Introduction

It has been hypothesized that amyloid β (Aβ) peptide, the major component of senile plaques, induces neurodegeneration in patients with Alzheimer's disease (Hardy and Selkoe, 2002). Several studies have found that the degree of aggregation of Aβ determines both Aβ toxicity and the nature of toxic manifestations (Pike et al., 1991; Eckman and Eckman, 2007). However, other reports have indicated that, in the brain of AD patients, there is a weak correlation between the severity of neuronal damage and plaque load (Lue et al., 1999; McLean et al., 1999). Accordingly, it has been suggested that soluble oligomeric Aβ mediates all of synaptic loss and dysfunction (Haass and Selkoe, 2007), and oxidative stress (Reddy, 2009), as early events in AD pathology.

Microglia are the major inflammatory cells of the brain, producing inflammatory cytokines and reactive oxygen species (ROS) in response to brain injury (Block, 2008; Perry et al., 2010). A consensus has emerged that downregulation of the clearance of Aβ and neurotoxins from activated microglia promote neuronal degeneration (Akiyama et al., 2000; McGeer and McGeer, 2002; Cameron and Landreth, 2010; Lue et al., 2010). In mixed neuron-microglia cultures, fibrillar Aβ (fAβ) stimulates microglia to produce superoxide by activating NADPH oxidase (Bianca et al., 1999), resulting in neurotoxicity (Qin et al., 2002). Activation of NADPH oxidase has also been detected in AD brains (Shimohama et al., 2000). Such findings suggest that microglia mediate the indirect cytotoxicity of Aβ by increasing the level of oxidative stress in the AD brain.

P2X7 receptor (P2X7R), one of the purinergic receptors, is an ATP-gated cation channel allowing Ca2+ and Na+ influx and K+ efflux (Burnstock, 2007). Recently, P2X7R has been identified as an important element in neuroinflammation and neurodegeneration (Skaper et al., 2010). P2X7R induces interleukin-1β (IL-1β) release from microglia by the activation of IL-1β converting enzyme, mediated, in turn, by the efflux of K+ from cells (Ferrari et al., 2006). In addition, prolonged activation of microglial P2X7R results in an increase in cytoplasmic Ca2+ level, NADPH oxidase activation, and superoxide generation, resulting in neuronal injury upon co-culture of neurons and microglia (Parvathenani et al., 2003; Skaper et al., 2006). Recently, it has been proposed that microglial P2X7R plays a role in the pathology of AD. Upregulation of P2X7R expression was seen both in microglia from AD human brains and in the microglia of Aβ-injected rat brains (McLarnon et al., 2006). However, the mechanisms responsible for Aβ-stimulated NADPH oxidase activation in the AD brain remain unclear. We have previously shown that ATP released from Aβ-stimulated microglia induced ROS production in culture via P2X7R-mediated NADPH oxidase activation (Kim et al., 2007). This observation prompted us to study the possibility that Aβ-mediated P2X7R activation in microglia is associated with Aβ-caused neuronal injury in the AD brain. Characterization of the timing and nature of P2X7R activation during AD development is important to develop an understanding of disease progression. Hence, we here examined age-related P2X7R expression, ROS production in microglia, and neuronal damage, in the APPswe/PS1dE9 (APP/PS1) mouse model of AD.

Results

Age-dependent Aβ accumulation and P2X7R expression

To correlate the P2X7R expression profile with the progress of AD pathogenesis, we first analyzed receptor expression in the cerebral cortex of APP/PS1 (Tg mice), from 1-12 months of age. Western blotting revealed a 75-kDa band corresponding to P2X7R, and the protein levels in Tg mice rose significantly when the animals were 3-12 months of age, compared to age-matched non-Tg littermates. Notably, protein levels increased also in non-Tg mice 12 months of age (Figures 1A and B). Thus, P2X7R levels progressively rose with age, in line with the development of AD pathogenesis. In agreement with the Western blot data, double labeling with thioflavin-S and P2X7R revealed that P2X7R expression increased in parallel with accumulation of thioflavin-S-positive fibrillar Aβ (fAβ). Interestingly, P2X7R expression was evident from 3 months preceding Aβ plaque formation, and strong P2X7R immunostaining was noted not only around Aβ plaques but also in regions distant from such plaques (Figure 1C). These findings prompted us to investigate the identity of P2X7R-expressing cells, and to seek to correlate the expression of P2X7R and Aβ both in the cores of plaques and also in regions where plaques were diffuse.

Figure 1
figure 1

Aβ accumulation and P2X7R expression in APP/PS1 mice. (A) Western blot analysis of P2X7R and actin in the cerebral cortex from 1, 3, 6, and 12 month-old Tg mice and non-Tg control. (B) P2X7R levels were measured at indicated points of time, normalized to those of β-actin and presented as means ± SEM (n = 3 animals for each condition). *P < 0.01 compared with age-matched non-Tg control, using ANOVA and Student-Newman-Keuls analyses. (C) Fluorescent photomicrographs of cerebral cortex section from 3, 6, 10, and 14 month-old Tg mouse labeled with the antibody against P2X7R and thioflavin-S. P2X7R (red) expression increased in parallel with accumulation of thioflavin-S-positive fibrillar Aβ (green). Abbreviations: N, non-Tg; T, Tg. Scale bar: 100 µm.

Full size image

P2X7R expression in CD11b-positive microglia

Triple immunofluorescence studies using antibodies against CD11b, P2X7R, and Aβ17-24 (4G8) were performed on brain sections from 3, 6, 10, and 14 month-old Tg mice. Double labeling with CD11b and P2X7R revealed that the vast majority of cortical cells showing P2X7R expression were CD11b-positive microglia, at both early and late timepoints. Double labeling with P2X7R and 4G8 showed that P2X7R immunostaining co-localized with microdeposits of Aβ both in the cores of plaques and in regions where plaques were diffuse, indicating that non-fibrillar Aβ-mediated P2X7R induction might occur in AD pathology. In support of this notion, triple labeling with CD11b, P2X7R, and 4G8 showed a P2X7R-negative microglial cell in contact with a plaque (Figure 2; from the brain of a 10 month-old mouse). In addition, the intensity of P2X7R staining increased with age, indicating that upregulation of P2X7R potentially played a role in AD progression (Figure 2). Additional experiment to quantify the co-localization of P2X7R in CD11b-postive cells in the frontal cortex and hippocampus was carried out, using 12 month-old Tg mice. The percentage of P2X7R+/CD11b+ in total P2X7R+ cells was 88.8, 86.5, 70.1, and 79.1% within the Layers III to V of frontal cortex, CA1 pyramidal layer, hilus, and granule cell layer of dentate gyrus respectively (see Table 1), indicating that major cell type showing P2X7R expression is CD11b-positive microglia in the frontal cortex and hippocampus of AD brain. In addition, we also observed that 77.7% of cells showing P2X7R expression were CD11b-positive microglia in the frontal cortex of 12 month-old non-Tg mice. These findings extend the previous studies showing the expression of P2X7R in microglia around Aβ plaques in the late stage of a mouse model of AD, and in human AD autopsy brain tissue (Parvathenani et al., 2003; McLarnon et al., 2006) by providing a comprehensive spatial and temporal analysis.

Figure 2
figure 2

P2X7R expression in CD11b-positive microglia. Fluorescent photomicrographs of the cerebral cortex from 3, 6, 10, and 14 month-old Tg mice labeled with antibodies against CD11b (green), P2X7R (red) and Aβ17-24 (4G8) (blue). Double labeling with CD11b and P2X7R showed the P2X7R expression in CD11b-positive microglia (yellow). P2X7R immunostaining co-localized with microdeposits of Aβ (arrowheads). Triple labeling with CD11b, P2X7R, and 4G8 showed a P2X7R-negative microglial cell in contact with a plaque in the brain of a 10 month-old mouse (arrow). Scale bar: 20 µm.

Full size image
Table 1 Colocalization of P2X7R and CD11b in APP/PS1 Tg mice Age: 12 mon; *: percentage of P2X7R+/CD11b+ in total P2X7R+ cells
Full size table

ROS production in P2X7R-positive microglia

To correlate the ROS production profile with expression of P2X7R, we analyzed the expression of the catalytic subunit of NADPH oxidase (gp91phox) in the cerebral cortex of Tg mice 1-12 months of age. Western blot analysis revealed a 58-kDa band corresponding to gp91phox, and the protein levels in Tg mice were significantly increased in the interval 6-12 months of age compared to age-matched non-Tg littermates (P < 0.01) (Figures 3A and B). Notably, the time course of gp91phox expression was similar to the kinetics of P2X7R expression, which began to rise 3 months before gp91phox was detected. To confirm that P2X7R-positive microglia could produce ROS at early stage of AD progression, triple labeling using either antibodies against CD11b, P2X7R, and gp91phox; or antibodies against CD11b and P2X7R, and the chemical hydroethidine (HE); were performed on brain sections of 6 month-old Tg mice. Triple labeling with CD11b, P2X7R, and gp91phox showed that most gp91phox staining could be merged with P2X7R-positive microglia. Finally, triple labeling with CD11b, P2X7R, and HE showed that P2X7R-positive microglia was merged with ethidium fluorescence, indicating that P2X7R-positive microglia could produce ROS in the Tg mouse model of AD (Figure 3C).

Figure 3
figure 3

Gp91phox expression and ROS production in P2X7R-positive microglia. (A) Western blot analysis of gp91phox and actin in the cerebral cortex from 1, 3, 6, and 12 month-old Tg mice and non-Tg control. (B) gp91phox levels were measured at indicated points of time, normalized to those of β-actin and presented as means ± SEM (n = 3 animals for each condition). *P < 0.01 compared with age-matched non-Tg control, using ANOVA and Student-Newman-Keuls analyses. (C) Fluorescent photomicrographs of cerebral cortex sections of 6 month-old Tg mice immunolabeled with either antibodies against CD11b (green), P2X7R (red), and gp91phox (blue) or antibodies against CD11b (green) and P2X7R (blue), and the chemical hydroethidine (HE) (red). Abbreviations: N, non-Tg; T, Tg. Scale bar: 20 µm.

Full size image

Postsynaptic density loss with P2X7R-positive microglia

It is known that synapse loss occurs in AD brains, and an association of Aβ peptides with dendritic spine loss and reductions in the levels of postsynaptic scaffold proteins, such as postsynaptic density-95 (PSD95) protein, have been observed in mouse models of AD (Pham et al., 2010). To determine the time point of PSD loss in Tg mice, we analyzed the levels of PSD95 in the cerebral cortex of Tg mice 1-12 months of age. Western blot analysis revealed a 95-kDa band corresponding to PSD95, and the protein levels in Tg mice were significantly decreased in the interval 6-12 months of age compared to age-matched non-Tg littermates (P < 0.01) (Figures 4A and B). To test whether ROS released from Aβ-stimulated microglia via P2X7R activation correlates with synapse loss, immunofluorescence studies using antibodies against PSD95, P2X7R, and Aβ17-24 (4G8) were performed on brain sections from 6 and 12 month-old Tg mice. First, the density of PSD95 staining in brain sections from such animals appeared to be reduced (Figure 4C, b, c) compared to that in 6 month-old control section (Figure 4C, a). Overall, the PSD95 loss profile was closely associated with P2X7R expression (Figure 4C, b, c). Red spots indicating P2X7R were correlated with sparse PSD95 staining, dendritic dystrophy, and neuronal cell body atrophy at 6 months (Figure 4C, b2), and such neuronal damages were augmented with increase of P2X7R expression at 12 months (Figure 4C, c3).

Figure 4
figure 4

Microglial P2X7R expression and their association with PSD loss in APP/PS1 mice. (A) Western blot analysis of PSD95 and actin in the cerebral cortex from 1, 3, 6, and 12 month-old Tg mice and non-Tg control. (B) PSD95 levels were measured at indicated points of time, normalized to those of β-actin and presented as means ± SEM (n = 3 animals for each condition). *P < 0.01 compared with age-matched non-Tg control, using ANOVA and Student-Newman-Keuls analyses. (C) Fluorescent photomicrographs of cerebral cortex sections of 6- and 12-month-old APP/PS1 (b, c) and 6-month-old non-Tg mice (a) immunolabeled with antibodies against PSD95 (green), P2X7R (red), and Aβ17-24 (4G8) (blue). a1, b2, and c3 are higher magnification images of the areas designated in a, b, and c. Abbreviations: N, non-Tg; T, Tg. Scale bars: a-f, 50 µm; a1-c3, 15 µm.

Full size image

Discussion

Although it has been proposed that activated microglia surrounding amyloid plaques cause neurodegeneration by producing neurotoxins (Akiyama et al., 2000; McGeer and McGeer, 2002), the mechanisms underlying microglia-mediated neuronal damage in the AD brain remain to be elucidated. In the present study, we observed that P2X7R expression and ROS production in Aβ-stimulated microglia were tightly associated with Aβ-mediated synapse loss in a mouse model of AD.

Here, we show that microglia were the major cell type demonstrating P2X7R-immunopositivity in the cerebral cortex of Tg mouse model of AD. Moreover, P2X7R expression was observed before Aβ plaque formation, and evident in regions with diffuse plaques, suggesting that non-fibrillar Aβ may be sufficient for P2X7R induction. Recently, soluble oligomers of Aβ have been suggested as potent stimulators of microglial activation (Hashioka et al., 2005; Jimenez et al., 2008), although most studies have considered fAβ to be a more powerful stimulus (McDonald et al., 1997; Combs et al., 2001). These findings suggest that P2X7R expression in microglia can be induced by soluble Aβ even very early in the course of AD pathogenesis. The ability of oligomeric Aβ to activate microglia provides crucial evidence that a soluble intermediate form of Aβ may cause oxidative damage prior to plaque formation in both humans and animal AD models (Pratico et al., 2001; Sonnen et al., 2008). Moreover, positron emission tomography studies of living AD patients, using the ligand PK11195, which specifically identifies activated microglia, have shown that microglial activation is evident even at early stages of the disease (Cagnin et al., 2001; Okello et al., 2009). Our finding that P2X7R expression precedes Aβ plaque formation strengthens this view. Moreover, a recent study demonstrated that overexpression of P2X7R alone, in the absence of any pathological insult, was sufficient to drive the activation and proliferation of microglia in rat primary hippocampal cultures (Monif et al., 2009). On the while, we observed minor population of P2X7R+/CD11b- cells in the frontal cortex and hippocampal areas of 12 month-old Tg mice, and recent study reported the altered neuronal expression of P2X7R in mouse models of Huntington's disease (Diaz-Hernández et al., 2009). Therefore, we can not rule out the possibility that P2X7R expression in other cell types including neuron is involved in the pathogenesis of AD.

Although it is unclear whether the production of inflammatory mediators in the AD brain enhances or retards disease progression, ROS synthesized by NADPH oxidase have been shown to mediate both Aβ-induced neuronal damage in co-cultures of neurons and microglia (Qin et al., 2002) and Aβ-caused cerebrovascular dysfunction (Park et al., 2005). As evidence that P2X7R is the missing link between Aβ and ROS production, ATP released from Aβ-stimulated microglia induced ROS production in culture via P2X7R-mediated NADPH oxidase activation (Kim et al., 2007), and a recent study showed that intrahippocampal injection of Aβ caused IL-1β accumulation in wild-type brain but not in that of P2X7R knock-out mice (Sanz et al., 2009). In the present study, Western blot analysis and triple immunostaining for CD11b, P2X7R, and gp91phox demonstrated that gp91phox expression and ROS production increased significantly from 6 months of age in the P2X7R-positive microglia (Figure 3). These results indicate that Aβ-stimulated microglia may be involved in the development of oxidative stress even during the early stages of AD, by inducing and activating P2X7R. In support of our findings, a recent study observed a sudden elevation in protein oxidation and lipid peroxidation in brains of APP/PS1 mice at about 6 months of age (Abdul et al., 2008).

Recently, the potential role played by neurotoxic Aβ oligomers in AD pathology has encouraged the development of a modified amyloid hypothesis. Soluble forms of Aβ induce synaptic damage and dendritic spine loss, and levels of synaptic proteins, including the presynaptic protein synaptophysin and the postsynaptic protein PSD95, are decreased in the AD brain (Gylys et al., 2004; Reddy et al., 2005). However, the detailed mechanisms of how Aβ causes such synaptic changes remain unclear. A recent report found that oligomeric Aβ surrounding plaques contributed to excitatory synapse loss in a mouse model of AD, and suggested that oligomeric Aβ may interact directly with synaptic proteins to cause dysfunction and eventual spine collapse (Koffie et al., 2009). On the other hand, recent studies found that oxidative stress was principally localized to synapses of the frontal cortex in the AD brain (Ansari and Scheff, 2010), and that microglia expressed high levels of gp91phox in a vulnerable brain region of mild cognitive impairment (MCI) patients (Bruce-Keller et al., 2010). In this regard, we found that cerebral cortex samples of Tg mice aged 6 months showed significant damage to PSD95-immunoreactive dendrites in regions staining positively for P2X7R (Figure 4). This finding suggests that P2X7R-positive microglia-mediated oxidative stress may cause synaptic loss. Moreover, we observed that most P2X7R and 4G8 immunostaining of Aβ co-localized (Figure 2). Therefore, our present work suggests that P2X7R-mediated ROS production in Aβ-stimulated microglia is one mechanism explaining oligomeric Aβ-mediated synaptotoxicity. Even though correlational studies such as the present work can only suggest contributions of microglial Aβ stimulation and synapse loss to AD pathology, recent studies in support of this hypothesis have shown that microglial activation and hippocampal synapse loss preceded tangle formation in the tau Tg mouse (Yoshiyama et al., 2007), and that inflammation triggered synaptic degeneration in mice with experimental autoimmune encephalomyelitis, an animal model of multiple sclerosis (Centonze et al., 2009). Interestingly, brain sections of 12 month-old non-Tg mice showed that the majority of cells showing P2X7R expression were CD11b-positive microglia. Microglial activation in healthy aging brain and the possibility of involvement in neuronal loss and decline of cognitive function have been reported (Sheng et al., 1998; Cagnin et al., 2001). These imply that microglial P2X7R expression in AD might be influenced to some extent by other aging factors besides Aβ, and contribute to AD pathology.

Together, the results suggest that Aβ-stimulated microglia may induce excessive production of ROS via the action of P2X7R, leading to neuronal damage even at early stages of AD pathogenesis, which may help to explain how microglia mediate neuronal damage in the AD brain.

Methods

Transgenic mouse model

Double transgenic mice expressing a chimeric mouse/human amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9) was obtained from Jackson Laboratory (Bar Harbor, ME) and maintained by crossing transgenic males with C57BL/6J × C3H/HeJ F1 female mice from Jackson Laboratory. APP/PS1 mice carry mouse APP with the double mutations (K595N and M596L) and human PS1 with a deletion of exon 9 found in familial AD patients. APP/PS1 mice used were 1-14 months old. For each experimental group, 3-4 female mice were used. All animal procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Ajou University School of Medicine.

Tissue preparation and immunohistochemistry

Animals were transcardially perfused with a saline solution containing 0.5% sodium nitrate and heparin (10 U/ml) followed by 4% paraformaldehyde dissolved in 0.1 M phosphate buffer (PB). Brains were removed from the cranium and postfixed for 1 h at 4℃, washed in 0.1 M PB and then immersed in 30% sucrose solution until they sank. Tissues were sectioned on a sliding microtome at a thickness of 35 µm, and then stored in stock solution [0.1 M phosphate buffer (pH 7.4), 30% (v/v) glycerol, 30% ethylene glycol] at 4℃ until use. Serial brain section was selected and processed for immunostaining as described previously (Wang et al., 2008). In brief, brain sections were incubated in PBS containing 0.2% Triton X-100 and 0.5% bovine serum albumin (BSA) for 30 min at room temperature (RT), rinsed two times in PBS containing 0.5% BSA, and then incubated overnight at 4℃ with the appropriate primary antibodies: rat anti CD11b (1:200; Serotec, Oxford, UK), rabbit anti P2X7 receptor (1:200; Sigma, St. Louis, MO), goat anti P2X7 receptor (1:200; Santacruz Biotechnology, Santa Cruz, CA), mouse monoclonal 4G8 (amino acids 17-24 of Aβ peptide, 1:200; Covance, Emeryville, CA), mouse anti gp91phox (1:200; BD Biosciences, San Diego, CA), or rabbit anti PSD95 (1:200; Abcam, Cambridge, UK). After incubation, sections were rinsed three times for 5 min in PBS and subsequently incubated with the corresponding secondary antibodies: FITC-labeled anti-rabbit IgG (1:200; Invitrogen, Carlsbad, CA), FITC-labeled anti-rat IgG (1:200; Invitrogen), Texas Red-labeled anti-rabbit IgG (1:200; Vector, Burlingame, CA), Texas Red-labeled anti-goat IgG (1:200; Invitorgen), AMCA-labeled anti-mouse IgG. (1:200; Vector), or AMCA-labeled anti-rabbit IgG. (1:200; Vector) for 2 h at RT. Tissues were mounted with Vectashield mounting medium (Biomed Corp., Foster City, CA) and viewed using an Olympus IX71 confocal laser scanning microscope (Olympus; Tokyo, Japan).

Thioflavin S staining

For thioflavin S staining, brain sections were mounted on gelatine-coated slides. After drying, slides were rinsed with water and incubated with thioflavin S (1% w/v; Sigma) for 5 min. Subsequently slides were rinsed in 70% ethanol for 5 min and briefly rinsed in water. For combination with P2X7 receptor (1:200; Sigma) immunofluorescence staining, sections were first fluorescence labeled as described above and subsequently stained with thioflavin S.

Counting of P2X7R+/CD11b+ cells

The percentage of P2X7R+/CD11b+ in total P2X7R+ cells within the Layers III to V of frontal cortex and hippocampus including CA1 pyramidal layer, hilus, and granule cell layer of dentate gyrus was quantified by counting the cells within a 250 × 250 µm square box, using an IX71 confocal laser scanning microscope (Olympus) at a magnification of 400 ×. The mean number of P2X7R+/CD11b+ cells was calculated from five areas of each section from 3 animals.

In situ detection of O2- and O2--derived oxidants

Hydroethidine histochemistry was performed for in situ visualization of O2- and O2--derived oxidants (Wu et al., 2003). Hydroethidine (1 mg/ml in PBS containing 1% dimethylsulfoxide; Molecular Probes, Eugene, OR) was administered intraperitoneally. After 15 min, brain sections were prepared and mounted on gelatin-coated slides, and the oxidized hydroethidine product, ethidium, was examined by confocal microscopy (Olympus).

Western blot analysis

Cerebral cortex tissues were isolated from transgenic mice and age-matched non-transgenic littermates at each time point. Tissues were homogenized with ice-cold lysis buffer containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM EDTA, 0.1% SDS, 0.5% sodium deoxycholate, supplemented with a protease inhibitors cocktail (Sigma). The homogenate was then centrifuged at 13,000 × g for 10 min and the supernatants collected. Protein concentration was determined using the DC protein assay kit (Bio-Rad, Hercules, CA). Total protein (40 µg) from each sample was separated on a 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, MA). After transfer, membranes were incubated in blocking buffer [tris-buffered saline (TBS) containing 0.5% skim milk] for 1 h at RT and then incubated overnight at 4℃ with following primary antibodies: rabbit anti P2X7 receptor (1:500; Sigma), mouse anti gp91phox (1:1000; BD Biosciences), rabbit anti PSD95 (1:1000; Abcam), and monoclonal anti β-actin (1:5000; Sigma). After several washes in 0.1% Tween-20 (Sigma) in TBS, membranes were incubated with the corresponding HRP-conjugated secondary antibodies (Amersham Biosciences, Buckinghamshire, UK) for 1 h at RT. After washing, blots were developed with ECL detection reagents (Amersham Biosciences), and exposed to X-ray film. The average intensity value of the pixels in a background selected region was calculated and was subtracted from each pixel in the samples using IamgeQuant software (Image Guage 4.0; Fuji Film, Tokyo, Japan). The densitometry values obtained were normalized with respect to the values obtained with anti-β-actin to ascertain the same amount of protein.

Statistical analysis

All values are represented as mean ± SEM. Statistical significance (P < 0.05 for all analyses) was assessed by analysis of variance (ANOVA) using Instat 3.05 (GraphPad Software, San Diego, CA), followed by Student-Newman-Keuls analyses.