Cerebral amyloid angiopathy (CAA) is a common cerebral small vessel disease that involves the accumulation of amyloid β-protein (Aβ) primarily in small- and medium-sized arteries and arterioles of the meninges and cerebral cortex as well as along the capillaries of the cerebral microvasculature [1
]. CAA is found, to varying degrees, in nearly 80% of elderly individuals [5
]. Since CAA results from cerebral vascular deposition of Aβ, it is not surprising that this condition commonly coexists in patients with Alzheimer’s disease (AD) [1
]. However, independent of AD, clinically CAA is a significant contributor to vascular-mediated cognitive impairment and dementia (VCID) [3
]. As a prominent small vessel disease CAA contributes to the cognitive decline in VCID in several ways by promoting perivascular neuroinflammation, impaired cerebral blood flow and ischemic infarcts, cerebral microbleeds, and larger hemorrhages, all of which can result in neuronal dysfunction, neuronal loss and white matter damage [3
Presently, the clinical diagnosis of CAA primarily relies on the detection of cerebral microbleeds and larger hemorrhages by imaging techniques [10
]. Previously, a set of criteria was established, known as the ‘Boston Criteria’, that determines a diagnosis of probable CAA based on the presence and anatomical location of cerebral microbleeds [11
]. An updated version of the “Boston Criteria” was proposed incorporating additional key imaging biomarkers that detect other cerebral vascular injuries associated with amyloid containing vessels including cortical superficial siderosis, enlarged perivascular spaces and white matter hyperintensities [12
]. Also, anatomical distribution patterns of amyloid accumulation in whole brain detected by positron emission tomography (PET) and Aβ specific radiotracer imaging studies has been informative [15
]. Although these additional imaging markers have improved the clinical diagnosis of probable CAA, they largely represent signs of late-stage disease that emerge after the extensive accumulation of cerebral vascular amyloid has become chronic. Markers that detect the earlier stages of CAA, prior to these late-stage imaging indications, would be beneficial to monitor disease progression in patients while at the same time avoiding therapeutic interventions which might trigger cerebral hemorrhage such as thrombolytic agents [17
]. Alternatively, biomarkers of early onset CAA disease would help to identify patients that could be enrolled in novel therapeutically directed clinical trials.
Recently, we generated a novel transgenic rat model of CAA designated rTg-DI that faithfully recapitulates many features of human capillary CAA type-1 [19
]. The rTg-DI rats express human AβPP in neurons harboring the familial CAA Dutch E693Q/IowaD694N mutations and produce chimeric Dutch/Iowa mutant Aβ in brain. This model exhibits early-onset and progressive microvascular/capillary amyloid accumulation in many forebrain brain regions including the cortex, hippocampus and thalamus that is largely composed of the Aβ40 peptide. However, there is little accumulation of fibrillar amyloid in larger cerebral vessels in this model. The deposition of cerebral microvascular amyloid is accompanied by robust perivascular neuroinflammation and associated with regular microbleeds [19
]. Accordingly, rTg-DI rats provide an invaluable preclinical platform to follow development of CAA pathologies.
In the present study, we show that rTg-DI rats exhibit consistent and progressive accumulation of primarily Aβ40 peptide in the form of microvascular CAA throughout the cortex, hippocampus and thalamus starting at about three months of age. In addition, rTg-DI rats develop consistent cerebral microbleeds that are readily detected by magnetic resonance imaging (MRI) beginning at six months of age with expansion to twelve months of age and confirmed by histological analysis of perivascular hemosiderin deposits at these ages. Measurements of cerebrospinal fluid (CSF) Aβ40 in one month old rTg-DI rats, that do not yet exhibit CAA, represent the homeostatic levels of this peptide prior to disease onset. However, at three months of age, when rTg-DI rats start exhibiting CAA but do not yet show microbleeds, the CSF levels of Aβ40 precipitously and uniformly dropped. At six months of age and later, with further accumulation of CAA and emergence of microbleeds in the brain, CSF levels of Aβ40 continued to drop. Parallel analysis of rTg-DI plasma Aβ40 showed similar trends with lower levels at the onset of CAA at three months and further reductions at six to twelve months. These findings strongly indicate that a reduction in Aβ40 levels in biological fluids represent an early disease related biomarker that correlates with the onset of CAA prior to the emergence of subsequent cerebral microbleeds that are detected by neuroimaging and histological confirmation. Finally, this study underscores the translational utility of rTg-DI rats as a valid preclinical model to further develop biomarkers and a platform to test therapeutic interventions for CAA.
CAA is a common cerebral small vessel disease of the elderly and a prominent comorbidity of AD that promotes and exacerbates VCID, yet our ability to diagnose CAA remains limited to late-stage neuroimaging markers for this condition. The neuroimaging markers used in the ‘Boston Criteria’ diagnosis is centered around vascular and perivascular changes that are associated with cerebral blood vessels impacted by the chronic and extensive accumulation of cerebral vascular amyloid. The major scoring factors for clinical diagnosis of CAA is the lobar presence of a large macrobleed, which occurs in very severe cases of disease, or the presence of multiple lobar microbleeds, which is more common and observed in chronic, late-stage disease with extensive cerebral vascular amyloid [11
]. Additional neuroimaging markers that aid clinical diagnosis include perivascular changes that occur in the vicinity around CAA-affected vessels. For example, cortical superficial siderosis indicates perivascular iron accumulation in pial arteries and arterioles reflecting cerebral microbleeds [12
]. Another noted perivascular change associated with CAA is the presence of dilated perivascular spaces, particularly around penetrating arterioles with amyloid deposition [8
]. Although the cause of this perivascular alteration remains unclear this space contains CSF and interstitial fluid (ISF) that under normal conditions plays an important role in Aβ clearance from brain [25
]. The presence of amyloid in these vessels promotes local perivascular inflammation and may disrupt normal CSF flow and exchange with ISF leading to enlargement of these spaces [4
]. Although these neuroimaging markers are useful in diagnosing chronic, late-stage disease with extensive CAA, they are not helpful in detecting early stages of emerging cerebral vascular amyloid accumulation that could be more amenable to therapeutic interventional strategies and allow for clearance of the emerging perivascular Aβ deposits.
Parenchymal amyloid plaques commonly found in AD primarily contain the longer Aβ42 species. Previous studies have shown that as brain parenchymal plaques accumulate in AD there is a significant decrease in the levels of Aβ42 detected in CSF as this peptide accumulates in brain [30
]. Accordingly, CSF Aβ42 levels have provided a surrogate biomarker for the parenchymal plaque burden in brain and have assisted in the clinical diagnosis of AD [32
]. In contrast to plaques, CAA deposits are largely composed of the shorter Aβ40 peptide [1
]. This suggests that a decrease in CSF Aβ40 levels could reflect the presence and burden of CAA. Indeed, studies have shown that patients diagnosed with CAA, based on late-stage neuroimaging biomarkers for cerebral microbleeds and cortical superficial siderosis, presented with lower levels of CSF Aβ40 [20
]. However, the utility of CSF Aβ40 levels to serve as a potential biomarker for earlier stage CAA, prior to the onset of microbleeds and other perivascular changes, remains unclear.
Most animal models to study CAA have largely involved the use of transgenic mouse lines that express human AβPP generally producing highly elevated levels of Aβ peptides in brain with or without familial CAA mutations [35
]. Although these mouse models have been helpful in studying the pathogenesis of CAA their usefulness in modeling the human disease has been met with limitations including the small size of the mouse brain, which has hampered neuroimaging studies, and the variable presentation of cerebral microbleeds. Our recently generated transgenic rat model rTg-DI more faithfully recapitulates many of the pathological features of human small vessel CAA including early-onset and progressive accumulation of cerebral microvascular fibrillar amyloid, perivascular neuroinflammation, consistent and robust development of cerebral microhemorrhages and small vessel occlusions and behavioral deficits [19
]. The rTg-DI rats used in this study provided a unique opportunity to evaluate the trajectories of Aβ40 in CSF and plasma in cohorts of animals that consistently progressed from the early-onset to late-stage disease pathologies. Indeed, from the accumulation of Aβ peptides in brain, to the development of CAA, to the emergence and expansion of microbleeds and to measurements of CSF/plasma Aβ40 the cohorts rTg-DI rats progressed through each stage of disease with uniformity.
The CSF compartment is an important exchange reservoir with the ISF compartment that surrounds the cellular components of the brain. This dynamic interaction plays an important clearance route for Aβ and other cellular metabolic waste products through glymphatic system transport and/or alternate intramural perivascular drainage pathways [25
]. Thus, the CSF compartment can reflect the ongoing clearance of Aβ and provide a snapshot of Aβ pathology and burden in the brain. For example, the chronic pathological accumulation of Aβ42 in parenchymal plaques in AD brain is reflected by decreased levels of this specific Aβ isoform in CSF and is associated with a decline in cognitive function [32
]. Similarly, our findings demonstrate that the selective accumulation of Aβ40 in CAA deposits is reflected by early onset decreased levels of this Aβ species in CSF.
Another important route for Aβ clearance from the CNS is across the blood-brain barrier of the cerebral endothelium into blood for peripheral removal [39
]. This clearance route is facilitated by endothelial Aβ transporters including low density lipoprotein receptor related protein 1 (LRP1) and p-glycoprotein [39
]. However, the levels of Aβ peptides in plasma are much lower than in CSF and can be more variable and, therefore, is less reliable in serving as a biomarker for AD [43
]. In the case of using rTg-DI rats these shortcomings are further compounded by the finding that compared with non-mutated wild-type Aβ the chimeric Dutch/Iowa CAA mutant Aβ poorly binds LRP1 and is much less effectively transported across the blood-brain barrier into the circulation [45
]. Indeed, the level of Aβ40 detected in plasma of rTg-DI rats was < 2% of what was measured in the CSF. Nevertheless, we still found that Aβ40 levels in plasma of rTg-DI rats dramatically drop with the onset of cerebral microvascular amyloid accumulation, thus mirroring what was observed in the CSF.
We recently reported that in CAA deposits in transgenic mice and humans, as well as in rTg-DI rats, the amyloid fibrils adopt a distinct anti-parallel configuration [19
]. The present findings suggest that once Aβ40 forms fibrillar amyloid deposits in cerebral microvessels and capillaries, this may act as a nidus for further seeding of additional soluble Aβ40 to expand vascular fibrillar amyloid with this distinct anti-parallel conformation and further impair its clearance through either the CSF or into the peripheral circulation.
The trajectory of disease markers in the rTg-DI model of microvascular CAA are summarized in Figure 7
. In rTg-DI rats the accumulation of Aβ peptides in brain and the onset of CAA began early, at about two to three months of age, and dramatically increased over twelve months. The emergence of cerebral microbleeds, confirmed by in vivo
MR imaging and histological evaluation, occur significantly later starting at around six months, with further expansion at nine and twelve months. The present findings show that at the onset of CAA, much earlier than the emergence of cerebral microbleeds, there is a precipitous drop in the CSF/plasma levels of Aβ40, the chief component of CAA deposits. The reductions in CSF Aβ40 levels in rTg-DI rats are consistent with prior studies using CSF collected from probable CAA patients with MRI confirmed cerebral microbleeds [20
]. However, the present findings show that reductions in CSF/plasma Aβ40 occur much earlier with the onset of CAA and prior to microbleeds. This underscores the value of this novel model to identify other possible biomarkers that correlate with disease state.
4. Materials and Methods
4.1. rTg-DI Rats
All animal experiments were approved by the local Institutional Animal Care and Use Committees at the University of Rhode Island (project #AN1718-008; approval dates 12/11/2017-12/10/2020) and Yale University (project #2019-20132; approval dates 11/1/2019-10/31/2022) and conducted in accordance with the United States Public Health Service’s Policy on Humane Care and Use of Laboratory Animals. rTg-DI rats were designed to express human AβPP (isoform 770) harboring the Swedish K670N/M671L, Dutch E693Q, and Iowa D694N mutations in neurons under control of the Thy1.2 promoter and produce chimeric Dutch/Iowa CAA mutant Aβ in their brains [19
]. Transgenic offspring were determined by PCR analysis of tail DNA. All subsequent analyses were performed with heterozygous transgenic rats.
4.2. CSF Collection
CSF was collected from the cisterna magna of rTg-DI rats at designated ages. Rats were deeply anesthetized with inhalation of isoflurane and then mounted on a stereotaxic unit. A midline incision was made beginning between the ear and ending approximately 2.5 cm caudally. The fascia was retracted and muscles dissected, exposing the atlanto-occipital membrane. Using a #11 scalpel, a small slit was made along the midline of the membrane and underlying dura under a surgical microscope. The CSF was collected through the dura slit by using a fine tip pipette and aliquoted into sterile Eppendorf tube and frozen at −80 °C. Approximately 150 µL of CSF was collected from each rat.
4.3. Plasma Isolation
Rat blood was collected by terminal cardiac puncture from anesthetized rTg-DI rats at the designated ages. Blood was collected in one tenth volume of 3.8% sodium citrate to prevent coagulation. Blood was centrifuged at 8000× g for 5 min at room temperature to remove platelets and cellular components. Plasma samples were stored at −80 °C until ELISA analysis.
4.4. Brain Tissue Collection and Preparation
Rats were euthanized at designated time points and perfused with cold-PBS, forebrains were removed and dissected through the mid-sagittal plane. One hemisphere was immersion-fixed with 70% ethanol overnight and subjected to increasing sequential dehydration in ethanol, followed by xylene treatment and embedding in paraffin. Alternatively, brains were fixed with 4% paraformaldehyde overnight at 4 °C and subjected to increasing concentrations (10%, 20%, 30%) of sucrose in PBS, then embedded in OCT compound (Sakura Finetek Inc., Torrance, CA, USA) and snap-frozen in dry ice. Other hemispheres were collected, frozen on dry ice and stored at –80 °C. Sagittal sections were cut at 10 µm thickness using a Leica RM2135 microtome (Leica Microsystems Inc., Bannockburn, IL, USA), placed in a flotation water bath at 40 °C, and then mounted on Colorfrost/Plus slides (ThermoFisher Scientific, Houston, TX, USA). In some cases, coronal sections were cut at 20 μm thickness from frozen brains using a Leica CM1900 cryostat (Leica Microsystems Inc.), stored in PBS with 0.02% sodium azide at 4 °C.
4.5. ELISA Quantitation of Aβ Peptides
The levels of soluble and insoluble Aβ40 and Aβ42 were determined by performing specific ELISAs as described [47
]. Briefly, brain hemispheres that were flash frozen and pulverized in liquid nitrogen. A soluble fraction was obtained by homogenizing tissue with 10 µL/mg of 1M sodium carbonate, 500 mM NaCl, pH 11.5 and 0.5 mm zirconium oxide beads in a bullet blender. Aliquots were spun at 1600× g
at 4 °C for 20 min. The supernatant was removed, which was the soluble fraction. The remaining pellet was suspended in 5 M guanidine-HCl, 50 mM Tris, pH 8.0 and rotated at room temperature for 3 h. Samples were centrifuged and the supernatant was collected, which was the insoluble fraction. For each of the two fractions, a sandwich ELISA was performed. Antibody reagents for the Aβ ELISAs were generously provided by Lilly Research Laboratories, Indianapolis, IN, USA. In the sandwich ELISAs Aβ40 and Aβ42 peptides were captured using the carboxyl-terminal specific antibodies m2G3 and m21F12, respectively, and biotinylated m3D6, specific for the N-terminus of human Aβ, was used for detection followed by streptavidin-HRP (Amdex RPN4401V; Fisher Scientific, Pittsburgh, PA, USA). Plates were developed using KPL SureBlue (SeraCare, Milford, MA, USA) and read using a Spectramax M2 plate reader (Molecular Devices, Sunnyvale, CA, USA). Each sample lysate was measured in triplicate and compared to linear standard curves generated with known concentrations of human Aβ. The same ELISA format was used to measure soluble Aβ levels in CSF and plasma collected from each rat.
4.6. Immunohistochemical Analysis
Antigen retrieval was performed by treating the tissue sections with proteinase K (0.2 mg/mL) for 10 min at 22 °C. Primary rabbit polyclonal antibody to collagen type IV was used to visualize cerebral microvessels (1:100; ThermoFisher, Rockford, IL, USA). Primary antibody was detected with Alexa Fluor 594-conjugated donkey anti-rabbit secondary antibody (1:1000). Staining for fibrillar amyloid was performed using thioflavin S. Prussian blue iron staining was performed to detect hemosiderin deposits reflecting signs of previous microhemorrhage.
4.7. Quantitative Histological Analysis of CAA Load and Microbleeds
The percent area amyloid coverage of cerebral microvessels and percent area iron staining in the thalamic region was determined in rats at each of the specified ages using stereological principles as previously described [19
4.8. Magnetic Resonance Imaging Analysis
For non-invasive MRI imaging, rTg-DI rats (N
= 4) and age-matched WT rats (N
= 4) were lightly anesthetized with dexmedetomidine (0.015 mg/kg/h) and low dose isoflurane 0.5–1% as previously described [49
]. MRI imaging was performed on a Bruker 9.4T MRI and images of the rat brain were acquired using a 40 mm volume transmit and receive coil. The 3D gradient echo imaging parameters were acquired with the following parameters: TR/TE/FA = 60 ms/2~32/15°, NEX = 6, resolution= 0.23 × 0.23 × 0.23 mm, scan time= 50 min. A proton density weighted anatomical MRI of each rat’s brain was acquired at the same spatial resolution. Following MRI, the anesthesia was discontinued, and the rats allowed to recover. Rats were scanned at three, six, and nine months of age.
Quantitative 3D T2* maps were calculated from the 3D multiple gradient echo (MGE TE = 2~32 ms) MRIs by assuming mono-exponential relationship between the signal and TEs [50
]. Deposition of paramagnetic containing blood product, such as ferritin, and T2* values have been reported to be linearly correlated. T2* values ≤ 20 ms within the thalamus was identified on the parametric T2* maps in each rat using the Amira software segmentation editor (Amira 6.4, ThermoFisher Scientific, Houston, TX, USA). The number of voxel with T2* ≤ 20 ms was converted into mm3
and used as an estimate of ‘total hemorrhagic load’ in the thalamus.
4.9. Statistical Analysis
Histological and biochemical data were analyzed by t-test at the 0.05 significance level.