Intraneuronal accumulation of β-amyloid (Aβ)42 is one of the earliest pathological events in humans and in animal models of Alzheimer's disease (AD). Apolipoprotein E 4 (APOE4) is the major identified genetic risk factor for late-onset AD, with Aβ deposition beginning earlier in apoE4-positive subjects. To directly determine the effects of APOE genotype on intraneuronal accumulation of Aβ1–42 at the onset of AD pathogenesis, we introduced lentiviral Aβ1–42 into the cortex of APOE targeted replacement (TR) mice at the age of 8–9 months. We demonstrated a significant isoform-dependent effect of human APOE, with dramatically enhanced intracellular Aβ1–42 deposits in the cerebral cortex of APOE4-TR mice 2 weeks after injection. Double-immunofluorescent staining showed that intracellular accumulation of lentiviral Aβ1–42 was mainly present in neurons, localized to late endosomes/lysosomes. This intraneuronal accumulation of Aβ1–42 correlated with increased tau phosphorylation and cell death in the ipsilateral cortex around the injection site. Aβ1–42 was also observed in microglia, but not in astrocytes. Quantitative analysis revealed more neurons with Aβ1–42 while less microglia with Aβ1–42 nearest to the injection site of Aβ1–42 lentivirus in APOE4-TR mice. Finally, apoE was present in neurons of the ipsilateral cortex of APOE-TR mice at 2 weeks after lentivirus injection, in addition to astrocytes and microglia in both the ipsilateral and contralateral cerebral cortex. Taken together, these results demonstrate that apoE4 tips the balance of the glial and neuronal Aβ toward the intraneuronal accumulation of Aβ.
The accumulation of extracellular plaques composed of β-amyloid (Aβ) is a hallmark pathological feature of Alzheimer's disease (AD). However, emerging evidence from transgenic mice and human patients indicates that this peptide can also accumulate intraneuronally. Intraneuronal accumulation of Aβ is one of the earliest pathological events in humans and in animal models of AD (1–4). Studies using specific antibodies have confirmed that Aβ42 is the form of Aβ present within neurons of human brains with AD pathology (5–7). Aβ42 represents ∼10% of synthesized Aβ, but is the earliest and predominant Aβ species deposited in plaques in AD (8). In vivo experiments using Aβ1–42 transgenic mice demonstrated that Aβ1–42 is sufficient to promote Aβ deposition (9). Little is known about the factors that initiate or modulate the onset of brain Aβ accumulation in sporadic, late-onset AD, which represents the vast majority of AD cases.
The strongest and best-established genetic risk factor for sporadic, late-onset AD is the apolipoprotein E (APOE) ɛ4 allele. The APOE ɛ4 allele dramatically increases AD risk and decreases age of onset, while the APOE ɛ2 allele decreases AD risk (10–13). In AD brains, apoE is a component of most plaques, and the APOE genotype influences the levels of Aβ deposition (11,14). Current evidence supports models in which apoE isoforms differentially modulate AD risk and onset through effects on Aβ aggregation and clearance (15–17).
In the present study, human APOE targeted replacement (TR) mice (18) were used to test the effects of APOE genetype on the early accumulation of intraneuronal Aβ42in vivo. APOE2-, APOE3- and APOE4-TR mice express the human APOE ɛ2, APOE ɛ3 and APOE ɛ4 alleles, respectively, under the mouse APOE promoter, and do not develop the plaques and tangles diagnostic of AD. To directly test whether different apoE isoforms affect the accumulation of intraneuronal Aβ1–42 at the very onset of AD pathogenesis, we generated a gene transfer animal model using lentiviral delivery of secreted Aβ1–42 into the brain, which results in the intracellular Aβ accumulation (19,20). In AD brains, Aβ deposits typically occur first in neocortical areas and appear to plateau relatively early in the disease process (21–23). Here, we found that APOE4 increased accumulation of intraneuronal Aβ1–42 in the cortex, compared with APOE2 and APOE3. Our results suggest that APOE ɛ4 allele can accelerate accumulation of intraneuronal Aβ early in the development of AD.
Microinjection and intracellular expression of lentiviral Aβ1–42 in the mouse cerebral cortex
We have previously generated a gene transfer animal model using lentivirus to express a secreted form of Aβ1–42 that resulted in intracellular, but not extracellular, Aβ accumulations in the targeted rat primary motor cortex (19). In control experiments to address the expression of Aβ1–42 in mouse brain, we injected Aβ1–42 lentivirus into the primary motor cortex of wild-type C57BL/6J mice. Figure 1A shows the schematic of the injection site of lentiviral Aβ1–42. Aβ1–42 distribution was assessed 2 weeks after gene delivery by immunohistological staining with MOAB2. MOAB2 (mouse IgG2b) is a pan-specific, high-titer antibody to Aβ residues 1–4, which did not detect amyloid precursor protein (APP) or amyloid precursor protein carboxy terminal fragments in cells overexpressing APP or in brain homogenates from transgenic mice overexpressing APP (24). By immunohistochemistry, MOAB2 co-localized with C-terminal antibodies specific for Aβ40 and Aβ42, but did not co-localize with either N- or C-terminal antibodies to APP (24). The immunofluorescence staining of MOAB2 was strong in the ipsilateral cortex near the injection area (Fig. 1D–F), but absent in the contralateral cerebral cortex (Fig. 1B and C) of brain sections from a wild-type C57BL/6J mouse. Furthermore, the expression of lentiviral Aβ1–42 was intracellular and revealed a punctate pattern in the higher magnifications (Fig. 1F).
Human APOE genotype affects the intracellular accumulation of lentiviral Aβ1–42
This gene transfer system was used to infect APOE-TR mice. Double immunohistochemistry for Aβ1–42 and a V5 tag indicative of lentiviral expression showed that the lentivirus was initially active in neurons of APOE2-, APOE3- APOE4-TR mice (data not shown). The expression of lentiviral Aβ1–42 in the cerebral cortex of APOE-TR mice was assessed with 3,3′-diaminobenzidine (DAB) immunohistochemical staining with MOAB2 2 weeks after lentivirus delivery. As with the immunofluorescence staining on brain sections from C57BL/6J mice, Aβ1–42 was also observed on brain sections from some APOE2- and APOE3-, and most APOE4-TR, mice. There was strong MOAB2-positive staining in the cortex near the injection site, but completely absent from the contralateral cortex. Figure 2A shows a representative image of the distribution of lentiviral Aβ1–42 in a coronal brain section from an APOE4-TR mouse.
Two sections within 0.5 mm of the injection site from each mouse were chosen for DAB staining to assess the levels of Aβ1–42. Two of 12 (17%) sections from APOE2-, 3 of 12 (25%) from APOE3- and 13 of 16 (81%) from APOE4-TR mice were positively stained for MOAB2 (Fig. 2B). The percentage of MOAB2 DAB-positive sections was significantly higher from APOE4-TR mice than that from APOE2- or APOE3-TR mice (P < 0.01, Fisher's exact test). There was no significant difference between the percentages of MOAB2-positive sections from APOE2- and APOE3-TR mice. All sections were MOAB2-negative in the contralateral cortex.
We further quantified the intensity of DAB staining of MOAB2 in the images at 10× magnification to analyze differences in Aβ1–42 between groups. The mean integrated density of MOAB2-positive staining in the ipsilateral cortex of APOE4-TR mice was significantly higher than that of the ipsilateral cortex of APOE2- or APOE3-TR mice [P < 0.01, two-way analyses of variance (ANOVA) followed by Tukey's post hoc analyses], as well as that of the corresponding cortex of the contralateral side of APOE4-TR mice (P < 0.01, two-way ANOVA followed by Tukey's post hoc analyses, Fig. 2C). Nissl-counterstaining after MOAB2 immunocytochemistry showed that Aβ1–42 staining (MOAB2, brown) showed an intracellular, punctate pattern. Aβ was observed both in neurons (arrows) and, to a lesser extent, in glia (arrowheads) (Fig. 2D). There was no observable MOAB2 immunopositive staining in the contralateral cortex (Fig. 2E).
Since accumulation of phopsho-tau also occurs in the brains of individuals with Alzheimer's disease, we performed immunohistochemistry with the phospho-tau (pSer199/202) antibody, AT8. AT8 immunopositive neural soma and neurites were present in the ipsilateral cortex near the injection site of Aβ1–42 lentivirus (Fig. 2F). Localized phosphorylated tau was also found in some areas without Nissl staining (see arrowheads), indicating the potential loss of a cell body in these areas (Fig. 2F). There was no observable AT8 immunopositive staining in the contralateral cortex and areas without Aβ immunostaining (Fig. 2G).
Lentiviral Aβ1–42 is present mainly in neurons, but also in microglia
To define the specific cell type that accumulated lentiviral Aβ intracellularly, we performed double-immunofluorescence histochemistry with MOAB2 and specific cell-type antibodies. Immunopositive staining for Aβ1–42 (MOAB2, red) was mainly present in cortical neurons (NeuN, green), and it displayed a punctate perinuclear localization (Fig. 3A). The accumulation of Aβ1–42 in neurons was observed in APOE2-, APOE3- and APOE4-TR mice (Fig. 3A). However, lentiviral Aβ1–42 immunoreactivity was also sometimes present in microglia, as evidenced by the co-localization of MOAB2 and Iba1 (Fig. 3B). In contrast, no co-localization was observed in astrocytes, despite their presence in areas with considerable Aβ immunostaining (Fig. 3C).
Aβ1–42 accumulates more in neurons and less in microglia in APOE4-TR mice
We tested whether APOE genotype affected the distribution of Aβ between neurons and microglia. As in Figure 3B, fluorescent immunolabeling revealed the Aβ1–42 marker MOAB2 (red) and the microglial marker Iba1 (green) co-localized to some extent in the ipsilateral cortex around the injection site of Aβ1–42 lentivirus in APOE2-, APOE3- and APOE4-TR mice. From these images, we counted the number of MOAB2-positive neurons, the number of Iba1-positive microglia and the number of MOAB2, Iba1-positive microglia. Consistent with Figure 2, the number of MOAB2-positive neurons in APOE4-TR mice (433 ± 36) was significantly more than in APOE2- (214 ± 69) or APOE3- (237 ± 68) TR mice (*P < 0.05, one-way ANOVA followed by Tukey's post hoc analyses) (Fig. 4A). There was no significant difference between the numbers of Iba1-labeled microglia of different groups (Fig. 4B). However, the number of MOAB2-positive microglia was significantly reduced in the analyzed cortex of APOE4-TR mice (Fig. 4C). The number of MOAB2-positive microglia was 16 ± 2 in APOE4-TR mice, significantly fewer than in APOE2-TR mice (25 ± 3, *P < 0.05, one-way ANOVA followed by Tukey's post hoc analyses) or APOE3-TR mice (29 ± 2, *P < 0.01, one-way ANOVA followed by Tukey's post hoc analyses). These results revealed increased neurons with Aβ1–42 and reduced microglia with Aβ1–42 in the cortex of APOE4-TR mice around the lentiviral injection site.
Intraneuronal Aβ1–42 is mainly localized to late endosomes/lysosomes
To investigate the intracellular localization of lentiviral Aβ1–42, we co-labeled APOE4 brain sections with MOAB2 (anti-Aβ1–42) (24) and ABL-93 (anti-LAMP-2, late endosomal/lysosomal marker) (25,26) or EEA1 (early endosomal marker) (27) antibodies. Intraneuronal deposits of lentiviral Aβ1–42 displayed a punctate pattern in the cytoplasm, which co-localized with the late endosomal/lysosomal marker, ABL-93 (Fig. 5A). The MOAB2 puncta did not obviously co-localize with the early endosomal marker EEA1 (Fig. 5B).
ApoE co-localizes with intraneuronal Aβ1–42
In the central nervous system (CNS), apoE is mainly secreted by glial cells, particularly astrocytes, although neuronal production of apoE has been observed under specific pathological conditions (28). In the present experiment, we observed apoE in glia-like cells in both the ipsilateral and contralateral brain cortex of Aβ1–42 lentivirus-injected APOE2-, APOE3- and APOE4-TR mice (Fig. 6A). In the ipsilateral cortex, much of the apoE was also present in cells of neuronal morphologies (Fig. 6A). ApoE-positive neurons were not observed in the contralateral cortex (Fig. 6A).
Double staining of APOE4 tissue with MOAB2 showed that, in the ipsilateral cortex near the injection area, apoE co-localized with lentiviral Aβ1–42 in neurons (Fig. 6B). Unlike the punctate staining pattern of Aβ1–42, intracellular staining of apoE was more homogeneous. The co-localization of apoE with Iba1 and GFAP demonstrated that apoE was present in both astrocytes and microglia in both the ipsilateral and contralateral cortex of mice regardless of APOE genotype (Fig. 6B).
Intraneuronal accumulation of Aβ42 is one of the earliest pathological events in humans and in animal models of AD (1,4). APOE4 is the major identified genetic risk factor for late-onset AD, with Aβ deposition beginning earlier in apoE4-positive subjects (10–12,29). In the current study, the effects of different APOE genotypes on intraneuronal accumulation of Aβ1–42 at the very early stages of AD pathogenesis were studied by directly introducing lentiviral Aβ1–42 into the cortical neurons of human APOE-TR mice at the age of 8–9 months. APOE-TR mice express either the human APOE2, APOE3 or APOE4 gene under the control of endogenous murine APOE regulatory sequences, resulting in physiological expression of human apoE (18,30). A significant isoform-dependent effect of human APOE on intracellular accumulation of Aβ1–42 was demonstrated, with dramatically enhanced Aβ1–42 in the cerebral cortex of APOE4-TR mice 2 weeks after direct delivery of lentiviral Aβ1–42. Aβ was mainly present in neurons, and to a lesser extent in microglia, but not in astrocytes. The presence of apoE4 favored the accumulation of Aβ in neurons over microglia compared with mice expressing apoE2 or apoE3. Intraneuronal Aβ1–42 was mainly localized to late endosomes/lysosomes and co-localized partially with apoE.
This approach to studying Aβ accumulation has several advantages. First, we can study processes that occur soon after the onset of Aβ generation. Secondly, we can examine the time course of Aβ accumulation from a defined point in time, i.e. when the lentivirus is injected. Thirdly, we can easily define the effects of specific genes without lengthy crosses that may occur with transgenic models. Here, we were able to quickly examine the effects of APOE genotype on early Aβ accumulation using commercially available mouse models. This system also focuses on Aβ produced by neurons, since our double immunohistochemistry for Aβ1–42 and a V5 tag showed the lentivirus was initially active in neurons.
With this system, we found that the presence of APOE4 allele significantly increased intraneuronal accumulation, and decreased microglial accumulation of Aβ1–42. These findings suggest that apoE2 and apoE3 may help decrease the intraneuronal accumulation of Aβ by promoting its uptake in microglia. The main Aβ clearance pathways include receptor-mediated uptake by neurons and glia, drainage into interstitial fluid or through the blood–brain barrier and proteolytic degradation by insulin-degrading enzyme and neprilysin (31). Previous studies have shown that apoE can regulate Aβ clearance through modulating the cellular uptake of an apoE–Aβ complex by receptor-mediated endocytosis (32–34), or removing Aβ from the brain by transport across the blood–brain barrier (17), or facilitating the cellular proteolytic degradation of Aβ within microglia and astrocytes (35,36). Whereas the low-density lipoprotein receptor mainly mediates uptake of APOE into microglia and astrocytes (37), low-density lipoprotein receptor-related protein 1 constitutes an uptake pathway for this apolipoprotein in neurons (17). A recent study showed that the pro-neurotrophin receptor sortilin is also a major neuronal receptor pathway for APOE-containing lipoproteins that counteracts Aβ accumulation in the brain (38). Sortilin-mediated endocytosis likely results in lysosomal degradation of Aβ in neurons (38). In cultured neuronal cells, apoE3 was more efficient than apoE4 at promoting extracellular Aβ internalization and delivery to lysosomes for degradation through Rab5- and Rab7-positive early and late endosomes (39). In neuronal–astrocyte co-cultures, apoE strongly increased the intraneuronal levels of both Aβ1–40 and Aβ1–42 by binding Aβ in the extracellular space and directing it to neurons, through highly efficient receptor-mediated endocytosis of apoE–Aβ complexes (40). Our findings suggest that APOE genotype affects the balance of neuronal and glial Aβ uptake mechanisms, and the intracellular degradation of Aβ once it has been internalized.
These results are in accordance with previous findings from transgenic mice overexpressing mutant human amyloid precursor protein V717F expressing human apoE isoforms (PDAPP mice expressing human apoE isoforms [PADPP/TRE mise]), which revealed that the brain levels of Aβ42 in young PDAPP/TRE4 mice were substantially elevated even at very early time points (41). In vivo microdialysis showed that APOE genotype did not influence Aβ production, but did affect interstitial Aβ levels, with apoE4 resulting in slower clearance from the hippocampus than apoE3 and apoE2 (42). These data support the hypothesis that it is Aβ clearance from the CNS, not production, that is impaired in individuals with late-onset AD (43–45). The presence of APOE4 allele increased intraneuronal accumulation and decreased microglial accumulation of Aβ1–42, suggesting apoE4 may inhibit Aβ proteolysis in neurons or inhibit Aβ internalization by microglia, compared with apoE3 and apoE2. This supposition is supported by the evidence that, in microglia, apoE3 promotes enzyme-mediated degradation of Aβ more efficiently than apoE4 (35). In addition, apoE promoted intraneuronal oligomerization of both Aβ1–40 and Aβ1–42 and impaired intraneuronal Aβ1–40 degradation (40), supporting intraneuronal effects of apoE4 leading to its accumulation there. Finally, apoE4-lipoproteins bind Aβ with lower affinity than do apoE3-lipoproteins (46), suggesting that apoE4 might be less efficient in mediating Aβ clearance specifically into microglia.
In a mouse model of inhibition of the Aβ-degrading enzyme neprilysin, intraneuronal oligomerization and accumulation of Aβ occurred more readily in mice that expressed the apoE4 isoform than the apoE3 isoform (47). Our in vivo data support the model of lysosomal degradation of intracellular Aβ1–42. We demonstrated that intraneuronal Aβ1–42 deposits partially co-localized to late endosomes/lysosomes. Aβ accumulation in lysosomes may cause loss of lysosomal membrane impermeability and leakage of lysosomal content (proteases and cathepsins) and contribute to neuronal toxicity (48–50). In vitro studies revealed that apoE4 potentiates lysosomal leakage induced by Aβ1–42 in cultured Neuro-2a cells (51,52). Inhibition of Aβ-degrading enzyme neprilysin in APOE3 and APOE4 mice led to an apoE4 isoform-specific accumulation of intraneuronal Aβ, accompanied by intracellular apoE and lysosomal activation (47). The release of lysosomal contents into the cytoplasmic compartments was considered one of the earliest events in intracellular Aβ-mediated neurotoxicity in vitro (48). It is possible that intraneuronal apoE4 and Aβ1–42 may work in concert to increase the susceptibility of lysosomal membranes to disruption, release of lysosomal enzymes into the cytosol and thus, trigger Aβ accumulation. Intraneuronal Aβ may also induce late endosomal multivesicular bodies and lysosomal damage, leading to leakage of Aβ from vesicles into the cytosol and activation of inflammatory mechanisms (53). This damage may be related to the induction in phospho-tau levels that we have observed in the presence of intraneuronal accumulation of Aβ1–42.
Microglia are mononuclear phagocytes of the innate immune system in the CNS and participate mostly in the clearance of Aβ via their ability to internalize and degrade Aβ (4). Soluble and aggregated Aβ uptake in microglia may be mediated by different mechanisms (4,54). Soluble Aβ internalization by microglia occurred through fluid phase macropinocytosis, with Aβ-containing macropinocytic vesicles fusing with late endosomes and lysosomes for degradation (55). Fibril/aggregated Aβ internalization by microglia proceeds by receptor-mediated endocytosis and receptor-mediated phagocytosis (56,57). In the present study, our observation that there is less intracellular Aβ in microglia in APOE4-TR mice may reflect that APOE4 inhibits Aβ internalization and degradation by microglia, thus favoring Aβ accumulation in neurons.
We also found that lentiviral Aβ1–42 caused accumulation of apoE in neurons. We found co-localization of apoE and Aβ1–42 in neurons, consistent with the findings that brain regions where neuronal apoE was present correlate with more severe Aβ deposits and are most vulnerable in developing neurofibrillary tangle pathology in AD (58,59). Here, accumulation of Aβ1–42 correlated with that of phospho-tau, confirming our previous finding that this intraneuronal Aβ is neurotoxic (19). Co-localization of apoE and Aβ was also found in a small population of apoE-positive synaptic terminals in both aged cognitive normal and AD mice (60). Intraneuronal apoE complexes may contribute to synaptic or neuronal toxicity and subsequent AD pathological changes.
In summary, we demonstrated that apoE4 facilitates intraneuronal Aβ1–42 deposits in cortical neuron at the early stage of disease pathogenesis, by directly introducing lentiviral Aβ1–42 into the cerebral cortex of APOE-TR mice. This lentiviral system for Aβ expression allows the definition of the time-dependent effects of apoE isoforms on Aβ accumulation. These results show that apoE4 facilitates intraneuronal Aβ deposits, thus influencing the risk of AD.
MATERIALS AND METHODS
Human APOE2-, APOE3- and APOE4-TR mice (on a C57BL/6J background), which express each of the human APOE alleles to replace the mouse APOE gene regulated by the endogenous murine apoE promoter (18), were bred under standard conditions with access to food and water ad libitum. Experiments were performed on age-matched female mice (8–9 months of age). All animal experiments were approved and conducted according to Georgetown University Animal Care and Use Committee. Every effort was made to reduce animal stress and to minimize animal usage.
We generated a gene transfer animal model using lentiviral delivery Aβ1–42 expression to specific brain regions that results in intracellular protein accumulation (19). The lentivirus expresses Aβ1–42 with a signal peptide cloned into a pLenti6-D-TOPO plasmid under the control of the cytomegalovirus promoter (20). Lentivirus was generated in HEK293FT cells, and the supernatants were collected, concentrated by centrifugation, resuspended in HBSS and aliquoted in autoclaved tubes. To assure that the same number and quality of viral particles were injected, we used the same preparation of virus and injected it into the APOE2, APOE3 and APOE4 mice over a period of 2 days.
Animals were anesthetized with intraperitoneal injection of a cocktail of ketamine and xylazine (100 and 10 mg/kg, respectively). Stereotaxic surgery was then performed to inject the lentivirus encoding Aβ1–42 into the left side of the primary motor cortex of 8–9-month-old female APOE2-, APOE3- and APOE4-TR mice (body weight 22–30 g), as described previously (61). The stereotaxic coordinates for the primary motor cortex of mice were 1.6 mm lateral (left), 1.6 mm ventral and 0.5 mm anterior. Viral stocks were injected through a microsyringe pump controller (Micro4) using total pump (World Precision Instruments, Inc.) delivery of 6 μl at a rate of 0.2 μl/min. The needle remained in place at the injection site for an additional minute before slow removal over 2 min.
Two weeks post-injection, animals were perfused transcardially with ice-cold phosphate-buffered saline (PBS; pH 7.4). Brains were removed and post-fixed in 4% paraformaldehyde at 4°C for 24 h, incubated in two sequential 30% sucrose solutions (in PBS) at 4°C for 24 h each, frozen on dry ice and coronal sections were cut at 40 μm thickness on a sliding microtome (Thermo Scientific Microm HM 430). Sections were stored in cryoprotectant (30% glycerin, 30% ethylene glycol, in 0.1 M PBS) at −20°C. A total of n = 20 mice (n = 6 for APOE2-TR mice, n = 6 for APOE3-TR mice and n = 8 for APOE4-TR mice) were used for immunohistochemistry.
Source of antibodies
The following primary antibodies were utilized: MOAB2 and biotinylated MOAB2 (anti-Aβ, mouse IgG2b, 0.5 mg/ml), goat anti-apoE antibody (Calbiochem), rabbit anti-NeuN antibody (Bioss, Inc.), rabbit anti-GFAP antibody (Invitrogen, CA, USA), rabbit anti-Iba1 antibody (Wako Pure Chemical Industries), rabbit anti-tau [pSpS199/202] phospho-specific antibody (AT8, Invitrogen), rat anti-LAMP-2 (ABL-93 concentrate, Iowa Hybridoma Bank) and rabbit anti-EEA1 (Cell Signaling Technology). The following secondary antibodies were used: for phosphorylated tau protein DAB staining we used biotinylated goat anti-rabbit (Vectastain, Vector Laboratories). Fluorescent secondary antibodies: AlexaFluor 594 donkey anti-mouse, AlexaFluor 488 donkey anti-mouse, AlexaFluor 594 donkey anti-rabbit, AlexaFluor 488 donkey anti-rabbit, AlexaFluor 488 donkey anti-goat, AlexaFluor 488 donkey anti-rat (all from Invitrogen Co.). The antibodies were diluted in Tris-buffered saline (TBS) containing 0.25%Triton X-100 + 2% bovine serum albumin + 0.005% sodium azide (NaN3).
Immediately prior to staining, brain sections from APOE2-, APOE3- and APOE4-TR mice were rinsed in 0.1 M PBS (3 × 10 min), washed in TBS (3 × 10 min), incubated in quench peroxidase (10% methanol, 3% hydrogen peroxide in 1× TBS) for 20 min, permeabilized with TBS containing 0.25%Triton X-100 (TBSX; 3 × 10 min) and blocked with 10% horse serum in TBSX for 1 h. Free-floating sections were subsequently incubated overnight at 4°C with a biotinylated anti-Aβ antibody, MOAΒ-2 (mouse, 1:1000 dilution of 0.5 mg/ml stock) or anti-tau [pSpS199/202] phospho-specific antibody (rabbit, 1:10 000 dilution), washed in TBS (3 × 10 min), incubated with biotinylated goat anti-rabbit secondary antibody for phosphorylated tau protein DAB staining (1:200) for 1 h, washed in TBS (3 × 10 min) and then incubated with avidin–biotin complex (Vector Laboratories) for 1 h. Sections were washed in TBS (2 × 15 min), rinsed in 0.1 M Tris–HCl (pH 7.5) for 3 min, and reaction products were visualized using 0.1 M Tris–HCl (pH 7.5) containing 0.05% 3/3′-diaminobenzidine tetrahydrochloride and 0.003% hydrogen peroxide. Sections were then washed in 0.1 M Tris–HCl (pH 7.5) buffer (3 × 5 min), transferred to TBS, mounted onto glass slides, air dried overnight, dehydrated through a series of graded alcohols, cleared in xylene and cover-slipped with permount.
Quantification of DAB immunostaining for Aβ1–42
Bright-field images at 10× magnification taken under light microscope of each tissue section were used for Digital Image Analysis. The areas examined include almost all MOAB2-positive staining at the lenti-Aβ1–42 microinjected cortex (two 10× microscopic fields) and the corresponding areas of the control cortex. Image J software was used to quantify the intensity of DAB staining (brown) of biotinylated antibody MOAB2 under a light microscope (Carl Zeiss, Germany) to evaluate the expression of Aβ1–42. Integrated density of each image was automatically quantified by the Image J software.
The coverslips were removed from some of DAB stained slides with xylene after the quantification of DAB immunostaining, rehydrated in descending series of ethanol (100, 100, 95 and 70%) and in distilled H2O. The sections were then treated with 0.1% cresyl violet solution for 5 min, dehydrated with ascending series of ethanol (70, 95, 100 and 100%), treated with xylene and cover-slipped with permount. Bright-field images were taken on a Zeiss Axiophot microscope (Carl Zeiss).
For immunofluorescent analyses, tissue sections were incubated with the following primary antibodies overnight: MOAB2 (1:1000), anti-apoE, (1:100 000), anti-GFAP (1:1000), anti-IbaI (1:100), anti-ABL93 (1:1000) and anti-EEA1 (1:1000). Then, sections were washed in TBSX (6 × 10 min), followed by Alexa 488 or Alexa 594 (Invitrogen) fluorophore-conjugated secondary antibodies at dilution of 1:1000. Images were captured on a Zeiss LSM 510 confocal microscope at 20×, 40× and 100× magnification.
Quantitative analysis of immunofluorescence-stained brain sections
The coronal brain sections containing the injection site of Aβ1–42 lentivirus were double immunofluorescence stained with MOAB2 (antibody for Aβ1–42) and Iba1 (antibody for microglia). The region for quantitative analysis was in the ipsilateral cortex nearest to the injection site of Aβ1–42 lentivirus, which contain most MOAB2-positive neurons. Images were captured at 40× magnification using a Zeiss LSM 510 confocal microscope across an area of 675 μm × 675 μm. The numbers of MOAB2-positive, Iba1-positive and MOAB2-/Iba1-double-positive cells were manually counted by a blinded investigator. The cell numbers were determined in three sections per animal, and the average of the counts thus obtained was recorded from APOE2 (n = 6), APOE3 (n = 6) and APOE4 (n = 8) animals.
Fisher's exact test was performed to test the frequency difference of MOAB2 immunostaining (positive versus negative) between groups of APOE2-, APOE3- and APOE4-TR mice. Two-way ANOVA and one-way ANOVA were used to examine differences in mean integrated densities of MOAB2 immunostaining and numbers of MOAB2-, Iba1- and MOAB2-/Iba1-double immunopositive cells among groups, followed by Tukey's post hoc analyses. Data were expressed as mean ± SEM unless otherwise specified. The statistical significance criterion of P-value was 0.05.
This work was supported by the National Institutes of Health (R01 AG035379 to G.W.R. and P01 AG030128 to G.W.R. and M.J.L.). W.Z. was supported by the China Scholarship Council-Georgetown University Fellowship Program.
The authors thank Dr Maia Parsadanian for her technical assistance of apoE and phospho-tau immunostaining. The ABL-93 antibody against lamp-2 developed by Dr August J. T., was obtained from the Developmental Studies Hybridoma Bank developed under auspices of the NICHD, National Institutes of Health, and maintained by the Department of Biology, University of Iowa, Iowa City, IA 52242.
Conflict of Interest statement. None declared.