Observations in APP Bitransgenic Mice Suggest that Diffuse and ...

The American Journal of Pathology, Vol. 178, No. 5, May 2011 Copyright © 2011 American Society for Investigative Pathology. Published by Elsevier Inc. All rights reserved. DOI: 10.1016/j.ajpath.2011.01.052

Neurobiology

Observations in APP Bitransgenic Mice Suggest that Diffuse and Compact Plaques Form via Independent Processes in Alzheimer’s Disease

Anna Lord,*† Ola Philipson,* Therése Klingstedt,‡ Gunilla Westermark,§ Per Hammarström,‡ K. Peter R. Nilsson,‡ and Lars N.G. Nilsson* From the Departments of Public Health and Caring Sciences/ Molecular Geriatrics * and Medical Cell Biology,§ Uppsala University, Uppsala; Bioarctic Neuroscience AB,† Stockholm; and Department of Chemistry,‡ IFM, Linköping University, Linköping, Sweden

Studies of familial Alzheimer’s disease suggest that misfolding and aggregation of amyloid-␤ (A␤) peptides initiate the pathogenesis. The Arctic mutation of A␤ precursor protein (APP) results in AD, and Arctic A␤ is more prone to form A␤ protofibrils and extracellular deposits. Herein is demonstrated that the burden of diffuse A␤ deposits but not compact plaques is increased when tg-Swe mice are crossed with tg-ArcSwe mice synthesizing low levels of Arctic A␤. The diffuse deposits in bitransgenic mice, which contain primarily wild-type A␤42, accumulate in regions both with and without transgene expression. However, APP processing, when compared with tg-Swe, remains unchanged in young bitransgenic mice, whereas wild-type A␤42 aggregation is accelerated and fibril architecture is altered in vitro and in vivo when a low level of Arctic A␤42 is introduced. Thus, the increased number of diffuse deposits is likely due to physical interactions between Arctic A␤ and wild-type A␤42. The selective increase of a single type of parenchymal A␤ deposit suggests that different pathways lead to formation of diffuse and compact plaques. These findings could have general implications for Alzheimer’s disease pathogenesis and particular relevance to patients heterozygous for the Arctic APP mutation. Moreover, it further illustrates how A␤ neuropathologic features can be manipulated in vivo by mechanisms similar to those originally conceptualized in prion research. (Am J Pathol 2011, 178:2286 –2298; DOI:

and extracellular compact plaques. The plaques are composed of various amyloid-␤ (A␤) peptides and stain with amyloid dyes such as Congo red. A␤ is generated via endoproteolysis of the amyloid precursor protein (APP). It aggregates and gives rise to compact plaques with an amyloid core, but also parenchymal diffuse A␤ deposits and cerebrovascular amyloid angiopathy in vessel walls. Discoveries of point mutations in the APP gene that caused familial AD provided major insights into the pathogenesis.1 The Arctic mutation (E693G), located within the A␤ domain of APP, segregates with clinical AD.2,3 Because Arctic A␤ (A␤Arc) is more prone to form protofibrils than is wild-type A␤ (A␤wt),3,4 this mutation was used in combination with the Swedish mutation (KM670/671NL) to develop an animal model rich in soluble oligomeric A␤ assemblies.5,6 Several founder lines were generated; however, only the line with the highest APP expression (line B, Table 1) has been reported to date.5,7–9 Its neuropathologic phenotypes were compared in detail with those of a transgenic model harboring only the Swedish mutation (tg-Swe; line A, Table 1).10 In other founder lines of tg-ArcSwe and tg-Swe, the synthesis of human APP Supported by grants from Uppsala University, Landstinget i Uppsala län, the Swedish Brain Fund, Alzheimerfonden, Demensfonden, Gamla Tjänarinnor, Gun och Bertil Stohnes Stiftelse, Magnus Bergvall, Åhlénsstiftelsen, Lars Hierta, Lundströms Minne, Frimurarstiftelsen, Svenska Läkaresällskapet, the Swedish Research Council (No. 2009-4389 to L.N.G.N.; No. 2008-4169 to P.H.; No. 2009-5343 to G.W.), the Swedish Foundation for Strategic Research and the Knut and Alice Wallenberg Foundation (P.H. and K.P.R.N.), Linköping University (P.H. and K.P.R.N.), the Swedish Foundation for Strategic Research (P.H. and K.P.R.N.), The European Union FP7 HEALTH (Project LUPAS) (K.P.R.N. and P.H.), the European Research Council (Project MUMID) (K.P.R.N.), and Astrid and Georg Olsson (P.H. and K.P.R.N.). A.L. and O.P. contributed equally to the study. Accepted for publication January 13, 2011.

10.1016/j.ajpath.2011.01.052)

Supplemental material for this article can be found at http://ajp. amjpathol.org or at doi: 10.1016/j.ajpath.2011.01.052.

The cardinal histopathologic lesions in brains affected with Alzheimer’s disease (AD) are neurofibrillary tangles

Address reprint requests to Lars N.G. Nilsson, Ph.D., Department of Public Health and Caring Sciences/Molecular Geriatrics, Uppsala University, Rudbeck Laboratory, Dag Hammarskjölds väg 20, SE-751 85 Uppsala, Sweden. E-mail: [email protected].

2286

Selective Increase in Diffuse A␤ Plaques 2287 AJP May 2011, Vol. 178, No. 5

Table 1.

Primary Characteristics of Various APP-Transgenic Lines

Model

Line

Human APP expression

tg-Swe tg-ArcSwe tg-ArcSwe tg-ArcSwe ⫻ tg-Swe

A B D Bitransgenic

6.0 2.0 0.85 6.85

1 1 1 1

Nontransgenic

0

1

⫹

Murine APP expression

⫽

APP overexpression* Sevenfold APPSwe Threefold APPArcSwe Twofold APPArcSwe ⬃Eightfold APP (APPArcSwe ⫹ APPSwe)

Age at onset of plaque deposition (months) 12 5–6 18 12

*APP expression compared with endogenous murine APP.

and A␤ was low, and A␤ deposition was scarce or undetectable even in very old animals. Herein we cross-bred a tg-ArcSwe founder line with very scarce A␤ pathology with a plaque-depositing tg-Swe line, in which human APP was expressed at a high level. We deemed that the experiment would be interesting for the following reasons. First, amyloid formation in human brain and APP transgenic mice is perceived to be seeded; that is, rapid fibril formation is thought to be preceded by a slow and rate-limiting reaction in which a nucleus is formed.11,12 Both in vitro and in vivo studies have clearly demonstrated that A␤Arc is more protofibrillogenic than A␤wt.4,5,13 It was hypothesized that simultaneous production of A␤Arc and A␤wt in bitransgenic mice would result in a prion-like aggregation mechanism14 in which the more aggregation-prone A␤Arc would act as a ␤-sheet template and accelerate A␤wt fibrillization. Second, both A␤Arc and A␤wt are produced in patients heterozygous for the Arctic mutation, and the pathogenic effects of such a mixture of A␤ peptides, if any, are unknown. By comparing the neuropathologic features of bitransgenic mice (tg-ArcSwe and tg-Swe) with single transgenic tg-Swe mice, whether aggregation-prone A␤Arc regulates A␤wt aggregation and/or formation of amyloid cores in the brain could be assessed. It was observed that the mixture of A␤Arc and A␤wt in bitransgenic mice led to more A␤42-immunoreactive (ir) diffuse deposits in areas with and without transgene expression in bitransgenic mice but did not lead to more compact plaques. These findings suggest that diffuse and compact plaques develop via independent processes.

Materials and Methods Transgenic Mice The expression cassette used to generate transgenic models and the procedures of genotyping have been described.5 All mice were kept at the animal facility at Uppsala University and housed under standard conditions with a 12-hour light/dark cycle and with free access to food and water. The tg-ArcSwe mice primarily used in this study were derived from a founder line with a low level of transgene APP expression. The phenotypes of this founder line differ markedly from the high-expressor tg-ArcSwe model.5,10 The various transgenic lines used

in the study are given in Table 1. All mice were genotyped using PCR with primers located at 1638 to 1655 and 2299 to 2281 in Y00264, followed by digestion with MboII and separation of bands on 2% agarose gel. The PCR product from a tg-ArcSwe mouse was then cleaved into two bands (456 and 205 bp), that from a tg-Swe mouse into three bands (364, 205, and 92 bp), and that from a bitransgenic tg-ArcSwe and tg-Swe mouse harboring both transgenes into four bands (456, 364, 205, and 92 bp). All mice used in the study were genotyped at age 3 weeks (ear biopsy) and again after death (tail biopsy). The experiments were approved by an ethical committee and performed in compliance with national and local animal care and use guidelines (protocol C223/8). Mice were anesthetized using 0.4 mL tribromoethanol (Avertin), 25 mg/mL, and intracardially perfused with 0.9% (w/v) NaCl solution. Brain hemispheres were frozen on dry ice for biochemical analyses. Alternatively, they were either directly frozen in metylbutane at ⫺25°C or immersed in 4% paraformaldehyde for 24 hours and used for immunohistochemical analyses.

Histologic and Image Analyses Fixed brain tissue was cryoprotected via sequential immersion in 10%, 20%, and 30% (w/v) sucrose for 24 hours. Coronal sections 25-␮m thick were prepared with a sledge microtome and stored at 4°C in PBS with 10 mmol/L NaN3. Two sections per individual, approximately 500 ␮m apart (bregma ⫺1.0 mm to ⫺3.0 mm), were selected for quantitative immunostaining. Fixed tissue sections were incubated in 25 mmol/L prewarmed citrate buffer (pH 7.3) for 5 minutes at 85°C, rinsed in deionized water, and immersed in 70% formic acid for 5 minutes. Endogenous peroxidase activity was quenched with 0.3% dH2O2 in Dako block (DakoCytomation, Glostrup, Denmark) for 15 minutes, and sections were permeabilized with 0.4% Triton X-100. The sections were incubated with 1 ␮g/mL polyclonal A␤40- or A␤42-specific antibodies,15 monoclonal A␤Arc-specific antibodies (mab27 at 1 ␮g/mL9), 0.5 ␮g/mL N-terminal A␤ antibody 6E10 (Covance, Princeton, NJ) or 82E1 (IBL International GmbH, Hamburg, Germany), 0.2 ␮g/mL goat anti-mouse poprotein J (AF2747; R&D Systems, Inc., Minneapolis, MN), goat anti-apolipoprotein E (ApoE) (AB947, diluted 1:5000; Chemicon International Inc., Temecula, CA), or 0.5 ␮g/mL monoclonal anti-human heparan sulfate

2288 Lord et al AJP May 2011, Vol. 178, No. 5

(H1890, against 10E4 epitope; US Biological, Swampscott, MA) in PBS-buffered saline solution with 0.1% Tween 20 at 4°C overnight. Biotinylated goat anti-rabbit, goat anti-mouse, or rabbit anti-goat antibodies (Vector Laboratories, Inc., Burlingame, CA) were applied for 30 minutes at room temperature. Sections were then incubated with streptavidin-coupled horseradish peroxidase (Mabtech AB, Nacka Strand, Sweden) for 30 minutes and developed using the Nova Red chromogen kit (Vector Laboratories, Inc.). For double staining with the A␤42specific antibody and either thioflavine T (ThT) or the luminescent conjugated oligothiophene (LCO) p-FTAA, the sections were treated as described, and a fluorescent secondary antibody [Alexa Fluor 594 goat anti-rabbit (Invitrogen Corp., Carlsbad, CA)] diluted 1:250 in PBS was used to visualize the A␤ immunostaining. The sections were washed in PBS and incubated with 3 ␮mol/L pFTAA16 in PBS or with 0.2% (w/v) ThT (Sigma-Aldrich Corp., St. Louis, MO) in PBS for 30 minutes with shaking at room temperature. After rinsing with PBS, the sections were washed in dH2O and mounted with Vectashield (Vector Laboratories, Inc.). Compact plaques were detected using Congo red (Sigma-Aldrich Corp.) as previously described.17 The burden of diffuse and compact deposits in the cerebral cortex was measured in two image fields per section at ⫻20 magnification. Images were captured using a Nikon microscope (DXM1200F; Nikon Instruments Inc., Melville, NY) equipped with a digital camera, converted to gray scale, and segmented with an autothreshold command (Image Pro-Plus; MediaCybernetics, Silver Spring, MD). Custom macros were used to measure the stained area of interest as a percentage of the total area.

p-FTAA and Fluorescence Microscopy Frozen brain sections 10-␮m thick from 18-month-old tg-ArcSwe (line B), tg-Swe (line A), and bitransgenic tgArcSwe (line D) and tg-Swe (line A) mice were fixed in absolute ethanol for 10 minutes and rehydrated in 50% ethanol followed by water. The sections were incubated in PBS for 10 minutes and stained with 3 ␮mol/L pFTAA16 in PBS for 30 minutes at room temperature. After rinsing with PBS, the sections were mounted with Dako fluorescence mounting medium (DakoCytomation A/S). The medium was allowed to solidify for 3 hours before the slide rims were sealed with nail polish. Spectra and fluorescence images were obtained using a fluorescence microscope (Zeiss Axioplan; Carl Zeiss Ltd., Jena, Germany) equipped with a SpectraCube (optical head) module (Applied Spectral Imaging, Ltd., Migdal Ha’Emek, Israel) using bandpass filters 405/30 (LP450), 470/40 (LP515), and 546/12 (LP590). p-FTAA bound to compact plaques (10 plaques per animal, three animals from each group) was excited at 405 and 546 nm. SpectraView 3.0 EXPO software (Applied Spectral Imaging, Ltd.) was used to merge the information from the double excitation and create full visible-light emission spectra (450 to 700 nm) for p-FTAA when combining those excitation wavelengths. Fluorescent spectral unmixing, which is a func-

tion within the program, was performed to separate the emitted light intensities according to wavelength. The emission spectrum for p-FTAA bound to compact plaques in tg-ArcSwe (line B) was used to define red emission, and the emission spectrum for p-FTAA bound to compact plaques in tg-Swe (line A) was used to define green emission. By performing spectral unmixing, the fraction of each defined spectrum could be determined for each compact plaque found in tg-ArcSwe (line B), tg-Swe (line A), and bitransgenic mice, respectively.

SDS-PAGE and Western Blot Analysis For analysis of A␤, multiphasic urea/SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis were performed. This gel system enables differentiation of A␤ of various lengths and also A␤wt and A␤Arc. Samples were subjected to electrophoresis using a 15% T, 5% C bicine-Tris separation gel [T, total concentration of acrylamide; C, bis-acrylamide cross-linker content (acrylamide/bis-acrylamide, 40%; Sigma-Aldrich Corp.)] containing 8 mol/L urea. They were run at 24 mA, 30 V. Gels and buffers have been described previously,18,19 with the distinction that in the present study, 8 mol/L urea was included in the stacking gel. For analysis of total brain A␤, tissues at bregma ⫺1.0 mm to ⫺3.0 mm were used. Tissue sections (30 ␮m; bregma 1.0 mm to ⫺0.1 mm) were dissected and analyzed for A␤ levels in striatum and cerebral cortex. Brain tissue was homogenized in a final concentration of 70% formic acid at a ratio of 1:10 (w/v) using a Dounce homogenizer, sonicated for 30 seconds, and centrifuged at 100,000g at 4°C for 60 minutes. The supernatants were stored at ⫺80°C before they were analyzed with enzyme-linked immunosorbent assay (ELISA) and SDS-PAGE Western blot analysis. For analysis of APP levels, brain tissues (bregma ⫺1.0 mm to ⫺3.0 mm) from 2-month-old mice were homogenized in Tris-buffered saline solution [20 mmol/L Tris and 137 mmol/L NaCl (pH 7.6)] supplemented with 2% (w/v) SDS and complete protease inhibitor (Hoffman-La Roche AG, Basel, Switzerland), sonicated for 30 seconds, and centrifuged at 100,000g at 4°C for 60 minutes. Proteins in the supernatant were separated on 4% to 20% Tris-tricine gels (Invitrogen Corp.). Protein transfer and Western blot analysis were performed as previously described.5

A␤ Enzyme-Linked Immunosorbent Assay Ninety-six–well plates were coated for 12 hours at 4°C with antibodies specific for A␤40 or A␤42 (200 ng per well) and used for measuring A␤1-40 or A␤1-42 levels. For the total A␤ assay, plates were coated with 100 ng per well of the N-terminal A␤-specific antibody 82E1 (IBL International GmbH) in PBS [130 mmol/L NaCl, 7 mmol/L NaH2PO4, and 3 mmol/L Na2HPO4 (pH 7.4)] and blocked with 1% bovine serum albumin in phosphate-buffered NaCl, 0.15% Kathon (Rohm and Haas Co., Philadelphia, PA). Formic acid extracts of mouse brain were neutralized in 1 mol/L Tris (pH 10) and diluted in ELISA incubation buffer (PBS with 0.1% bovine serum albumin, 0.05% Tween 20). Samples or standards (recombinant wt/Arctic

Selective Increase in Diffuse A␤ Plaques 2289 AJP May 2011, Vol. 178, No. 5

A␤1-40 or 1-42) were added to the plates in triplicate and incubated for 2 hours at room temperature. Biotinylated mAb27, 0.5 ␮g/mL, with an epitope in the mid-domain of Arctic A␤,9 was used to measure total A␤Arc. Biotinylated 4G8, 0.5 ␮g/mL (Covance, Inc.), was used to detect total A␤, and biotinylated mAb1C3, 0.5 ␮g/mL,7 was used to quantify A␤1-40 or A␤1-42. Both biotinylated detection antibodies and streptavidin-coupled horseradish peroxidase (Mabtech AB), diluted 1:2000, were incubated for 1 hour in successive steps. K-blue aqueous (ANL-Produkter AB, älvsjö, Sweden) was used as horseradish peroxidase substrate, the reaction was stopped using 1 mol/L H2SO4, and the optical density was measured at 450 nm with a spectrophotometer (SpectraMax 190; Molecular Devices, Inc., Sunnyvale, CA). Wells were washed three times between each step in ELISA washing buffer (phosphate-buffered NaCl with 0.1% Tween 20 and 0.15% Kathon).

mAb158 Protofibril/Oligomer ELISA The mAb158 ELISA7 detects oligomers larger than 100 kDa, which is consistent with the original description of oligomeric assemblies of A␤.20 In brief, each well in 96well plates was coated with 200 ng mAb158 in PBS overnight at 4°C and blocked with 1% bovine serum albumin. The samples were centrifuged at 17,900 ⫻ g for 5 minutes before incubation to ensure that they were devoid of any insoluble A␤ aggregates. Samples were added to plates in duplicate and incubated for 2 hours at room temperature. Biotinylated mAb158, 0.5 ␮g/mL, was added and incubated for 1 hour at room temperature, followed by incubation with streptaviden-coupled horseradish peroxidase for 1 hour at room temperature. K-blue aqueous (ANL-Produkter AB) horseradish peroxidase substrate was used, and the reaction was stopped using 1 mol/L H2SO4. After the blocking step, plates were washed three times between each step. All samples and antibodies were diluted in ELISA incubation buffer (PBS with 0.1% bovine serum albumin and 0.02% Tween 20). The standard was generated by incubating A␤1-42wt (American Peptide Co., Inc., Sunnyvale, CA) in 50 ␮mol/L PBS for 30 minutes at 37°C, followed by centrifugation at 16,000 ⫻ g for 5 minutes. The supernatant was loaded on a Superdex 75 10/300 GL column, resulting in a single high-molecular-weight peak, which was collected.

Kinetic Measurement of A␤ Fibrillation Recombinant A␤1-42wt and A␤1-42Arc (American Peptide Co., Inc.) were treated with 100% hexafluoroisopropanol (Sigma-Aldrich Corp.), reaching a final concentration of 1 mmol/L. Hexafluoroisopropanol was allowed to evaporate in a chemical fume hood, and A␤ was then solubilized as a 400-␮mol/L stock solution in 10 mmol/L NaOH and stored at ⫺80°C. Immediately before use, A␤1-42wt and A␤1-42Arc were diluted in PBS [50 mmol/L phosphate buffer and 100 mmol/L NaCl (pH 7.4)] to 20 ␮mol/L and added to wells of a nonbinding 96-well microtiter plate (Greiner Bio-One GmbH, Frickenhausen, Germany). ThT (Sigma Corp.) and p-FTAA16 were pre-

pared as 15-␮mol/L stock solutions in PBS (pH 7.4) and added to the wells at a final concentration of 1.2 ␮mol/L. Plates were incubated at room temperature for up to 22 hours under agitation. A␤ aggregation was recorded with a fluorometer (Wallac Victor 2 multilabel counter; Perkin Elmer Inc., Waltham, MA) as described for ThT21 and p-FTAA.16 The wavelengths of excitation and emission were 440 and 485 nm, respectively. Experiments were performed in triplicate, and fluorescence from free ThT or p-FTAA, incubated without A␤ in parallel, was subtracted from the readings.

Electron Microscopy The ultrastructure of A␤ aggregates was examined by collecting samples of A␤-solutions, 20 ␮mol/L, from the kinetic experiment. Solutions of A␤ were diluted to 1 ␮mol/L in mH2O and spotted on grids coated with carbon-formvar. Excess fluid was removed after 30 seconds, and 2.5% uranyl acetate, 5 to 10 ␮L, was added to the grids, and again excess fluid was quickly removed. The negatively stained specimen were examined and photographed using an electron microscope (JEOL 1230; JEOL Ltd., Tokyo, Japan) at 100 kV.

Statistical Analyses Data were analyzed using the Student’s t-test or with oneor two-way analysis of variance followed by Tukey’s multiple comparison post hoc test (GraphPad Software, Inc., San Diego, CA). Results were presented as values representing the group mean and standard error of the mean (SEM). The relationship between A␤40-ir and A␤42-ir burden was investigated using linear regression, and slopes were compared using the F-test. P ⬍ 0.05, P ⬍ 0.01, and P ⬍ 0.001 were considered significant.

Results Transgene Expression in Founder Lines of Tg-Swe and Tg-ArcSwe Mice Intraneuronal A␤ accumulates in young tg-ArcSwe mice, and extracellular deposition of compact plaques begins at age 5 to 6 months, whereas the frequency of diffuse A␤ deposits remains extremely low.5 This highexpressing founder line, with threefold APP expression compared with the level of murine APP, is hereafter referred to as tg-ArcSwe line B. In tg-Swe line A, hereafter referred to as tg-Swe, APP expression is sevenfold higher than in nontransgenic mice, and extracellular A␤ deposition emerges at approximately 12 months of age.10 Several additional founder lines of tg-ArcSwe with lower transgene expression were generated. It is not known whether a small amount of highly aggregation-prone peptides such as A␤Arc can influence the comparably slower aggregation of A␤wt. To investigate this, tg-Swe mice were crossed with tg-ArcSwe (line D), which express human APPArcSwe at a low level (total APP is less than twofold compared with murine

2290 Lord et al AJP May 2011, Vol. 178, No. 5

Figure 1. High-level APP-expressing tg-Swe mice (line A) produced seven times more human APP than did tg-ArcSwe mice (line D). A: Experimental design of the study is schematically illustrated. B: APP protein synthesis of transgenic models, tg-Swe (line A), and nontransgenic mice, tg-ArcSwe (lines B and D), was analyzed using Western blot analysis and antibodies 6E10 (human APP), 22C11 (human plus endogenous APP), and ␤-actin. Anatomical distribution of human APP detected with 6E10-immunostaining in 2-month-old tg-Swe, line A (C and F), low-expressing founder line of tg-ArcSwe, line D (D and G), and a nontransgenic mouse (E and H). Scale bars ⫽ 1 mm.

APP) (Figure 1A; Table 1). The total level of APP protein, human plus murine, in 2-month-old mice was analyzed using Western blot analysis. Human APP was sevenfold more abundant in young tg-Swe than in age-matched tg-ArcSwe (line D) as measured at densitometry. For comparison, a tg-ArcSwe mouse (line B)5 with threefold APP expression and a nontransgenic mouse were included in the analyses (Figure 1B; Table 1). In tg-Swe, with sevenfold APP expression, an abundance of human APP was observed in the hippocampal pyramidal cell layers CA1 and CA3, in the lower layers of cerebral cortex, and to some extent in the upper cortical lamina (Figure 1, C and F). The anatomical distribution was similar in tg-ArcSwe (line D), although the laminar expression pattern in the cerebral cortex was attenuated (Figure 1, D and G). As expected, in a nontransgenic mouse, only a faint nonspecific background signal was observed (Figure 1, E and H).

Burden of Compact Plaques in Tg-Swe Is Not Altered by a Minute Amount of A␤Arc Whether A␤Arc, at a level that seemed insufficient on its own to generate extracellular A␤ deposition, could

instigate or modify the neuropathologic conditions produced by an excess of A␤wt was investigated. Therefore, tg-ArcSwe (line D) and tg-Swe were crossed, and the neuropathologic features of bitransgenic mice were compared with those of singly tg-Swe mice. The neuropathologic conditions of singly tg-ArcSwe (line D) mice were reassessed in two aged individuals generated from the cross-breeding experiment at the study end point (ie, 18-month-old mice). Six tissue sections distributed along the rostrocaudal axis of the brain were stained with different A␤-antibodies (6E10, 82E1, and A␤42-specific). In one of the 18-month-old mice, a single plaque was observed, and in the other mouse, three A␤-immunopositive plaques were observed in six tissue sections altogether. Thus, A␤ deposition and intraneuronal A␤ immunostaining in aged tg-ArcSwe (line D) was limited. Biochemical analyses confirmed that age-dependent A␤ accumulation was modest. The mean (SEM) level of A␤1-40 was 9.9 (1.2) pmol/g (n ⫽ 2) at 2 months, and 19.5 (1.7) pmol/g at 18 months (n ⫽ 2) in tg-ArcSwe (line D). This was in contrast to findings in 11-month-old tg-ArcSwe (line B, n ⫽ 2) with a cortical A␤-burden of 2.5% (0.3%), total A␤1-40 levels of 8982 (770) pmol/g, and strong intraneuronal A␤ immunostaining (see Supplemental Figure S1 at http://ajp. amjpathol.org). These results were not surprising because it is well known that A␤ deposition will occur only if a sufficiently high level of A␤ synthesis is attained and that age at onset depends on the level of transgene expression.22 If the effect of transgenes on pathologic features were simply additive, bitransgenic mice, tg-ArcSwe (line D) and tg-Swe, and singly tgSwe would be expected to exhibit essentially the same type and quantity of A␤ deposits. Brains from bitransgenic and tg-Swe mice were first examined with Congo red and an A␤40-specific antibody. Both of these markers are associated with compact plaques in tg-Swe. At age 12 months, A␤40 burden was not higher in bitransgenic mice [mean (SEM), 1.2% (0.3%); n ⫽ 17] than in tg-Swe [1.0% (0.2%); n ⫽ 16; P ⫽ 0.55; Figure 2, A–C]. Likewise, Congo red burden was not increased in bitransgenic mice [0.25% (0.09%); n ⫽ 17] compared with singly tg-Swe [0.19% (0.07%); n ⫽ 16; P ⫽ 0.57; see Supplemental Figure S2, A–C, at http://ajp.amjpathol.org]. At age 18 months, again A␤40-ir burden was not higher in bitransgenic compared with tg-Swe mice [3.6% (0.6%); n ⫽ 7, and 2.8% (0.6%); n ⫽ 8; P ⫽ 0.43; Figure 2C]. Likewise, Congo red burden was not elevated in bitransgenic mice [1.3% (0.6%); n ⫽ 7] compared with agematched tg-Swe [1.0% (0.3%); n ⫽ 8; P ⫽ 0.61; see Supplemental Figure S2C at http://ajp.amjpathol.org]. Sections were then stained using an LCO, p-FTAA,16 a conformational-dependent probe with a flexible structure that can serve as an optical tool to reveal heterogeneity among A␤ and prion deposits in a manner similar to luminescent conjugated polythiophenes.10,23,24 As expected from previous imaging experiments using Drosophila models,25 the p-FTAA emission spectrum of compact plaques in tg-ArcSwe (line B; Figure 2D) and tg-Swe

Selective Increase in Diffuse A␤ Plaques 2291 AJP May 2011, Vol. 178, No. 5

A␤42-ir plaques was more than fourfold higher in bitransgenic mice [4.1% (0.9%); n ⫽ 17] than in tg-Swe [0.9% (0.3%); n ⫽ 16; P ⬍ 0.01 Figure 3, A, D, and G]. In 18-month-old mice, A␤42 burden was still increased more than threefold in bitransgenic mice [9.9% (1.3%); n ⫽ 7] compared with tg-Swe mice [3.1% (0.8%); n ⫽ 8; P ⬍ 0.001; Figure 3, B, E, and G]. Groups of bitransgenic (r ⫽ 0.86; P ⬍ 0.0001) and tg-Swe mice (r ⫽ 0.90; P ⬍ 0.0001) fitted well into separate linear regression models when A␤40 burden was plotted against A␤42 burden for each animal. The slope of linear regression for bitransgenic mice [2.3 (0.3)] and tg-Swe mice differed [1.1 (0.1); F1,44 ⫽ 14.8; P ⬍ 0.001]. Thus, introduction of a small amount of A␤Arc, produced by low expression of the tg-ArcSwe transgene, strongly and selectively increased the burden of diffuse A␤42-immunopositive deposits (Figure 3H). Next, consecutive sections were stained to investigate which A␤ peptides were located in diffuse and compact plaques. A␤42, A␤40, A␤Arc, and Congo red staining co-localized in compact plaques in bitransgenic mice. There was an additional population of diffuse A␤42-ir deposits in the cerebral cortex in these mice, but also in the striatum, where there is essentially no transgene expression

Figure 2. A␤Arc did not change A␤40 burden but influenced p-FTAA fluorescence of individual compact plaques in bitransgenic mice. Representative 12-month-old bitransgenic (A) and singly tg-Swe (B) mice. C: There was no statistically significant difference in cortical A␤40-burden between 12- and 18-month-old bitransgenic and singly tg-Swe mice. Fluorescence images of p-FTAA bound to compact plaques and excited at 470 and 546 nm. Images illustrate a variation in p-FTAA emission spectrum between compact plaques found in tg-ArcSwe, line B (D), singly tg-Swe (E), and bitransgenic mice (F), respectively. The resulting emission when the excitation wavelengths (470 and 546 nm) are separated into two channels is shown below. Spectral unmixing of p-FTAA emission when bound to compact plaques in bitransgenic mice revealed a shift toward that of compact plaques in tg-ArcSwe compared with singly tg-Swe. G and H, Fraction of red spectrum of 10 individual plaques from three animals in each group (filled and open boxes and diamonds). Scale bars: 500 ␮m (B); 30 ␮m (D–F).

mice (Figure 2E) differed. The p-FTAA fluorescence of compact plaques in bitransgenic mice (Figure 2F) resembled that in tg-Swe; however, small areas within the deposits were more similar to compact plaques in tg-ArcSwe (line B). Spectral analyses of emitted fluorescence from multiple individual compact plaques confirmed that signals in bitransgenic mice were intermediate to those of compact plaques in tg-ArcSwe and tg-Swe (Figure 2, G and H). Thus, a minute amount of A␤Arc in bitransgenic mice did not influence burden of compact plaques, but did result in a different LCO spectral profile in individual compact plaques.

Increased Accumulation of Diffuse A␤42-Immunoreactive Deposits in Bitransgenic Mice Tissue sections were immunostained with an A␤42-specific antibody, which in brains affected with AD typically detects both compact plaques and diffuse A␤ deposits, although diffuse A␤ deposits are rarely observed in tgSwe. At age 12 months, the mean (SEM) burden of

Figure 3. A␤Arc favored an increased A␤42-burden in bitransgenic mice. A␤ burden in the cerebral cortex of 12-month-old (A and D) and 18-month-old (B, C, E, and F) bitransgenic (A–C) and singly tg-Swe (D–F) mice. A␤42 burden in representative 12-month-old (A) and 18-month-old (B) bitransgenic mice and in 12-month-old (D) and 18-month-old (E) tg-Swe. A␤40 burden is shown from semiadjacent sections of 18-month-old bitransgenic (C) and tg-Swe (F) mice. G: A␤42 burden was investigated in all animals using quantitative image analysis. H: There was a linear correlation between A␤40 burden and A␤42 burden among individuals within the experimental groups, but a substantial difference between bitransgenic and singly tg-Swe mice. Scale bar ⫽ 900 ␮m.

2292 Lord et al AJP May 2011, Vol. 178, No. 5

Figure 4. Increased A␤42 burden in bitransgenic mice was due to diffuse deposits that were not recognized by an antibody specific for A␤Arc. A␤-deposition in 18-month-old bitransgenic (A–D and I–L) and singly tg-Swe mice (E–H and M–P). Adjacent sections (10 ␮m) were immunostained with A␤42-specific antibody (A, E, I, and M), A␤Arc-specific antibody (B, F, J, and N), Congo red (C, G, K, and O), and A␤40-specific antibody (D, H, L, and P). Arrows in A and E indicate areas shown at higher magnification (I and M, respectively), and arrows in I indicate some diffuse A␤42-immunoreactive deposits. Scale bars: 1 mm (A–H); 200 ␮m (I–P). The spots in (F) are unspecific staining that are due to slight damage to parts of this tissue section.

(Figure 4, A–D and I–L). Neither using a protocol for silver intensification of Ni-DAB product,26 nor using unfixed tissue section and a 10-fold higher concentration of mAb279 resulted in staining of the diffuse deposits with the A␤Arc antibody (data not shown). In tg-Swe, A␤40, A␤42, and Congo red staining were most often co-localized, and the compact plaques, as expected, did not stain with an A␤Arc antibody (Figure 4, E–H and M–P). It is important to note that the selectivity of the A␤Arc-specific antibody, mAb27, depends on a unique conformation around glycine 22 and not on the linear amino acid sequence. Thus, mAb27 is selective for A␤Arc at ELISA analysis but not, for example, at denaturing Western blot analysis.9 The selectivity of mAb27 immuno-

staining was determined using tissue sections from plaquebearing tg-ArcSwe, line B (positive control) and tg-Swe (negative control) mice that were prepared and stained in parallel to those from bitransgenic mice. Factors were then searched that could further explain how diffuse and compact plaques could form via independent processes. ApoE resides in extracellular deposits and regulates A␤ deposition in vivo.27 Other plaque-associated factors such as apolipoprotein J and heparan sulfate selectively affect burden of compact plaques.28,29 Therefore, we searched for selective markers of diffuse deposits and compact plaques. Among those, only heparan sulfate was specifically associated with compact plaques, whereas

Selective Increase in Diffuse A␤ Plaques 2293 AJP May 2011, Vol. 178, No. 5

Figure 5. A␤Arc led to increased accumulation of A␤42wt and more diffuse deposits in the striatum in bitransgenic mice. Urea/SDS-PAGE and Western blot analysis (with 6E10 antibody) of extracts of 70% formic acid (total A␤) of plaque-bearing bitransgenic and tg-Swe mice at ages 12 and 18 months. A: For comparison, tissue from a 6-monthold tg-ArcSwe (line B; high-expressor line) with plaques and synthetic A␤ peptides were included in the experiments. B: Immunostaining of a brain section from an 18-month-old bitransgenic mouse with an A␤42-specific antibody before and after the striatum (caudate putamen) had been dissected. Extracts of 70% formic acid from dissected striatum (caudate putamen) and cerebral cortex layers I to VI in an 18-month-old bitransgenic mouse at the level of the frontal cortex (bregma 1.0 mm to ⫺0.1 mm). C: Localization of A␤ peptides differed markedly, and A␤1-42wt was the predominant A␤ peptide in the striatum, whereas both A␤1-40wt and A␤1-42wt were abundant in the cerebral cortex. D: Further analyses of the same formic acid extracts of dissected tissues at ELISA (pmol/g tissue) demonstrated that A␤Arc was present but represented only a small fraction of total A␤ in both cerebral cortex and striatum (n ⫽ 4). Scale bars: 1 mm (B); 100 ␮m in enlarged images.

ApoE and apolipoprotein J were found in both diffuse and compact plaques (see Supplemental Figure S3 at http://ajp.amjpathol.org).

A␤Arc Facilitates Striatal A␤1-42wt Accumulation and Formation of Diffuse Deposits A␤ deposition in bitransgenic and tg-Swe mice was biochemically analyzed using urea/SDS-PAGE and Western blot analysis. Extracts of 70% formic acid (total A␤) from bitransgenic and tg-Swe mouse brain with early- and late-stage plaque (ages 12 and 18 months) were investigated. The system was calibrated with synthetic A␤ peptides (lanes 1 to 4, Figure 5A), and a 6-month-old tg-ArcSwe (line B) with high level of APP expression was used as reference material (lane 7).10 In 18-month-old bitransgenic mice, A␤1-38wt, A␤1-40wt, and A␤1-42wt were predominant, with low levels of A␤Arc (A␤1-40Arc and/or A␤1-42Arc) (lane 5, Figure 5A). These A␤ species were present at 12 months in bitransgenic mice (lane 8), although the levels of all A␤ species were much lower and A␤1-40wt was the predominant species. As expected, high levels of A␤1-40wt were observed in 18month-old tg-Swe, and A␤1-38wt and A␤1-42wt were also detected (lane 6). Brain tissue from a tg-ArcSwe line B mouse with plaque deposition contained A␤Arc peptides, primarily A␤1-40Arc and A␤1-42Arc and presum-

ably A␤1-38Arc (lane 7), as previously described.10 An immunohistochemical and a biochemical technique were combined to further determine the identity of A␤ in the diffuse deposits in bitransgenic mice. In thin tissue sections, the striatum, which contained only diffuse A␤ deposits, was separated from the cerebral cortex, in which both diffuse and compact deposits were observed (Figure 5B). Extracts of 70% formic acid in striatal and cerebral cortical tissue (layers I to VI) from a representative 18-month-old bitransgenic mouse were analyzed using urea/SDS-PAGE. There was a marked region-specific difference, with A␤1-42wt as the predominant A␤ peptide in the striatum, whereas A␤1-42wt and A␤1-40wt with some A␤1-38wt were present in the cerebral cortex (Figure 5C). Because the sensitivity of SDS-PAGE and Western blot analysis is somewhat limited, ELISA was used to measure A␤Arc levels (82E1/mAb27 sandwich) and total A␤ levels (82E1/4G8 sandwich) in the cerebral cortex and striatum. The mean (SEM) levels of A␤Arc and total A␤ in cerebral cortex were 160 (24) pmol/g and 2300 (190) pmol/g (n ⫽ 4), respectively; that is, approximately 7% of total A␤ was A␤Arc. In the striatum, the levels of A␤Arc and total A␤ were 16 (4) pmol/g and 770 (50) pmol/g, respectively (n ⫽ 4); that is, 2% of total A␤ was A␤Arc (Figure 5D). These findings suggest that A␤Arc/APP facilitates formation of diffuse A␤ deposits, which consist primarily of A␤1-42wt and possibly small amounts of A␤Arc. In old (age 24

2294 Lord et al AJP May 2011, Vol. 178, No. 5

months) tg-Swe mice, diffuse A␤42-immunoreactive deposits were also observed in areas both with and without transgene expression, for example, in the striatum (see Supplemental Figure S4 at http://ajp.amjpathol.org). Thus, A␤Arc strongly facilitated a process whereby diffuse deposits formed. It did not create a unique process of A␤ aggregation that does not normally occur in tg-Swe mice.

APPArcSwe Does Not Alter APPSwe Processing in Bitransgenic Mice APPArcSwe could possibly interact with APPSwe and, thereby, quantitatively or qualitatively influence its endoproteolysis, for example, through APP dimerization.30 This could explain different A␤ neuropathologic features in aged bitransgenic and tg-Swe mice. If so, alterations in the steady-state level of A␤ would be expected in young bitransgenic mice without extracellular A␤ deposition. The total A␤ level tended to be higher in formic acid extracts from brains of 2-month-old bitransgenic mice [mean (SEM), 8.0 (0.5) pmol/g A␤1-40 and 6.7 (0.3) pmol/g A␤1-42; n ⫽ 5] compared with age-matched tgSwe [7.0 (0.6) pmol/g A␤1-40; P ⫽ 0.28, and 6.0 (0.3) pmol/g A␤1-42; n ⫽ 6; P ⫽ 0.14] (see Supplemental Figure S5A at http://ajp.amjpathol.org); however, the A␤42/A␤40 ratio in the bitransgenic mice [0.84 (0.04)] and tg-Swe [0.85 (0.04)] did not differ (P ⫽ 0.75) (see Supplemental Figure S5B at http://ajp.amjpathol.org). When a total A␤Arc ELISA (A␤1-xArc) was used, the level of A␤Arc in bitransgenic mice was 1.60 (0.20) pmol/g (n ⫽ 5), and, as expected, was not detectable in tg-Swe (n ⫽ 5) (see Supplemental Figure S5C at http://ajp.amjpathol. org). Thus, total A␤ levels were higher in young bitransgenic mice than in age-matched tg-Swe mice; however, this increase was not more than what would be expected by the increased transgene expression (ie, the additional approximately one-eighth processing of APPArcSwe and A␤Arc synthesis in bitransgenic mice). Thus, it was concluded that APPArcSwe did not selectively increase A␤1-42wt synthesis from the APPSwe transgene in bitransgenic mice.

A␤Arc Facilitates Formation of Prefibrillar and Fibrillar Assemblies of A␤wt in Vitro The increased levels of diffuse deposits in bitransgenic mice could perhaps be explained by the ability of A␤Arc to facilitate refolding and aggregation of A␤wt. To investigate this, an in vitro system was designed to evaluate the influence of A␤1-42Arc on A␤1-42wt aggregation. LCOs enable staining and analysis of prefibrillar A␤ assemblies.16 Thus, the fluorescence of an LCO, p-FTAA, was used as a readout of prefibrillar A␤ aggregates, and that of ThT was used as a measure of amyloid fibril formation. Recombinant A␤1-42wt and A␤1-42Arc were treated with hexafluoroisopropanol to dissociate possible preformed A␤ aggregates that could seed fibril formation. The monomeric A␤ peptides were allowed to form fibrils, and the kinetics was monitored with ThT (Figure 6A) and p-FTAA (Figure 6B). A␤1-42wt and A␤1-42Arc were incubated

Figure 6. A mixture of A␤Arc and A␤wt accelerated prefibrillar A␤ aggregate formation in vitro. Aggregation of recombinant A␤1– 42 was monitored with ThT (A) and p-FTAA (B). Fibrillation kinetics of A␤1-42wt (open circles; 20 ␮mol/L), A␤1-42wt and A␤1-42Arc in a ratio of 7:1 (open triangles; 20 ␮mol/L total A␤), A␤1-42wt and A␤1-42Arc in a ratio of 1:1 (solid triangles; 20 ␮mol/L total A␤), and A␤1-42Arc (solid circles; 20 ␮mol/L). p-FTAA detects early events in the aggregation process.16 With both probes, the lag phase was shorter when A␤Arc was present. Each symbol represents the mean (SEM) of two or three experiments. Ultrastructural analysis of A␤ aggregates at electron microscopy. Samples of A␤1-42Arc (20 ␮mol/L; C, F, and I), A␤1-42wt and A␤1-42Arc at a ratio of 7:1 (20 ␮mol/L total A␤; D, G, and J), and A␤1-42wt (20 ␮mol/L; E, H, and K) were analyzed at 0 minutes (C–E), 100 minutes (F–H), and 1200 minutes (I–K). Scale bars: 100 nm (C–K).

either alone or together in ratios of 7:1 or 1:1 (A␤wt/ A␤Arc). The total concentration of A␤ (ie, A␤1-42wt and/or A␤1-42Arc either alone or in combination) was always 20 ␮mol/L. A lag phase was followed by an exponential growth phase and then a plateau phase. With both ThT and p-FTAA, the lag phase was shorter when A␤1-42wt and A␤1-42Arc were mixed before polymerization, at 7:1 or 1:1. The effect seemed to depend on the presence of A␤Arc but not on the percentages of A␤Arc and A␤wt. When mixing A␤1-42wt and A␤1-42Arc, the intensity of ThT fluorescence was noticeably lower in the plateau phase; however, this was not observed when using p-FTAA. The ultrastructure of in vitro preparations was then investigated. The starting material for all samples contained only small aggregates but no fibrils (Figure 6, C–E). An almost evenly distributed carpet of proteins was present at time 0 when A␤1-42wt and A␤1-

Selective Increase in Diffuse A␤ Plaques 2295 AJP May 2011, Vol. 178, No. 5

protofibrils, which are large soluble A␤ oligomers, were more abundant in Tris-buffered saline solution extracts in 12-month-old bitransgenic tg-ArcSwe and tg-Swe [mean (SEM), 163 (20) pmol/L; n ⫽ 11] than in age-matched tg-Swe [83 (13) pmol/L; n ⫽ 9; P ⬍ 0.01].

Discussion

Figure 7. Diffuse A␤ deposits in bitransgenic mice were stained using p-FTAA. Tissue sections 25 ␮m thick from a bitransgenic mouse were fixed in paraformaldehyde and stained with an A␤42-specific antibody together with either p-FTAA (A) or ThT (B). The area at the tip of the open arrow in A is magnified (C–E), and the area at the tip of the solid arrow in B is enlarged (F–H). Both dyes enabled visualization of compact plaques, and p-FTAA also stained numerous diffuse A␤ deposits (F–H). Scale bars: 1 mm (A and B); 100 ␮m (C–H).

42Arc were mixed (Figure 6D). At 100 minutes, when p-FTAA-fluorescence was increased, small clusters of thin fibrils were observed in samples consisting only of A␤1-42Arc (Figure 6F), whereas a few fibrillar aggregates with varying morphologic appearance were present at 100 minutes when A␤1-42wt and A␤1-42Arc were mixed (7:1; Figure 6G). In contrast, in samples that contained only A␤1-42wt peptide, p-FTAA fluorescence intensity was low (Figure 6B), and both fibrils and low-molecularweight aggregates were rare at this time point (Figure 6H). At 1200 minutes, incubation of A␤1-42Arc alone gave rise to an abundance of tightly aggregated thin fibrils (Figure 6I). In contrast, mixtures of A␤1-42wt and A␤1-42Arc contained large numbers of massive fibrillar aggregates. These mixtures of peptides formed thin fibrils that seemed to be composed of globular subunits (Figure 6J). Their appearance was different from the straighter fibrils observed with A␤1-42wt (Figure 6K).

Analyses of A␤ Protofibrils, p-FTAA Staining, and A␤42-Immunoreactivity in Bitransgenic Mice Staining with ThT or p-FTAA was combined with A␤ IHC on paraformaldehyde-fixed sections. Many but not all diffuse A␤42-immunoreactive deposits in the cerebral cortex and striatum were also stained using p-FTAA in an 18-month-old bitransgenic mouse (Figure 7, A, C–H). In contrast, there was no overlap between ThT staining and A␤42 immunoreactivity in the striatum of bitransgenic mice (see Supplemental Figure S6 at http://ajp.amjpathol. org). The same results were observed when thioflavine S was used instead of ThT (data not shown). Moreover, A␤

Enhanced deposition of small and diffuse A␤42-immunopositive deposits were observed in regions with and without transgene expression in bitransgenic mice producing both A␤wt and A␤Arc. In contrast, the burden of compact plaques was similar in bitransgenic and singly tg-Swe mice. Transgenic mice express only mutant human APP, for example, with the Arctic and/or Swedish mutations,5,31 whereas patients are heterozygous for the Arctic mutation or any other APP mutation and, therefore, produce a mixture of both A␤wt and A␤mutant peptides. The Arctic mutation results in a neuropathologic condition consisting of cerebral amyloid angiopathy and an abundance of parenchymal A␤ deposits lacking amyloid cores, and the symptoms are those of typical AD.2 In the present experiments, a mixture of A␤wt and A␤Arc in bitransgenic mice facilitated diffuse A␤ deposition. Recently, in a similar experimental approach, transgenic models APPDutch and APP23 were cross-bred. The APPDutch model mimics hereditary cerebral hemorrhage with amyloidosis–Dutch type,32,33 with cerebral amyloid angiopathy and diffuse A␤ deposits. APP23/APPDutch mice produced a mixture of A␤wt and A␤Dutch, and because A␤ levels were substantially higher in young APP23 mice than in age-matched APPDutch mice, the conditions were comparable to those of the present study. Parenchymal A␤ deposition was reduced in APP23/APPDutch, and vascular load with microhemorrhages was increased, which suggests that A␤Dutch interacted with A␤wt and altered the pathogenic phenotypes.34 Thus, direct physical interactions between A␤Arc/A␤wt and A␤Dutch/A␤wt may be relevant pathogenic mechanisms in patients heterozygous for either of the mutations, both of which are located in APP at the same amino acid position. The conditions in human brain affected with AD unquestionably differ from those of transgenic mouse brain because N- and C-terminal truncated A␤ that have undergone posttranslational modifications are far more prevalent in human brain.35,36 In the transgenic models used, the population of A␤ species was more homogeneous and dominated by a few types of A␤ peptides.10 If it is assumed that A␤Arc and A␤wt are produced in equimolar amounts in Arctic human brain, an interesting future experiment would be to crossbreed mice with equal amounts of A␤Arc and A␤wt. The effect, if any, of murine A␤ on extracellular A␤ deposition is likely small because A␤ pathologic features in APP23 mice with or without the murine APP gene were indistinguishable at the light and ultrastructural levels.37 Even overexpression of murine APP did not alter the extent of parenchymal A␤ deposition in APPSwe and PS1dE9 mice (line 85), but did increase the solubility of A␤wt42 and led to more cerebral amyloid angiopathy.38 Nonetheless,

2296 Lord et al AJP May 2011, Vol. 178, No. 5

ideally, bitransgenic mice should be further bred on an APP-knockout background. In general, considering the differences to human brain, experiments in transgenic models must be interpreted cautiously with respect to patients with AD, for example, those with the Arctic mutation. Parenchymal A␤ deposits in human brain are classified according to morphologic features, amyloid tinctorial characteristics, and neuritic component of the corona. Whether one plaque type converts to another and, if so, whether plaque evolution reflects disease progression have been long-standing questions.39 Diffuse A␤ deposits are observed in patients without dementia symptoms at death and in primates and dogs. They are commonly assumed to be less pathogenic or even inert, and little is known about their genesis. Few animal models are rich in diffuse A␤ deposits. PDAPP mice with a high metabolic A␤42 level are an exception.40 It has been suggested that aggregated A␤ in diffuse plaques form an amyloid core that becomes surrounded by dystrophic neurites (ie, emergence of a mature plaque), which are then removed via phagocytosis, leading to a “burnt out” plaque. The hypothesis has been challenged in cross-sectional studies of brains postmortem using IHC and various silver staining protocols.41 However, the conclusions rest on the assumption of AD as a homogeneous pathogenic entity. The strongest experimental evidence for the “plaque evolution theory” comes from investigations of patients with Down’s syndrome, who invariably develop dementia and neuropathologic features of AD due to trisomy of chromosome 21. Postmortem brains of different ages have, therefore, been investigated to understand disease progression. Extracellular diffuse deposits of A␤ ending at amino acid 42 (A␤x-42) were observed in 12-year-old children, whereas compact plaques with A␤ ending at amino acid 40 (A␤x-40) were not evident until age 30 years.42,43 Because diffuse A␤ deposits predate compact plaques in brains in patients with Down’s syndrome, it has been thought that with age they transform into compact plaques. In essence, formation of an amyloid core is a slow process involving extracellular A␤ accumulation in tissues, followed by structural modification, isomerization, and truncation of assembled A␤ peptides. In the present study, accelerated accumulation of diffuse deposits was observed that occurred independent of compact plaque formation. It was interpreted that diffuse and compact plaques can form via distinct processes, at least in transgenic mice. Consistent with this hypothesis, rapid formation of compact plaques has been demonstrated in transgenic models, which are known to develop few diffuse A␤ deposits.44,45 A␤ begins to accumulate inside neurons.46 The composition of intraneuronal A␤ aggregates could depend on the subcellular location, facilitators (eg, certain lipids or amyloid-associated factors), and origin (eg, A␤ metabolism in a particular neuronal subtype). Stable A␤ aggregates or A␤ monomers, which have been released from neurons, would then guide downstream events, leading to a specific type of extracellular A␤ deposit. The existence of prion strains illustrates that conformational differences can generate distinct pathologic phenotypes.47

Axonal transport and synaptic release are clearly involved in extracellular plaque formation.48,49 Thus, diffuse deposits in the striatum of bitransgenic mice were likely due to corticostriatal connections, resulting in axonal transport of APP/A␤ aggregates and/or extracellular aggregation after synaptic release of A␤ monomers. No compact plaques were observed in the striatum, at least at 18 months of age, again arguing that diffuse and compact plaques form via different mechanisms. An important question is whether A␤Arc co-aggregate with A␤wt in diffuse deposits or whether they behave more like enzymes and facilitate a conformational change in A␤wt without becoming a part of the final A␤wt aggregates. The latter idea is reminiscent to that of pathologic chaperones, heterologous proteins such as ApoE and ␣1-antichymotrypsin, both of which regulate formation of A␤ deposits26,27 and are thought to induce or stabilize the ␤-sheet conformation.50,51 Although both ApoE and antichymotrypsin stain A␤ deposits, they do not necessarily co-polymerize with A␤ in the fibrils.52 Data from the present study suggest that A␤Arc behaved as a pathologic chaperone; that is, diffuse deposits were not A␤Arcimmunoreactive. However, because small amounts of A␤Arc were detected in striatal tissue extracts at ELISA analysis, that A␤Arc to some extent became building blocks of the diffuse deposits cannot be ruled out. Heparan sulfate was found only in compact plaques, possibly suggesting that they might selectively favor formation of an amyloid core,28 whereas ApoE and antichymotrypsin facilitate both diffuse and compact extracellular A␤ deposition.26,27 Selective association of heparan sulfate with compact plaques is consistent with previous observations in brains affected with AD 53 and with the dependence of heparin-binding properties of A␤ on its aggregation state.54 Cross-breeding of bitransgenic mice with animals in which the biologic characteristics of heparan sulfate proteoglycans have been manipulated28 would be needed to further challenge the hypothesis of different pathways leading to either diffuse A␤ deposits or compact plaques. Mixtures of A␤Arc/A␤wt seemed to inhibit fibrillar but not prefibrillar growth in vitro compared with A␤Arc or A␤wt alone, possibly suggesting that the interaction occurred at an early stage. Interpreting differences in ThT signal in the plateau phase is not straightforward. The signal should reflect amyloid fibril concentration; however, it also depends on the number of available binding sites and the affinity of ThT for those sites in a heterogeneous sample, and on quantum yield.55 The percentage of A␤Arc of total A␤ in aged bitransgenic mice (⬍10%; Figure 5D) was lower than metabolic A␤Arc levels in young mice (⬎10% (Figure S5), although A␤Arc alone is much more prone to aggregate and form amyloid fibrils than is A␤wt, both in vitro and in vivo.4,5,13 Apparently, end products consisting of mixtures of A␤wt and A␤Arc did not easily form, perhaps because fibrils have an ordered structure. As demonstrated by others, A␤40wt interacted with A␤42wt and inhibited fibril formation both in vitro and in vivo.56,57 Moreover, the efficiency of seeded elongation of A␤40wt by A␤wt40 fibrils was superior to that of A␤40wt by A␤40Arc fibrils in vitro.58 Consistent with this,

Selective Increase in Diffuse A␤ Plaques 2297 AJP May 2011, Vol. 178, No. 5

it was observed that co-incubation of A␤Arc with A␤wt affected fibril structure in vitro and in vivo, as assessed at p-FTAA spectral analysis and electron microscopy. Others have reported that interactions between monomeric and protofibrillar A␤ interfered with the assembly and elongation of amyloid fibrils and stabilized A␤ protofibrils,59 and the present study detected higher A␤ protofibrillar levels in bitransgenic mice. Heterologous interactions between A␤ variants selectively increased diffuse A␤ deposits. That is, A␤Arc behaved as a facilitator; however, in principle, it illustrates that A␤ deposition is tractable to A␤ fragments or other binders serving as decelerators.

Acknowledgments We thank the Uppsala University Transgenic Facility for assistance in developing APP transgenic models and Paul O=Callaghan, B.Sc., for English corrections.

References 1. Hardy J: A hundred years of Alzheimer’s disease research. Neuron 2006, 52:3–13 2. Basun H, Bogdanovic N, Ingelsson M, Almkvist O, Näslund J, Axelman K, Bird TD, Nochlin D, Schellenberg GD, Wahlund LO, Lannfelt L: Clinical and neuropathological features of the Arctic APP gene mutation causing early-onset Alzheimer disease. Arch Neurol 2008, 65:499 –505 3. Nilsberth C, Westlind-Danielsson A, Eckman CB, Condron MM, Axelman K, Forsell C, Stenh C, Luthman J, Teplow DB, Younkin SG, Näslund J, Lannfelt L: The “Arctic” APP mutation (E693G) causes Alzheimer’s disease by enhanced Abeta protofibril formation. Nat Neurosci 2001, 4:887– 893 4. Johansson AS, Berglind-Dehlin F, Karlsson G, Edwards K, Gellerfors P, Lannfelt L: Physiochemical characterization of the Alzheimer’s disease–related peptides A beta 1-42Arctic and A beta 1-42wt. FEBS J 2006, 273:2618 –2630 5. Lord A, Kalimo H, Eckman C, Zhang XQ, Lannfelt L, Nilsson LN: The Arctic Alzheimer mutation facilitates early intraneuronal Abeta aggregation and senile plaque formation in transgenic mice. Neurobiol Aging 2006, 27:67–77 6. Stenh C, Englund H, Lord A, Johansson AS, Almeida CG, Gellerfors P, Greengard P, Gouras GK, Lannfelt L, Nilsson LN: Amyloid-beta oligomers are inefficiently measured by enzyme-linked immunosorbent assay. Ann Neurol 2005, 58:147–150 7. Englund H, Sehlin D, Johansson AS, Nilsson LN, Gellerfors P, Paulie S, Lannfelt L, Pettersson FE: Sensitive ELISA detection of amyloidbeta protofibrils in biological samples. J Neurochem 2007, 103: 334 –345 8. Sahlin C, Lord A, Magnusson K, Englund H, Almeida CG, Greengard P, Nyberg F, Gouras GK, Lannfelt L, Nilsson LN: The Arctic Alzheimer mutation favors intracellular amyloid-beta production by making amyloid precursor protein less available to alpha-secretase. J Neurochem 2007, 101:854 – 862 9. Lord A, Englund H, Söderberg L, Tucker S, Clausen F, Hillered L, Gordon M, Morgan D, Lannfelt L, Pettersson FE, Nilsson LN: Amyloidbeta protofibril levels correlate with spatial learning in Arctic Alzheimer’s disease transgenic mice. FEBS J 2009, 276:995–1006 10. Philipson O, HammarströM P, Nilsson KP, Portelius E, Olofsson T, Ingelsson M, Hyman BT, Blennow K, Lannfelt L, Kalimo H, Nilsson LN: A highly insoluble state of Abeta similar to that of Alzheimer’s disease brain is found in Arctic APP transgenic mice. Neurobiol Aging 2009, 30:1393–1405 11. Jarrett JT, Lansbury PT Jr: Seeding “one-dimensional crystallization” of amyloid: a pathogenic mechanism in Alzheimer’s disease and scrapie? Cell 1993, 73:1055–1058

12. Gajdusek DC: Transmissible and nontransmissible dementias: distinction between primary cause and pathogenetic mechanisms in Alzheimer’s disease and aging. Mt Sinai J Med 1988, 55:3–5 13. Cheng IH, Palop JJ, Esposito LA, Bien-Ly N, Yan F, Mucke L: Aggressive amyloidosis in mice expressing human amyloid peptides with the Arctic mutation. Nat Med 2004, 10:1190 –1192 14. Prusiner SB: Molecular biology of prion diseases. Science 1991, 252:1515–1522 15. Näslund J, Haroutunian V, Mohs R, Davis KL, Davies P, Greengard P, Buxbaum JD: Correlation between elevated levels of amyloid beta-peptide in the brain and cognitive decline. JAMA 2000, 283: 1571–1577 16. Åslund A, Sigurdson CJ, Klingstedt T, Grathwohl S, Bolmont T, Dickstein DL, Glimsdal E, Prokop S, Lindgren M, Konradsson P, Holtzman DM, Hof PR, Heppner FL, Gandy S, Jucker M, Aguzzi A, Hammarström P, Nilsson KP: Novel pentameric thiophene derivatives for in vitro and in vivo optical imaging of a plethora of protein aggregates in cerebral amyloidoses. ACS Chem Biol 2009, 4:673– 684 17. Nilsson LN, Bales KR, DiCarlo G, Gordon MN, Morgan D, Paul SM, Potter H: Alpha-1-antichymotrypsin promotes beta-sheet amyloid plaque deposition in a transgenic mouse model of Alzheimer’s disease. J Neurosci 2001, 21:1444 –1451 18. Wiltfang J, Arold N, Neuhoff V: A new multiphasic buffer system for sodium dodecyl sulfate–polyacrylamide gel electrophoresis of proteins and peptides with molecular masses 100,000 –1000, and their detection with picomolar sensitivity. Electrophoresis 1991, 12:352–366 19. Klafki HW, Wiltfang J, Staufenbiel M: Electrophoretic separation of betaA4 peptides (1– 40) and (1– 42). Anal Biochem 1996, 237:24 –29 20. Walsh DM, Lomakin A, Benedek GB, Condron MM, Teplow DB: Amyloid beta-protein fibrillogenesis: detection of a protofibrillar intermediate. J Biol Chem 1997, 272:22364 –22372 21. LeVine H III: Thioflavine T interaction with synthetic Alzheimer’s disease beta-amyloid peptides: detection of amyloid aggregation in solution. Protein Sci 1993, 2:404 – 410 22. Mucke L, Masliah E, Yu GQ, Mallory M, Rockenstein EM, Tatsuno G, Hu K, Kholodenko D, Johnson-Wood K, McConlogue L: High-level neuronal expression of abeta 1– 42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation. J Neurosci 2000, 20:4050 – 4058 23. Sigurdson CJ, Nilsson KP, Hornemann S, Manco G, Polymenidou M, Schwarz P, Leclerc M, HammarströM P, Wuthrich K, Aguzzi A: Prion strain discrimination using luminescent conjugated polymers. Nat Methods 2007, 4:1023–1030 24. Nilsson KP, Åslund A, Berg I, Nystrom S, Konradsson P, Herland A, Inganäs O, Stabo-Eeg F, Lindgren M, Westermark GT, Lannfelt L, Nilsson LN, HammarströM P: Imaging distinct conformational states of amyloid-beta fibrils in Alzheimer’s disease using novel luminescent probes. ACS Chem Biol 2007, 2:553–560 25. Berg I, Nilsson KP, Thor S, Hammarström P: Efficient imaging of amyloid deposits in Drosophila models of human amyloidoses. Nat Protoc 2010, 5:935–944 26. Nilsson LN, Arendash GW, Leighty RE, Costa DA, Low MA, Garcia MF, Cracciolo JR, Rojiani A, Wu X, Bales KR, Paul SM, Potter H: Cognitive impairment in PDAPP mice depends on ApoE and ACTcatalyzed amyloid formation. Neurobiol Aging 2004, 25:1153–1167 27. Bales KR, Verina T, Dodel RC, Du Y, Altstiel L, Bender M, Hyslop P, Johnstone EM, Little SP, Cummins DJ, Piccardo P, Ghetti B, Paul SM: Lack of apolipoprotein E dramatically reduces amyloid beta-peptide deposition. Nat Genet 1997, 17:263–264 28. Li JP, Galvis ML, Gong F, Zhang X, Zcharia E, Metzger S, Vlodavsky I, Kisilevsky R, Lindahl U: In vivo fragmentation of heparan sulfate by heparanase overexpression renders mice resistant to amyloid protein A amyloidosis, Proc Natl Acad Sci USA 2005, 102:6473– 6477 29. DeMattos RB, O’Dell MA, Parsadanian M, Taylor JW, Harmony JA, Bales KR, Paul SM, Aronow BJ, Holtzman DM: Clusterin promotes amyloid plaque formation and is critical for neuritic toxicity in a mouse model of Alzheimer’s disease, Proc Natl Acad Sci USA 2002, 99: 10843–10848 30. Scheuermann S, Hambsch B, Hesse L, Stumm J, Schmidt C, Beher D, Bayer TA, Beyreuther K, Multhaup G: Homodimerization of amyloid precursor protein and its implication in the amyloidogenic pathway of Alzheimer’s disease. J Biol Chem 2001, 276:33923–33929

2298 Lord et al AJP May 2011, Vol. 178, No. 5

31. Rönnbäck A, Zhu S, Dillner K, Aoki M, Lilius L, Näslund J, Winblad B, Graff C: Progressive neuropathology and cognitive decline in a single Arctic APP transgenic mouse model. Neurobiol Aging 2011, 32:280 – 292 32. Levy E, Carman MD, Fernandez-Madrid IJ, Power MD, Lieberburg I, van Duinen SG, Bots GT, Luyendijk W, Frangione B: Mutation of the Alzheimer’s disease amyloid gene in hereditary cerebral hemorrhage, Dutch type. Science 1990, 248:1124 –1126 33. Herzig MC, Winkler DT, Burgermeister P, Pfeifer M, Kohler E, Schmidt SD, Danner S, Abramowski D, Sturchler-Pierrat C, Burki K, van Duinen SG, Maat-Schieman ML, Staufenbiel M, Mathews PM, Jucker M: Abeta is targeted to the vasculature in a mouse model of hereditary cerebral hemorrhage with amyloidosis. Nat Neurosci 2004, 7:954 –960 34. Herzig MC, Eisele YS, Staufenbiel M, Jucker M: E22Q-mutant Abeta peptide (AbetaDutch) increases vascular but reduces parenchymal Abeta deposition. Am J Pathol 2009, 174:722–726 35. Mori H, Takio K, Ogawara M, Selkoe DJ: Mass spectrometry of purified amyloid beta protein in Alzheimer’s disease. J Biol Chem 1992, 267:17082–17086 36. Kalback W, Watson MD, Kokjohn TA, Kuo YM, Weiss N, Luehrs DC, Lopez J, Brune D, Sisodia SS, Staufenbiel M, Emmerling M, Roher AE: APP transgenic mice Tg2576 accumulate Abeta peptides that are distinct from the chemically modified and insoluble peptides deposited in Alzheimer’s disease senile plaques. Biochemistry 2002, 41: 922–928 37. Calhoun ME, Burgermeister P, Phinney AL, Stalder M, Tolnay M, Wiederhold KH, Abramowski D, Sturchler-Pierrat C, Sommer B, Staufenbiel M, Jucker M: Neuronal overexpression of mutant amyloid precursor protein results in prominent deposition of cerebrovascular amyloid. Proc Natl Acad Sci USA 1999, 96:14088 –14093 38. Jankowsky JL, Younkin LH, Gonzales V, Fadale DJ, Slunt HH, Lester HA, Younkin SG, Borchelt DR: Rodent A beta modulates the solubility and distribution of amyloid deposits in transgenic mice. J Biol Chem 2007, 282:22707–22720 39. Armstrong RA: Beta-amyloid plaques: stages in life history or independent origin? Dement Geriatr Cogn Disord 1998, 9:227–238 40. Johnson-Wood K, Lee M, Motter R, Hu K, Gordon G, Barbour R, Khan K, Gordon M, Tan H, Games D, Lieberburg I, Schenk D, Seubert P, McConlogue L: Amyloid precursor protein processing and A beta42 deposition in a transgenic mouse model of Alzheimer disease. Proc Natl Acad Sci USA 1997, 94:1550 –1555 41. Gibson PH: Relationship between numbers of cortical argentophilic and congophilic senile plaques in the brains of elderly people with and without senile dementia of Alzheimer type. Gerontology 1985, 31:321–324 42. Lemere CA, Blusztajn JK, Yamaguchi H, Wisniewski T, Saido TC, Selkoe DJ: Sequence of deposition of heterogeneous amyloid betapeptides and APO E in Down syndrome: implications for initial events in amyloid plaque formation. Neurobiol Dis 1996, 3:16 –32 43. Iwatsubo T, Mann DM, Odaka A, Suzuki N, Ihara Y: Amyloid beta protein (A beta) deposition: a beta 42(43) precedes A beta 40 in Down syndrome. Ann Neurol 1995, 37:294 –299

44. Meyer-Luehmann M, Spires-Jones TL, Prada C, Garcia-Alloza M, de Calignon A, Rozkalne A, Koenigsknecht-Talboo J, Holtzman DM, Bacskai BJ, Hyman BT: Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer’s disease. Nature 2008, 451:720 –724 45. Yan P, Bero AW, Cirrito JR, Xiao Q, Hu X, Wang Y, Gonzales E, Holtzman DM, Lee JM: Characterizing the appearance and growth of amyloid plaques in APP/PS1 mice. J Neurosci 2009, 29:10706 –10714 46. Philipson O, Lannfelt L, Nilsson LN: Genetic and pharmacological evidence of intraneuronal Abeta accumulation in APP transgenic mice. FEBS Lett 2009, 583:3021–3026 47. Prusiner SB: Prions. Proc Natl Acad Sci USA 1998, 95:13363–13383 48. Lazarov O, Lee M, Peterson DA, Sisodia SS: Evidence that synaptically released beta-amyloid accumulates as extracellular deposits in the hippocampus of transgenic mice. J Neurosci 2002, 22:9785–9793 49. Sheng JG, Price DL, Koliatsos VE: Disruption of corticocortical connections ameliorates amyloid burden in terminal fields in a transgenic model of Abeta amyloidosis. J Neurosci 2002, 22:9794 –9799 50. Wisniewski T, Frangione B: Apolipoprotein E: a pathological chaperone protein in patients with cerebral and systemic amyloid. Neurosci Lett 1992, 135:235–238 51. Potter H, Wefes IM, Nilsson LN: The inflammation-induced pathological chaperones ACT and Apo-E are necessary catalysts of Alzheimer amyloid formation. Neurobiol Aging 2001, 22:923–930 52. Söderberg L, Bogdanovic N, Axelsson B, Winblad B, Naslund J, Tjernberg LO: Analysis of single Alzheimer solid plaque cores by laser capture microscopy and nanoelectrospray/tandem mass spectrometry. Biochemistry 2006, 45:9849 –9856 53. O’Callaghan P, Sandwall E, Li JP, Yu H, Ravid R, Guan ZZ, van Kuppevelt TH, Nilsson LN, Ingelsson M, Hyman BT, Kalimo H, Lindahl U, Lannfelt L, Zhang X: Heparan sulfate accumulation with Abeta deposits in Alzheimer’s disease and Tg2576 mice is contributed by glial cells. Brain Pathol 2008, 18:548 –561 54. Watson DJ, Lander AD, Selkoe DJ: Heparin-binding properties of the amyloidogenic peptides Abeta and amylin: dependence on aggregation state and inhibition by Congo red. J Biol Chem 1997, 272: 31617–31624 55. LeVine H III: Quantification of beta-sheet amyloid fibril structures with thioflavin T. Methods Enzymol 1999, 309:274 –284 56. Hasegawa K, Yamaguchi I, Omata S, Gejyo F, Naiki H: Interaction between A beta(1– 42) and A beta(1– 40) in Alzheimer’s beta-amyloid fibril formation in vitro. Biochemistry 1999, 38:15514 –15521 57. Kim J, Onstead L, Randle S, Price R, Smithson L, Zwizinski C, Dickson DW, Golde T, McGowan E: Abeta40 inhibits amyloid deposition in vivo. J Neurosci 2007, 27:627– 633 58. O’Nuallain B, Williams AD, Westermark P, Wetzel R: Seeding specificity in amyloid growth induced by heterologous fibrils. J Biol Chem 2004, 279:17490 –17499 59. Lashuel HA, Hartley DM, Petre BM, Wall JS, Simon MN, Walz T, Lansbury PT Jr: Mixtures of wild-type and a pathogenic (E22G) form of Abeta40 in vitro accumulate protofibrils, including amyloid pores. J Mol Biol 2003, 332:795– 808