ORIGINAL PAPER NMR metabolomics for assessment of exercise effects with mouse biofluids Laurence Le Moyec & Laurence Mille-Hamard & Mohamed N. Triba & Carole Breuneval & Hélène Petot & Véronique L. Billat Received: 4 March 2012 / Revised: 29 May 2012 / Accepted: 29 May 2012 # Springer-Verlag 2012 Abstract Exercise modulates the metabolome in urine or blood as demonstrated previously for humans and animal models. Using nuclear magnetic resonance (NMR) metab- olomics, the present study compares the metabolic conse- quences of an exhaustive exercise at peak velocity (Vp) and at critical velocity (Vc) on mice. Since small-volume sam- ples (blood and urine) were collected, dilution was neces- sary to acquire NMR spectra. Consequently, specific processing methods were applied before statistical analysis. According to the type of exercise (control group, Vp group and Vc group), 26 male mice were divided into three groups. Mice were sacrificed 2 h after the end of exercise, and urine and blood samples were drawn from each mouse. Proton NMR spectra were acquired with urine and depro- teinized blood. The NMR data were aligned with the icoshift method and normalised using the probabilistic quotient method. Finally, data were analysed with the orthogonal projection of latent-structure analysis. The spectra obtained with deproteinized blood can neither discriminate the con- trol mice from exercised mice nor discriminate according to the duration of the exercise. With urine samples, a signifi- cant statistical model can be estimated when comparing the control mice to both groups, Vc and Vp. The best model is obtained according to the exercise duration with all mice. Taking into account the spectral regions having the highest correlations, the discriminant metabolites are allantoin, ino- sine and branched-chain amino acids. In conclusion, metab- olomic profiles assessed with NMR are highly dependent on the exercise. These results show that urine samples are more informative than blood samples and that the duration of the exercise is a more important parameter to influence the metabolomic status than the exercise velocity. Keywords NMR . Mouse . Exercise . Urine . Blood . Exhaustion . Peak velocity . Critical velocity Introduction Recently, the metabolic pathways involved during physical exercise were investigated by the metabolomic methods. Metabolomics is highly sensitive to environmental factors including age, pharmacology and toxicology (for review, see [1]). Therefore, the phenotypic responses of a subject to exercise may be shown using this method. Among the analytic techniques implemented for metab- olomics, mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy are the most common ones and pro- vide values for a high number of parameters. Through mass spectrometry, a highly sensitive technique, metabolites pres- ent at low concentration may be detected. On the other hand, the method needs a complex sample preparation protocol. For several reasons, including the chromatographic and the ionization methods, a low reproducibility of the measure- ments is the main pitfall of mass spectrometry. With proton NMR spectroscopy, the metabolites detected must be at relatively high concentration in biofluids (more than 10 μM). This limitation is balanced by the high reproduc- ibility of the technique associated to the linear analytical response. Moreover, proton NMR sensitivity does not L. Le Moyec (*) : L. Mille-Hamard : C. Breuneval : H. Petot : V. L. Billat UBIAE INSERM U902, Université Evry Val d’Essonne, Bat Maupertuis, rue du Père Jarlan, 91025 Evry Cedex, France e-mail: laurence.lemoyec@univ-evry.fr M. N. Triba CSPBAT UMR 7244, Université Paris 13, 74, rue Marcel Cachin, 93000 Bobigny, France Anal Bioanal Chem DOI 10.1007/s00216-012-6165-6