Recovery Correction Technique for NMR Spectroscopy of Perchloric Acid Extracts using DL-valine-2,3-d2: Validation and Application to 5-Fluorouracil-induced Brain Damage

PURPOSE: We explored a recovery correction technique that can correct metabolite loss during perchloric acid (PCA) extraction and minimize inter-assay variance in quantitative (1)H nuclear magnetic resonance (NMR) spectroscopy of the brain and evaluated its efficacy in 5-fluorouracil (5-FU)- and saline-administered rats. METHODS: We measured the recovery of creatine and dl-valine-2,3-d2 from PCA extract containing both compounds (0.5 to 8 mM). We intravenously administered either 5-FU for 4 days (total, 100 mg/kg body weight) or saline into 2 groups of 11 rats each. We subsequently performed PCA extraction of the whole brain on Day 9, externally adding 7 µmol of dl-valine-2,3-d2. We estimated metabolite concentrations using an NMR spectrometer with recovery correction, correcting metabolite concentrations based on the recovery factor of dl-valine-2,3-d2. For each metabolite concentration, we calculated the coefficient of variation (CEV) and compared differences between the 2 groups using unpaired t-test. RESULTS: Equivalent recoveries of dl-valine-2,3-d2 (89.4 ± 3.9%) and creatine (89.7 ± 3.9%) in the PCA extract of the mixed solution indicated the suitability of dl-valine-2,3-d2 as an internal reference. In the rat study, recovery of dl-valine-2,3-d2 was 90.6 ± 9.2%. Nine major metabolite concentrations adjusted by recovery of dl-valine-2,3-d2 in saline-administered rats were comparable to data in the literature. CEVs of these metabolites were reduced from 10 to 17% before to 7 to 16% after correction. The significance of differences in alanine and taurine between the 5-FU- and saline-administered groups was determined only after recovery correction (0.75 ± 0.12 versus 0.86 ± 0.07 for alanine; 5.17 ± 0.59 versus 5.66 ± 0.42 for taurine [µmol/g brain tissue]; P < 0.05). CONCLUSION: A new recovery correction technique corrected metabolite loss during PCA extraction, minimized inter-assay variance in quantitative (1)H NMR spectroscopy of brain tissue, and effectively detected inter-group differences in concentrations of brain metabolites between 5-FU- and saline-administered rats.