In vivo neurometabolic profiling in patients with spinocerebellar ataxia types 1, 2, 3, and 7

Abstract

Spinocerebellar ataxias (SCAs) belong to polyglutamine repeat disorders and are characterized by a predominant atrophy of the cerebellum and the pons. Proton magnetic resonance spectroscopy ((1) H MRS) using an optimized semiadiabatic localization by adiabatic selective refocusing (semi-LASER) protocol was performed at 3 T to determine metabolite concentrations in the cerebellar vermis and pons of a cohort of patients with SCA1 (n=16), SCA2 (n=12), SCA3 (n=21), and SCA7 (n=12) and healthy controls (n=33). Compared with controls, patients displayed lower total N-acetylaspartate and, to a lesser extent, lower glutamate, reflecting neuronal loss/dysfunction, whereas the glial marker, myoinositol (myo-Ins), was elevated. Patients also showed higher total creatine as reported in Huntington's disease, another polyglutamine repeat disorder. A strong correlation was found between the Scale for the Assessment and Rating of Ataxia and the neurometabolites in both affected regions of patients. Principal component analyses confirmed that neuronal metabolites (total N-acetylaspartate and glutamate) were inversely correlated in the vermis and the pons to glial (myo-Ins) and energetic (total creatine) metabolites, as well as to disease severity (motor scales). Neurochemical plots with selected metabolites also allowed the separation of SCA2 and SCA3 from controls. The neurometabolic profiles detected in patients underlie cell-specific changes in neuronal and astrocytic compartments that cannot be assessed by other neuroimaging modalities. The inverse correlation between metabolites from these two compartments suggests a metabolic attempt to compensate for neuronal damage in SCAs. Because these biomarkers reflect dynamic aspects of cellular metabolism, they are good candidates for proof-of-concept therapeutic trials. © 2015 International Parkinson and Movement Disorder Society.

Trial registration: ClinicalTrials.gov NCT01470729.

Keywords: NMR spectroscopy; biomarker; movement disorders; neurochemical profile; spinocerebellar ataxia.

© 2015 International Parkinson and Movement Disorder Society.

Figures

Figure 1

Mean concentrations of metabolites that showed significant differences in the vermis and the pons of patients with SCA1, 2, 3, 7 vs. controls. p values represent Dunnett-corrected statistically significant differences between patients and controls differences (* p myo-Ins and higher energy marker tCr. Error bars represent standard deviations (SDs).

Figure 2

Correlation between clinical scores and neurochemical concentrations in the vermis and the pons of patients with SCAs. SARA scores correlated with (A) tCr in SCA1 vermis, (B) tNAA in SCA7 vermis, (C) myo-Ins in SCA3 pons, (D) tCr in SCA3 pons, (E) tNAA in SCA3 pons, and (F) tNAA in SCA7 pons. Metabolites that showed Dunnett significance when each SCA type was compared to controls were included in the correlation analysis. p values of the correlations have been corrected for multiple testing with step-down Bonferroni method.

Figure 3

Principal component analysis of metabolites of interest and disease characteristics of patients with SCAs. The first two components accounted for 57.4% and 64% variation in the vermis and the pons respectively. PCA was able to separate the neuronal markers – tNAA and Glu – from the energetic marker (tCr), the glial marker (myo-Ins) and the SARA score.

Figure 4

Separation between patients with SCAs and controls by plotting the concentrations of neurochemicals against each other. The concentrations of metabolites that showed significant differences in patients – tNAA, tCr, myo-Ins and Glu – were plotted against each other to determine the ratio that could separate subjects into patient and control groups, with almost no overlap. SCA2 has the best separation with at most one dataset overlapping with controls.