Samples were centrifuged (15 min, 3730 g, 4°C), rested on ice for 10 min, and supernatants were pooled, divided into 20 aliquots (500 µL; 5 per protocol), and stored at −80°C. Sample preparation: 4 protocols AmiconUltra® 0.5 mL 3 kDa filters were washed five times with 500 µL Milli-Q water (5 min, 13 790 g, cooling temperature). Each protocol used 500 µL saliva per aliquot and samples were diluted with phosphate buffer containing TSP and calcium formate to obtain identical final calcium formate concentrations. Protocol 1: Thawed aliquots were mixed with 125 µL phosphate buffer (0.5 mM TSP) and 50 µL 3.5 mM deuterated calcium formate and analysed by 1H-NMR in 5 mm tubes. Protocol 2: Thawed aliquots were filtered (3 kDa, 90 min, 13 790 g, 4°C). Filtrates (385 µL) were mixed with 240 µL phosphate buffer (0.2 mM TSP) and 50 µL calcium formate and analysed in 5 mm tubes. Protocol 3: Thawed aliquots were filtered in same conditions as protocol 2, freeze-dried (22 h), reconstituted with 278 µL phosphate buffer (0.075 mM TSP) and 22 µL calcium formate, and analysed in 3 mm tubes. Protocol 4: Thawed aliquots were freeze-dried (22 h), reconstituted as in protocol 3, then filtered (3 kDa, 90 min, 13 790 g, 4°C) and analysed in 3 mm tubes. NMR spectroscopy Spectra were recorded at 298 K on a Bruker® Neo spectrometer (500.13 MHz) equipped with a 5 mm TCI cryoprobe. 1H-NMR spectra were acquired using CPMG (protocol 1) or 1D NOESY presat (protocols 2–4) sequences with 64 transients. Buckets Spectra were processed with MestReNova® (v14.3.1), corrected for phase and baseline, normalized to the formate peak and binned (0.04 ppm, 0.5–9.0 ppm). Water (4.5-5.1 ppm), TSP (-0.5-0.5 ppm) and glycerol (3.53-3.60 ppm and 3.63-3.68 ppm) regions were excluded. Data were imported into SIMCA® (v16.0.1); calcium formate buckets were removed, UV scaling applied and PCA performed. Standard deviations of PC scores (PC1–PC3) from five replicates were calculated per method. Quantification Spectra were analysed with Chenomx® NMR Suite (v9.02). Based on the formate signal (8.46 ppm), 44 metabolites were identified and quantified. Intra- and inter-day precision of protocol 4 Sample collection and preparation Saliva from four individuals was collected as described above, pooled after centrifugation, divided into 15 aliquots (500 µL; 5 per day) and stored at −80°C before preparation according to protocol 4. Quantification NOESY presat spectra were analysed with Chenomx® NMR Suite, allowing quantification of 41 metabolites. For each metabolite, ANOVA was used to assess intra- and inter-day precision by extracting SDs and CVs with R (v4.3.1).
Institutions
- Universite de LiegeLiege