Early brainstem [18F]THK5351 uptake is linked to cortical hyperexcitability in healthy aging

Abstract

BACKGROUNDNeuronal hyperexcitability characterizes the early stages of Alzheimer's disease (AD). In animals, early misfolded tau and amyloid-β (Aβ) protein accumulation - both central to AD neuropathology - promote cortical excitability and neuronal network dysfunction. In healthy humans, misfolded tau and Aβ aggregates are first detected, respectively, in the brainstem and frontomedial and temporobasal cortices, decades prior to the onset of AD cognitive symptoms. Whether cortical excitability is related to early brainstem tau - and its associated neuroinflammation - and cortical Aβ aggregations remains unknown.METHODSWe probed frontal cortex excitability, using transcranial magnetic stimulation combined with electroencephalography, in a sample of 64 healthy late-middle-aged individuals (50-69 years; 45 women and 19 men). We assessed whole-brain [18F]THK5351 PET uptake as a proxy measure of tau/neuroinflammation, and we assessed whole-brain Aβ burden with [18F]Flutemetamol or [18F]Florbetapir radiotracers.RESULTSWe found that higher [18F]THK5351 uptake in a brainstem monoaminergic compartment was associated with increased cortical excitability (r = 0.29, P = 0.02). By contrast, [18F]THK5351 PET signal in the hippocampal formation, although strongly correlated with brainstem signal in whole-brain voxel-based quantification analyses (P value corrected for family-wise error [PFWE-corrected] < 0.001), was not significantly associated with cortical excitability (r = 0.14, P = 0.25). Importantly, no significant association was found between early Aβ cortical deposits and cortical excitability (r = -0.20, P = 0.11).CONCLUSIONThese findings reveal potential brain substrates for increased cortical excitability in preclinical AD and may constitute functional in vivo correlates of early brainstem tau accumulation and neuroinflammation in humans.TRIAL REGISTRATIONEudraCT 2016-001436-35.FUNDINGF.R.S.-FNRS Belgium, Wallonie-Bruxelles International, ULiège, Fondation Simone et Pierre Clerdent, European Regional Development Fund.

Keywords: Aging; Alzheimer’s disease; Neuroimaging; Neuroscience.

Figure 1. Cortical excitability assessment and PET value extraction in early deposition sites.

(A) Cortical excitability over the frontal cortex was assessed using neuronavigation-based TMS coupled to EEG. TMS-EEG target area was located in the superior frontal gyrus. Darker and brighter yellow areas represent the range of stimulation targets across participants projected onto the averaged normalized T1 structural volume and the median TMS-EEG stimulation hotspot over the sample, respectively. (B) Butterfly plot of TMS-evoked EEG response over the 60 electrodes (–100 ms before TMS to 300 ms after TMS; average of approximately 250 trials). Cortical excitability was computed as the slope (μV/ms, dotted line on inset) of the first component of the TEP response at the electrode closest to the stimulation hotspot. (C) Automatic brainstem segmentation methods were used to extract ^[18F]THK5351 SUVR in the brainstem monoaminergic gray matter (bmGM; top row). Aβ burden (^[18F]Flutemetamol or ^[18F]Florbetapir centiloid units) was extracted in the earliest aggregation sites (7) using bilateral medial superior frontal, inferior temporal, and fusiform regions (bottom row).

Figure 2. Associations between cortical excitability and early ^[18F]THK5351 PET signal, as well as Aβ burden.

(A) Significant and positive association between cortical excitability values and mean ^[18F]THK5351 standardized uptake value ratio (SUVR) in the brainstem monoaminergic gray matter (bmGM; Pearson’s correlation: r = 0.29, P = 0.02; GLMM: F_1,57 = 4.76, P = 0.03, R²_β* = 0.08). (B) No significant association between cortical excitability and mean ^[18F]Flutemetamol/^[18F]Florbetapir (centiloid units) in an ROI covering the earliest Aβ aggregation sites (Pearson’s correlation: r = –0.20; P = 0.11; GLMM: F_1,57 = 1.41, P = 0.24). Simple regressions are displayed, and full GLMM outputs are reported in Table 2. Dotted lines represent 95% CI for these simple regressions. Inset images illustrate the ROI used to extract PET values in each analysis.

Figure 3. ^[18F]THK5351 regional correlates of brainstem ^[18F]THK5351 uptake and association between cortical excitability and mean MTL ^[18F]THK5351 SUVR.

(A) VBQ analyses revealed that mean ^[18F]THK5351 SUVR in the brainstem monoaminergic gray matter (bmGM) region was positively correlated to gray matter ^[18F]THK5351 PET signal within most of the medial temporal lobe (MTL, P_{FWE-corrected} < 0.001). (B) VBQ analyses further showed that mean ^[18F]THK5351 SUVR in the bmGM region was positively associated with ^[18F]THK5351 SUVR in the white matter tracts connecting the brainstem to the MTL, but also in widespread projections toward the whole cortex (P_{FWE-corrected} < 0.001). (C) Paired t test analysis showed that mean ^[18F]THK5351 SUVR in the bmGM was significantly lower than in the MTL (t = 10.14, P < 0.0001). (D) Average ^[18F]THK5351 SUVR in the MTL was, however, not significantly associated with cortical excitability values (Pearson’s correlation: r = 0.14, P = 0.25; GLMM: F_1,57 = 1.89, P = 0.17). Simple regression is displayed, and full GLMM outputs are reported in Table 4. Dotted lines represent 95% CI for this simple regression.