Rat strain‐dependent variations in brain metabolites detected by in vivo¹H NMR spectroscopy at 16.4T

Abstract

Localized in vivo (1) H NMR spectroscopy is playing an increasing role in preclinical studies, because of its ability to quantify the concentrations of up to 20 metabolites in rat brain. To assess the differences between often-used rat strains, the neurochemical profiles of Sprague-Dawley, Wistar and Fischer rats were determined at ultrashort TE at 16.4 T. To ascertain high-qualitative quantification, a first experiment examined the dependence of the measuring time on the quantification results and precision by precisely the number of averages between 16 and 320. It was shown that most metabolites can be quantified accurately within a short scan time, yielding Cramér-Rao lower bounds below 20% and stable concentrations for 16 metabolites with as few as 32 or 64 averages in the thalamus and hippocampus, respectively. Interstrain differences in metabolite concentrations were shown to be moderate, with taurine varying significantly between Sprague-Dawley and Wistar rats, and slightly more distinct differences from Fischer rats, including variations in glutamate and myo-inositol. The high spectral quality and quantification precision of all data again demonstrated the potential of in vivo (1)H NMR spectroscopy at ultrahigh field.

Quantification issues of in vivo1H NMR spectroscopy of the rat brain investigated at 16.4 T

“…Especially for relatively abundant metabolites, the quantitative value found was not dependent to a high degree on the SNR, which was the main factor determining the quantification results for almost all metabolites in a previous study in the human brain at 4 T . This indicates that an increase in SNR only improves quantification to a certain degree, in agreement with a previous study demonstrating a decreasing effect of averaging on the CRLBs . As a result of the high SNRs that are reached regularly for most peaks at high field strengths, this limit may already have been reached and further improvements in the quantification accuracy may require a reduction in the linewidth.…”

“…For the heavily overlapping metabolite pairs Cr/PCr, GPC/PCh, Gln/Glu, Gly/mI and NAA/NAAG, the correlation coefficients were calculated to quantify the interactions between the peaks. For Cr/PCr, the concentration of NAA was used as reference and assumed to be 8.39 µmol/g .…”

Determination of regional variations and reproducibility in in vivo 1H NMR spectroscopy of the rat brain at 16.4 T

“…In this study, the goal of differentiating between signals from different brain structures required decreasing the voxel sizes by a factor of up to 3.5 compared to previous studies (7, 8) to avoid partial volume effects. Combined with the well‐established double localization scheme, combining OVS with STEAM volume selection, this made it possible to acquire spectra with negligible lipid contamination as well as minimal partial volume effects in all examined brain regions.…”

“…Macromolecular components were determined by implementing prior knowledge, derived by parameterization of a metabolite‐nulled spectrum in LCModel. The individual macromolecular components were included in the fit for each spectrum based on the same parameterization, as it was shown that these components did not vary significantly between rat brain regions (16) and rat strains (8). Concentration values with CRLB > 50% were excluded from the analysis.…”

“…Although we recently have shown the possibility to quantify up to 20 metabolites in the rat brain at 16.4 T (7, 8), the relatively large voxel sizes used in that study imply contributions from different tissue types in each spectrum. With the goal of obtaining data with minimal partial volume effects, this study was designed to show that spectra from much smaller voxels can be acquired without significant loss in information.…”