Cerebral Amyloid Angiopathy and Parenchymal Amyloid Deposition in Transgenic Mice Expressing the Danish Mutant Form of Human BRI2

Abstract Familial Danish dementia (FDD) is an autosomal dominant neurodegenerative disease clinically characterized by the presence of cataracts, hearing impairment, cerebellar ataxia and dementia. Neuropathologically, FDD is characterized by the presence of widespread cerebral amyloid angiopathy (CAA), parenchymal amyloid deposition and neurofibrillary tangles. FDD is caused by a 10‐nucleotide duplication‐insertion in the BRI2 gene that generates a larger‐than‐normal precursor protein, of which the Danish amyloid subunit (ADan) comprises the last 34 amino acids. Here, we describe a transgenic mouse model for FDD (Tg‐FDD) in which the mouse Prnp (prion protein) promoter drives the expression of the Danish mutant form of human BRI2 . The main neuropathological findings in Tg‐FDD mice are the presence of widespread CAA and parenchymal deposition of ADan. In addition, we observe the presence of amyloid‐associated gliosis, an inflammatory response and deposition of oligomeric ADan. As the animals aged, they showed abnormal grooming behavior, an arched back, and walked with a wide‐based gait and shorter steps. This mouse model may give insights on the pathogenesis of FDD and will prove useful for the development of therapeutics. Moreover, the study of Tg‐FDD mice may offer new insights into the role of amyloid in neurodegeneration in other disorders, including Alzheimer disease.


INTRODUCTION
Familial Danish dementia (FDD) was first described in members of a single Danish family in the Djursland peninsula (31). The disease was originally named heredopathia ophthalmo-oto-encephalica because of the presence of cataracts, hearing problems and neurological disease (30,31). Gradual loss of vision, caused by posterior sub-capsular cataract and retinal neovascularization, seems to be the first symptom of the disease which starts before the age of 30. Impaired hearing tends to appear 10-20 years later. Ataxia of a cerebellar type starts after the age of 40. Patients develop slurred speech as well as staggering and swaying gait. Paranoid psychosis usually develops after the age of 50, followed by dementia. Most patients die within the sixth and seventh decade of life (30,31).
Neuropathologically, FDD is characterized by the presence of cerebral amyloid angiopathy (CAA) in vessels of the retina and leptomeninges as well as in vessels of the gray and white-matter of the central nervous system (CNS) (15,26,34). Although CAA is one of the most prominent features of the disease, the incidence of cerebral hemorrhage is rare. All regions of the hippocampus show extensive diffuse amyloid peptide deposition with the presence of large (60-120 mm) cotton-wool-like plaques, abnormal neurites, that are found in the vicinity of blood vessels with amyloid, and neurofibrillary tangles (NFTs) composed of tau paired helical filaments (15,34). Western blot analysis of the insoluble tau isolated from a patient with FDD showed a pattern similar to that seen in cases of familial British dementia (FBD) (33) and Alzheimer disease (AD) (14,15). Molecular genetic analysis in patients with FDD revealed the presence of a 10-nucleotide duplication insertion (BRI 2795-796InsTTTAATTTGT) in the 3'-end of the coding region of the BRI2 gene (also known as ITM2B (7)), located on the long arm of chromosome 13 (33,34). The wild-type BRI2 gene encodes a 266 amino acids type-II single transmembrane domain protein of unknown biological function (35). The mutation in BRI2 causes a frame-shift in the BRI2 sequence, generating the Danish amyloid precursor protein of 277 amino acids (ADanPP), of which the~4 kDa Danish amyloid subunit (designated ADan) comprises the last 34 amino acids (34). Furin and other subtilisin-like pro-protein convertases (PCs) (24) can process both the normal and mutated precursor protein to produce the C-terminal fragments of 23 and 34 amino acids, respectively (17,18). In addition, all cases of FDD examined so far show the presence of amyloid b (Ab) deposited in combination with ADan or alone in blood vessels or brain parenchyma (15,34). The signifi-cance of the co-deposition of both types of amyloid is not clear, although Ab has been found co-deposited with other cerebral amyloids such as cystatin C (37) and prion protein (3).
Herein, we report the generation and neuropathological characterization of a transgenic animal model for FDD (Tg-FDD) in which the Danish mutant form of human BRI2 is expressed under the control of the mouse prion protein promoter. Tg-FDD mice show significant vascular amyloid deposition, parenchymal ADan deposition, amyloid associated gliosis, intracellular and extracellular deposition of oligomeric forms of ADan as well as tau immunoreactive deposits in neuropil. Tg-FDD mice may be considered as a new model of cerebral amyloidosis, which may be useful in further elucidating the pathogenesis of FDD and for testing of diagnostic and therapeutic strategies.

Construction of the MoPrP-BRI2795InsTTTAATTTGT transgene and generation of transgenic mice
The 10-nucleotide duplication (TTTAATTTGT) (34) was introduced in the BRI2 complementary DNA (cDNA) sequence by polymerase chain reaction (PCR), and the resulting BRI2795InsTTTAATTTGT cDNA sequence was PCR-amplified using oligonucleotide primers containing an XhoI site, a Kozak consensus sequence and a stop codon. The BRI2795InsTTT AATTTGT cDNA was inserted into the previously described pBS/ MoPrP.Xho vector (2) and the orientation verified by DNA sequencing. The resulting construct was linearized and gel-purified prior to injection into hybrid C3HeB/FeJ mouse embryos. Transgenic lines were established using standard methods at the Indiana University Transgenic and Knock-out Mouse Core Facility. Lines were crossed to non-transgenic C57BL/6J mice for over 10 generations and maintained by crossing transgenic animals to nontransgenic C57BL/6J mice. For genotyping, DNA was extracted from tail clips by proteinase K digestion followed by ethanol precipitation. The transgene was detected by PCR amplification of a 372-basic pair (bp) product using a forward primer (5'-GAT GCC CCA GCT GCT CTC TAC CAG-3') and a reverse primer (5'-GTA AGT TTC CTT GTC ATG AC-3') located in the human BRI2 cDNA sequence (33). The care and use of animals in this study were in accordance with institutional guidelines.

Histological and immunohistochemical studies
Mice were anesthetized and transcardially perfused with cold 0.9% saline. After perfusion, the animals were decapitated; the skulls opened and the brains removed and kept at 4°C in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2. Eight-micrometer-thick sections were stained with the Hematoxylin-Eosin (H&E) and thioflavin S (ThS) methods. Immunohistochemical labeling was carried out following published protocols (1). For immunohistochemistry, sections were incubated overnight at 4°C with the primary antibodies. Immunostaining was visualized using the avidin-biotin system (Vectastain; Vector Laboratories, Burlingame, CA, USA) and 3,3'-diaminobenzidine (Sigma) as the chromogen. The sections were counterstained with cresyl violet or H&E. Co-localization experiments were done by double-labeling immunofluorescence. Antigen retrieval was performed in a microwave oven in 0.01 M sodium citrate (pH 6.0). Sections were permeabilized in phosphate buffered saline (PBS) containing 0.3% Triton X-100 and unspecific binding was blocked in 10% goat serum, 2% bovine serum albumin, 0.1% Triton X-100 diluted in PBS for 1 h at room temperature. Primary Abs were incubated overnight at 4°C followed by appropriate secondary antibodies conjugated with Alexa-Fluor 488 and Alexa-Fluor594 (Molecular Probes Inc, Eugene, OR, USA). Sections were then washed three times in PBS, mounted with ProLong anti-fade (Molecular Probes Inc) and then viewed under a Leica DM4000B microscope with fluorescence attachment and digital camera. Fluorescence images were acquired with a Zeiss LSM 510 confocal microscope equipped with argon and helium/neon lasers using an inverted 63¥ NA 1.2 waterimmersion objective.

Immunoblot analysis
Brain tissue was homogenized in 10 vol of 50 mM Tris (pH 8,0), 150 mM NaCl and 5 mM EDTA containing a proteinase inhibitor cocktail (Complete, Roche Molecular Biochemicals, Indianapolis, IN, USA). Amyloid was extracted by incubation with 99 % formic acid (Sigma) for 2 h at room temperature and the formic acidsoluble material was dried under a N2 atmosphere as described (34).

Generation of transgenic mice expressing the Danish mutant form of BRI2 (Tg-FDD)
We generated transgenic mice by microinjecting a construct containing a cDNA encoding the Danish mutant form of human BRI2 ( Figure 1A) to oocytes of C3HeB/FeJ hybrids. Five independent transgenic lines were obtained, three of which produced stable lines of mice that express the transgene. Expression of the transgene in the CNS was driven by the mouse prion protein (moPrnp) promoter (2) ( Figure 1B). The presence of the transgene was confirmed by PCR amplification of a 372-bp fragment, specific for the human BRI2 cDNA sequence (33). All three lines developed accumulation of amyloid in the brain with similar regional deposition and were studied together. ADan amyloid was detectable in animals older than 7-8 month of age by western blot analysis. ADan was present in the formic acid soluble fraction mostly as a~4 kDa monomer ( Figure 2A, lane 1, see arrow). ADan polymers, in particular trimers and tetramers, were also seen by western blot analysis using Ab 1700. No immunoreactivity was seen in control mice (lane 2) or using Ab 1705, against the ABri peptide (not shown). The specificity of Ab 1700 was confirmed by western blot analysis using recombinant proteins (Figure 2A, inset).
As the animals aged, Tg-FDD mice displayed an abnormal grooming behavior. One-year-old transgenic mice showed an arched back and walked with a wide-based gait and shorter steps. Feet clasping was observed upon suspension of the mice by their tails. Transgenic mice would flex their front and hind limbs inward, with paws clasped together and drawn in toward the body. In contrast, wild-type mice demonstrate normal limb posture when suspended by their tails.

Vascular amyloid deposition in Tg-FDD mice
Brains from 34 heterozygous and homozygous mice and 12 agematched non-transgenic littermate controls (1-24 months old) were neuropathologically examined. Between 1 and 6 months of age, no obvious pathology was detected. At~7 months of age, transgenic animals consistently began to exhibit CAA primarily in pial (leptomeningeal) cerebellar vessels. As the animals aged, (Ն8-9 months), large and medium-sized parenchymal and penetrating vessels of the cerebrum (neocortex, hippocampus, thalamus and olfactory bulb), the brain stem and the spinal cord also showed amyloid angiopathy (Figures 3-5). CAA was present in all transgenic mice studied and amyloid deposition was seen to increase with age. As in patients with CAA (36), individual vessels had a varying extent of amyloid deposition, sometimes seen in the form of focal globular clumps ( Figure 3K). In the hippocampus, vascular amyloid deposition was seen in the walls of large and medium size vessels ( Figure 4G,H) and in the wall of vessels of the hippocampal fissure ( Figure 4B,N). Vessels affected by CAA showed thickened, eosinophilic walls ( Figure 4K-N). Abs against the ADan peptide immunolabeled vascular deposits in Tg-FDD mice but did not label vessels of non-transgenic littermates (Figures 3-5). Parenchymal vessels of the spinal cord that contained amyloid deposits were also immunolabeled by Ab 1700. Vascular deposits were also immunostained using Abs against P-component (not shown) and ApoE ( Figure 5F). Preliminary attempts to identify Ab deposits were unsuccessful. Nontransgenic littermates showed none of these neuropathological changes. Accumulation of ADan amyloid in vessels was accompanied by a loss of smooth muscle cells (SMCs) in the vessel wall as shown by the depletion of vascular SMC a-actin ( Figure 2B-D) and a decrease in the number of SMC nuclei. Vessels that were not affected by amyloid deposition showed a continuous rim of SMCs. Despite the massive occurrence of amyloid angiopathy in Tg-FDD mice, in none of the old mice (21 months) we analyzed did we detect hemorrhages.

Parenchymal accumulation of ADan in Tg-FDD mice
Parenchymal amyloid deposition was seen in brain areas affected by CAA (Figures 3 and 4) with the exception of the cerebellum, where it was relatively uncommon to observe the presence of ThS fluorescent material that was not associated with vessels ( Figure 5). In the cerebral cortex we observed the presence of ThS-fluorescent deposits of~11 mm in diameter and some small punctate deposits of~3 mm in diameter ( Figure 3A). These deposits were immunolabeled by Ab 1700. Widespread diffuse deposition of ADan was also observed between the same regions ( Figure 3C,G-I). Immunoreac-tivity was not seen in non-transgenic littermates ( Figure 3F) or using Ab 1705 (not shown). Some of the Ab 1700-immunopositive parenchymal deposits appeared to be within oligodendrocyes ( Figure 3G-I). Double-labeling of neurons with Abs against NeuN and Ab 1700 showed co-localization of both immunoractivities ( Figure 6). Abs against GFAP and activated microglia failed to reveal co-localization when used in combination with Ab 1700.
Parenchymal amyloid deposition was also seen throughout the hippocampus and was most prominent in the CA3 and CA2 regions and the hilus ( Figure 4A,B). Amyloid deposits were seen as small plaques (7-10 mm in diameter) that sometimes showed a weak ThS fluorescent surrounding ( Figure 4G,H) and in the form of punctate structures ( Figure 4H). Ab 1700 immunolabeled amyloid deposits ( Figure 4I,J,O,P), and also showed the presence of abundant ADan diffuse deposits ( Figure 4I,J). As in the cerebral cortex, some of the immunopositive deposits detected using Ab 1700 appeared to be intracellular ( Figure 4I,O,P). Furthermore, Ab 1700 revealed the presence of diffuse parenchymal deposition of ADan in the spinal cord (not shown). Analysis of H&E-stained sections occasionally revealed neurons with a pyknotic appearance, in particular in cortical layer II of the cerebral cortex, and neuronal shrinkage and loss of Purkinje cells in the cerebellum. Non-transgenic littermates showed none of these neuropathological changes.   Figure 3D). Only a weak astrocyte immunostaining was observed in all examined brain regions of age-matched wild-type mice ( Figure 3E). The presence of numerous GFAP-positive reactive astrocytes was also noted in the hippocampus of Tg-FDD mice ( Figure 4D,E) but not in nontransgenic littermates ( Figure 4F). In general, activated astrocytes in the hippocampus were seen in areas affected by ADan amyloid deposition ( Figure 7A-D) but not in areas devoid of amyloid. In addition to astrocytes, we observed 5D4-immunopositive microglia located in the close vicinity of the vascular and parenchymal amyloid deposits. Microglial cell processes were seen surrounding amyloid deposits ( Figure 7E-H), but were not observed in association with diffuse deposits.

Deposition of oligomeric forms of ADan and presence of abnormal tau deposits in Tg-FDD mice
A robust immunostaining was observed throughout the cerebrum ( Figure 8A) using Ab A11, which recognizes Ab oligomers and oligomers of other proteins involved in neurodegenerative diseases (16). Immunostaining was seen both intra-and extra-cellularly in Tg-FDD mice but not in wild-type controls ( Figure 8B). In the cerebellum, Ab A11 immunostaining was seen in cell bodies and dendrites of Purkinje cells ( Figure 8C). Staining of adjacent sections with ThS revealed lack of co-localization of A11 immunostaining and ThS fluorescence, strongly suggesting that the deposits recognized by Ab A11 represent ADan oligomers. Immunohistochemistry using Ab AT8 revealed the presence of numerous tau-immunopositive deposits throughout the neuropil of the cerebral cortex and hippocampus ( Figure 8D). No tangle formation was seen by ThS fluorescence.

DISCUSSION
We generated transgenic mice expressing the Danish mutant form of BRI2 (34). As in FDD, CAA is one of the main neuropathological findings in this model with vascular amyloid leading to a loss of SMCs. We also observed the presence of parenchymal (mostly diffuse) deposition of ADan and accumulation of oligomeric forms of the peptide. Some ThS fluorescent parenchymal amyloid deposits were also seen in the cortex and hippocampus. In Tg-FDD mice, amyloid burden appears to increase with age implying a progressive ADan deposition that exceeds its clearance. As in FDD, amyloid deposition leads to a neuroninflammatory response (15). Although we found tau-positive structures, we did not detect NFTs. Thus, Tg-FDD mice may be considered a new model of cerebral amyloidosis.
One of the main neuropathological features of FDD is the presence of widespread CAA (15,34). In Tg-FDD mice, expression of the Danish mutant form of BRI2 driven by the moPrnp promoter is sufficient for the development of CAA. CAA and parenchymal plaques have been also observed when the expression of the mutant BRI2 sequence is driven by the Syrian hamster Prnp promoter (6). Thus, the presence of CAA and the localized loss of SMCs seen in the Tg-FDD mouse model strongly suggest that neural cells are the source of the cerebrovascular amyloid in FDD. In addition, it denotes the vasculotropic nature of the ADan peptide. The solubility properties of ADan may permit this molecule to diffuse, resulting in ADan being deposited around leptomeningeal vessels, as it has been suggested for Ab1-40 (38). Our data also suggest that loss of SMCs does not require intracellular ADan production and that it may be induced in vivo by extracellular amyloid peptides, as it has been suggested for transgenic mice models expressing mutant forms of the amyloid b precursor protein (AbPP) (8,13,22,32). Although CAA is considered an important cause of intracerebral hemorrhage as well as ischemic stroke (27) astrocytes and activated microglia. Neuroinflammation associated with amyloid deposition has been observed in patients with FDD (15,29), AD (28) and in mutant AbPP transgenic mice (13,22). Whether the inflammatory response, characterized by the released of potentially neurotoxic substances and cytokines, may cause neurodegenerative changes in neurons, or may have a neuroprotective role by stimulating the clearance of the amyloid deposits, remains to be determined (23). Importantly, treatments designed to reduce CAA-induced neuroinflammation appear to have lessened the dementia in affected individuals (12,25). Moreover, the reduction of the cerebral microvascular deposition of Ab in an animal model expressing a human Swedish/Dutch/Iowa mutant AbPP has been shown to diminish regional neuroinflammation (21). As seen in FDD patients (34), Tg-FDD mice showed presence of abundant diffuse ADan deposits in the cerebral parenchyma. In addition, we observed the presence of small, dot-like, fluorescent ThS structures. Larger plaque-like amyloid deposits were sporadically observed, and whether these deposits are related to blood vessels remains to be determined. These plaque-like amyloid deposits were seen associated with GFAP-labeled astrocytes and activated microglia. This finding suggests that ADan amyloid deposition in Tg-FDD may cause the activation of an inflammatory process, similarly to what has been reported in animal models of AD (9,11,12,21,23,25). We also found intracellular Ab 1700immunoreactivity in close proximity to the cell nucleus in Tg-FDD mice. This finding may reflect intracellular accumulation of the amyloid peptide or the precursor protein as the residues recognized by Ab 1700 are present in both the ADan peptide and the ADanPP sequences (34). Processing of the precursor protein by PCs (17,18) and by ADAM 10 and SPPL2a/b (19) seems to be initiated upon the arrival of the protein to the Golgi (4). After processing, aminoterminal proteolytic fragments accumulate intracellularly and the carboxyl-terminal processing products are secreted via a regulated secretory pathway (4). In vitro studies using transfected cell lines have suggested that while the ABri peptide found in patients with FBD (33) is detected both intracellularly and in the medium, the ADan peptide accumulates predominantly in intracellular compartments (18). Whether intracellular accumulation of ADan peptides results in neuronal dysfunction leading to neurodegeneration remains to be seen. In contrast to Tg-FDD mice, transgenic mice expressing a BRI 2 protein-Ab42 fusion construct that generates only Ab42 in the brain show little evidence of intraneuronal Ab accumulation (20).
The presence of ADan-oligomers detected with Ab A11 is highly indicative that a soluble pool (intracellular and extracellular) of ADan may play a role in neurodegeneration. In vitro studies have shown that fresh solutions containing soluble oligomers of ADan are more toxic to neuronal cells than aged solutions containing insoluble aggregates (10). Interestingly, the study of Ab oligomers which do not co-localize with the Ab fibril plaques in mouse models have shown that Ab oligomers can impair synaptic function and spatial memory and may contribute to some of the early cognitive dysfunction in AD (16). Thus, the study of oligomeric forms of ADan in Tg-FDD mice may allow us to determine the exact roles of intracellular and extracellular ADan peptides in neuronal toxicity and to investigate whether similar mechanisms of toxicity exist between FDD and AD in vivo. Our results suggest that Tg-FDD mice may provide a unique model of cerebral amyloid deposition that will allow drawing parallels with mouse models of AD. Moreover, this model may provide important information on whether amyloid peptides, regardless of their primary amino acid sequence, share common pathogenic mechanisms.