An Intracellular Amyloid-β/AβPP Epitope Correlates with Neurodegeneration in those Neuronal Populations Early Involved in Alzheimer’s Disease

Abstract

The main histopathological hallmarks of Alzheimer’s disease (AD) are the extracellular deposition of neuritic amyloid plaques, composed of amyloid-β (Aβ) peptide, and the intracellular accumulation of neurofibrillary tangles, composed of hyperphosphorylated tau. Both traits are emulated in the 3xTg-AD mouse model. Because the relevance of this model in the bibliography and the main role of Aβ in neuronal impairment, here we have detailed the brain Aβ/AβPP distribution to subsequently quantify cellular density and intracellular burden for specific neuronal populations in the early stages of the disease. 6E10 immunoreactivity was evident in the deep layers of the cerebral cortex, in the pyramidal cell layer of the hippocampus, in the basolateral amygdala nucleus, and in the deep cerebellar nuclei macroneurons; whereas the specific neuronal populations with decreased cellular density were the large pyramidal neurons from the layers V-VI in the cerebral cortex, the pyramidal neurons from the CA_2-3 region in the hippocampus, and the large neurons from the basolateral nucleus in the amygdala, apart from the already reported deep cerebellar nuclei. Interestingly, we found a strong correlation between intracellular Aβ/AβPP burden and cellular density in all these populations. In addition, behavioral testing showed the functional consequences of such a neuronal depletion. Concretely, anxious-like behavior is manifested in the corner and open-field tests, as well as cognitive functions shown to be impaired in the novel object recognition test and Morris water maze paradigm. To our knowledge, this is the first deep characterization of the specific neuronal populations affected in the 3xTg-AD mouse model.

Keywords

3xTg-AD Alzheimer’s disease amyloid-β AβPP behavioral alterations 6E10 immunoreactivity neuronal loss

INTRODUCTION

Alzheimer’s disease (AD) is the most common neurodegenerative disorder worldwide. It is clinically characterized by cognitive decline, learning and memory impairment, and emotional alterations such as anxiety or phobia, being all of these symptoms manifested several years after the first neuropathological changes [1]. Specifically, AD is associated with the extracellular accumulation of amyloid-β (Aβ) peptide as toxic oligomeric species, which shall eventually aggregate as senile plaques. In addition, hyperphosphorylated tau protein is found intracellularly aggregated as neurofibrillary tangles. Despite many efforts focused on understanding the causes of AD, the exact mechanisms triggering the pathology still remain elusive. More than two decades ago, the amyloid cascade hypothesis designated Aβ peptide accumulation as the cause triggering the disease [2]. This concept has been modulated over time; apart from accumulating Aβ peptide, other degenerative processes occur, such as calcium disruption, mitochondrial failure, oxidative stress, metabolism alterations, vascular injury, inflammatory response, synaptic damage, and neuronal death [3, 4].

Neurotoxicity in AD was initially associated with the presence of amyloid plaques, but soluble oligomeric forms of Aβ have been demonstrated to be the most toxic species and to better correlate with cognitive decline [5]. In addition, intracellular accumulation of Aβ peptide has been observed in brains from AD patients as well as from several AβPP transgenic mice [6], some of which also display some extent of neuronal loss [7, 8].

The 3xTg-AD mouse reproduces amyloid and tau pathologies as regional and temporal patterns analogous to AD patients [9, 10], with Aβ appearing first in the cerebral cortex and then spreading out to the hippocampus, so it is extensively used to decipher Aβ-downstream molecular, histological, or behavioral alterations, as well as to predict outcomes from potential pharmacological interventions. A detailed description of amyloid pathology progression through primary motor cortex, amygdala, CA₁ region from the hippocampus (Cornus Ammonis), and entorhinal cortex from 3xTg-AD males showed that intracellular Aβ accumulation is already evident at very early stages (∼3 mo), whereas extracellular Aβ accumulation appears at the late stages of the disease (∼18 mo) [11]. Intracellular amyloid pathology in 3xTg-AD was also directly related to the onset of cognitive and behavioral impairments [12], which are widely assessed in this model [13]. However, despite the extensive research performed in the 3xTg-AD mouse, intracellular amyloid pathology has not been completely elucidated in terms of the specific neuronal populations affected in the early stages of the disease.

The goal of this work is to provide an accurate picture of which neuronal populations display intracellular Aβ/AβPP accumulation in 5-month-old 3xTg-AD females, as well as to demonstrate its correlation with neuronal loss. Functional assessment of behavioral and psychological symptoms of dementia (BPSD)-like symptoms, learning, and recognition and spatial memory testing complement these results.

METHODS

Animals

The triple-transgenic mouse model of AD (3xTg-AD) was initially engineered at the University of California, Irvine, by introducing AβPP_Swe and MAPT_P301_L transgenes into a PS1_M146V homozygous knock-in single-cell embryo [9, 10]. AβPP_Swe and PS1_M146V are mutations found in familial AD, and MAPTP301L mutation is associated with a familial form of frontotemporal dementia rather than AD.

Five-month-old 3xTg-AD females from the colony established by our group at the Servei d’Estabulari from the Universitat Autònoma de Barcelona (UAB), and the corresponding non-transgenic (NTg) mice with the same genetic background (B6129SF2), were used. Founder animals were provided by The Jackson Laboratory, Bar Harbor, ME, USA. Animals were maintained under standard laboratory conditions (temperature of 22±2°C and relative humidity of 55±5%, 12-h light:dark cycle starting at 08:00 a.m., wood chips for bedding, and food and waterad libitum). All the experiments were approved by the UAB Animal Research Committee and the Government of Catalonia.

Experimental design

Fourteen 3xTg-AD females were distributed randomly in two groups for histological analyses (n = 4) and behavioral studies (n = 10), and were compared to the same number of gender- and age-matched NTg mice. The temporal sequence of behavioral tests was performed based on the degree of stress in each test, with the most stressful ones at the end. First, the corner test (CT) and open-field test (OFT) were carried out on the day 1. Novel object recognition test (NORT) test was assessed on the day 2. Learning abilities and spatial memory were evaluated by the Morris water maze (MWM) on the days 3 to 9.

Histological procedures

Mice were anesthetized by intraperitoneal injection of ketamine/xylazine (87.5/12.5 mg/kg body weight) and transcardially perfused with 4% paraformaldehyde in PBS-buffer (pH 7.4). Brains were removed from the skulls, weighted, dissected on ice, rinsed in cold PBS and rapidly immersed for 36 h in the same fixative at 4°C. Then, samples were rinsed again in PBS, dehydrated by immersion in a battery of increasing concentrations of ethanol solutions (50%, 70%, 96%, 100%) and a last step in xylene. Paraffin embedding was carried out following the regular procedures from our laboratory. Finally, the cerebrum was serially sectioned in the coronal plane whereas the cerebellum was cut in the sagittal plane (all the sections were 10 μm thick). Representative slides from the seriated regions were stained with hematoxylin-eosin and cresyl-violet.

Immunohistochemistry

Sections were deparaffined in xylene, hydrated in decreasing concentrations of ethanol solutions, and extensively washed in distilled water. Endogen peroxidase activity was blocked by placing the slides in 3% H₂O₂ in methanol for 10 min. Antigen retrieval was performed by formic acid immersion (70% in distilled water for 10 min). Non-binding sites were blocked using 5% bovine serum albumin in PBS containing 0.05% Tween-20 and 5% normal goat serum at room temperature for 1 h. Then, slides were incubated overnight at 4°C with the 6E10 monoclonal antibody (mAb) (Covance Signet, Princeton, NJ, USA) and revealed by the Mouse ExtrAvidin Peroxidase Staining Kit antibody produced in goat and DAB (3-3’ Diaminobenzidine) (both products provided by Sigma-Aldrich, St. Louis, MO, USA). Finally, sections were immersed in hematoxylin for 10 s, extensively washed, dehydrated, immersed in xylene, and cover-slipped in DPX mounting medium.

Immunofluorescence

Deparaffined sections were pre-treated with sodium citrate buffer (0.01M sodium citrate, 0.05% Tween-20, pH 6) for antigen retrieval, blocked and incubated overnight at 4°C with the specific primary antibody (as detailed in Table 1), except for the anti-neurofilament H and the anti-cathepsin D mAbs that were incubated for 48 h. Then, the corresponding fluorophore-conjugated secondary antibody was incubated at room temperature for 1 h. Finally, sections were cover-slipped for microscopy observation using DAPI (4’,6-diamino-2-fenilindol)-containing Vectashield (Vector Laboratories, Burlingame, CA, USA) solution, for fluorescent nuclei staining. When double immunodetections were performed, both primary antibodies, as well as both secondary ones, were incubated at a time.

Table 1

Primary and secondary antibodies used for Immunofluorescence. List of primary and secondary antibodies used for IF, working dilution, commercial source, and product reference. FITC, fluorescein isothiocyanate; Cy3, cyanine 3

	Dilution	Source	Cat.
Primary antibodies
Mouse anti-amyloid-β (1–16) (6E10)	1:100	Covance Signet	SIG-39320
Rabbit polyclonal anti-NeuN	1:200	Merck Millipore	ABN78
Rabbit polyclonal anti-neurofilament H	1:100	Merck Millipore	ABN76
Rabbit polyclonal anti-calretinin	1:500	Merck Millipore	AB5054
Rabbit polyclonal anti-GFAP	1:200	DAKO	Z033401-2
Rabbit polyclonal anti-Iba1	1:100	Abcam	AB108539
Goat polyclonal anti-Cathepsin D	1:100	Santa Cruz Biotech	sc-6487
Secondary antibodies
Goat anti-rabbit IgG, FITC¹ conjugated	1:100¹	Chemicon, Millipore	AP132F
Goat anti-mouse IgG, Cy3 conjugated	1:100	Chemicon, Millipore	AP124C
Donkey anti-goat IgG, FITC conjugated	1:100	Jackson Immunores.	705-095-147

¹FITC-conjugated anti-rabbit secondary antibody was used at 1:200 when anti-NeuN and anti-GFAP antibodies were immunodetected.

Image processing

Images were captured in a Zeiss Axiophot microscope with a ProgRes C10 plus camera (Jenoptik, Jena, Germany) with the objective lens Plan-Neofluar from 2.5x to 63x. Confocal imaging was performed by the spectral Fluoview-1000 (Olympus, Tokyo, Japan) by capturing 7–10 optical sections, separated by 1 μm, with the objective lens UPLSAO0 from 40x to 60x. Images were processed with Adobe Photoshop (v. 7.0, Adobe Inc., San Jose, CA, USA) or Imaris software (v. 7.2.1, Bitplane, Belfast, UK).

Quantitative analysis

Intensity of fluorescence signal was measured by using Image J software (v. 1.43 u, NIH, Bethesda, MD, USA). In brief, 6E10 immunoreactivity was considered in each cell individually by manually delimitating somas, except for those populations arranged as dense rows or strips (granular cells from the olfactory bulb, hippocampus, and cerebellum), whose fluorescence intensity was measured by delimitating the layer. GFAP (glial fibrillary acidic protein) intensity was measured in the entireimage.

The total number of cells of each specific neuronal population was counted per section (except for the granular cells from the olfactory bulb, dentate gyrus from the hippocampus and the cerebellar cortex, which were counted in 5–8 representative images per section) and replicated in four sections per animal (40 μm apart), and areas occupied by these populations were measured. Then, quantifications were normalized by dividing the number of cells by the corresponding area and results were expressed as cellular density (cells/mm²).

All the analyzed sections from the olfactory bulb corresponded to the coordinates of Fig. 3 (interaural 7.36 mm and Bregma 3.55 mm) from Franklin and Paxinos [14]. The neocortex, hippocampus, and amygdala were examined in the range of coordinates between Figure 43 (interaural 2.34 mm and Bregma –1.46 mm) and Figure 48 (interaural 1.74 mm and Bregma –2.06 mm). In the case of the cerebellum, the sagittal sections analyzed corresponded to the coordinates of Figure 107 (lateral 0.72 mm).

Fig.1

6E10 immunodetection through the olfactory bulb, neocortex, hippocampus, amygdala and cerebellum. Low-magnification view of each analyzed region and zoom-in to those neuronal population containing intracellular Aβ/AβPP. A, B) Olfactory bulb and mitral cells; C, D) Neocortex and large pyramidal neurons from the deeper layers; E, F) Hippocampus and PCL from the region 1 of the Cornus Ammonis; G, H) Basolateral nucleus in the amygdala and its macroneurons; I-K) Paravermal section of the cerebellum, with Purkinje neurons and macroneurons from the DCN. Coronal sections, except for the cerebellum (sagittal). Scale bar in A, 400 μm; B, 30 μm; C, 500 μm; D, 50 μm; E, 100 μm; F-H, 50 μm; I, 400 μm; J, K, 200 μm.

Fig.2

Double immunofluorescence detection of Aβ/AβPP and specific neuronal markers. A) Aβ/AβPP and neurofilament H colocalization within the large pyramidal neurons from the layers V-VI from the neocortex. B) Aβ/AβPP and calretinin distribution through the neocortex. C) Aβ/AβPP colocalization with the neuronal marker NeuN in the pyramidal cells layer (PCL) from the region 1 of Cornus Ammonis (CA₁) in the hippocampus. D) Aβ/AβPP colocalizes with the neuronal marker NeuN also in the basolateral amygdalar (BLA) nucleus. In green, the specific neuronal marker. In red, 6E10 mAb antibody fluorescent signal. In blue, DAPI staining. Scale bars, 100 μm.

Fig.3

Confocal visualization of Aβ/AβPP distribution. A) Granular deposition of intracellular Aβ/AβPP and the corresponding orthogonal views. White arrow indicates Aβ/AβPP accumulation along the cytoplasmic prolongation. Scale bar, 20 μm. B) 3D-composition of the same region. C) Zoom-in to an individual cell showing the intracellular distribution of Aβ/AβPP mainly through the cytoplasm. White arrowheads point to a discrete immunodetection within the nucleus. D) Granular deposition of intracellular Aβ/AβPP colocalizes with cathepsin D.

Behavioral and cognition tests

All the tests were recorded by a digital USB camera (The Imaging Source, Germany) and processed by ANY-maze software (v. 5.14, Stoelting Europe, Ireland). All the material involved in the behavioral testing was cleaned thoroughly between trials to ensure the absence of olfactory cues.

Corner test

The CT was performed by placing the mice in a standard home-cage (makrolon, 22×22×14.5 cm) for 30 s. Neophobia to the new home-cage was evaluated by the number of corner visits, latency to the first corner visit, number of rearings, and latency to the first rearing.

Open-field test

The OFT consisted of an arena (white polyethylene, 42×42×50 cm) where the animals were placed for 15 min. Anxiety was measured by locomotor activity (total distance traveled across time), as well as other behavioral patterns such as crossings through a central area virtually delimited (number of entries, distance traveled in the center versus distance traveled in the periphery, time in the center and time in the periphery), rearings (number of rearings and latency at the first one), self-groomings (number of self-groomings and latency to the first one), and number of defecations.

Novel object recognition test

The NORT was performed in the same arena as the OFT, so no habituation was required. Animals were familiarized to two identic objects (A and A’) for 5 min (familiarization phase). Then, the recognition phase was assessed by exposing the animals to one of the familiar objects (A) and a novel one (B), also for 5 min. To properly evaluate short-term recognition memory, the interval between familiarization and the recognition test was 15 min [15]. Exploratory behavior was considered as total explorations number, time exploring and distance traveled. Recognition memory was evaluated as the preference of the animal to explore a familiar or a novel object. Differences in the accumulated time exploring both objects were expressed by the discrimination index (DI), calculated as (B–A)/(B+A). DI can vary between +1 and –1, with positive values when the novel object is the preference, negative values when the familiar object is the preference, and 0 when the preference in null [16].

Morris water maze test

The MWM paradigm tests the ability to find a hidden platform in a pool (Ø, 120 cm) full of opaque water, stained with non-toxic white tempera. Animals performed 4 initial trials, in which the platform (Ø, 11 cm) was highlighted by a colored flag for contrasting to the white water and walls. Then, the platform was hidden for the 5 consecutive days of acquisition phase (4 trials per day, with an inter-trial period of 10 min), in which the animals should reach the platform being oriented by proximal and distal cues. Mice were manually guided to the platform if they were unable to find it in 60 s. A final probe, consisting of 60 s of navigation in absence of the platform, was performed 24 h after the last acquisition stage.

Starting points were different in each trial, and the sequence of trials varied every day. In the final probe, animals started from the more distant point to the usual platform location. Once removed from the water, animals were located under a heating lamp to prevent hypothermia. Mean swimming speed and distance traveled were measured across trials during the acquisition stages. Latency to the platform, platform crossings, and path efficiency, as well as time and distance traveled within a virtual zone centered on former platform location (normalized by areas), were measured in the final probe.

Data analysis

Results were statistically evaluated with the non-parametric Mann-Whitney U-test (significance was set to 95% of confidence). Measured values are expressed as medians (IQR), whereas percentages (6E10-ir and neuronal depletion) were calculated from means.

A simple linear regression analysis between the Aβ/AβPP signal and the logarithm of neuronal depletion was performed.

RESULTS

Aβ/AβPP distribution through the brain

Although intracellular Aβ accumulation at the cerebral cortex, hippocampus, and amygdala of young 3xTg-AD mice is already reported [12], the precise neuronal populations affected are not known. In this work, we first assessed the intracellular Aβ/AβPP distribution through several regions of the central nervous system by immunohistochemistry with 6E10 mAb, and then identified which neuronal types are 6E10-immunoreactive (ir) by colocalization with specific neuronal markers.

Figure 1 illustrates the Aβ/AβPP distribution along the rostrocaudal axis, focusing on three representative levels: the olfactory bulb (the most rostral part); neocortex, hippocampus, and amygdala (in an intermediate location); and cerebellum (the most caudal part). 6E10 immunodetection was almost negligible in the olfactory bulb whereas was evident in specific populations from the cortex, hippocampus, and amygdala, as well as in thecerebellum.

To assess the involvement extent of each neuronal population, intracellular Aβ/AβPP burden was semi-quantified by measuring the fluorescent 6E10 signal (Table 2) and those cells with evident signal were colocalized with specific neuronal markers (Fig. 2). No 6E10 immunoreactivity was found in the main nor the accessory olfactory bulbs, except for a faint detection in the mitral cells (Fig. 1A-B). In the case of the neocortex, a strong 6E10 signal was detected within the large neurons located in layers V-VI (Fig. 1C-D). This signal colocalized with the anti-heavy neurofilament (NFH) polypeptide antibody (Fig. 2A), which is a type of intermediate filament typically restricted to neurons with a complex cytoskeletal structure such as the large pyramidal neurons [17]. On the other hand, the anti-calretinin antibody was used as a marker of cortical interneurons (concretely, bipolar cells and double bouquet cells; with large basket cells, martinotti cells, and bitufted cells occasionally expressing calretinin [18]) and 6E10 signal was not colocalized (Fig. 2B).

Table 2

Semi-quantification of the intracellular Aβ/AβPP burden in the olfactory bulb, neocortex, hippocampus, amygdala and cerebellum. Structures, regions and neuronal populations analyzed. CA₁ and CA_2-3, regions 1 and 2-3 from the Cornus Ammonis. PCL, pyramidal cells layer. GCL, granular cells layer. The second column contains the intracellular Aβ/AβPP burden measured by fluorescent intensity. Symbols represent the different intensity gradients of relative fluorescent signal: –, no signal; –/+, ambiguous; +, faint; ++, medium; +++, strong intensity

Neuronal population	Intracellular Aβ/AβPP burden
Olfactory bulb
Main olfactory bulb
Glomeruli¹	–
Mitral cells	–/+
Granular cells	–
Neocortex
Large pyramidal neurons	+++
Calretinin-ir neurons	–
Hippocampus
Cornus Ammonis
CA₁ PCL cells	+++
CA_2-3 PCL cells	+++
Dentate Gyrus
GCL cells	+/–
Amygdala
Basolateral amygdalar nucleus
Large pyramidal neurons	+++
Cerebellum
Cerebellar cortex
Paravermal Purkinje neurons	–/+
Paravermal granular cells	–
Deep cerebellar nuclei
Fastigial macroneurons	++
Interpositus macroneurons	++
Dentate macroneurons	++

¹Glomeruli is an exception from the neuronal categorization, since it is an interneuronal connection structure.

In the hippocampus, Aβ/AβPP staining was observed in the pyramidal cells layer (PCL) from the Cornus Ammonis, at both CA₁ and CA_2-3 regions (Fig. 1E, F). Besides, 6E10 signal colocalized with the anti-NeuN neuronal nuclear marker (Fig. 2C). No evident 6E10 signal was detected in the dentate gyrus.

In the amygdala, a strong 6E10 signal was detected in the basolateral amygdalar (BLA) nucleus (Fig. 1G, H), from the basolateral group of nuclei, which colocalized with large NeuN-ir neurons (Fig. 2D). Finally, Aβ/AβPP distribution through the cerebellum was analyzed in the cerebellar cortex and deep cerebellar nuclei (DCN). Purkinje neurons showed a faint Aβ/AβPP staining whereas no 6E10 signal was detected in the internal granular layer (IGL). Macroneurons from the DCN exhibited a robust Aβ/AβPP staining (Fig. 1I-K).

Even though Aβ/AβPP burden varies among cells from the same neuronal population, it displayed a similar intracellular pattern of distribution through all the 6E10-ir neuron types studied, being mainly located in the perikaryon and, when cells contained an elevated Aβ/AβPP burden, also in the cytoplasmic prolongations (Fig. 3). Nuclear 6E10 signal was poorly detected. Interestingly, 6E10 immunodetection follows a punctuate pattern that provides a dotted aspect to the Aβ/AβPP-containing cells, in accordance to the accumulation of Aβ peptide within endosomes and lysosomes, or multivesicular bodies [8, 19]. Figure 3D confirms colocalization of Aβ/AβPP with cathepsin D, a marker for intracellular vesicular structures.

Neuronal depletion

Neuronal depletion has been widely described in AD human brain [20]. In AD transgenic mice, obvious neuronal loss has only been sporadically reported despite the numerous publications about other pathologic features such as amyloid and tau pathologies, synaptic disruption, inflammation, and behavioral and cognitive impairments [21]. Thus, since an exhaustive analysis of neuronal depletion in transgenic mice is required, we assessed the loss as general mass reduction of 3xTg-AD brains, and further examined cellular depletion of a wide variety of neuronal populations.

Weights from 3xTg-AD brains were significantly reduced when compared to the NTg ones (U-test, p = 0.029, Fig. 4A). Those slides stained with hematoxylin-eosin and cresyl-violet were analyzed and cytoarchitectonic alterations were not observed in any of the encephalic regions studied. Neuronal loss was assessed by cellular counting in anatomically matched sections from each animal. Supplementary Table 1 and Fig. 4B-F compile the overall values obtained (medians and IQR as descriptive measures) and the statistical analysis by the non-parametric Mann-Whitney U-test. Table 3 details the percentage of depletion (as calculated from the means) in those populations with significant reduction of neuronal densities. Glomeruli, mitral cells, and granular cells were examined in the olfactory bulb, but no significant depletion was found (Fig. 4B). In the cerebral cortex, cellular density of those NFH-ir cells was reduced 29% in the 3xTg-AD mice as compared to the NTg ones (U-test, p = 0.029), whereas no significant reduction was observed in calretinin-ir neuronal populations (Fig. 4C). In the hippocampus, thecellular density in the PCL at the CA was not significantly different when the whole region was considered, but the 3xTg-AD mice achieved a marginally significant reduction of 13% when CA_2-3 was individually considered (U-test, p = 0.057) (Fig. 4D). Cellular density was also evaluated in the granular cells layer from the dentate gyrus, but no significant differences were found. In the amygdala, cellular density was reduced 45% in the BLA nucleus (U-test, p = 0.029) (Fig. 4E). Reduction of cellular densities in the cerebellum was already considered in a previous work of the group [22]. Interestingly, macroneuron density from the 3xTg-AD mice was reduced about 32% and 22% compared to the NTg ones in the fastigial and interpositus DCN, respectively (U-test, p = 0.029 and p = 0.029) (Fig. 4F).

Fig.4

Neuronal densities in the olfactory bulb, cerebral cortex, hippocampus, amygdala and cerebellum. A) Brain weight expressed in mg, is compared between the NTg and 3xTg-AD mice. B) Glomeruli, mitral cells and granular cells from the olfactory bulb. C) Neurofilament H-immunoreactive (NFH-ir) and calretinin-ir cells from the neocortex. D) Cells in the Pyramidal Cell Layer (PCL) from the entire Cornus Ammonis (CA_entire), and regions 1 (CA₁) and 2-3 (CA_2-3) separately. E) Macroneurons from the basolateral amygdalar (BLA) nucleus. F) Paravermal Purkinje and granular cells, and macroneurons from the fastigial, interpositus and dentate deep cerebellar nuclei from the cerebellum. Densities (cells/mm²) are expressed as box plots, and whiskers represent minimum and maximum values. ^#indicates the marginally significant differences between the 3xTg-AD and NTg mice with p≤0.1, *significant differences with p≤0.05, **p≤0.01. ¹indicates glomeruli as an exception of the neuronal categorization, since it is an interneuronal connection structure. ²indicates those neuronal populations arranged as a monolayer of cells (density expressed as cells/mm of length).

Table 3

Neuronal depletion in young 3xTg-AD females. Percentage of depletion in those populations with significant reduction of neuronal densities. Means of neuronal densities from both conditions (NTg and 3xTg-AD) were used to determine neuronal depletion in the 3xTg-AD mice. NFH, neurofilament H. CA_2-3, regions 2-3, from the Cornus Ammonis. PCL, pyramidal cells layer. BLA, basolateral amygdala

Structure	Neuronal population	Depletion (%)
Neocortex	NFH-ir large pyr neurons	28.91
Hippocampus	CA_2-3 PCL cells	13.15
Amygdala	BLA macroneurons	45.28
Cerebellum	Fastigial nucleus	31.79
Cerebellum	Interpositus nucleus	21.57

Conclusively, these results evidence that neuronal depletion at first stages of the disease mainly occurs on glutamatergic neuronal populations. Moreover, the higher neuronal loss was found in the amygdala, followed by the cerebral cortex and, finally, by the hippocampus. At the same time, neuronal loss in the cerebellum follows the already reported mediolateral gradient in the DCN macroneurons [22].

Severity of the intracellular Aβ/AβPP pathology

The number of 6E10-ir cells was determined in those neuronal populations in which evident intracellular signal was detected. This data, when related to the cellular counting, revealed the proportion of neurons containing Aβ/AβPP (Fig. 5A). In the layer V-VI of the neocortex from 3xTg-AD mice, 79% of large pyramidal neurons presented intracellular Aβ/AβPP. In the CA from the hippocampus, the signal was detected in 25% of cells in the PCL, whereas proportions were around 31% and 14% when considering both regions CA₁ and CA_2-3 individually. In the amygdala, 87% of large neurons from the BLA nucleus presented an evident signal. Finally, in the DCN from the cerebellum, 81%, 48%, and 79% of the macroneurons from the fastigial, interpositus, and dentate nuclei presented intracellular Aβ/AβPP. Hence, these results suggest that the plurality of neuronal populations involved in the pathology presents different levels of severity.

Fig.5

Proportion of neurons exhibiting intracellular Aβ/AβPP and correlation with neuronal depletion. A) Proportion of 6E10-ir cells in each neuronal population where evident signal was detected in 3xTg-AD mice. Proportions are expressed by percentages of the means (6E10-ir cells/total cells number) and error bars represent SEM. CA, Cornus Ammonis; CA₁ or CA_2-3, Cornus Ammonis regions 1 or 2-3, respectively. PCL, pyramidal cells layer. BLA nucleus, basolateral amygdalar nucleus. DCN, deep cerebellar nucleus. B) Linear regression analysis between the proportion of 6E10-ir cells and the logarithm of neuronal depletion (calculated in Table 3).

Since the proportion of Aβ/AβPP-containing neurons followed the same regional vulnerability pattern than neuronal depletion, we tried to find a correlation and found a linear association between the 6E10 signal and the logarithm of neuronal depletion (Fig. 5B). This means that a small change in the proportion of cells containing intracellular amyloid induces a huge change in their density.

Finally, we detected the presence of astroglia once intracellular Aβ/AβPP accumulation and neuronal depletion are occurring. Glial fibrillary acidic protein (GFAP, a typical astrocytic marker) was immunodetected and fluorescent signal measured in the hippocampus. GFAP fluorescence intensity values were higher in 3xTg-AD mice in comparison to the NTg ones, although differences did not reach statistical significance (Fig. 6A). Interestingly, astrocytes and microglia appeared wrapping the soma of some Aβ/AβPP-containing neurons (Fig. 6B), even though both types of glia were also observed in contact with non-reactive 6E10 neurons.

Fig.6

Astroglia and microglia colocalization with Aβ/AβPP-containing neurons. A) Astroglia immunodetection by anti-GFAP antibody (green) and fluorescence intensity (RFU) quantification in the hippocampus. Scale bars, 100 μm. B) Double immunofluorescence detection of cortical astrocytes (anti-GFAP and 6E10 mAbs, in the left), or cortical and hippocampal microglia (Iba1 and 6E10, in the middle and the right), in 3xTg-AD mice. Scale bars, 10 μm.

BPSD-like symptoms

3xTg-AD females already manifested evident BPSD-like symptoms in the CT and OFT at 5 months (Supplementary Table 1). When the horizontal activity was considered in the CT, no differences were found in the total number of corner visits between 3xTg-AD and NTg mice, but latencies to the corner were reduced in the 3xTg-AD mice when compared to the NTg ones (0.65 s and 3.10 s, U-test, p = 0.048), evidencing an anxious-like behavior probably induced by neophobia to the cage (Fig. 7A). However, no differences were found in the vertical activity. Anxious-like behavior was also observed in the OFT since the 3xTg-AD mice exhibited reduced exploratory activity (Fig. 7B), as shown in the distance traveled (a total of 18.6 m by the 3xTg-AD mice and 46.1 m by the NTg ones, U-test, p < 0.001), the number of rearings (12 rearings by the 3xTg-AD mice and 71.5 by the NTg ones, U-test, p < 0.001) and the latency to the first one (58 s and 17 s in the 3xTg-AD mice and NTg ones, respectively, U-test, p = 0.010). Other emotional traits were analyzed, as the number of self-groomings, the latency to the first one, and the number of defecations. Significant differences between 3xTg and NTg mice were reached in the number of self-groomings (3 and 4.5, respectively, U-test, p = 0.034) and in the latency to the first one (276 s and 111 s, respectively, U-test, p = 0.017).

Fig.7

BPSD-like symptoms. A) Corner visits (n), latency to the corner (s), number of rearings (n), and first rearing latency (s) in the CT. B) Total distance traveled (m), number of rearings and latency to the first one, self-groomings (n) and the respective latency (s), and number of defecations (boli) in the OFT. Values are expressed by box plots, and whiskers represent minimum and maximum value. *indicates significant differences between the 3xTg-AD and the NTg mice with p≤0.05, **p≤0.01, ***p≤0.001.

Cognition and memory disabilities

Results obtained from the NORT are detailed in Fig. 8A and Supplementary Table 2. In summary, DI evidenced that recognition memory was impaired in the 3xTg-AD mice whereas the NTg ones manifested preference to the novel object (0.03 and 0.72, respectively, U-test, p = 0.004). Interestingly, the accumulated time exploring the objects presented higher differences than the number of explorations, so that the quality of explorations was the most suitable parameter to assess recognition memory in the NORT [13]. Also, and consistent with the CT and OFT results, 3xTg-AD mice exhibited anxious-like behavior since exploration during the test was reduced in comparison to the NTg ones, with a total distance traveled of 1.62 m and 9.04 m respectively (U-test, p < 0.001), 4 and 18.5 approximations to the objects (U-test, p = 0.005), and 1.35 s and 42.1 s of accumulated time exploring the objects (U-test, p = 0.002).

Fig.8

Learning and memory impairment. A) Distance traveled (s), total explorations (n), total time exploring (s) and its discrimination index (DI) in the NOR test. B) Mean swimming speed (m/s) and distance traveled (m) across the acquisition stages of the Morris water maze (MWM) test. C) Platform crossings (n), latency to the platform (s), and path efficiency, as well as normalized time (s) and distance (m) through virtual concentric areas around the platform (first segment of the test, 0–30 s). Values are expressed by box plots, and whiskers represent minimum and maximum values, except for acquisition stages (graphs express medians and IQR). ¹Time and distance are normalized by the area of each zone. ^#refers to the marginally significant differences between the 3xTg-AD and the NTg mice with p≤0.1, *significant differences with p≤0.05, **p≤0.01, ***p≤0.001.

Regarding the MWM, Fig. 8B illustrates the results obtained in the 5 consecutive acquisition stages. At the first acquisition stage, distance traveled to reach the platform was significantly higher in the 3xTg-AD mice when compared to the NTg ones (9.6 m and 5.4 m, respectively, U-test, p = 0.003). However, 3xTg-AD mice experienced a learning process since the path to the platform was gradually decreasing as acquisition trainings were performed, to finally get similar values to the NTg ones. By contrast, the NTg mice did not change their path to the platform, probably because the animals were conscious of its location from the beginning (values did not decrease, but were still lower than those for the 3xTg-AD mice). Mean swimming speed by the 3xTg-AD mice was higher than that by the NTg ones in all the acquisition stages, corroborating the anxious-like behavior mentioned above.

Figure 8C compiles the results obtained in the first segment of the final probe (0-30 s), which better represent the test, since mice behaved different once they became aware of the platform absence. The number of platform crossings was reduced in the 3xTg-AD mice, as well as the latency to the platform occurring later, but differences did not reach significance. Path efficiency performed by 3xTg-AD mice was significantly inferior than that of the NTg ones (0.06 and 0.16, respectively, U-test, p = 0.017). Moreover, to evaluate the precision of the animals in searching the platform, concentric zones around the platform were considered. Interestingly, evaluation of time spent and distance traveled in these zones showed graduallydecreasing values in the NTg group (as more distant the zone, the lower the level) suggesting that these mice learned where the platform was placed, while 3xTg-AD mice presented a random pattern.

DISCUSSION

Intraneuronal Aβ in the 3xTg-AD mouse model

The intraneuronal accumulation of Aβ is a key topic in the pathomechanism of AD, but specificity of the 6E10 mAb widely used to detect the peptide still remain controversial [23]. Albeit 6E10 mAb is directed against the N-terminus of the Aβ peptide (epitope within 3–10), it also detects full-length AβPP and the 1st product of amyloidogenic AβPP processing, β-secretase-cleaved βCTF, which has the same N-terminus as the peptide [24]. In 2011, it was claimed that intraneuronal AβPP, not free Aβ peptides, were present in 3xTg-AD mice, but the article was retracted by the editors in 2015 because evidence of data misrepresentation [25]. The most recent agreement is that intraneuronal accumulation of the Aβ peptide in the 3xTg-AD mouse model, as detected with the 6E10 mAb and first described by LaFerla [9 , 26], is indeed present [27 –30]. The 6E10 signal is the same than that with the specific-Aβ-mAb MOAB-2 and colocalizes with cathepsin-D, a marker for acidic organelles, which is further evidence of the presence of intraneuronal Aβ, distinct from Aβ associated with the cell membrane, in this mouse model. In addition, preliminary data from our laboratory show that 6E10 detection in this work should mostly correspond to the Aβ peptide since the signal is very much reduced when the 3xTg-AD mouse is treated with scFv-h3D6. ScFv-h3D6 is a single chain variable fragment designed at our laboratory [31 –33], and derived from humanized mAb-3D6 (bapineuzumab). Bapineuzumab is known to recognize the 1-5 N-terminus of the Aβ peptide, and also βCTF, but not the full-length AβPP [34 –36]. Therefore, detection with 6E10 mAb in this work should mostly correspond to intracellular Aβ, although additional work would be needed to unequivocallyprove it.

The 3xTg-AD mouse develops molecular, histological, and behavioral alterations reproducing human AD progression [10 , 13]. Brain accumulation of Aβ peptide in AD patients has been demonstrated since the first description of the pathology [37, 38]. Even the initial idea that amyloid plaques were causing the synaptic disruptions and neuronal death, evoking the clinical symptoms of AD, evidence from the last two decades of research converged in that Aβ toxicity is caused byintracellular accumulation rather than extracellular [39]. As is the case of the 3xTg-AD mouse, intraneuronal accumulation of toxic Aβ may be associated with endoplasmic reticulum stress in the brains of AD patients at an early stage [26, 40]. On the other hand, Down syndrome patients develop AD in middle age as a consequence of the trisomy of chromosome 21, where the AβPP gene is located, and provide a unique situation in which to study the early and sequential development of these changes [41]. These studies demonstrate that intracellular accumulation of Aβ in neurons and astrocytes precedes extracellular deposition. Similarly, intraneuronal Aβ in Down syndrome appears to trigger a pathological cascade leading to oxidative stress and a neurodegeneration typical of AD [42]. For an overall view on the recent progress in AD pathology, seereview [43].

Glutamatergic neurons vulnerability in young 3xTg-AD females

In the 3xTg-AD mice, intraneuronal Aβ has previously been noticed in the neocortex, subiculum, CA in the hippocampus, or amygdala [9 , 12] in both genders and at different ages. In this work, we detected Aβ/AβAPP in the neocortex (large pyramidal neurons from deep layers), hippocampus (CA₁ and CA_2-3), amygdala (macroneurons from the BLA nucleus), and cerebellum (DCN) of 5-month-old females. Also, a faint signal was detected in mitral cells from the olfactory bulb and Purkinje neurons from the cerebellar cortex.

Neuronal loss has also been one of the most extensively described hallmarks in AD patients, especially in the hippocampus [44 –46] and entorhinal cortex [47]. However, proper reproduction of neuronal loss in AD models has been the main hurdle for modeling. This feature is almost negligible in most of the AβPP-transgenic mice although it has been observed in some AβPP/PS1-transgenic models [48 –51]. For instance, neuronal depletion has been described in9-month-old 5xFAD mice, which contains five additive mutations causing Aβ overproduction and accumulation already at 1.5 months [8]. The extent to which neuronal loss occurs in the 3xTg-AD mice is not yet detailed in depth, but some authors qualitatively describe neuronal atrophy and depletion through the cerebral cortex and hippocampus, in association with behavioral and cognitive alterations [52, 53]. In the present work, a reduction of the neuron density was found within specific Aβ/AβPP-containing populations: NFH-ir large pyramidal neurons from layers V-VI in the neocortex, cells from the PCL on the CA_2-3 region in the hippocampus, and large pyramidal neurons from the BLA nucleus in the amygdala. Different grades of depletion occurred through the affected regions, with the higher neuronal loss found in the amygdala, followed by the neocortex, and, finally, by the hippocampus. As previously reported, macroneurons from the DCN in the cerebellum were depleted following a mediolateral gradient [22]. Taken together, these results suggest that glutamatergic neurons, rather than the GABAergic ones, are affected by intracellular Aβ/AβPP pathology and neuronal depletion at early stages of the disease in 3xTg-ADfemales.

Intraneuronal Aβ/AβPP and cellular density correlation in AD-involved neuronal populations

As mentioned above, intraneuronal Aβ accumulation is considered one of the key features of AD and is correlated with behavioral and cognitive decline [12 , 28–30]. The pathogenic role of intraneuronal Aβ has been explained by several mediators such as oxidative stress, calcium dysregulation, intracellular signaling mechanisms, and tau hyperphosphorylation, which contribute to neuronal dysfunction and neurodegeneration [26 , 54–56]. Also, synaptic loss has been proposed to be mediated by the failure of live neurons to maintain functional axons and dendrites or by neuron death [57]. According to the evident pathological role that intraneuronal Aβ plays on the early neurodegeneration and neuronal loss in AD, we found that a small change in the proportion of cells containing intracellular Aβ/AβPP promotes a huge change in their density.

Interestingly, when astroglia was considered, we found higher levels of GFAP intensity in the 3xTg-AD mice in comparison with the NTg ones, although statistical significance was not achieved. This reflects the role of these glial cells in assisting neurons to prevent neuronal death and to recover the issue after death has occurred. Lysosomal degradation in brain parenchyma cells is one of the mechanisms by which Aβ is cleared from the brain [58], with astrocytes being reported to readily ingest smaller sized Aβ species [59]. This reflects the relevance of astroglia at the first stages of the disease, although as the disease progresses these cells, initially resistant to Aβ cytotoxicity, become severely affected by high loads of Aβ and their functions impair [60]. In this work, astrocytes and microglia were co-immunodetected with 6E10-ir cells and we visualized both glia types wrapping the soma of some Aβ-containing neurons [61, 62]. The presence of microglia is also in consonance with its protective role before death, apart from its further role in clearing dead cells [63].

Behavioral and cognitive implications

Several behavioral and cognitive alterations in the 3xTg-AD mouse model have been described [55], and their correlation with the intraneuronal amyloid pathology across time was demonstrated [12]. Here, we further proved the relationship between neuron vulnerability, in terms of Aβ/AβPP pathology and cell depletion of specific neuronal populations, and the manifestation of behavioral and cognitivealterations.

On the one hand, anxious-like behavior was evidenced by the neophobia (in terms of reduced exploratory activity and freezing behavior) exhibited once the animals were placed on the cage (CT) or the white and illuminated open arena (OFT). BPSD-like symptoms were supported by the reduced exploratory activity also found in the NORT. Any locomotor alteration was discarded since the 3xTg-AD mice exhibited an even faster locomotion (mean swimming speed) than the NTg mice in the MWM test. Precisely, this hyperactive swimming pattern is also considered a common BPSD-like symptom, supporting the elevate level of anxiety in this mouse model[32 , 65].

On the other hand, in agreement with previous reports [13 , 66], cognitive disturbances in young 3xTg-AD females were exhibited in both NORT and MWM paradigm. In the first test, recognition memory impairment was manifested by the inability todiscriminate between objects. Results from the MWM demonstrated that 3xTg-AD mice learn, although more slowly than the NTg mice, where the platform was, but long-term spatial memory was impaired, in both the ability to reach the area and precision to locate it.

As a general conclusion, intracellular Aβ/AβPP accumulation in 5-month-old 3xTg-AD females correlates with neuronal loss and is accompanied by BPSD and cognitive impairment.

Footnotes

ACKNOWLEDGMENTS

The authors acknowledge Dr. Enrique Claro for providing capthepsin D mAb, Dr. Jaume Ferrer for advice with behavioral testing, and the Servei d’Estadística at the UAB for advice on data treatment.

This work was supported by the Instituto de Salud Carlos III/FEDER [FIS-PI113-01330], Generalitat de Catalunya [SGR-GRC-2014-00885], and Generalitat de Catalunya/FEDER [2014-PROD00032]. PIF-UAB student grants (GEC).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Sperling

, Aisen

, Beckett

, Bennett

, Craft

, Fagan

, Iwatsubo

, Jack

, Kaye

, Montine

, Park

, Reiman

, Rowe

, Siemers

, Stern

, Yaffe

, Carrillo

, Thies

, Morrison-Bogorad

, Wagster

, Phelps

(2011) Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 280–292.

Hardy

, Higgins

(1992) Alzheimer’s disease: The amyloid cascade hypothesis. Science 256, 184–185.

Querfurth

, LaFerla

(2010) Alzheimer’s disease. N Engl J Med 362, 329–344.

Scheltens

, Blennow

, Breteler

MMB

, de Strooper

, Frisoni

, Salloway

, Van der Flier

(2016) Alzheimer’s disease. Lancet 388, 505–517.

Walsh

, Selkoe

(2007) A beta oligomers - a decade of discovery. J Neurochem 101, 1172–1184.

Gouras

, Tampellini

, Takahashi

, Capetillo-Zarate

(2010) Intraneuronal β-amyloid accumulation and synapse pathology in Alzheimer’s disease. Acta Neuropathol 119, 523–541.

Oakley

, Cole

, Logan

, Maus

, Shao

, Craft

, Guillozet-Bongaarts

, Ohno

, Disterhoft

, Van Eldik

, Berry

, Vassar

(2006) Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: Potential factors in amyloid plaque formation. J Neurosci 26, 10129–10140.

Eimer

, Vassar

(2013) Neuron loss in the 5XFAD mouse model of Alzheimer’s disease correlates with intraneuronal Aβ42 accumulation and Caspase-3 activation. Mol Neurodegener 8, 2.

Oddo

, Caccamo

, Shepherd

, Murphy

, Golde

, Kayed

, Metherate

, Mattson

, Akbari

, LaFerla

(2003) Triple-transgenic model of Alzheimer’s disease with plaques and tangles: Intracellular Aβ and synaptic dysfunction. Neuron 39, 409–421.

10.

Oddo

, Caccamo

, Kitazawa

, Tseng

, LaFerla

(2003) Amyloid deposition precedes tangle formation in a triple transgenic model of Alzheimer’s disease. Neurobiol Aging 24, 1063–1070.

11.

Mastrangelo

, Bowers

(2008) Detailed immunohistochemical characterization of temporal and spatial progression of Alzheimer’s disease-related pathologies in male triple-transgenic mice. BMC Neurosci 9, 81.

12.

Billings

, Oddo

, Green

(2005) Intraneuronal Aβ causes the onset of early Alzheimer’s disease-related cognitive deficits in transgenic mice. Neuron 45, 675–688.

13.

Stover

, Campbell

, Van

Winssen CM

, Brown

(2015) Early detection of cognitive deficits in the 3xTg-AD mouse model of Alzheimer’s disease. Behav Brain Res 289, 29–38.

14.

Franklin

KBJ

, Paxinos

(2012) The Mouse Brain in Stereotaxic Coordinates, Academic Press, Sydney.

15.

Leger

, Quiedeville

, Bouet

, Boulouard

, Schumann-Bard

, Freret

(2013) Object recognition test in mice. Nat Protoc 8, 2531–2537.

16.

Ennaceur

, Delacour

(1988) A new one-trial test for neurobiological studies of memory in rats. 1: Behavioral data. Behav Brain Res 31, 47–59.

17.

DeFelipe

, Fariñas

(1992) The pyramidal neuron of the cerebral cortex: Morphological and chemical characteristics of the synaptic inputs. Prog Neurobiol 39, 563–607.

18.

Markram

, Toledo-Rodriguez

, Wang

, Gupta

, Silberberg

, Wu

(2004) Interneurons of the neocortical inhibitory system. Nat Rev Neurosci 5, 793–807.

19.

Langui

, Girardot

, El Hachimi

, Allinquant

, Blanchard

, Pradier

, Duyckaerts

(2004) Subcellular topography of neuronal Abeta peptide in APPxPS1 transgenic mice. Am J Pathol 165, 1465–1477.

20.

Morrison

, Hof

(1997) Life and Death of Neurons in the Aging Brain. Science 278, 412–419.

21.

Bayer

, Wirths

(2010) Intracellular accumulation of amyloid-Beta - a predictor for synaptic dysfunction and neuron loss in Alzheimer’s disease. Front Aging Neurosci 2, 8.

22.

Esquerda-Canals

, Marti

, Rivera-Hernández

, Giménez-Llort

, Villegas

(2013) Loss of deep cerebellar nuclei neurons in the 3xTg-AD mice and protection by an anti-amyloid β antibody fragment. MAbs 5, 660–664.

23.

Wirths

, Dins

, Bayer

(2012) AβPP accumulation and/or intraneuronal amyloid-β accumulation? The 3xTg-AD mouse model revisited. J Alzheimers Dis 28, 897–904.

24.

Horikoshi

, Sakaguchi

, Becker

, Gray

, Duff

, Aisen

, Yamaguchi

, Maeda

, Kinoshita

, Matsuoka

(2004) Development of Abeta terminal end-specificantibodies and sensitive ELISA for Abeta variant. Biochem Biophys Res Commun 319, 733–737.

25.

Winton

, Lee

, Sun

, Wong

, Leight

, Zhang

, Trojanowski

, Lee

VM-Y

(2011) Intraneuronal APP, not free Aβ peptides in 3xTg-AD mice: Implications for tau versus Aβ-mediated Alzheimer neurodegeneration. J Neurosci 31, 7691–7699.

26.

Soejima

, Ohyagi

, Nakamura

, Himeno

, Iinuma

, Sakae

, Yamasaki

, Tabira

, Murakami

, Irie

, Kinoshita

, LaFerla

, Kiyohara

, Iwaki

, Kira

(2013) Intracellular accumulation of toxic turn amyloid-β is associated with endoplasmic reticulum stress in Alzheimer’s disease. Curr Alzheimer Res 10, 11–20.

27.

Youmans

, Tai

, Kanekiyo

, Stine

, Michon

S-C

, Nwabuisi-Heath

, Manelli

, Fu

, Riordan

, Eimer

, Binder

, Bu

, Yu

, Hartley

, LaDu

(2012) Intraneuronal Aβ detection in 5xFAD mice by a new Aβ-specific antibody. Mol Neurodegener 7, 8.

28.

Rasool

, Martinez-Coria

, Wu

, LaFerla

, Glabe

(2013) Systemic vaccination with anti-oligomeric monoclonal antibodies improves cognitive function by reducing Aβ deposition and tau pathology in 3xTg-AD mice. J Neurochem 126, 473–482.

29.

Guzmán

, Bouter

, Richard

, Lannfelt

, Ingelsson

, Paetau

, Verkkoniemi-Ahola

, Wirths

, Bayer

(2014) Abundance of Aβ_5-x like immunoreactivity in transgenic 5XFAD, APP/PS1KI and 3xTG mice, sporadic and familial Alzheimer’s disease. Mol Neurodegener 9, 13.

30.

, Zhao

, Lu

, Zhang

, Hua

(2016) Intracellular Aβ and its pathological role in Alzheimer’s disease: Lessons from cellular to animal models. Curr Alzheimer Res 13, 621–630.

31.

Marín-Argany

, Rivera-Hernández

, Martí

, Villegas

(2011) An anti-Aβ (amyloid β) single-chain variable fragment prevents amyloid fibril formation and cytotoxicity by withdrawing Aβ oligomers from the amyloid pathway. Biochem J 437, 25–34.

32.

Giménez-Llort

, Rivera-Hernández

, Marín-Argany

, Sánchez-Quesada

, Villegas

(2013) Early intervention in the 3xTg-AD mice with an amyloid β-antibody fragment ameliorates first hallmarks of Alzheimer disease. MAbs 5, 665–677.

33.

Montoliu-Gaya

, Martínez

, Villegas

(2017) Understanding the contribution of disulphide bridges to the folding and misfolding of an anti-Aβ scFv. Protein Sci 26, 1138–1149.

34.

Bard

, Cannon

, Barbour

, Burke

, Games

, Grajeda

, Guido

, Hu

, Huang

, Johnson-Wood

, Khan

, Kholodenko

, Lee

, Lieberburg

, Motter

, Nguyen

, Soriano

, Vasquez

, Weiss

, Welch

, Seubert

, Schenk

, Yednock

(2000) Peripherally administered antibodies against amyloid beta-peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nat Med 6, 916–919.

35.

Gouras

, Tampellini

, Takahashi

, Capetillo-Zarate

(2010) Intraneuronal β-amyloid accumulation and synapse pathology in Alzheimer’s disease. Acta Neuropathol 119, 523–541.

36.

Bouter

, Lopez Noguerola

, Tucholla

, Crespi

GAN

, Parker

, Wiltfang

, Miles

, Bayer

(2015) Abeta targets of the biosimilar antibodies of Bapineuzumab, Crenezumab, Solanezumab in comparison to an antibody against N-truncated Abeta in sporadic Alzheimer disease cases and mouse models. Acta Neuropathol 130, 713–729.

37.

Alzheimer

(1906) Über einen eigenartigen schweren Erkrankungsprozeβ der Hirnrincle. Neurol Cent 25, 1134.

38.

Aho

, Pikkarainen

, Hiltunen

, Leinonen

, Alafuzoff

(2010) Immunohistochemical visualization of amyloid-beta protein precursor and amyloid-beta in extra- and intracellular compartments in the human brain. J Alzheimers Dis 20, 1015–1028.

39.

LaFerla

, Green

, Oddo

(2007) Intracellular amyloid-beta in Alzheimer’s disease. Nat Rev Neurosci 8, 499–509.

40.

Oka

, Leon

, Sakumi

, Ide

, Kang

, LaFerla

, Nakabeppu

(2016) Human mitochondrial transcriptional factor A breaks the mitochondria-mediated vicious cycle in Alzheimer’s disease. Sci Rep 6, 37889.

41.

Gyure

, Durham

, Stewart

, Smialek

, Troncoso

(2001) Intraneuronal abeta-amyloid precedes development of amyloid plaques in Down syndrome. Arch Pathol Lab Med 125, 489–492.

42.

Lott

, Head

, Doran

, Busciglio

(2006) Beta-amyloid, oxidative stress and down syndrome. Curr Alzheimer Res 3, 521–528.

43.

Hane

, Lee

, Leonenko

(2017) Recent progress in Alzheimer’s disease research, part 1: Pathology. J Alzheimers Dis 57, 1–28.

44.

Price

, Ko

, Wade

, Tsou

, McKeel

, Morris

(2001) Neuron number in the entorhinal cortex and CA1 in preclinical Alzheimer disease. Arch Neurol 58, 1395–1402.

45.

Rössler

, Zarski

, Bohl

, Ohm

(2002) Stage-dependent and sector-specific neuronal loss in hippocampus during Alzheimer’s disease. Acta Neuropathol 103, 363–369.

46.

West

, Kawas

, Stewart

, Rudow

, Troncoso

(2004) Hippocampal neurons in pre-clinical Alzheimer’s disease. Neurobiol Aging 25, 1205–1212.

47.

Gómez-Isla

, Price

, McKeel

, Morris

, Growdon

, Hyman

(1996) Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer’s disease. J Neurosci 16, 4491–4500.

48.

Casas

, Sergeant

, Itier

J-M

, Blanchard

, Wirths

, van der Kolk

, Vingtdeux

, van de Steeg

, Ret

, Canton

, Drobecq

, Clark

, Bonici

, Delacourte

, Benavides

, Schmitz

, Tremp

, Bayer

, Benoit

, Pradier

(2004) Massive CA1/2 neuronal loss with intraneuronal and N-terminal truncated Abeta42 accumulation in a novel Alzheimer transgenic model. Am J Pathol 165, 1289–1300.

49.

LaFerla

, Oddo

(2005) Alzheimer’s disease: Aβ, tau and synaptic dysfunction. Trends Mol Med 11, 170–176.

50.

Lomoio

, López-González

, Aso

, Carmona

, Torrejón-Escribano

, Scherini

, Ferrer

(2012) Cerebellar amyloid-β plaques: Disturbed cortical circuitry in AβPP/PS1 transgenic mice as a model of familial Alzheimer’s disease. J Alzheimers Dis 31, 285–300.

51.

Brasnjevic

, Lardenoije

, Schmitz

, Kolk

, Dickstein

, Takahashi

, Hof

, Steinbusch

, Rutten

(2013) Region-specific neuron and synapse loss in the hippocampus of APPSL/PS1 knock-in mice. Transl Neurosci 4, 8–19.

52.

Bittner

, Fuhrmann

, Burgold

, Ochs

, Hoffmann

, Mitteregger

, Kretzschmar

, LaFerla

, Herms

(2010) Multiple events lead to dendritic spine loss in triple transgenic Alzheimer’s disease mice. PLoS One 5, e15477.

53.

Schaeffer

, Catanozi

, West

, Gattaz

(2017) Stereological investigation of the CA1 pyramidal cell layer in untreated and lithium-treated 3xTg-AD and wild-type mice. Ann Anat 209, 51–60.

54.

Tseng

, Kitazawa

, LaFerla

(2004) Amyloid beta-peptide: The inside story. Curr Alzheimer Res 1, 231–239.

55.

Giménez-Llort

, Blázquez

, Cañete

, Johansson

, Oddo

, Tobeña

, LaFerla

, Fernández-Teruel

(2007) Modeling behavioral and neuronal symptoms of Alzheimer’s disease in mice: A role for intraneuronal amyloid. Neurosci Biobehav Rev 31, 125–147.

56.

Lipton

, Rezaie

, Nutter

, Lopez

, Parker

, Kosaka

, Satoh

, McKercher

, Masliah

, Nakanishi

(2016) Therapeutic advantage of pro-electrophilic drugs to activate the Nrf2/ARE pathway in Alzheimer’s disease models. Cell Death Dis 7, e2499.

57.

Bloom

(2014) Amyloid-β and tau: The trigger and bullet in Alzheimer disease pathogenesis. JAMA Neurol 71, 505–508.

58.

Kanekiyo

, Xu

, Bu

(2014) ApoE and Aβ in Alzheimer’s disease: Accidental encounters or partners? Neuron 81, 740–754.

59.

Nielsen

, Mulder

, Beliën

JAM

, Musters

RJP

, Eikelenboom

, Veerhuis

(2010) Astrocytic Aβ1-42 uptake is determined by Aβ-aggregation state and the presence of amyloid-associated proteins. Glia 58, 1235–1246.

60.

Söllvander

, Nikitidou

, Brolin

, Söderberg

, Sehlin

, Lannfelt

, Erlandsson

(2016) Accumulation of amyloid-β by astrocytes result in enlarged endosomes and microvesicle-induced apoptosis of neurons. Mol Neurodegener 11, 38.

61.

Maragakis

, Rothstein

(2006) Mechanisms of disease: Astrocytes in neurodegenerative disease. Nat Clin Pract Neurol 2, 679–689.

62.

Sofroniew

, Vinters

(2010) Astrocytes: Biology and pathology. Acta Neuropathol 119, 7–35.

63.

Fuhrmann

, Bittner

, Jung

CKE

, Burgold

, Page

, Mitteregger

, Haass

, LaFerla

, Kretzschmar

, Herms

(2010) Microglial Cx3cr1 knockout prevents neuron loss in a mouse model of Alzheimer’s disease. Nat Neurosci 13, 411–413.

64.

Giménez-Llort

, García

, Buccieri

, Revilla

, Suñol

, Cristofol

, Sanfeliu

, ol

, Cristofol

, Sanfeliu

(2010) Gender-specific neuroimmunoendocrine response to treadmill exercise in 3xTg-AD mice. Int J Alzheimers Dis 2010, 128354.

65.

García-Mesa

, López-Ramos

, Giménez-Llort

, Revilla

, Guerra

, Gruart

, Laferla

, Cristòfol

, Delgado-García

, Sanfeliu

(2011) Physical exercise protects against Alzheimer’s disease in 3xTg-AD mice. J Alzheimers Dis 24, 421–454.

66.

Filali

, Lalonde

, Theriault

, Julien

, Calon

, Planel

(2012) Cognitive and non-cognitive behaviors in the triple transgenic mouse model of Alzheimer’s disease expressing mutated APP, PS1, and Mapt (3xTg-AD). Behav Brain Res 234, 334–342.