Longitudinal evaluation of proton magnetic resonance spectroscopy metabolites as biomarkers in Huntington’s disease

Abstract

Proton Magnetic resonance spectroscopy (1H-MRS) is a non-invasive method of exploring cerebral metabolism. In Huntington’s disease, altered 1H-MRS-determined concentrations of several metabolites have been described; however, findings are often discrepant and longitudinal studies are lacking. 1H-MRS metabolites may represent a source of biomarkers, thus their relationship with established markers of disease progression require further exploration to assess prognostic value and elucidate pathways associated with neurodegeneration. In a prospective single-site controlled cohort study with standardised collection of CSF, blood, phenotypic and volumetric imaging data, we used 3T 1H-MRS in conjunction with the linear combination of model spectra method to quantify seven metabolites (total n-acetylaspartate, total creatine, total choline, myo-inositol, GABA, glutamate and glutathione) in the putamen of 59 participants at baseline (15 healthy controls, 15 premanifest and 29 manifest Huntington’s disease gene expansion carriers) and 48 participants at 2-year follow-up (12 healthy controls, 13 premanifest and 23 manifest Huntington’s disease gene expansion carriers). Intergroup differences in concentration and associations with CSF and plasma biomarkers; including neurofilament light chain and mutant Huntingtin, volumetric imaging markers; namely whole brain, caudate, grey matter and white matter volume, measures of disease progression and cognitive decline, were assessed cross-sectionally using generalized linear models and partial correlation. We report no significant groupwise differences in metabolite concentration at baseline but found total creatine and total n-acetylaspartate to be significantly reduced in manifest compared with premanifest participants at follow-up. Additionally, total creatine and myo-inositol displayed significant associations with reduced caudate volume across both time points in gene expansion carriers. Although relationships were observed between 1H-MRS metabolites and biofluid measures, these were not consistent across time points. To further assess prognostic value, we examined whether baseline 1H-MRS values, or rate of change, predicted subsequent change in established measures of disease progression. Several associations were found but were inconsistent across known indicators of disease progression. Finally, longitudinal mixed effects models revealed glutamine + glutamate to display a slow linear decrease over time in gene expansion carriers. Altogether, our findings show some evidence of reduced total n-acetylaspartate and total creatine as the disease progresses and cross-sectional associations between select metabolites, namely total creatine and myo-inositol, and markers of disease progression, potentially highlighting the proposed roles of neuroinflammation and metabolic dysfunction in disease pathogenesis. However, the absence of consistent group differences, inconsistency between baseline and follow-up, and lack of clear longitudinal change suggests that 1H-MRS metabolites have limited potential as Huntington’s disease biomarkers.

Longitudinal study of neurochemical, volumetric and behavioral changes in Q140 & BACHD mouse models of Huntington’s disease

“…A machine learning algorithm based upon logistic regression of multiple measures from longitudinal human HD studies can predict risk of progression for participants entering future clinical trials, reducing sample size and expected outcome variability (Koval et al, 2022). Alignment of disease severity by ‘Huntington Age’ (Koval et al, 2022) might reduce some of the variability among both human and mouse HD 1 H MRS studies (Lowe et al, 2022). Human 1 H MRS would also benefit from determination by optimal protocols in higher field strength magnets, as used in mice studies.…”