This work should interest people studying neurodegenerative diseases and particularly Huntington’s disease using ¹H-spectroscopy and CEST imaging.Huntington’s disease (HD) is an inherited neurodegenerative disease characterized by motor, cognitive and psychiatric symptoms¹. Atrophy of the striatum is currently the best biomarker of disease progression in HD gene carriers. However, there is an urgent need to identify novel functional biomarkers of disease progression to better understand pathological processes and to monitor HD patients in clinical trials. Changes in brain metabolites have been also consistently seen in HD patients and animal models using MRS², but metabolite measurements are generally limited to a single voxel. Thus, novel methods that could measure the metabolic defects in the entire brain, and with a precise anatomical resolution, the metabolic defects would be of major interest. Therefore, we propose to perform Chemical Exchange Saturation Transfer imaging of glutamate (gluCEST³) to map glutamate distribution in the brain of genetic mouse and rat models of HD, and to evaluate the relevance of gluCEST in the context of HD.

Mouse model: 12 months knock-in mice expressing mouse/human exon 1 containing 140 CAG repeats inserted in the HTT gene were used (CAG 140 KI)⁴. Three cohorts were compared: Wild Type (+/+, n=5), heterozygous (+/Tg, n=5) and homozygous mice (Tg/Tg, n=5) for the HTT gene.

Rat model: 12 months transgenic rats obtained using a human Bacterial Artificial Chromosome containing 97 CAG/CAA repeats were used (BAC-HD⁵). Two cohorts were compared: Wild Type (+/+, n=9) and homozygous (+/Tg, n=8) rats.

NMR: MRS data were acquired on a horizontal 11.7T Bruker magnet in a voxel positioned in the left striatum (8 and 43µL for mice and rats respectively). A LASER sequence was used with TR/TE=5000/20ms. Glutamine (Gln), total choline (tCho), myo-inositol (Ins), glutamate (Glu), total N-Acetyl-Aspartate (tNAA) and Taurine (Tau) concentrations were calculated relative to total Creatine (tCr) with good precision (Cramér-Rao lower bounds <5%) using LCModel⁶.

GluCEST: GluCEST images were acquired using TSE sequence preceded by a frequency-selective continuous wave saturation pulse of 1s with a B₁ intensity of 5µT applied at frequencies ranging from -5 to 5ppm by 0.5ppm steps. B₀ inhomogeneity was corrected using WASSR⁷. GluCEST images were calculated using asymmetric Magnetization Transfer Ratio (MTRasym) at ±3ppm.

Mice: Figure 1 shows typical spectrum acquired in mouse striatum (Fig.1a) and metabolic profiles for each genotype (Fig.1b). Notable decreases were measured, particularly for tNAA (-17.4% and -24.3% for +/Tg and Tg/Tg, respectively) and Glu (-13% and -14.9% for +/Tg and Tg/Tg, respectively). A significant increase of Gln (+27%) was measured in Tg/Tg. Figure 2 shows examples of gluCEST images acquired in each cohort. Lower gluCEST contrasts in HD mice were seen, especially for homozygous (Fig.2, bottom panel). Mean MTRasym spectra measured in striatum (Fig.3a) and corpus callosum (Fig.3b) of +/+ (solid lines) and Tg/Tg (dotted lines) confirmed the decrease of gluCEST contrast. In order to perform regional analysis of gluCEST contrast, several regions of interest were drawn (Fig.4, top panel) and mean MTRasym were calculated. Variations of mean gluCEST contrast at 3ppm were calculated between +/+ and +/Tg (Fig.4, bottom left) and between +/+ and Tg/Tg (Fig.4, bottom right). Homozygous mice exhibited decreased glutamate across a majority of the brain especially in the cortex and the striatum. Surprisingly, the most affected structure in both heterozygous and homozygous was the corpus callosum (-22% (p-value=0.092) and -28% (p-value=0.045) respectively).

Rat: MRS results showed similar trend in metabolic modifications, but did not reach significance (data not shown). However, variation map of gluCEST contrast showed a decrease in the whole brain and especially in the corpus callusom (Fig.5) as observed in mice.

The decrease in [Glu] and [tNAA], two neuronal metabolites, suggests neuronal alterations in HD animals. The increase in [Gln], an astrocytic metabolite, may reflect a slight inflammation or astrocytic reactivity. GluCEST confirmed the decrease in [Glu] in the striatum in HD mice. In addition, good spatial resolution of gluCEST over ¹H-MRS allowed identification of other brain regions with a reduction in [Glu]. Glutamate loss was less pronounced in heterozygous as compared to homozygous, which is consistent with a faster progression of the disease in this latter group. Interestingly, the corpus callosum was the most affected structure in both rodent models. This finding is in agreement with MRI studies showing early alterations of white matter in HD patients^1,8. In this study, we evaluated for the first time gluCEST as a potential biomarker of HD. Variation maps of glutamate levels could be a valuable tool to follow HD progression.