Imaging and quantifying microglial activation in vivo and ex vivo using diffusion MRI – with validation by immunohistochemistry

Abstract

AbstractBackgroundAlzheimer disease (AD) is the most common type of dementia.There is growing evidence that the amyloid plaques in postmortem AD human brains are surrounded by activated microglia. Using diffusion basis spectrum imaging (DBSI), we have previously observed increased cellular diffusivity in AD patients, suggesting increased microglia activation as a key pathology in AD. To better understand the underlying mechanism and further establish the noninvasive and nonradioactive DBSI biomarkers of microglia activation, we conducted a realistic Monte Carlo (MC) simulation to model diffusion MRI signals associated with microglia in various activation states.MethodFirst, realistic microglia diffusion signals were generated using MC simulation with real 3D microglia geometry obtained from neuromorpho.org (Fig. 1A). We simulated different pathophysiological scenarios to validate DBSI’s capability in quantitatively detecting microglia activation (Fig. 1B). Second, we conducted and ex vivo MRI scans on human postmortem AD brain samples, followed by immunohistochemical (IHC) staining of microglia (Iba‐1) in aligned sections from the same tissue. DBSI cell apparent diffusion coefficients (ADC), reflecting microglia activation, were estimated and compared to quantitative histological assessments of microglia activation. Finally, in vivo DBSI was conducted on 200 participants who were classified into healthy control, preclinical AD, and early symptomatic AD groups based on cognitive assessments and measures of CSF amyloid (Aβ42) and pTau181.ResultTwo exemplary regions in grey matter (GM) and white matter (WM) of a postmortem AD brain were selected (Fig. 1C). We found that GM had significantly higher DBSI‐derived cell ADC than WM (Fig. 1D). Iba‐1 IHC revealed that WM had significantly larger average microglial area (p<0.005) and bigger max microglia larea (Fig. 1E‐F). MC Simulation was conducted on 700 microglia models. Microglial activation would result in an increased DBSI cell ADC (Fig. 1G). According to MC simulation, increased DBSI‐derived cell ADC in preclinical and early symptomatic AD indicated microglial activation in preclinical and early symptomatic AD (Fig. 1H).ConclusionDBSI‐derived cell ADC increases with microglial activation as assessed by quantitative Iba‐1 IHC. This finding is supported by MC simulation using real 3D microglia geometry. The study indicates that DBSI has the potential to provide non‐invasive MRI biomarkers of microglia activation.