Evaluation of plasma metabolomic profile by exploiting the synergy between MS and NMR

Metabolomics plays a unique role in the study of biological systems related to health, disease and the environment [1] by studying all metabolites of a biological system [2], or even the end products of the expression of its gene. Due to the complexity of the metabolome, to obtain a comprehensive, sensitive and reliable information, is essential to combine different analytical technique. So, the aim of this project is to integrate NMR and HRMS on plasma samples obtained from patients undergoing treadmill test evaluating the changes occurred in metabolites before and after physical activity. Mass spectra were acquired on an Orbitrap Q-Exactive mass spectrometer equipped with a HESI source operating in positive and negative ionization mode; LC separation was performed with a RP and a HILIC column to detect as many metabolites as possible. Each sample was also acquired in NMR with a Bruker Avance 700 MHz spectrometer equipped with a triple-resonance TXI probe and a SampleXpress Lite autosampler, using a noesypr1d pulse sequence for water suppression. The intensities normalization for all spectra is mandatory, due to the large variability in the analytes total concentration within the different samples. Firstly, PCA was used to obtain an unsupervised evaluation of the results; already from this, a difference emerged between the samples before and after physical activity. Then a reduction in the dimensionality of the LC-MS data was performed, with the application of successive OPLS-DA models; the number of variables was reduced to obtain best Q2 value. So, the reduced dataset was added to the NMR one with an increase in predictivity, compared to the latter. The results obtained from the OPLS-DA, of the NMR/HRMS model, lead us to identify the main metabolic differences in basal samples versus stress samples. The second approach performed was SYNHMET [3] (SYnergic use of NMR and HRMS for METabolomics), the intensities obtained from the MS are linearly correlated with the NMR signals, resulting in standard-free identification of mass hits and conversion of intensities into concentrations. The correlation parameter was the exact mass; a good correlation is essential to ensure that the analyte in question is the same in both techniques. The correlation that does not exist in one sample exists in all of them as a group because, in this case, is the distribution of intensities that determines whether a given chemical shift belongs to a molecule with a specific m/z. PCA results demonstrate how all samples can be divided based on the subject gender. The evaluation of two OPLS-DA models showed different significant metabolites for each separation group. So, the applicability of the presented procedure was demonstrated.