The objective of this study was to compare ex vivo proton high-resolution magic angle spinning magnetic resonance spectra of intact tissue with those spectra obtained by solution (1)H NMR of brain extracts of the same sample. Sixteen brain tissue samples from simian immunodeficiency virus-infected rhesus macaques from both frontal cortex and putamen were evaluated by comparing brain metabolite quantities of N-acetylaspartate (NAA), choline-containing compounds (Cho), myo-inositol (MI), creatine (Cr), lactate (Lac), glutamate (Glu) and acetate (Ace). The ratios of the individual NMR peak areas of all metabolites relative to the creatine peak area were calculated. Linear regression analysis revealed significant correlations between measurements using the two methods. The strength of the correlations varied depending on the metabolite studied. We found highly significant correlations for NAA/Cr (r2 = 0.77; p < 0.0001), NAA + Ace/Cr (r2 = 0.73; p < 0.0001) and MI/Cr (r2 = 0.75; p < 0.0001). We observed somewhat less strong correlations for Glu/Cr (r2 = 0.54; p < 0.002) and Lac/Cr (r2 = 0.54; p < 0.002). There was a substantially weaker correlation for Cho/Cr (r2 = 0.32; p = 0.02). When plotting the metabolite ratios obtained by 1H HRMAS NMR of the intact tissue sample on the ordinate vs 1H NMR of the tissue extract on the abscissa, most metabolites exhibited a slope close to unity, and a positive intercept probably due to macromolecular contributions to the MAS spectra. The slope for Cho/Cr was substantially less than unity. Generally, samples from the frontal cortex showed a better correlation between intact and extracted tissue samples than putamen. This is most prominent in the cases of NAA/Cr and Cho/Cr. We conclude that both methods provide substantially the same information for most major brain metabolites, with the exception of the Cho resonance.
Copyright 2005 John Wiley & Sons, Ltd