Brains from non-Alzheimer’s individuals containing amyloid deposits accelerate Aβ deposition in vivo

AD (79 years old, female, AD clinical diagnosis that was confirmed post-mortem), NDAN (81 years old, male, non-demented diagnosis) and aged control (59 years old, male, non-demented diagnosis) brain samples were obtained from the National Disease Research Interchange (Philadelphia, PA, USA). MCI (95 years old, male, MCI diagnosis by neuropsychological test) sample was kindly provided by Dr. Eliezer Masliah (University of California at San Diego, California, USA). All areas analyzed corresponded to the cingulate cortex. Research on human samples was performed following The Code of Ethics of the World Medical Association (Declaration of Helsinki). Informed consent was obtained for experimentation with human subjects. Samples were manipulated following the universal precautions for working with human samples and as directed by the Institutional Review Board of The University of Texas Medical School at Houston.

APP_Swe/PSEN1_∆E9 transgenic mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). These mice over-express the human version of amyloid precursor protein (APP) harboring the Swedish double mutation (K670M and N671L) and the human presenilin-1 protein with the DeltaE9 mutation (PSEN1-∆E9) [33]. Treated animals were housed in groups of up to 5 in individually ventilated cages under standard conditions (22°C, 12 h light–dark cycle) receiving food and water ad libitum. All animal manipulation was in agreement with NIH guidelines and approved by the Animal Welfare Committee of the University of Texas – Medical School at Houston. 5–8 animals per experimental group were used as indicated in each section. Males and females were indistinctly used (overall 43% males, 57% females).

Frozen cingulate cortex sections of AD, MCI, NDAN and aged non-demented individuals were homogenized at 10% (w/v) in ice-cold PBS containing a cocktail of protease inhibitors (Roche Diagnostics GmbH, Mannheim, Germany). Resulting homogenates were stored at −80°C until used for animal injection. In order to measure the amount of insoluble Aβ in each case, 200 μL aliquots of each sample were centrifuged at 100,000×g for 1 h and 4°C using a L100K Beckman-Coulter ultracentrifuge (Beckman-Coulter, Brea, CA, USA). Supernatants were recovered and saved as “PBS fractions” and pellets were resuspended in 200 μL of a 2% SDS solution by pipetting and sonication. Samples were centrifuged as explained above and the supernatants were diluted 40 times in EC buffer (0.02 M phosphate buffer, pH 7, 0.4 M NaCl, 2 mM bovine serum albumin, 0.05% CHAPS and 0.05% sodium azide). Pellets were resuspended in 200 μL of 70% formic acid (FA) and centrifuged for 30 minutes using the same temperature and speed previously described. Resulting supernatants ("FA fractions") were diluted 20 times in 1 M Tris buffer (pH 11) to adjust pH. All resulting samples were stored at −80°C until measured by an ELISA kit able to specifically detect Aβ₄₀ and Aβ₄₂ (Invitrogen, Carlsbad, CA, USA). Histological characterization was performed as described below.

~30 days-old APP_Swe/PSEN1_∆E9 mice were intra-cerebrally injected with 10 μL of 10% (w/v) human cingulate cortex homogenate. Briefly, mice were anesthetized using isofluorane. Skin was incised and a small hole was drilled in the skull. Samples (brain homogenates and PBS) were injected into the right hippocampus using the following coordinates as measured from bregma: antero-posterior (AP), -1.8 mm; medio-lateral (ML), -1.8 mm; dorso-ventral (DV), -1.8 mm. When finishing treatment, skin was closed using surgical suture. Animals were placed on a thermal pad until recovery and monitored daily for several days. Mice were sacrificed by CO₂ inhalation at ~150 days after treatment. Brains were removed and the right (injected) hemisphere was stored in 10% formalin for histological studies.

10-μm-thick serial slices from all animal groups (n = 5-8/group; 5 sections/stain/animal) were processed in parallel for histological analyses. For immunohistochemistry, sections were deparaffinazed and the endogenous peroxidase activity was blocked with 3% H₂O₂/10% methanol in PBS for 20 min. Then, brain sections were incubated in formic acid 85% for 5 min for antigen retrieval. The sections were incubated overnight at room temperature with the mouse anti-Aβ antibody at a 1:1000 dilution (Covance, Princeton, NJ). After washing, sections were incubated for 1 h with an HRP-linked secondary sheep anti-mouse antibody at a 1:1000 dilution (GE Healthcare, Little Chalfont, UK). Peroxidase reaction was visualized using DAB Kit (Vector) following the manufacturer’s instructions. Finally, sections were dehydrated in graded ethanol, cleared in xylene, and cover-slipped with DPX mounting medium (Innogenex, San Ramon, CA). A similar approach was used for the detection of hyper-phosphorylated tau in human samples (AT8 antibody, 1:100 dilution; Pierce, Rockford, IL, USA). For ThS staining, sections were incubated in ThS (Sigma) solution (0.025% in 50% ethanol) for 5–10 min after deparaffinization. Sections were dehydrated in graded ethanol, cleared in xylene, and cover-slipped with DPX mounting medium (Innogenex, San Ramon, CA).

Briefly, sagittal slices of all animal groups (n = 5-8 per group) were examined under a microscope (DMI6000B, Leica, Buffalo Grove, IL, USA) and image analyses quantification was performed using the ImagePro software (Rockville, MD, USA). 4G8 immunohistochemistry and ThS staining were quantified in every tenth sections, in a total of 5 sections. All 4G8 and ThS reactive plaques (parenchymal and vascular, induced and naturally occurring) were included in the analyses. Burden was defined as the area of the brain labeled per total area analyzed (data analyses in cortex included all cortical areas). Burden quantification was performed by an investigator blinded to the experimental groups.

Data were expressed as means ± standard error (SEM). Skewness/kurtosis statistic test was used to confirming normal distribution of the data. In all the groups, one-way analysis of variance (ANOVA) followed by a Tukey's multiple comparison analysis was used to determine differences among the groups. The values are expressed as means ± SEM. Data was analyzed using the Graph Pad prism software, version 5.0. *p < 0.05, **p < 0.01, ***p < 0.001. Statistical differences were considered significant for values of p < 0.05.

The aggregation and deposition of proteins is one of the main pathological features of AD [3]. To compare the neuropathological changes in MCI and NDAN with the ones observed in a classical AD case, we characterized brain samples from individuals affected by these conditions using different histological techniques. Formalin-fixed samples from the cingulate cortex were stained with thioflavin-S (ThS) to identify amyloid structures, and immunostained using the 4G8 and AT8 antibodies to detect Aβ deposits and hyperphosphorylated-tau, respectively (Figure 1). Extensive accumulation of ThS reactive Aβ deposits was found in the brain of MCI and NDAN individuals (Figure 1a, upper and middle panels), comparable to the one found in the brain from a confirmed AD case. For all three cases, a wide variety of Aβ arrangements, such as parenchymal mature plaques, diffuse aggregates, vascular deposits, and intracellular structures were found. However, established tau pathology was identified only in the AD brain (Figure 1a, lower panels); MCI brains did not display AT8-positive neurons and NDAN samples showed only a couple of reactive neurons in several slices analyzed. As expected, all these pathological characteristics were absent in the brain of a non-demented individual (Figure 1a, right panels).

To further characterize these specimens, we quantified the amount of soluble/insoluble Aβ after serial extraction in phosphate buffered saline (PBS), sodium dodecyl sulfate (SDS), and formic acid (FA) coupled to human-Aβ ELISA. Whereas the levels of insoluble Aβ₄₀ (SDS and FA fractions) in the MCI brain were higher compared to the other samples (Figure 1b), the concentration of this molecule in the NDAN preparation showed similar levels to the one observed in a non-demented aged individual, used as control. Interestingly, the concentration of insoluble Aβ₄₂ was similar among the AD, MCI and NDAN cases (Figure 1c). Only the brain sample obtained from the aged control showed lower levels of this peptide. The amount of SDS-insoluble/formic acid-soluble Aβ₄₂ (representing the most aggregated Aβ species) in AD, MCI and NDAN was ~60-fold higher than in the aged control brain (Figure 1c).

Several AD-mouse models have been used to study the transmissibility of Aβ misfolding. Most of the transgenic lines previously used involve overexpression of either the wild type or mutant form of the human amyloid precursor protein (APP). Before starting our study, we investigated the appropriate time-frame to assess an exogenous Aβ seeding effect in the double transgenic mouse model used in this study. For this purpose, we followed the time-course of Aβ aggregation and deposition in the brain of APP_Swe/PSEN1_∆E9 mice (at 5, 6, 7 and 8 months of age, n = 5-8) by immunohistochemistry. Previous reports suggest that these transgenic animals first start exhibiting amyloid deposits at 4–5 months of age and have extensive plaque deposition at 12 months old [34]. We confirmed these results by observing a time dependent increase of Aβ deposition in both cortex (Figure 2a and 2c) and hippocampus (Figure 2b and 2d). Our data clearly shows that the inflection point in terms of amyloid deposition for this transgenic line is between 6 and 7 months old. The Aβ burden at 7 months old was ~5-fold (cortex) and ~4-fold (hippocampus) higher than in 6 months old mice. Brain samples from male and female subjects showed comparable amounts of Aβ deposits within the same group.

To test the hypothesis that the sole presence of Aβ aggregates, regardless of the clinical signs, is required to accelerate Aβ deposition in AD-mice, brain extracts from patients affected by MCI or NDAN were intra-cerebrally (i.c.) injected into APP_Swe/PSEN1_∆E9 mice at ~1 month old. Additionally, untreated APP_Swe/PSEN1_∆E9 mice, as well as mice injected with PBS or brain homogenates obtained from either an AD patient or an aged non-demented individual, were used as controls. All animals were sacrificed ~5 months after inoculation (6 months old), and brains were collected for histopathological analyses. Since seeding effects strongly depends on the incubation periods, the time point selected to analyze the animals was chosen to provide sufficient time to observe an effect and avoid a possible masking due to the endogenous appearance of brain amyloidosis. Brain slices were stained with antibodies against Aβ and the burden of this peptide was measured in hippocampus (injection site). In order to assess if seeding expands from its original placement, we also checked brain cortex, which is an area strongly affected by Aβ lesions in AD patients. As expected, we observed a higher amount of amyloid load in mice injected with the AD extract compared to 6 month old untreated animals and mice injected with PBS or an aged-control brain (Figures 3c and 4c). Strikingly, we found that animals inoculated with the MCI extract had a substantially higher amount of Aβ deposits compared to the control groups (Figures 3 and 4). The Aβ burden was ~5-fold (cortex, Figure 3c) and ~11-fold (hippocampus, Figure 4c) higher compared to that observed in 6 months-old untreated animals. The amyloid load in the cortex of these animals was similar to the one observed in mice inoculated with AD brain (Figure 3c), but the amount in hippocampus was higher (Figure 4c). Interestingly, brains from animals inoculated with the NDAN extract also showed an increased deposition of amyloid aggregates, displaying a high amount of Aβ plaques in the cortical area (Figure 3c) in a similar way as observed for the animals injected with the AD and MCI inocula. The induction of Aβ aggregation by NDAN in the cortex was ~4-fold higher than in 6 months old untreated animals (Figure 3c); however, this treatment did not show any statistically significant difference in the hippocampal area (Figure 4). These results demonstrate that MCI and NDAN brain extracts can accelerate Aβ pathology in a similar manner as AD samples. Nevertheless, the presence of extensive-diffuse accumulation of Aβ reactive deposits in the corpus callosum (Figure 4a) was different from the typical aggregates naturally generated in this model (Figure 2a and 2b).

To specifically evaluate the formation of fibrillar amyloid deposits, ThS staining was performed in brains of animals from all experimental and control groups (Figure 5). ThS-reactive deposits were observed mostly in the corpus callosum and fimbria of animals injected with AD or MCI brain extracts (Figure 5a and 5b). Interestingly, the extent of ThS burden in AD and MCI injected animals was lower compared to the Aβ burden measured by antibody binding (Figures 3 and 4), and only reached a statistical significant increase in the hippocampus of animals inoculated with the MCI extract (Figure 5c).