Quantitative determination of endogenous sorbitol and fructose in human nerve tissues by atmospheric-pressure chemical ionization liquid chromatography/tandem mass spectrometry.

Attachment of anions to sorbitol and fructose has been shown to enhance sensitivity in both electrospray ionization (ESI) and atmospheric-pressure chemical ionization (APCI) mass spectrometry. The post-column addition of CHCl3 produced Cl-adducts of sorbitol and fructose but their signals were suppressed due to the elevated background. Different chlorinated compounds and different additive methods were systematically investigated to form more abundant Cl-adduct precursor ions and deprotonated product ions. The major causes of the high background were explored and effective methods were developed to improve the signal-to-noise ratios and reproducibility. The compositions of mobile phase, percentages of organic modifiers (MeCN, MeOH and water), columns, oven temperature, flow rates and different gradients were investigated to separate sorbitol from fructose along with their isomers including glucose, galactose, mannose, sorbose, mannitol, and dulcitol. The optimized separation was achieved on a Luna 5 mu NH2 100A column (150 x 4.6 mm) using a mobile phase containing MeCN with 0.1% of CH2Cl2 and 50% MeOH in water at a flow rate of 800 microL/min and an oven temperature of 40 degrees C using a gradient liquid chromatography (LC) system. Human nerve tissue samples were extracted by protein precipitation followed by mixed-mode solid-phase extraction. The LC/ESI-MS/MS method produced higher peak intensities than LC/APCI-MS/MS. However, there were matrix effects from extracted tissues in LC/ESI-MS/MS but not in LC/APCI-MS/MS. Consequently, APCI proved to be the more effective method of ionization. Then the LC/APCI-MS/MS method was fully validated and successfully applied to analysis of clinical samples. The concentrations of endogenous sorbitol and fructose were determined using calibration curves employing sorbitol-13C6 and fructose-13C6 as surrogate analytes. The method has provided excellent intra- and inter-assay precision and accuracy with linear ranges of 0.2-80 ng/mg for sorbitol and 1-400 ng/mg for fructose in human nerve tissues.