Abstract
412
Objectives: 11C-UCB-J is a specific PET ligand for synaptic vesicle glycoprotein 2A (SV2A) and has been proposed as a potential biomarker for synaptic density (1-3). Previously, we have found significant reduction (~ 40%) of hippocampal SV2A binding, indicating synaptic density loss, in Alzheimer’s disease (AD) participants compared to age-matched cognitively normal (CN) elderly participants (4). The goal of this study is to investigate the relationship between synaptic density measured by 11C-UCB-J PET and neuronal function measured by FDG PET in the same AD and CN participants.
Methods: Five AD and 6 CN participants were enrolled for both 11C-UCB-J and FDG PET scans with the HRRT. AD participants were all confirmed Aβ+ by 11C-PiB PET and spanned the disease stages from amnestic mild cognitive impairment (MCI, n=3) to mild dementia (n=2). CN participants were all confirmed Aβ-. For 11C-UCB-J PET, arterial samples were used to measure the metabolite-corrected input function. VT parametric images were produced with the one-tissue compartment model. Multiple regions of interest (ROIs) were taken from the AAL template and individual MR images: caudate, cerebellum (CB), anterior cingulate cortex, posterior cingulate cortex, frontal cortex, hippocampus, occipital cortex, parietal cortex, pulvinar nuclei, putamen, temporal cortex, and thalamus. BPND was calculated with the centrum semiovale (CS) as reference region. For FDG PET, influx constant (Ki) parametric images were produced using the Patlak plot with a population input function scaled by individual venous blood samples. The same ROIs were applied to the FDG Ki images. Comparison of Ki and Ki ratios (normalized to CB) in the ROIs between AD and CN groups was performed with two-tailed, unpaired student’s t-test with P<0.05 for significance. Additional analyses examined the relationship between FDG (Ki or Ki ratios) and SV2A (K1 and BPND) with Pearson correlation coefficients, two-tailed, with P<0.05 for significance.
Results: In this small sample, there was no statistically significant difference of regional FDG Ki between AD and CN. Consistent with known pattern of hypometabolism in AD, there were significantly lower FDG Ki ratios (normalized to CB) in the hippocampus (26%, P=0.007) and posterior cingulate cortex (24%, P=0.005) as well as in the temporal, frontal, and parietal cortices (P<0.03) in AD compared to CN. There was overall good correlation (R> 0.6 with P<0.05, n=11) between examined FDG and SV2A parameters in the regions of hippocampus and posterior cingulate cortex. These correlations included hippocampal FDG Ki and SV2A K1 (R=0.83, P=0.002), hippocampal FDG Ki ratio and SV2A BPND (R=0.84, P=0.001), FDG Ki ratio and SV2A BPND in the posterior cingulate cortex (R=0.79, P=0.004). Statistically significant correlation between FDG Ki ratio and SV2A BPND was also noted in the regions of caudate (R=0.71, P=0.014) and pulvinar nuclei(R=0.61 and P=0.047). There was no statistically significant correlation between FDG Ki ratio and SV2A BPND found in the other ROIs (R<0.6 and P>0.05). Similar results were found in Ki ratios normalized to CS. Conclusions: There was overall good correlation between neuronal metabolism assessed by FDG PET and synaptic density by SV2A PET in the hippocampus, posterior cingulate cortex, caudate and pulvinar nuclei in this pilot study. It is noteworthy that these regions all receive extensive projections from temporal, parietal, and/or prefrontal association cortices known to be hypometabolic in AD. Further exploration in other projecting brain regions in a large scale cohort study is needed to better elucidate these relationships and their significance in AD pathophysiology. Funding: The Dana Foundation, Pilot grant from Yale ADRC NIH P50AG047270, and NIH R01AG52560-01A1 References: 1. Nabulsi et al., 2016, JNM, 57:5:777-84. 2. Finnema et al., 2016, Sci Transl Med, 8:348ra96. 3. Finnema et al., 2017, JCBFM, epub. 4. Chen et al., abstracts for CTAD 2017 and HAI 2018