Spatial Relationship Between Synapse Loss and β-Amyloid Deposition in Tg2576 Mice

Abstract

Although there is evidence that β-amyloid impairs synaptic function, the relationship between β-amyloid and synapse loss is not well understood. In this study, we assessed synapse density within the hippocampus and the entorhinal cortex of Tg2576 mice at 6–18 months of age using stereological methods at both the light and electron microscope levels. Under light microscopy, we failed to find overall decreases in the density of synaptophysin-positive boutons in any brain areas selected, but bouton density was significantly decreased within 200 μm of compact β-amyloid plaques in the outer molecular layer of the dentate gyrus and layers II and III of the entorhinal cortex at 15–18 months of age in Tg 2576 mice. Under electron microscopy, we found overall decreases in synapse density in the outer molecular layer of the dentate gyrus at both 6–9 and 15–18 months of age, and in layers II and III of the entorhinal cortex at 15–18 months of age in Tg 2576 mice. However, we did not find overall changes in synapse density in the stratum radiatum of the CA1 subfield. Furthermore, in the two former brain areas, we found a correlation between lower synapse density and greater proximity to β-amyloid plaques. These results provide the first quantitative morphological evidence at the ultra-structure level of a spatial relationship between β-amyloid plaques and synapse loss within the hippocampus and the entorhinal cortex of Tg2576 mice.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disease and the most common cause of dementia in the elderly (Forsyth and Ritzline, 1998). The neuropathology of AD is characterized at first by the deposition of senile plaques and neurofibrillary tangles, and later by the loss of neurons and their processes (Price, 1986; DeKosky et al., 1996; Hardy and Selkoe, 2002). Cognitive impairment appears to be most closely correlated in time with the loss of neurons and neuronal processes (Price and Morris, 1999). However, the relationship between plaque and tangle deposition and the neuronal degeneration that follows it is not well understood. Transgenic animal models, such as mice that overexpress amyloid precursor protein (APP) (Hsiao et al., 1996, 2001; Janus and Westaway, 2001), provide a valuable opportunity to study the pathogenesis of AD, and in particular, the relationships between various components of the disease process.

The Tg2576 mouse, which expresses the human “Swedish” mutation, is one of the most well characterized strains of APP transgenic animals (Hsiao et al., 1996). It overexpresses APP, which leads to increased levels of Aβthat eventually polymerize to form β-amyloid plaques similar to those observed in AD (Hsiao et al., 1996). However, tangles and neuron loss are not commonly observed in the Tg2576 mice (Irizarry et al., 1997). While the cellular basis of learning and memory deficits in Tg2576 mice is not well understood (Hsiao et al., 1996, 2001; Barnes et al., 2005), Jacobsen et al. (2006) recently detected a relationship between decreased spine density in the outer molecular layer of the dentate gyrus and the declines in LTP and contextual fear conditioning.

Synapse loss induced by Aβ remains a likely basis for the behavioral deficits observed in Tg2576 mice, since Aβ has been shown to disrupt neuronal function (Stern et al., 2004) and decrease dendritic spine density (Knowles et al., 1998; Lanz et al., 2003; Moolman et al., 2004; Tsai et al., 2004; Spires et al., 2005). However, studies using synaptophysin immunohistochemical staining under light microscopy have not produced consistent findings of synapse loss in Tg2576 mice. (Takeuchi et al., 2000; King and Arendash, 2002; Spires et al., 2005). A shortcoming of the work done to date on synapse morphology in APP transgenic mice is that it has been primarily restricted to the light microscope level. At the light microscope level, decreases in the labeling of pre-synaptic and post-synaptic terminals using synaptophysin and postsynaptic density-95 (PSD–95) may represent decreases in the expression of these proteins rather than the actual number of synapses. For assessing the type and number of synapses, electron microscopy is optimal.

In the present study, we assessed synapse density within the hippocampus and the entorhinal cortex of Tg2576 at 6–9 and 15–18 months of age in Tg2576 mice using stereological methods at both the light and electron microscope level. We found a decrease in synapse density in the outer molecular layer of the dentate gyrus at both 6–9 and 15–18 months of age and in layers II and III of the entorhinal cortex at 15–18 months of age in transgenic mice as compared to age-matched non-transgenic controls. Moreover, decreases in synapse density were correlated with proximity to compact β-amyloid plaques. These results are the first to provide evidence of a spatial relationship between synapse loss and β-amyloid plaques at the ultra-structural level in Tg2576 mice.

Material and methods

A total of 43 mice of both sexes were used for this study (Hsiao et al., 1996). Nineteen mice were divided into cohorts that were 6–9 months old (5 Tg + and 4 Tg-) and 15–18 months old (5 Tg + and 5 Tg-) and used for Nissl staining, synaptophysin and Aβ immunohistochemical studies at the light microscope level. Another 24 mice were divided into cohorts that were 6–9 months old (6 Tg+ and 6 Tg-) and 15–18 months old (6 Tg+ and 6 Tg-) and used for electron microscope studies of synapse density. All experiments were done in accordance to the National Institutes of Health and Institutional Guidelines. The breeding and maintenance of Tg2576 mice were done in consultation with the veterinary staff in the Department of Comparative Medicine at Washington University School of Medicine.

Tg2576 mice were derived from C57B6/SJL x C57B6 crosses (Hsiao et al., 1995, 1996), and contained the double mutation Lys670-Asn, Met671-Leu (K670N, M 671L). This mutation is driven by a hamster prion protein gene promoter in C57B6j x SJL. Levels of APP in the brains of transgenic animals (Tg+) are more than 4 times higher than APP levels in control mice and Aβ levels are 5–14 times higher than Aβ levels in control mice (Hsiao et al., 1995, 1996). The presence or absence of the human APP gene was demonstrated in individual animals by post-weaning tail biopsy and DNA genotyping. After DNA extraction, PCR analysis was performed using primers as previously described (Hsiao et al., 1996). Non-transgenic littermates of the Tg2576 mice (Tg-) were used as control animals.

Tissue Preparation and Section Sampling

For light microscope studies, animals were deeply anesthetized using a 3.3 ml/kg ketamine/xylazine mixture (86:13mg/ml) and perfused transcardially with 1% heparinized 0.01M phosphate buffer (PBS) for 2 minutes and then 4% paraformaldehyde for 25–30 minutes. Brains were removed and post-fixed overnight at 4° C using the same fixative with 30% sucrose. Next, the brains were dissected and embedded in Tissue-Tek embedding medium (Electron Microscopy Sciences, Hatfield, PA), and cut into 35μm thick sections in the coronal plane using a cryostat (Leica CM 1850 UV, Nussloch, Germany). The cortex was nicked in the left hemisphere prior to sectioning to serve as a marker of orientation and the sections were mounted onto slides with a consistent orientation. Approximately 80 sections from each animal were divided into equal quarters (~20 sections) for the stereological counting of synapses using a systematic-uniform-random (SUR) manner. Every second section selected from the initial sampling was sub-sampled with a random start to give a final SUR sample of 10–11 sections (West et al, 1991). Of the 6 series of sections (12–15 sections each), we used one series for synaptophysin staining, one series for synaptophysin and β-40 double labeling, one series for Thioflavin S staining, and one series for Nissl staining. The remaining 2 series were stored for future studies.

For the electron microscope studies, animals were deeply anesthetized with 3.3 ml/kg of a ketamine/xylazine (86:13mg/ml) mixture, and then perfused transcardially with 0.01M PBS containing heparin sodium for 2 minutes, followed by a 30 minute perfusion with 2.0% paraformaldehyde, 2.0% glutaraldehyde and 4% sucrose in 0.1M PBS. After perfusion, the animals were placed in sealed plastic bags and kept in a refrigerator for 2 hours at 4° C. The brains were then removed and post-fixed in the same fixative at 4° C for at least 24 hours. Sectioning was performed in the coronal plane using a vibratome to obtain 200 μm thick sections. The resultant fifteen sections from each brain (encompassing the whole hippocampus and entorhinal cortex) were rinsed in cold 0.1 M PBS, treated with 2% OsO4 in 0.1M PBS for 90 min at 4° C and rinsed again in 0.1 M PBS at room temperature. The sections were then dehydrated in a graded series of ethanol solutions, followed by propylene oxide, and left overnight in a mixture (1:1) of propylene oxide-Polybed 812 (Electron Microscopy Sciences, Hatfield, PA). Finally, the sections were flat embedded in Polybed 812 in an oven at 60° C for 48–72 hours. From the 15 embedded sections, five were selected for semi-thin and ultra-thin section cutting. The selection was made to include representative areas of dorsal, medial and ventral hippocampus and lateral and medial entorhinal cortex (one section/per area).

Selected sections were trimmed and sectioned using a Reichert Ultracut E ultramicrotome (Austria). Semi-thin (1–2 μm) sections were cut as reference sections before and immediately after ultra-thin sectioning, and stained with Nissl (see Figure 2A). Ultra-thin (75–90 nm) sections containing the stratum radiatum of CA1 and the outer molecular layer of the dentate gyrus (dorsal blade), or the lateral and medial entorhinal cortex, were mounted on 400 mesh grids (every mesh grid is 62X62 μm², Electron Microscopy Sciences, Hatfield, PA). The sections were stained using 3% uranyl acetate for 20 min followed by lead citrate for 5 min and then were examined using a JEOL 100 CX electron microscope (Japan).

Figure 2.

Panel A: Semi-thin sections of the hippocampus used to identify hippocampal subregions for ultra-thin section identification. Compact β-amyloid plaques (P) in semi-thin sections are distinguishable with Nissl staining. SO: stratum oriens; SP: stratum pyramidale; SR: stratum rediatum; LM: stratum lacunosum-moleculare; ML: molecular layer of dentate gyrus; DG: dentate granule cell layer. Panel B: Electron micrographs used for synapse counting. Images were scanned into a computer dataset and stereological methods were used for synapse counting. The red dots indicate counted synapses.

Immunohistochemical and Nissl Staining

For synaptophysin immunohistochemical staining, previously described methods were employed (Phinney, et al., 1999). Briefly, SUR selected sections were pre-incubated in Tris-buffered saline (TBS, 0.05M) with 0.2% Triton X –100 (TBS-T, pH 7.4) containing 5% normal horse serum (NGS; Sigma, St. Louis MO) at room temperature for 30 minutes. The sections were then incubated with a monoclonal antibody against synaptophysin (clone SVP38, 1:50, Monosan Inc, Frontstraat, Netherlands) in TBS-T overnight at 4° C. After washing with TBS, the sections were incubated with biotinylated secondary antibody (anti-mouse IgG, 1:200) for 2 hours at room temperature. The sections were rinsed again in TBS, and then incubated with an avidin-biotin complex for 1 hour at room temperature (Vector Laboratories, Burlingame, CA). Synaptophysin immunoreactivity was visualized using a DAB staining kit (Vector Laboratories, Burlingame, CA). sections were washed with PBS, mounted on gelatin-coated slides, dehydrated, cleared and cover-slipped using Permount.

For immunohistochemical staining of β-amyloid plaques, SUR selected sections were rinsed with 0.1 M PBS (pH 7.4), and incubated in a blocking solution of 5% normal goat serum for 1 hour. Sections were then incubated overnight in the primary antibody Aβ-40 at 4° C (rabbit polyclonal pan antibody, 1:1000, Biosource Camarillo, CA). After PBS washing, the sections were incubated in biotinylated anti-rabbit secondary antibody for 2 hours at room temperature, then in an avidin-biotin complex for 1 hour at room temperature (Vector Laboratories, Burlingame, CA). Aβ-like immunoreactivity was visualized using a DAB kit (Vector Laboratories, Burlingame, CA). Aβ-40 stains both diffuse and compact plaques. However, compact plaques stained with Aβ-40 were distinguished from diffuse plaques by the characteristic presence of a densely stained and compact central core. To confirm the presence of compact (fibrillar) β-amyloid plaques, thioflavine S was used to stain floating sections in a 1% thioflavine S aqueous solution for 5 minutes, and then differentiated in 70% alcohol 3–5 minutes (Guntern et al., 1992)

To perform β-amyloid and synaptophysin double-labeling, anti-rabbit antibody conjugated to cyanine 3(Cy3) secondary antibody (1:100; Jackson ImmunoResearch, West Grove, PA) and anti-mouse antibody conjugated to Alexa 488 secondary antibody were used (1:100, Molecular Probes, Eugene, Oregon). Selected stained sections were examined using Olympus Fluoview 1000 confocal microscopy (Olympus America INC. Melville, NewYork).

In order to test the specificity of the antibodies, three different control studies were performed. For synaptophsin or Aβ-like immunoreactive labeling only, the primary antibody was omitted. For synaptophysin and Aβ double labeling experiments, the primary antibodies and secondary antibodies were mismatched and the specificity of the mouse and rabbit IgG’s was tested for the appearance of fluorescent labels in the inappropriate confocal channel. Finally, we tested the specificity of the primary antibody by pre-absorbing the rabbit anti-Aβ anti-serum with its control peptide (Aβ 1-40) to detect if the antibody activity were blocked.

For Nissl staining, SUR selected sections were transferred into ddH2 O for 3 minutes and then into 50 % EtOH for 3 minutes. The slides were bathed in 0.3% cresyl violet in 50% EtOH until staining was dense (about 1 minute); they were then moved to 50% EtOH with glacial acetic acid and destained to the desired intensity (about 1 minute). The sections were dehydrated in an ethanol gradient for 5 minutes (50%, 70%, 90%, 3 times in 100%), cleared in three 5 minutes baths of xylenes, and then cover-slipped. The density of CA1 pyramidal neurons and their soma size were quantified using Nissl stained sections.

Synaptophysin-Positive Bouton and Neuron Counting

The Stereologer^TM system (Olympus Denmark A/S, Albertslund, Denmark) was used for stereological counting. This system includes a light microscope equipped with a color video camera and high-resolution color video card with output to a color monitor and a motorized stage controlled by joystick to record movement in the x-y-z axes. The system also includes closed-loop feedback control to ensure accurate stage movement in the z-axis, and system software (CAST-GRID program), which overlays counting frames onto the video image.

Synaptophysin-positive boutons were quantified using stereological methods previously described (Calhoun et al., 1998; Phinney et al., 1999; Rutten et al., 2005). Two sampling methods were used for synaptophysin-positive bouton analysis. Initial sampling was at low magnification in the stratum radiatum of the CA1 subfield, the outer molecular layer of the dentate gyrus and layers II and III of the entorhinal cortex. Boundaries were identified according to cellular structures. Synaptophysin-positive bouton counting was done using an optical fractionator method (West et al., 1991) under a 100X objective. The density of synaptophysin-positive boutons (boutons per mm³ ) was calculated by dividing the total number of boutons counted by the volume sampled (Bonthius et al., 1992, 2004). The sampled volume was determined as the number of disectors multiplied by the volume of one disector (312.5 μm³ with 1.5 μm guard height). Secondary sampling was performed using sections double-labeled for synaptophysin and β-amyloid plaques. Areas 200 μm adjacent to compact amyloid plaques were then selected for synaptophysin-positive bouton counting using the same method as initial sampling. Twenty areas in each animal (12 areas for hippocampus and 8 areas for entorhinal cortex) were randomly chosen from the sections for this propose.

The counting of pyramidal neurons (i.e., cells with a visible nucleolus) was performed in the pyramidal cell layer of the dorsal CA1 subfield of the hippocampus. The border of the CA1 subfield with the CA2 subfield was defined as the region where the neurons increased in size and became more loosely packed, while the border between the CA1 subfield and the subiculum was defined by the bundle of perforant path fibers. Neuron counting was done using an optical fractionator method (West et al., 1991) and the density of pyramidal neurons (cells per mm³ ) was calculated by dividing the number of cells counted by the total volume sampled (Bonthius et al., 1992, 2004). The sampled volume was determined as the number of disectors multiplied by the volume of one disector (37.8 mm³ ). Neuron size was estimated using the rotator measurement (Gundersen et al., 1997). Rotator measurement provides estimates of individual cell size without assumptions regarding the shape or orientation of the cells, provided that the cells are either isotropic in nature or that isotropic sampling is applied, in this case through the use of a vertical design. A unique reference point was associated with each cell (i.e., the nucleolus). Rotator measurements were obtained using the CAST-GRID program.

Assessment of Synapse Number by Electron Microscope

Synapse density was determined within the stratum radiatum and the outer molecular layer of the dorsal, medial and ventral CA1 and dentate gyrus, and layers II and III of the lateral or medial entorhinal cortex. Two sampling approaches were used for this analysis. Initial sampling was performed using landmarks from adjacent semi-thin sections as a reference. At low magnification, the boundaries of the stratum radiatum of the CA1 subfield, the outer molecular layer of dentate gyrus and layers II and III of the entorhinal cortex boundaries were identified according to their characteristic cellular structures (see Figure 2 A).Then, 8–15 photographs from each section were taken systematically at 800x magnification using alternate grid squares. Secondary sampling was performed using landmarks and β-amyloid plaques observed in adjacent semi-thin sections (compact plaques were distinguished with Nissl staining by the presence of a compact core of amyloid surrounded by dystrophic neuritis, see Figure 2A). Compact plaques were identified under electron microscopy by the characteristic presence of a dense core of amyloid, dystrophic neuritis and a variable microglial and astrocytic process (see Figure 3 A; Wegiel et al., 2001). Neuropil areas adjacent to amyloid plaques were then systematically traversed and photomicrographs were taken at 8,000 X from the edges of observed plaques to alternate grid squares (i.e., total 20 photographs/ section, 5 photographs from each direction, 4 directions/ plaque (see Figure 1). Five sections (3 for the hippocampus and 2 for the entorhinal cortex) from each animal, including the dorsal, medial and ventral hippocampus and the lateral and medial entorhinal cortex, were assessed. A total of 1531 photographs were taken from the hippocampus and 696 photographs were taken from the entorhinal cortex for analysis. Synapses were identified under electron micrographs that were enlarged photographically to a final magnification of 29,000X. A magnification standard (grating replica) was used for each series of electron micrographs.

Figure 3.

Amyloid and synaptophysin double-labeling in entorhinal cortex of Tg2576 mice at 18 months. Panel A: β-Amyloid plaque labeled with Cy3. Panel B: Synaptophysin labeled with alex 488. Panel C: Intense synaptophysin immunostaining on the periphery of the plaque (arrow). Bar in C also for A,B = 50 μm.

Figure 1.

Sampling method for electron microscopy. Figure 1 represents an ultra-thin section mounted on a 400 mesh grad. The red circle in the center represents an amyloid plaque. From the edge of the plaque, one photograph per grid (i.e., 1, 2, 3….) was taken, randomly choosing one grid from 4 different directions. The black circle line indicates proximity to the nearest plaque. The distance between grid bars is 62 μm.

A stereologic disector technique was used to measure synapse density (West and Gundersen, 1990; Geinisman et al., 2000). Each disector consisted of micrographs of two adjacent ultrathin sections, a reference section and a look-up section immediately above it. Synapses were identified on photographs by the presence of synaptic vesicles and postsynaptic densities. Both asymmetrical (i.e. excitatory glutamatergic synapses) and symmetrical synapses (i.e. inhibitory GABAergic synapses) (Watson, 1988; Kennedy, 2000; Lund et al., 2001) were counted. Only those synapses that occurred in the reference but not in the look-up section were counted (see Figure 2 B). The area of the unbiased counting frame was 247 μm², the disector height was 0.085 μm, and the disector volume is 20.99μm³. The latter value was used to calculate the synapse density (synapses per unit volume) as the quotient of the mean number of synapses counted per disector and the mean volume of the disectors.

Data analysis

Anatomic variables were compared across groups using two-way ANOVA. When genotype effects (i.e., transgenic (Tg+) mice versus non-transgenic (Tg-) controls), time effects, or genotype x time interactions were found, post-hoc analyses were performed using a Fisher’s Protected Least Squares Design (PLSD) tests. Repeated measures ANOVA was used to examine the effects of plaque proximity on synapse density. For ANOVA, statistical significance was accepted for p-values less than 0.05. In order to estimate the precision of stereologic counting in individual subjects, coefficients of error were calculated by dividing the standard error by the mean value of the synapse densities (Bonthius et al., 2004; Jacobs et al., 2005).

Results

Light microscope analysis of synaptophysin-positivec bouton density

Synaptophysin immunostaining was dense in the neuropil of the hippocampus and the cortex in both Tg+ and Tg- mice. Within the hippocampal formation, the pyramidal cell layer and granule cell layer of the dentate gyrus showed very light staining. However, the mossy fiber area showed intense clusters of staining. Synaptophysin immunostaining in the entorhinal cortex appeared more evenly distributed except in areas occupied by neuronal soma. At higher magnification, the synaptophysin immunostaining appeared distinctly granular and synaptic boutons were apparent (see Figure 3B). At the periphery of β-amyloid plaques, synaptophysin immunostaining occurred in large clusters, but the cores of such plaques exhibited little affinity for the synaptophysin antibody (see Figure 3C).

There was no significant effect of genotype (F=2.94, df=1,15, p>0.05) or age (F=1.24, df=1,15, p>0.05), nor was there a significant genotype by age interaction (F=0.17, df=1,15, p>0.05), on synaptophysin-positive bouton density within the stratum radiatum of CA1 (see Table 1). There was also no significant effect of genotype (F=2.73, df=1,15, p>0.05) or age (F=1.24, df=1,15, p>0.05 ), nor was there a significant genotype by age interaction (F=0.31, df=1,15, p>0.05) on synaptophysin-positive bouton density within the outer molecular layer of the dentate gyrus. Finally, there was no significant effect of genotype (F=2.95, df=1, 15, p>0.05) or age (F=2.58, df=1,15, p>0.05), nor was there a significant genotype by age interaction (F=0.46, df=1,15, p>0.05), on synaptophysin-positive bouton density within layers II and III of the entorhinal cortex.

Table 1.

Synaptophysin-positive bouton density in the stratum radiatum of CA1, the molecular layer of the dentate gyrus and layers II and III of the entorhinal cortex (per mm³ X10⁹)

	Hippocampus				Entorhinal Cortex
	SR of CA1		ML of Dentate Gyrus		Layer II and III
Age (N)	Tg+	Tg−	Tg+	Tg−	Tg+	Tg−
6 – 9m (9)	0.512±0.064 (0.056)	0.545±0.053 (0.047)	0.536±0.091(0.076)	0.562± 0.026 (0.023)	0.484±0.039 (0.040)	0.502±0.054 (0.034)
15–18m (10)	0.473±0.047 (0.044)	0.527±0.056 (0.047)	0.459±0.049(0.048)	0.524± 0.104 (0.088)	0.444±0.033 (0.034)	0.486±0.026 (0.025)

SR= stratum radiatum; ML=molecular layer of dentate gyrus; N= the number of animals. Data displayed in table are group means of bouton density ± SD. The numbers in parentheses are coefficients of error.

Electron microscope analysis of synapse density

There was no significant effect of genotype (F=0.29, df=1,20, p >0.05) or age (F=1.38, df=1,20, p>0.05), nor was there a significant genotype by age interaction (F = 0.68, df=1,20, p>0.05), on synapse density within the stratum radiatum of CA1. However, there was a significant overall effect of genotype (F=26.22, df=1.20, p<0.001) and age (F=9.43, df=1,20, p<0.01), but no significant genotype by age interaction (F=0.53, df=1,20, p>0.05), on synapse density in the outer molecular layer of the dentate gyrus. Post-hoc analyses showed that synapse density was significantly decreased at 6–9 months of age and 15–18 months of age in Tg+ as compared to Tg- mice (see Table 2).There was also a significant overall effect of genotype (F=6.34, df=1,20, p<0.05), but not of age (F=1.43, df=1,20, p>0.05), and a significant genotype x age interaction (F=4.69, df=1,20, p<0.05), on synapse density within layers II and III of the entorhinal cortex. Synapse density was significantly decreased at 15–18 months of age in Tg+ mice as compared to Tg- mice (see Table 2).

Table 2.

Synapse density in the stratum radiatum of CA1, the molecular layer of the dentate gyrus and layers II and III of the entorhinal cortex (per mm³ X10⁹)

	Hippocampus				Entorhinal Cortex
	SR of CA1		ML of Dentate Gyrus		Layer II and III
Age (N)	Tg+	Tg−	Tg+	Tg−	Tg+	Tg−
6–9 m (12)	3.09±0.61 (0.093)	3.04±0.26 (0.036)	2.38±0.39 (0.067) *	3.00±0.31 (0.042)	2.14±0.09 (0.042)	2.18±0.14 (0.064)
15–18m (12)	2.74±0.41 (0.058)	2.98±0.39 (0.053)	1.85±0.36 (0.079) *	2.67±0.31 (0.046)	1.76±0.10 (0.056) *	2.29±0.11 (0.048)

SR= stratum radiatum; ML=molecular layer of dentate gyrus; N= the number of animals. Data displayed in table are group means of synapse density ± SD. The numbers in parentheses are coefficients of error.

^*Post-hoc analyses of Tg + versus Tg − animals at the indicated age, p <0.05.

Spatial relationship between synapse density and β–amyloid plaques

We examined synapse density as a function of distance from compact plaques using both light and electron microscopy in 15–18 month old animals. Under light microscopy, we counted synaptophysin-positive boutons within 200 μm of compact plaques and in plaque-free regions in the hippocampus (the outer molecular layer of the dentate gyrus) and the entorhinal cortex (layers II and III) in Tg+ mice. Under light microscopy, there was a significant effect of plaque proximity on synaptophysin-positive bouton density in the hippocampus (F=9.52, df=1,13, p<0.05) and in the entorhinal cortex (F=25.22, df=1,13, p<0.001) in Tg+ mice. Under electron microscopy, we examined synapse density in sampling regions 62, 124, 186 and 248 μm from the edges of compact β-amyloid plaques in the hippocampus (the outer molecular layer of the dentate gyrus) and the entorhinal cortex (layers II and III) in Tg+ mice and compared it to analogous regions of the same brain areas in Tg+ and Tg- mice (see Figure 4). There was a significant effect of sampling region (F=5.83, df = 4,36, p<0.01) and a significant sampling region by genotype interaction (F=4.94, df = 8,36, p<0.001) on synapse density in the hippocampus. There was also a highly significant effect of sampling region (F=28.22, df=4, 48, p<0.0001), as well as a significant sampling region by genotype interaction (F=13.28, df=8,48, p<0.0001), on synapse density in the entorhinal cortex (see Figure 5).

Figure 4.

Electron microscope photographs of the entorhinal cortex of Tg2576 mice at 18 months of age. Panel A: Compact β-amyloid plaque and dystrophic neuritis. Panel B: Neuropil at 62 μm from a β-amyloid plaque, synapse density is significantly decreased in association with degenerated neurites. Panel C: Neuropil at 124 μm from the edge of a β-amyloid plaque, showing decreased synape density number as compared to an area distant from the plaque (>300 μm, Panel D). The red dots indicate counted synapses. P: the core of β-amyloid plaque; DN: dystrophic neuritis. Bar in D also for A,B,C = 1.5 μm.

Figure 5.

Synapse density in proximity to compact β–amyloid plaques in the hippocampus and the entorhinal cortex in Tg+ mice at 15–18 months of age (see text for statistics). Tg+P: In this group, sampling regions were located from 62 to 248 μm from plaques (P) in Tg+ mice; Tg+NP: In this group, sampling regions were located in areas with no plaques (NP) in Tg+ mice; Tg-: In this group, sampling regions were located in analogous areas as those sampled in Tg+P and Tg+NP mice.

Neuron density and size in the CA1 subfield of the hippocampus

There was no significant effect of genotype (F=1.02, df=1,16, p>0.05) or age (F=3.42, df = 1,16, p >0.05) on neuron density, nor was there a significant genotype by age interaction (F=1.41, df=1,16, p>0.05). Also, there was no effect of genotype (F=0.23, df=1, 10, p>0.05) or age (F=1.59, df=1,10, p>0.05) on neuron size.

Control staining and stereology techniques

When the primary antibodies were omitted and the sections were incubated with secondary antibodies and stained with DAB, no DAB-labeled synaptophysin-positive boutons or plaques were observed. When primary and secondary antibodies were mismatched, no signals were detected in the opposite channel indicating there was no cross-reactivity between the mouse and rabbit IgG’s. When control peptide Aβ 1–40 was pre-absorbed with polyclonal rabbit anti-Aβ, the tissue appeared totally unlabeled.

As shown in Tables 1 and 2, the coefficients of error (CE) for measurements on most individual subjects were less than the recommended target CE of 0.06. These results suggest that the stereology techniques employed in this study produced precise estimates of the synapse density.

Discussion

In the present study, we found a spatial relationship between synapse density and β-amyloid plaques using both light and electron microscopy in Tg2576 mice. Using quantitative light microscopy, we did not find a significant change in overall synaptophysin-positive bouton density in the stratum radiatum of CA1, in the outer molecular layer of the dentate gyrus or in layers II and III of the entorhinal cortex in Tg+ animals as compared to Tg- controls. However, we did find a decrease in synaptophysin-positive bouton density in the vicinity (i.e., within 200 μm) of individual β-amyloid plaques in the hippocampus and the entorhinal cortex. Using quantitative electron microscopy, we found a significant overall decrease in synapse density in the outer molecular layer of the dentate gyrus at both 6–9 months of age and 15–18 months of age and in the layers II and III of entorhinal cortex at 15–18 months of age in Tg+ mice as compared to age-matched Tg- controls. Consistent with our light microscope results, we also found a correlation between synapse density and proximity to compact plaques in the hippocampus and the entorhinal cortex.

The hippocampus and entorhinal cortex play major roles in learning and memory (Knowles et al., 1992; Alvarez et al., 1995; Bannerman et al., 2001; Buckmaster et al., 2004). Also, they are primary sites for β-amyloid deposition in APP transgenic mice (Su and Ni, 1998; Reilly et al., 2003). The perforant path that projects from the entorhinal cortex to the dentate gyrus is among the most vulnerable pathways in cortex with respect to both aging and AD pathology (Scheff et al., 1996; Gazzaley et al., 1997). In addition, there have been recent reports of decreased spine density and alterations in dendritic branching in the outer molecular layer of the dentate gyrus in both Tg2576 mice and PDAPP mice (Wu et al., 2004; Jacobsen et al., 2006). These findings and our quantitative electron microscope results support the hypothesis that the outer molecular layer of the dentate gyrus is especially vulnerable to the effects of APP overexpression. We also found numerous diffuse and compact plaques in the outer molecular layer of the dentate gyrus and the entorhinal cortex in Tg+ mice at 15–18 months of age, while there were relatively few plaques in the stratum radiatum of CA1. This phenomenon may explain the relative absence of synapse loss in the stratum radiatum of the CA1 subfield in Tg+ mice in this study. The spatial association between β-amyloid deposition and synapse loss in Tg2576 mice suggests a causal role for Aβ in the neurodegeneration associated with AD.

β-Amyloid has been proposed to play a key role in inducing the process of neurodegeneration in AD (Selkoe et al., 1991; Hardy and Selkoe, 2002). Knowles et al. (1998) used three-dimensional triple immunofluorescent confocal microscopy to investigate the structural relationship between β-amyloid deposits and neuritis in postmortem cortical tissue from AD patients and found that β-amyloid deposition decreased SMI32 (a marker for neurofilament H) immunoreactivity, which correlates closely with neuron loss and dementia. Tg2576 mice, which overexpress human APP, develop age-related β-amyloid plaques and behavioral impairments (Hsiao et al., 1996), but little if any neuron loss has been observed (Irizarry et al., 1997; Takeuchi et al., 2000). However, our results, as well as the results of other studies (Urbanc et al., 2002; Spires et al., 2005) suggest that β-amyloid deposition may be associated with sublethal neuronal damage (i.e., synapse loss). Also, intracellular recordings from pyramidal neurons in Tg2576 mice show that β-amyloid plaques are associated with defects in cortical synaptic integration (Stern et al., 2004) and morphological studies in Tg2576 mice show that β-amyloid plaques are related to decreases in dendritic spine density (Lanz et al., 2003; Moolman et al., 2004; Tsai et al., 2004; Spires et al., 2005). Finally, Urbanc et al. (2002) found that thioflavin-S positive plaques (fibrillar plaques) are selectively neurotoxic. Consistent with these studies, we found the most dramatic decreases in synaptic density in the vicinity of compact β-amyloid plaques in the hippocampus and entorhinal cortex of Tg2576 mice.

β-Amyloid peptide accumulates in the brains of AD patients and in animal models of the disease (e.g., APP transgenic mice) in a number of forms, ranging from soluble Aβ oligomeric entities to insoluble fibrillar plaques. In transgenic mouse models of AD, β-amyloid deposits have been classified as diffuse or compact, with compact deposits being more frequently associated with neuritic changes (Reilly et al., 2003). Recent studies have shown that soluble, low-molecular-weight (8-24 kDa) Aβ oligomers, referred to as β-amyloid-derived diffusible ligands, may be capable of producing neuronal dysfunction (Lambert et al., 1998; Walsh et al., 2002; Gong et al., 2003; Mattson, 2004). Morphological studies also indicate that Aβ oligomers can accumulate within neuronal processes (Takahashi et al., 2004; Kokubo et al., 2005). Several studies have demonstrated that soluble Aβ can disrupt synaptic function and impair LTP (Larson et al., 1999; Moechars et al., 1999; Chapman et al., 1999; Cleary et al., 2005). In support of the hypothesis that soluble Aβ can be neurotoxic, we also found decreases in synapse density in the outer molecular layer of the dentate gyrus in Tg2576 mice at an age when soluble Aβ is typically increased (i.e., 6–9 months of age), but when plaque deposition is minimal (Holcomb et al., 1998). However, it remains difficult to determine whether synapse loss in proximity to β -amyloid plaques is due to some type of plaque-related tissue disruption or to increased concentrations of Aβ oligomers. Plaques could be a source of specifically toxic species of oligomeric Aβ (Walsh et al., 2002). It also has been hypothesized that Aβ may be toxic through an excitotoxic cascade (i.e. deregulation of calcium homeostasis inducing oxidative stress and abnormal increases in glutamate release), which is facilitated by the proximity of plaques (Mattson, 1997; El Khoury et al., 1998; Mclellan et al., 2003; Hynd et al., 2004).

There were notable discrepancies between the results of our light microscope and electron microscope studies. For example, the data from our electron microscope studies, but not our light microscope studies, suggested generalized decreases in synapse density in the outer molecular layer of the dentate gyrus and in the entorhinal cortex of Tg2576 mice. This discrepancy might be due to the fact that synaptophysin labels a specific protein within the presynaptic terminal; thus, allowing us to count only synapses with a functional presynaptic terminal. β-Amyloid plaques may destabilize synaptic elements without completely destroying whole axons or dendrites. Thus, Spires et al. (2005) found decreases in dendritic spines but not axons in proximity to β-amyloid plaques. Our observation of synaptophysin clusters at the margins of β-amyloid plaques (see Figure 3) supports this hypothesis. The presence of synaptophysin positive clusters in the vicinity of β-amyloid plaques is consistent with previous work (King et al., 2002). However, the molecular basis of these clusters is not clear. Such clusters may represent the rearrangement of presynaptic structural elements after the degeneration of post-synaptic elements.

In an attempt to relate β-amyloid deposition to neuronal shrinkage, which might have influenced neuron counting, we measured soma size in addition to neuronal density in Tg2576 mice and controls. However, we did not find an effect of genotype on soma size or neuron density in the CA1 pyramidal layer at 6–9 or 15–18 months of age. These results are consistent with the findings of previous studies that failed to find neuron loss despite β-amyloid deposition in Tg 2576 mice (Irizarry et al., 1997; Takeuchi et al., 2000).

The neurodegeneration observed in AD is mainly restricted to cell bodies and the dendrites of glutamatergic neurons in layers III and IV of the neocortex (Albin and Greenamyre, 1992). Furthermore, the densities of NMDA receptors are decreased (Wakabayashi et al., 1999; Paneggress et al., 2002). Recent studies suggest that glutamate receptors may be involved in Aβ-induced synaptic dysfunction (Snyder et al., 2005; Tyszkiewicz and Yan, 2005; Floden at al., 2005), and soluble Aβ-40 can induce NMDA-dependent degradation of a postsynaptic density-95 protein within glutamatergic synapses (Roselli et al., 2005). In future studies, it would be important to determine whether specific synapse types (glutamatergic versus GABA) are differentially reduced in proximity to β-amyloid plaques.