RAGE Does Not Affect Amyloid Pathology in Transgenic ArcA Mice

Background: Alzheimer’s disease (AD) is characterized by brain accumulation of the amyloidpeptide (A ) that triggers a cascade of biochemical and cellular alterations resulting in the clinical phenotype of the disease. While numerous experiments addressed A toxicity, the mechanisms are still not fully understood. The receptor for advanced glycation end products (RAGE) binds A and was suggested to be involved in the pathological processes of AD. Objective: Our purpose was to assess the effect of RAGE deletion on A -related pathology. Methods: We crossed RAGE knockout (RAGE –/– ) mice with transgenic mice harboring both the Swedish and Arctic A precursor protein mutations (arcA mice). We assessed A levels, A brain deposition, A -degrading enzyme activities, A precursor protein expression and processing, number and morphology of microglia as well as cognitive performance of 6and 12-month-old RAGE –/– /arcA , RAGE –/– , arcA and wild-type mice. Results: RAGE –/– /arcA mice had significantly lower levels of SDSand formic-acid-extracted A in the cortex and hippocamReceived: September 9, 2009 Accepted after revision: November 18, 2009 Published online: February 10, 2010 D i s e a s e s Ivana Vodopivec Division of Psychiatry Research, University of Zurich August Forel-Strasse 1 CH–8008 Zurich (Switzerland) Tel. +41 44 634 88 86, Fax +41 44 634 88 74, E-Mail ivana.vodopivec @ bli.uzh.ch © 2010 S. Karger AG, Basel 1660–2854/09/0066–0270$26.00/0 Accessible online at: www.karger.com/ndd D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 1 :3 8: 54 A M RAGE Does Not Affect Amyloid Pathology in Transgenic ArcA Mice Neurodegenerative Dis 2009;6:270–280 271 end products (RAGE), a multiligand receptor belonging to the immunoglobulin superfamily [1] . To determine the phenotype of brain -amyloidosis in the absence of RAGE, we developed a mouse model, RAGE –/– /arcA . ArcA mice, which overexpress the human  -amyloid precursor protein (APP) with both the Swedish and the Arctic mutations, show first cognitive deficits at the age of 6 months [2] . At that age, the animals have abundant A intracellular punctate deposits in the cortex and hippocampus with no apparent A plaque load. At 12 months of age, A plaques and cerebral amyloid angiopathy (CAA) are prominent features. Therefore, we analyzed the RAGE –/– /arcA mice at 6 and 12 months of age, and tested their cognitive abilities, biochemical and histopathological changes, including brain A levels, plaque load, CAA and microglial morphology. RAGE –/– /arcA animals showed significantly lower levels of SDSand formic-acid-extractable A in the cortex and hippocampus presumably related to concomitant increases in insulin-degrading enzyme (IDE) activity, as compared with arcA mice at the age of 6 months. However, these differences did not result in better cognitive performance and were no longer detectable 6 months later when massive A accumulation occurs. Materials and Methods Animals RAGE knockout (RAGE –/– ) mice [3] were crossed with arcA mice [2] to generate the RAGE –/– /arcA mouse model. The arcA model is based on the overexpression of human APP 695 with the Swedish (K670N/M671L) and the Arctic (E693G) mutations in a single construct under the control of the prion protein promoter. There were 4 genotypes investigated in parallel: RAGE –/– / arcA mice, their respective RAGE –/– littermates, arcA and their respective wild-type littermates. Sixand 12-month-old mice, 8–9 animals per genotype and age, were analyzed. Each animal group was balanced for gender. The mice were kept under standard housing conditions and a reversed 12-hour light/dark cycle, with free access to food and water. The animal experiments were approved by the Cantonal Veterinary Authority of Zurich. The abrogation of RAGE expression in RAGE –/– animals was demonstrated by Western blot of lung tissue, both directly by detecting no RAGE signal and indirectly by demonstrating reporter EGFR expression (online supplementary fig. 1, www. karger.com/doi/10.1159/000261723). Y Maze The spontaneous alternation rate was assessed using a Yshaped plastic maze, with 40 ! 20 ! 10 cm arm sizes. During 5-min sessions, the sequences of arm entries were recorded; alternation was defined as successive entries into 3 arms, in overlapping triplet sets. The percent alternation was calculated as the ratio of actual to possible alternations (defined as the total number of arm entries – 2) ! 100%. The behavioral testing was performed in the dark phase. Tissue Preparation Within 1 week of Y maze testing the mice were anesthetized with ketamine-xylazine cocktail (10 mg/kg, i.p.) and flush-perfused transcardially with ice-cold phosphate-buffered saline. Brains were rapidly removed and divided sagittally. One hemibrain was postfixed in phosphate-buffered 4% paraformaldehyde and paraffin-embedded; the other hemibrain was dissected into cortex, hippocampus and cerebellum snap-frozen and stored at –80 ° C for protein analysis. Histology Five-micrometer-thick sagittal sections were cut with a Leica RM 2135 microtome. The sections were pretreated with citrate buffer (20 min in microwave at 85 ° C) followed by 95% formic acid (FA) [(5 min at room temperature (RT)], and blocked with 4% bovine serum albumin, 5% goat serum and 5% horse serum at RT for 1 h. The sections were subsequently incubated with primary antibodies overnight at 4 ° C at the following dilutions: mouse 6E10 (1: 500); rabbit anti-Iba1 (1: 500). Secondary fluorophoreconjugated antibodies were used for 2 h at RT. Thioflavin S staining was done according to standard protocol. Automated plaque and microglia counting was performed on 6 sections from each animal, spaced 50 m apart, using the software Image J (http:// rsb.info.nih.gov/ij/). The measurements were limited to the cortex, which was analyzed in the frontal and parietal regions, and the hippocampus. Three-Step A Extraction Protocol Cortices were homogenized in 10 tissue volumes (w/v) of 1% Triton X-100 buffer (50 m M of Tris-HCl, pH 8.0/1% Triton X-100/ complete protease inhibitor cocktail, Roche) with Teflon glass homogenizer (40 strokes) and subsequently centrifuged at 100,000 g for 1 h at 4 ° C. The supernatant was retained as the 1% Triton X100 soluble fraction. The resulting pellet was homogenized in 10 tissue volumes (w/v) of 2% SDS buffer (50 m M of Tris-HCl, pH 8.0/2% SDS/2 m M EDTA/complete protease inhibitor cocktail EDTA-free, Roche) and centrifuged at 100,000 g for 1 h at 8 ° C. The ensuing supernatant was collected as the 2% SDS soluble fraction. Given the small amount of tissue obtained per hemibrain, mouse hippocampi were homogenized in 10 tissue volumes (w/v) of 1% Triton X-100, 2% SDS buffer (50 m M of Tris-HCl, pH 8.0/1% Triton X-100/2% SDS/complete protease inhibitor cocktail, Roche) with Teflon glass homogenizer (40 strokes) and subsequently centrifuged at 100,000 g for 1 h at 8 ° C. Two percent SDS pellets from cortices and hippocampi were further extracted with a minimum of 225  l of 70% FA and spun at 100,000 g for 1 h at 8 ° C. The FA supernatant was collected by aspiration, avoiding the surface lipid layer and the acid insoluble bottom pellet. Aliquots of the 70% FA soluble fraction were lyophilized overnight in order to remove the acid and stored at –80 ° C. Immunoblots Tissue extracts were separated by SDS-PAGE using 10–20% Tricine gels (Invitrogen) and transferred onto nitrocellulose membranes. After an optional epitope retrieval and blocking in 5% milk TBS-T, the membranes were probed with 6E10 (1: 500, Signet), 22C11 (1: 1,000, Chemicon International), anti-APP C-terminal D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 1 :3 8: 54 A M Vodopivec /Galichet /Knobloch /Bierhaus / Heizmann /Nitsch Neurodegenerative Dis 2009;6:270–280 272 (1: 4,000, Sigma), anti-mouse/rat RAGE (1: 500, R&D Systems; 1: 1,000 rabbit polyclonal antibodies against RAGE V, C1 and C2 domains [4] ) and anti-GFP (1: 1,000, Roche) antibodies. Immunopositive bands were visualized by chemiluminescence (ECL, Amersham Biosciences) and subsequently quantified by densitometric analysis of their intensity levels. -Tubulin was used as loading control (1: 10,000, Sigma). The results were expressed as signal intensities normalized to values pertaining to arcA animals. A 40 and A 42 ELISA of Cerebral A The A 40 and A 42 quantities in the above-mentioned fractions were determined using hAmyloid 40 and hAmyloid  42 ELISA kits (The Genetics Company AG, Switzerland). Enzymatic Assays Cortical tissue from 4 mice per genotype was pooled. The pooled tissue was homogenized in a 4-fold (w/v) volume of 20 m M Tris-HCl (pH 7.4), 0.5% Triton X-100, 10 M PMSF. Homogenate was centrifuged at 13,000 rpm for 15 min at 4 ° C. Supernatant was used as a sample for the assay. The protein concentration was determined with the BCA Protein Assay (Pierce) and measured against bovine serum albumin standards. IDE Activity Assay. The IDE activity in the supernatant was determined with the fluorometric Innozyme Insulysin/IDE Immunocapture Activity Assay Kit (Calbiochem, Merck) according to the manufacturer’s instructions. The reaction time was 1 h. The IDE activity was expressed in nanograms/milliliter of crude cortical homogenates that were normalized to 2 mg/ml protein concentration. Angiotensin-Converting Enzyme Activity Assay. The angiotensin-converting enzyme (ACE) activity in the supernatant against the synthetic substrate N-hippurylL -histidylL -leucine was determined using an ACE colorimetric kit (Buhlmann Laboratories AG, Schönenbuch, Switzerland). The reaction time was 3 h. Neprilysin Activity Assay. The neprilysin enzyme activity was measured as described previously [5, 6] , using a fluorometric assay for the generation of free dansyl-D-Ala-Gly (DAG) from N dansyl-Ala-Gly-D-nitro-Phe-Gly (DAGNPG; Sigma), a fluorogenic substrate for neprilysin. Substrate solutions containing 1 m M of DAGPNG and 10 M of enalapril (Sigma) in 20 m M of Tris HCl with and without the addition of 10 M of NEP inhibitor phosphoramidon were prepared and preincubated at 37 ° C for 10 min. Enalapril, an ACE inhibitor, was added to prevent ACE-mediated DAGNPG cleavage. Fifty micorliters of the sample were incubated for 1 h with 100  l of each substrate solution. The reaction was stopped by boiling for 10 min at 90 ° C. The samples were then spun for 5 min at 10,000 rpm. The fluorescence of the supernatants was measured at an emission wavelength of 562 nm and an excitation wavelength of 342 nm. The NEP activity was calculated as a difference in fluorescence between samples incubated with and without phosphoramidon, and expressed as relative fluorescence units. Statistical Analysis The data were analyzed using SPSS 14.0 (SPSS Inc., Chicago, Ill., USA). Differences between the means were assessed by Student’s t test or 1-way ANOVA followed by Fisher’s LSD or Tukey’s post-hoc tests. Correlation studies were performed by parametric correlation and linear regression analysis. The null hypothesis was rejected at the 0.05 level. Results Absence of RAGE Reduced the Amount of Insoluble A in 6-Month-Old ArcA Mice In order to evaluate the influence of RAGE deletion (online suppl. fig. 1) on A deposition, we examined the A peptide levels in the brain regions relevant to cognition, i.e. the hippocampus and cortex. The A levels were assessed with respect to the extraction conditions and the variant length (A 40 and A 42) using Western blot and C-terminus-specific ELISA. The brains were extracted sequentially in 1% Triton X-100, 2% SDS and 70% FA, to allow for the isolation of A species that differ in the aggregation state and cellular/extracellular localization, with SDS-extractable A pertaining to the membraneenriched fraction and FA-extractable (insoluble) A representing extracellular amyloid. Both the Western blot and ELISA approaches indicated significantly lower overall levels of accumulated A in the cortex and hippocampus at the age of 6 months in RAGE –/– /arcA animals compared to arcA mice with normal RAGE expression. Interestingly, the difference was due to the low A levels extracted into SDS and FA, but not caused by Triton X100-requiring peptide ( tables 1 and 2 ). Despite the clear differences in the amount of extractable A found by biochemical methods, histological analysis did not reveal any difference in A -occupied brain areas at 6 months ( fig. 1 a), possibly due to the fact that at this age, visible plaque formation has not yet started in arcA mice [2] . At the age of 12 months, when A deposition has progressed much more, biochemical and histopathological approaches revealed no more difference in the total A load, nor A 40 and A 42 variants, in any of the fractions between RAGE –/– /arcA and arcA mice ( tables 1 and 2 ; fig. 1 b). However, the A 40 level was significantly decreased in the serum of RAGE –/– /arcA animals. Given the possibility that the genetic manipulation might not only influence A deposition and quantities but also morphological features of deposits, we examined both plaques and CAA but found no structural differences between RAGE –/– /arcA and arcA mice ( fig. 2 and 3 ). Plaques in RAGE –/– /arcA animals, detected at 12 months ( fig. 3 a, g), had the same dense-core morphology as arcA mice ( fig. 3 b, d, h, j) and were also accompanied by severe CAA ( fig. 3 c, i). RAGE Did Not Influence APP Expression and Processing To determine whether the decreased amount of A could be due to altered APP expression or processing D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 1 :3 8: 54 A M RAGE Does Not Affect Amyloid Pathology in Transgenic ArcA Mice Neurodegenerative Dis 2009;6:270–280 273 caused by the absence of RAGE, we examined the levels of full-length APP (APP), soluble APP (sAPP) as well as C-terminal fragments produced by  -secretase (CTF) by Western blot using 6E10, 22C11 and anti-APP C-terminal antibodies. There was no difference in APP, sAPP and CTF levels according to the densitometric analysis of immunoblots upon normalization to -tubulin in the cortex and the hippocampus between the 2 transgenic groups ( fig. 4 ). RAGE –/– /ArcA Mice Showed Increased A -Degrading Enzymatic Activity Our finding that the cerebral levels of SDSand FAsoluble A were reduced in the brains of 6-month-old mice without any change in APP processing suggested that the activity of 1 or more A -degrading proteases might be increased. To test this hypothesis, we examined the enzymatic activity of 3 proteases in the cortices of 6and 12-month-old animals. Analysis of IDE activity revealed significant main effects of the APP transgene [F(1,16) = 28.218, p = 0.001] and of the RAGE knockout condition [F(1,16) = 37.102, p ! 0.001], with no interaction, indicating their additive effect. At the age of 6 months, IDE in RAGE –/– /arcA had a significantly higher activity than in arcA mice (p = 0.05), RAGE –/– mice (p = 0.11) and wild-type animals (p ! 0.001; fig. 5 a). However, at the age of 12 months the difference between RAGE –/– /arcA and arcA was not statistically significant; RAGE –/– /arcA animals had an increased activity of the enzyme only in comparison with the wild-type animals (p = 0.043). There was no significant difference in the activity of 2 other A -degrading enzymes, ACE and neprilysin, among the 4 groups ( fig. 5 b, c). Absence of RAGE Did Not Ameliorate Cognitive Impairment To clarify whether the reduced amount of A in RAGE –/– /arcA mice at 6 months would result in less seTable 1. A levels measured by ELISA specific for A x-40 and A x-42 Age months Genotype A Cortex Hippocampus Serum pg/ml 1% Triton ng/ml 2% SDS ng/ml 70% FA ng/ml 1% Triton, 2% SDS, ng/ml 70% FA ng/ml 6 RAGE–/–/arcA A 40 4.5280.02 4.6280.09 1.5980.48* 3.4780.13 ND ND A 42 6.3780.10 4.5980.13* 1.3080.50* 2.2480.04 ND ND arcA A 40 4.6180.03 6.0280.62 11.3483.53 3.5680.21 ND ND A 42 6.6880.15 6.8880.87 6.5282.02 1.7680.33 ND ND 12 RAGE–/–/arcA A 40 27.7388.93 143.48841.50 380.488111.04 20.3884.33 65.18818.37 258.51832.12* A 42 5.9281.07 50.1486.49 101.47842.07 5.5680.65 25.2785.35 ND arcA A 40 15.7188.83 62.41812.58 402.398140.74 11.8082.88 96.35833.93 432.70859.93 A 42 4.0181.19 44.3885.90 97.64833.33 5.2980.92 36.13810.80 ND Values are means 8 SE. ND = Not detectable. * p < 0.05. Table 2. Monomeric A levels determined by densitometric analysis of Western blot assay with 6E10 antibody Age months Genotype Cortex Hippocampus 1% Triton 2% SDS 70% FA 1% Triton, 2% SDS 70% FA 6 RAGE–/–/arcA 96.0832.5 10.383.7** 7.783.9*** 67.0823.2 3.3681.8* arcA 100.086.5 100.0835.0 100.0823.1 100.0830.1 100.0855.3 12 RAGE–/–/arcA 222.1888.8 164.7848.6 101.6824.7 76.2818.9 83.6825.7 arcA 100.0833.0 100.0812.4 100.0818.3 100.0815.7 100.0820.5 Data are mean relative intensity 8 SE. * p < 0.05; ** p <0.01; *** p < 0.001. D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 1 :3 8: 54 A M Vodopivec /Galichet /Knobloch /Bierhaus / Heizmann /Nitsch Neurodegenerative Dis 2009;6:270–280 274 vere cognitive deficits that had previously been described in arcA mice, the different genotype groups were tested in the Y maze, a working memory test. There was no significant difference in Y maze performance between the RAGE –/– /arcA and arcA animal groups at the age of 6 and 12 months. Although they had less SDSand FAsoluble A load at 6 months, the RAGE –/– /arcA mice had the same extent of cognitive deficits as the arcA animals. This impairment was not due to the lack of RAGE, since the RAGE –/– mice performed equally well as the wild-type animals ( fig. 6 ). Cognitive Impairment Correlated with Triton-X-Extractable A Given the cognitive deficits in the animals harboring the APP transgene, we evaluated whether the extent of accumulated A was related to the degree of cognitive performance in individual mice. There was a significant negative correlation between the percent alternation in the Y maze and monomeric A cortical levels present in 1% Triton X-100 fraction determined by densitometric analysis of Western blot with 6E10 antibody but not with the A levels in other fractions or in the hippocampus. Additionally, there was no significant correlation between Y-maze performance and the levels of 2 length variants, A 40 or A 42. The association between soluble A load in the cortical area and working memory impairment was significant at the age of 6 months for both RAGE –/– /arcA (p = 0.0146) and arcA mice (p = 0.043) ( fig. 7 a), and the r values obtained for the 2 groups were not significantly different. At 12 months of age, the negative correlation was only significant for the RAGE –/– / arcA mice (p = 0.020; fig. 7 b). 0 0.05 O cc up ie d br ai n ar ea (% )