Pitfalls to be considered on the metabolomic analysis of biological samples by HR-MAS.

Abstract

HR-MAS (High-Resolution Magic AngleSpinning) is considered a powerfultechnique for metabolomic studies ofbiological samples that provides “intact”tissue spectra (Cheng et al., 1998; Waterset al., 2000; Sitter et al., 2002; Martinez-Bisbal et al., 2004; Payne et al., 2006;Coen et al., 2007; Bathen et al., 2010).The performance of HR-MAS, fol-lowed by quantitative histopathology hasdemonstrated that, despite some changes,HR-MAS can preserve approximately thetissue histopathologic features producingwell-resolved spectra of cellular metabo-lites (Cheng et al., 2000). Nevertheless,therearesomeaspectsarousedintheliter-ature about the possible biochemical andstructural changes that can occur duringthe sample storage and handling or mea-surement, and that should be consideredin the use of HR-MAS for the metaboli-cal characterization of biological samples.Ignoring these aspects might hamper therationale use of HR-MAS and a properinterpretation of the results, while con-sidering these factors may contribute toimprove the spectral quality and signifi-cance and would help to obtain coherentand reliable information of the sample.In the case of samples of biomedicalinterest, as those obtained from humanpatients, when NMR installations arein close proximity to the surgical unit,HR-MAS spectrum of fresh tissue canbe obtained within minutes of excision(Kinross et al., 2011). Frequently this isnot the case and, in the general prac-tice, the sample needs to be preserveduntil the moment of the HR-MAS study.The preservation is the key in biologi-cal samples, because it is needed to berepresentative of the alive system. Then,the sample handling and storage condi-tions are very relevant. The postmorteminterval, the time before being preservedand the surgical procedure itself, maypromote changes in the tissue metabo-lite concentration (Cheng et al., 1997;Opstad et al., 2008). This degradationcould be observed for example in humanbrain tissue by the decomposition of NAAto Acetate and Aspartate (Cheng et al.,1997). Therefore, the total sum of NAAand Acetate could represent an estimationof the initial content of NAA, althoughAcetate could be produced also by otherdegradationprocesses(Chengetal.,1997).In normal rat brain significant changesduetocontrolleddelaysinsamplefreez-ing were observed in metabolites associ-ated with glycolysis, as Alanine, Glucose,and Lactate (Opstad et al., 2008). Usually,the tissue is preserved by snap-freezing orultrafreezing until the acquisition of HR-MAS experiments. But freezing and thaw-ingislikelytocauseunpredictableamountof cell damage and lysis, with physicaldisruption of the cellular compartmen-talization and changes in the molecu-lar composition (Middleton et al., 1998;Waters et al., 2000; Bourne et al., 2003).Denaturation of proteins, together withthe effect of cellular breakdown, will havethe combined effect to reduce the num-ber of non-specific binding sites for smallendogenous solutes, causing an incre-ment in the concentration of NMR visi-ble metabolites (Middleton et al., 1998).These alterations have been observed alsoworking with other tissues. Thus, thespectra of rat renal tissue after freezingby both, snap-freezing or using liquidnitrogen, compared with spectra of freshtissue, showed variation in the signalintensities of water soluble metabolites (asAlanineandGlycine)andlipids,andcaus-ing a release of specific amino acids asLeucine, Valine and Isoleucine (releasedfrom specific binding sites of storageforms), while also reducing the free con-centration of renal osmolytes (probablydue to enzymatic conversion, binding tomacromolecules of sequestration by lipidvesicles) (Middleton et al., 1998; Waterset al., 2000). These changes likely dependonthetypeoftissueandorgan,sincefollowing snap-freezing procedures, littlechanges were observed in rat liver tissue(Waters et al., 2000). Preserving the sam-ples in a buffer and frozen until the NMRstudy or flushing the samples with D