Spatio-Temporal Characterization of Brain Inflammation in a Non-human Primate Stroke Model Mimicking Endovascular Thrombectomy

Abstract

Reperfusion therapies in acute ischemic stroke have demonstrated their efficacy in promoting clinical recovery. However, ischemia/reperfusion injury and related inflammation remain a major challenge in patient clinical management. We evaluated the spatio-temporal evolution of inflammation using sequential clinical [¹¹C]PK11195 PET-MRI in a non-human primate (NHP) stroke model mimicking endovascular thrombectomy (EVT) with a neuroprotective cyclosporine A (CsA) treatment. The NHP underwent a 110-min transient endovascular middle cerebral artery occlusion. We acquired [¹¹C]PK11195 dynamic PET-MR imaging at baseline, 7 and 30 days after intervention. Individual voxel-wise analysis was performed thanks to a baseline scan database. We quantified [¹¹C]PK11195 in anatomical regions and in lesioned areas defined on per-occlusion MR diffusion-weighted imaging and perfusion [¹⁵O₂]H₂OPET imaging. [¹¹C]PK11195 parametric maps showed a clear uptake overlapping the lesion core at D7, which further increased at D30. Voxel-wise analysis identified individuals with significant inflammation at D30, with voxels located within the most severe diffusion reduction area during occlusion, mainly in the putamen. The quantitative analysis revealed that thalamic inflammation lasted until D30 and was significantly reduced in the CsA-treated group compared to the placebo. In conclusion, we showed that chronic inflammation matched ADC decrease at occlusion time, a region exposed to an initial burst of damage-associated molecular patterns, in an NHP stroke model mimicking EVT. We described secondary thalamic inflammation and the protective effect of CsA in this region. We propose that major ADC drop in the putamen during occlusion may identify individuals who could benefit from early personalized treatment targeting inflammation.

Keywords: Endovascular thrombectomy; Inflammation; PET-MRI; Stroke; TSPO.

Fig. 1

Example of individual dataset from each group of the study. From top to bottom: diffusion-weighted imaging per-occlusion and [¹⁵O]H₂O perfusion PET per-occlusion; fluid attenuation inversion recovery MRI at day 7 and overlay of [¹¹C]PK11195 DVR; fluid attenuation inversion recovery MRI at day 30 and overlay of [¹¹C]PK11195 DVR

Fig. 2

A SPMs showing the spatial distribution of significant voxels (p < 0.05 corrected) where [¹¹C]PK11195 DVR increase after MCAO in NHP compared to the baseline database. The first column represents the corresponding [¹¹C]PK11195 DVR map at day 30 (see color scale on the left), the second one the individual FLAIR post-gadolinium imaging at day 30. Clusters of significant voxels are superimposed on corresponding individual FLAIR post-gadolinium imaging at day 30 (third column). Z score in color scale on the right. B ADC maps during occlusion of the four individuals identified by persistent inflammation. The first column represents the ADC maps at occlusion (with the corresponding color scale). ADC maps are superimposed on M. fascicularis atlas in the second column. The third column displayed the localization of SPM identified clusters of [¹¹C]PK11195 DVR increase at D30 compared to baseline database

Fig. 3

Quantification in lesion core and penumbra. Individual values of absolute differences in [.¹¹C]PK11195 DVR. A Differences between the baseline and day 7 (Δ DVR_{D7−baseline}). B Differences between the baseline and day 30 (Δ DVR_{D30−baseline}). Placebo n = 5, CsA n = 6. Non-parametric Mann–Whitney test, * p < .05

Fig. 4

Quantification in the thalamus. Individual values of absolute differences in [¹¹C]PK11195 DVR. A Differences between the baseline and day 7 (Δ DVR_{D7−baseline}). B Differences between the baseline and day 30 (Δ DVR_{D30−baseline}). Placebo n = 5, CsA n = 6. Non-parametric Mann–Whitney test, ** p < .05

Fig. 5

Post-mortem immunohistochemistry of corresponding in vivo PET-MRI (fluid attenuation inversion recovery MRI at day 30 and overlay of [¹¹C]PK11195 DVR) (NHP #9). Tissue sections were double immunostained with CD68 for phagocytic cells (white arrow head) and TSPO for inflammation response (yellow arrow head). Tissue sections were immunostained with Iba1 for microglial response (red arrow head) or GFAP for reactive astrocytes (pink arrow head). All sections were counterstained with 4, 6-diamidino-2-phenylindole (DAPI, blue) for cell nuclei. 1: peri-lesion area. 2: thalamus. Scale bar: 50 m