3223
Diffuse white matter disease revealed in novel rat model of cerebral amyloid angiopathy type 1 by multi-modality MRI at 9.4T.1Department of Anesthesiology, Yale University, New Haven, CT, United States, 2Department of Biomedical and Pharmaceutical Sciences, University of Rhode Island, Kingston, RI, United States
Synopsis
A novel cerebral amyloid angiopathy type 1 rat model which robustly develops microvascular amyloid beta deposits is studied. We report that this animal model develops white matter atrophy, cerebral micro-bleed, and loss of white matter integrity. These findings are corroborated by immunohistochemistry showing axonal disruption and vacuolization.
Introduction
The majority of people suffering from dementia are reported to have a cerebral small vessel disease (SVD) component and currently no effective treatment exists. SVDs are a group of disorders resulting from pathological alteration of the small blood vessels and the most common ones are cerebral amyloid angiopathy (CAA) and hypertensive SVD. Significant challenges exist in both prognosis and diagnosis due to the overlapping SVD subtype pathologies and the multitude of MRI diagnostic hallmarks including white matter disease, lacunes, cerebral microbleeds (CMB), and dilated perivascular spaces [1]. A major obstacle in dissecting the biological mechanism(s) and contributing pathology to SVD is related to a lack of animal models which faithfully reflect human clinical disease progression. We recently created a novel CAA type 1 rat model [2] which robustly develops microvascular amyloid beta (Ab) deposits and CMBs. Here we report on MR image acquisition and analysis strategies to characterize white matter (WM) changes as well as CMBs in the novel CAA type 1 rat model using multi-modality MRI to evaluate the extent its resemblance to clinical CAA-SVD.Method
All imaging acquisitions were performed on a Bruker 9.4T/16 MRI equipped with a volume transmit/receive or surface receive only coils interfaced with Paravision 6. In-vivo scans: 11 months old CAA rats (N=11) and wild-type rats (WT N=11) were scanned while anesthetized with dexmedetomidine supplemented with isoflurane [3]. For voxel based morphometry (VBM), a 3D proton density weighted (PDW) based on SPGR sequence was implemented (TR/TE/FA=50ms/4ms/7° 0.23x0.23x0.23mm) and analyzed via custom tissue probability maps using SPM12 [4]. For CMB detection, a 3D MGE technique was implemented (TR/TE/FA=60ms/2°~32°/15 0.23x0.23x0.23mm) for calculating 3D T2* map which was spatially normalized using the warping parameters derived from the VBM analysis. In-vitro scan: 11 months old CAA rats (N=11) and WT rats (N=15) were transcardially perfusion fixed using active MR staining [5] and skull intact brains were scanned in-vitro using a 3D PGSE-DTI technique (TR/TE=300ms/22ms 0.16x0.16x0.16mm 6 directions). Non-diffusion weighted T2W (b=0) images were utilized to derive the spatial normalization parameters and applied onto FA and MD maps for voxel-wise statistical analyses using SPM12. Histology: Immunohistochemistry (IHC) was performed to characterize WM pathology in corpus callosum. Antibodies to pan-axonal and synaptophysin were used to identify axons and synaptic vesicles, respectively, and DAPI was used to label cell nuclei.Results
Focal hypointense signals in thalamus and cortex were discernable on the population averaged in-vivo T2* maps from CAA rats when compared to WT rats (Fig. 1a) representative of CMBs [1]. The morphometric analysis revealed significant total WM volume loss (567±20 mm3 vs 493±21 mm3 P<0.001) in CAA rats compared to WT rats even after adjusting for the total intracranial volume (0.300±0.007 vs 0.270±0.004 P<0.001) when compared to WT. Topographically, the distribution of WM volume loss in CAA rats (when compared to WT rats) was observed in the brain stem, fimbria and corpus callosum (CC) (Fig. 1b), which are areas not affected by CMBs. The whole brain WM fractional anisotropy (FA) was significantly lower in CAA rats compared to WT rats (0.41±0.01 vs 0.39±0.01 P<0.001) and the voxel-wise analysis revealed significantly reduced FAs in several WM areas including CC, fimbria and thalamus, which in large part coincided with areas of WM volume loss (Fig. 1c). WM degeneration in CC among CAA was further corroborated by IHC which revealed loss of coherence in axonal orientation, fragmentation, vacuolization and loss of synaptophysin compromising network of neuronal processes,as shown in Fig.2.Conclusions
The novel CAA type 1 rat model exhibit some of clinical imaging features of SVD including loss of WM volume and integrity and CMB [6]. The MR data acquisitions and voxel-wise image analyses developed in this study were sufficiently sensitive to reveal the distribution of CMBs and diffuse WM disease in a novel CAA type 1 rat model demonstrating its relevance to clinical SVD-CAA. 3D PDW SPGR sequence is a time-efficient alternative to the conventional T2W spin echo and attains good signal to noise ratio for VBM. The voxel-wise analysis of the T2* maps revealed clusters of CMBs in regions with confirmed high levels of microvascular Ab deposits known to enhance the vulnerability to CMB [2]. Areas with FA reductions were also globally diffuse and coincided with WM loss as detected by VBM hence confirming that both microstructural and morphological changes occur simultaneously in 11 months old CAA type 1 rats. Histology revealed extensive axonal disruption, vacuolization and loss of axons in CC of CAA rats. However, WM histopathology did not always co-localize with CMBs, suggesting that WM loss may not be directly associated with CMBs in CAA and other pathophysiological mechanisms may be responsible for the WM loss.
Acknowledgements
The present work was supported by National Institutes of Health RF-AG053991, RF-AG057705, R01-NS100366, and Foundation Leducq Transatlantic Network of Excellence (16/CVD/05).References
1. Wardlaw JM, et.al. Lancet Neurol. 2013 Aug;12(8):822-38.
2. Davis J, et.al. The American Journal of Pathology, In Press (2018)
3. Benveniste H, et.al. Anesthesiology. 2017 Dec;127(6)
4. Ashburner J, et.al. Neuroimage. 2005 Jul 1;26(3):8
5. Johnson GA, et.al. Neuroimage. 2012 Sep;62(3):1848-56.6. Ter Telgte A, et al. Nat Rev Neurol. 2018 Jul;14(7):387-398
Figures
Population averaged (a) T2*, (b) WM volumes, and (c) FA maps of 11 months old WT and CAA. Results two sample t-tests (WT>CAA) are shown in colormaps, representing p-values, which are superimposed onto the representative anatomical scan.
