Evaluation of human plasma sample preparation protocols for untargeted metabolic profiles analyzed by UHPLC-ESI-TOF-MS

Abstract

Eight human plasma preparation protocols were evaluated for their suitability for metabolomic studies by ultra-high-performance liquid chromatography coupled with electrospray ionization time-of-flight mass spectrometry: organic solvent protein precipitation (PPT) with either methanol or acetonitrile in 2:1 and 3:1 (v/v) ratios with plasma; solid-phase extraction (SPE) using C18 or HybridSPE cartridges; and a combination of PPT and SPE C18 cartridges and microextraction by packed sorbent. A study design in which the order of injection of the samples was not randomized is presented. The analyses were conducted in a BEH C18 column (1.7 μm, 2.1 mm × 100 mm) using a linear gradient from 100 % water to 100 % methanol, both with 0.1 % formic acid, in 21 min. The most reproducible protocol considering both the univariate and the multivariate analysis results was PPT with acetonitrile in a 2:1 (v/v) ratio with plasma, offering a mean coefficient of variation of the area of all the detected features of 0.15 and one of the best clusterings in the principal component analysis plots. On the other hand, the highest number of extracted features was achieved using methanol in a 2:1 (v/v) ratio with plasma as the PPT solvent, closely followed by the same protocol with acetonitrile in a 2:1 (v/v) ratio with plasma, which offered only 1.2 % fewer repeatable features. In terms of concentration of remaining protein, protocols based on PPT with acetonitrile provided cleaner extracts than protocols based on PPT with methanol. Finally, pairwise comparison showed that the use of PPT- and SPE-based protocols offers a different coverage of the metabolome.