Myo-inositol as a spectroscopic marker of neuroinflammation in frontotemporal lobar degeneration

Background and ObjectivesCost-effective and mechanistically-informative biomarkers are needed to advance novel therapeutics for dementia. We focus on neuroinflammation in disorders associated with frontotemporal lobar degeneration (FTLD). Commonly used assays include positron emission tomography (PET) imaging and peripheral blood markers. However, PET is limited by radioactivity, which impedes repeated measurements, while peripheral markers are indirect measures of neuropathology. This study investigates the potential of proton magnetic resonance spectroscopy (MRS) - a non-invasive technique that can be deployed alongside structural imaging sequences - to quantify myo-inositol, an inflammatory metabolite found in astroglial and microglial cells. We test the following hypotheses (1) myo-inositol is elevated in brain regions affected by FTLD disorders; (2) myo-inositol in affected brain regions correlates with peripheral immune markers; and (3) myo-inositol in affected brain regions correlates with cognitive decline. Methods54 people with FTLD-associated disorders (29 behavioural variant frontotemporal dementia, 25 progressive supranuclear palsy) and 73 age/sex-matched controls underwent single-voxel 7 Tesla MRS. 2cm isotropic voxels were placed over the right inferior frontal gyrus (R IFG, a disease-affected region of interest), and right occipital lobe (R OCC, a control region). Serum inflammatory markers were measured using nucleic acid-linked immuno-sandwich assays (NULISA). Cognition was assessed with the revised Addenbrookes Cognitive Examination (ACE-R). Mixed-effects and general linear models were used to evaluate diagnostic differences in regional myo-inositol and its association with cognition. Partial least squares regression identified two serum inflammatory components that covaried with myo-inositol; these components were subsequently tested in confounder-adjusted second-level linear models. ResultsMyo-inositol was elevated in the R IFG ({beta} = 0.20, t = 5.10, p = 8.8 x 10^-7) but not in the R OCC ({beta} = 0.05, t = 1.36, p = 0.18), in people with FTLD versus controls. In FTLD, myo-inositol in the R IFG was positively associated with a peripheral inflammatory profile characterised by IL-6, IL-1{beta}, GFAP, CSF2, CRP, and CCL22 ({beta} = 0.39, t = 5.12, p = 3.60 x 10^-6). R IFG myo-inositol was negatively associated with ACE-R performance in patients ({beta} = -44.59, t = -4.12, p = 0.00016). ConclusionMRS-derived myo-inositol is a promising neuroimaging biomarker of inflammation in FTLD. Elevated levels in disease-affected regions were associated with peripheral markers of inflammation (reflecting astrocytosis and microgliosis) and with cognitive impairment.