Address correspondence and reprint requests to Greg. M. Cole, Greater Los Angeles Veterans Affairs Healthcare System, Alzheimer Research-151, Building 7, Room A101, 16111 Plummer Street North Hills, CA 91343, USA. E-mail: email@example.com
Extracellular-signal regulated kinase (ERK) signaling is critical for memory and tightly regulated by acute environmental stimuli. In Alzheimer disease transgenic models, active ERK is shown to first be increased, then later reduced, but whether these baseline changes reflect disruptions in ERK signaling is less clear. We investigated the influence of the familial Alzheimer’s disease transgene APPsw and β-amyloid peptide (Aβ) immunoneutralization on cannulation injury-associated (i.c.v. infusion) ERK activation. At both 12 and 22 months of age, the trauma-associated activation of ERK observed in Tg− mice was dramatically attenuated in Tg+. In cortices of 22-month-old non-infused mice, a reduction in ERK activation was observed in Tg+, relative to Tg− mice. Intracerebroventricular (i.c.v.) anti-Aβ infusion significantly increased phosphorylated ERK, its substrate cAMP-response element-binding protein (CREB) and a downstream target, the NMDA receptor subunit. We also demonstrated that Aβ oligomer decreased active ERK and subsequently active CREB in human neuroblastoma cells, which could be prevented by oligomer immunoneutralization. Aβ oligomers also inhibited active ERK and CREB in primary neurons, in addition to reducing the downstream post-synaptic protein NMDA receptor subunit. These effects were reversed by anti-oligomer. Our data strongly support the existence of an APPsw transgene-dependent and Aβ oligomer-mediated defect in regulation of ERK activation.
Alzheimer’s disease (AD) is clinically characterized by progressive cognitive impairment. Cognitive deficits in AD are widely believed to result from progressive synaptic dysfunction and neurodegeneration initiated by aggregated β-amyloid peptide 1–42 (Aβ42). Recently, soluble Aβ oligomers, including dimers, trimers, and dodecamers (12-mer or Aβ*56) have been proposed as key forms of Aβ causing cognitive deficits in animal models (Cleary et al. 2005; Lesne et al. 2006). Soluble Aβ concentrations predict synaptic changes (Lue et al. 1999) and correlate with the severity of dementia in AD patients (McLean et al. 1999). Antibody-mediated Aβ oligomer reduction can correct defects in synaptic plasticity (Klyubin et al. 2005) and reduce tau phosphorylation (Klyubin et al. 2005; Ma et al. 2006), another pathological hallmark of AD.
Although multiple mechanisms for Aβ peptide-induced memory deficits have been hypothesized including effects on tau phosphorylation and tangle formation, there is considerable evidence implicating a role for Aβ-induced disruption of kinases critical for memory (Zhu et al. 2002). For example, studies of human mental retardation syndromes have shown that extracellular-signal regulated kinases (ERKs) are critical for human learning (Costa et al. 2002). They are also known to contribute to molecular information processing in dendrites, to stabilize structural changes in dendritic spines and to interact with scaffolding and structural proteins at the synapse (Sweatt 2004). However, aberrant over-expression of ERK can lead to cell death (Zhuang and Schnellmann 2006). ERK shows stage-dependent abnormalities in mRNA and protein expression in AD (Webster et al. 2006) and AD models (Dineley et al. 2001). Although transient ERK activation plays important roles in memory-related processes, persistent activation can mediate NMDA-related excitotoxicity (Amadoro et al. 2006). Therefore, either hyper- or hypoactivation of ERK could contribute to disease pathways.
It has been reported that in early AD, there is extensive activation of ERK in astroglial cells in the white matter, while in advanced AD, there is reduced activation showing a strong inverse correlation with Braak stage and the Blessed score for cognition (Webster et al. 2006). However, ERK activation is not necessarily protective. For example, active ERK is a tau kinase that is elevated during the initial stages of neurofibrillary degeneration in the projecting neurons in the transentorhinal region. Nevertheless, ERKs causal role in neurofibrillary tangle formation remains unclear as many neurons with the highest amounts of ERK/mitogen-activated protein kinase immunoreactivity do not appear vulnerable to neurofibrillary tangles (Hyman et al. 1994). Interestingly, at late stages (although some ERK is still apparent in astrocytes), ERK activation is suppressed relative to early stage and normal controls in neuronal cell bodies and dystrophic neurites (Webster et al. 2006). In summary, studies in AD brain suggest stage-dependent ERK activation followed by loss of active ERK.
Similar to AD, the amyloid precursor protein (APP) transgenic Tg2576 model shows early stage ERK activation, while at later stages ERK is reduced (Dineley et al. 2001). Sustained activation of ERK was also observed in brains of an APP transgenic rat model with significant intraneuronal Aβ accumulation (Echeverria et al. 2004). In vitro, it has been demonstrated that sustained ERK activation is related to cell death after Aβ treatment in neurons and PC12 cells (Dineley et al. 2001; Jang and Surh 2005). Conversely, ERK activation can be reduced by γ-secretase inhibitor treatment of APPsw transfected cells (Echeverria et al. 2005). Abnormally sustained activation may contribute to cell death in other amyloid diseases like familial amyloidotic polyneuropathy (Monteiro et al. 2006) and work via multiple mechanisms, including pro-apoptotic events upstream of caspase 3 or via suppression of the anti-apoptotic signaling molecule Akt (Zhuang and Schnellmann 2006).
The known importance of ERK on dendrites and memory is consistent with the hypothesis that ERK hypoactivation in AD contributes to cognitive decline. Under certain conditions, Aβ or its fragments have been shown to inhibit ERK (or downstream cAMP-response element-binding protein; CREB) in neuroblastoma cells (Daniels et al. 2001), endothelial cells (Magrane et al. 2006), hippocampal neuron preparations (Xie 2004), or after Aβ25–35 i.c.v. injection into rats (Jin et al. 2005). Different Aβ aggregates may have different effects depending on the system, for example high levels of fibrillar Aβ in PC12 cells inhibited ERK, while other forms of Aβ increased ERK activation (Echeverria et al. 2005). In contrast, fibrillar Aβ stimulated ERK, while low (100 nmol/L) soluble oligomers initially stimulated but later down-regulated ERK in hippocampal slice cultures (Bell et al. 2004).
In view of the conflicting results with Aβ’s stimulatory or inhibitory effects on ERK in culture systems with different Aβ preparations and biphasic effects, we have focused on chronic in vivo studies. We decided to assess ERK regulation using injury which is known to stimulate ERK. Trauma or ischemia can induce ERK activation (Shackelford and Yeh 2006), possibly via growth factors that are induced by injury including transforming growth factor beta1 (Logan et al. 1992), insulin-like growth factor (IGF), and brain derived neurotrophic factor (Kazanis et al. 2004) all which are known to stimulate ERK (Willaime-Morawek et al. 2005; Johnson-Farley et al. 2006). To determine if AD models showed disruption in regulation of ERK, we examined whether trauma induction of ERK was altered by APPsw transgene and corrected by Aβ oligomer neutralization. We also investigated the impact of Aβ oligomer on downstream CREB and NMDA receptor (NMDAR) subunit (NR2B).
Chemicals and reagents
Unless otherwise noted, all reagents were obtained from Sigma (St Louis, MO, USA). Monoclonal anti-Aβ antibody (IgG2b) as anti-Aβ1–15 was raised against Aβ1–42 and characterized in our lab (Yang et al. 1994). Primary antibodies were: polyclonal phospho-p44/42 mitogen-activated protein kinase (Thr202/Tyr204) (phosphorylated ERK, pERK; BioSource, Camarillo, CA, USA); polyclonal anti-total ERK2 (Santa Cruz Biotechnology Inc, Beverly, MA, USA); polyclonal anti-phospho-CREB and monoclonal anti-post-synaptic density-95 (PSD-95) (phosphorylated CREB, pCREB; Upstate, Lake Placid, NY, USA); monoclonal anti-β-actin, anti-synaptophysin, anti-NR2B, and polyclonal anti-MAP2 (Chemicon International, Temecula, CA, USA); anti-Aβ monoclonal 6E10 (Signet Labs Inc., Dedham, MA, USA); monoclonal anti-Drebrin (MBL, Woburn, MA, USA); Rabbit polyclonal A11 oligomer-specific antibody, a generous gift from Dr C. Glabe, was produced and characterized as described (Kayed et al. 2003). A11 antibody recognizes oligomer species but not monomer and fibrils. IGF-1 was obtained from Peprotech Inc (Rochy Hill, NJ, USA). ABC reagent was from Vector Labs (Burlingame, CA, USA), metal enhanced diaminobenzidene (DAB) was from Pierce Biotechnology (Rockford, IL, USA), Aβ1–42 peptide was obtained from American Peptide (Sunnyvale, CA, USA). The CytoTox 96 Non-Radioactive Cytotoxicity Assay Kit was obtained from Promega (Madison, WI, USA).
Animals and treatment
Surgical and animal procedures were carried out with strict adherence to the current guidelines set out in the NIH Guide for the Care and Use of Laboratory Animals by the Association for Assessment and Accreditation of Laboratory Animal Care International which has accredited the Greater Los Angeles Healthcare System. All experiments involving animals were approved by the appropriate UCLA and VA Institutional Animal Care and Use Committee, Institutional Biosafety, and Research Development (R&D) Committees. Tg2576 (Tg+) or transgene negative littermates (Tg−) mice were used for surgical experiments. Intracerebroventricular infusion of monoclonal anti-Aβ antibody (IgG2b, 10G4) as anti-Aβ1–15 and control IgG2b were previously described (Ma et al. 2006). Briefly, custom length, stainless steel i.c.v. catheters (Plastics One, Roanoke, VA, USA) were stereotaxically implanted into the lateral ventricle, using the following coordinates from Bregma: 2.7 mm ventrally (from skull), +1.0 mm laterally and 0.5 mm posteriorly. The catheters were secured to skull using acrylic dental cement (Stoelting Corporation, Wood Dale, IL, USA) and a machine screw partially inserted into the skull. The i.c.v. cannula was connected to an Alzet osmotic mini-pump (#1002; Durect Corporation, Cupertino, CA, USA, releasing 0.25 μL/h) through polyethylene tubing into ventricular cannula. Control and anti-Aβ IgG2b (10G4) were purified using a protein A column followed by an endotoxin-removing column (Pierce Biotechnology) and dialyzed overnight against phosphate-buffered saline (PBS). After dialysis, the antibody concentration used to fill the osmotic mini-pumps was 20 μg/100 mL. The control group received equimolar infusions of non-specific mouse IgG2b. After 14 days, mice were deeply anesthetized with pentobarbital (50 mg/kg bw) and cardiac perfused with protease inhibitor-containing buffer (pH 7.2) as previously described (Lim et al. 2000). All infusions were intracerebroventricular and unilateral with biochemical analysis of unfixed tissue on the contralateral side (western and ELISA) and immunocytochemistry on the fixed (cannulated) side where the needle track was examined to confirm placement in the right lateral ventricle.
Tissue samples were processed in Tris-buffered saline (TBS) (soluble fraction) and lysis buffer (membrane fraction) containing protease inhibitors and phosphatase inhibitors as previously described (Calon et al. 2004). Mixed cortices were used for pERK, pCREB, NR2B, and other endpoints analysis.
Cell culture and treatment
Human SH-SY5Y neuroblastoma cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 2 mmol/L l-glutamine and 10% (v/v) fetal calf serum; 5 × 105 cells were plated on six-well plates and grown to 80% confluency at 37°C in a humidified 5% CO2 atmosphere incubator. Cultured hippocampal neurons were prepared from embryonic 18 day Sprague–Dawley rat fetuses as described previously (Zhao et al. 2004). Before treatment with A11 oligomer-specific antibody (2 ng/mL), cells were rinsed once with serum-free Dulbecco’s modified Eagle’s medium media for SH-SY5Y cells or Neurobasal medium without glutamate and B27 for primary neurons. A11 antibodies and serum-free medium for SH-SY5Y cells or A11 antibodies and Neurobasal medium without glutamate and B27 for primary neurons were added to cells simultaneously for one hour, followed by addition of 250 nmol/L Aβ42 oligomers and incubation for 1.5–24 h at 37°C.
Preparation of Aβ oligomers
A modified protocol was used for the preparation of Aβ oligomers (Kayed et al. 2003). Briefly, 1 mg of Aβ42 peptide was dissolved in 250 μL of 1,1,1,3,3,3-hexafluoro-2-propanol in the vial and incubated at 25°C until dissolved completely. The 1,1,1,3,3,3-hexafluoro-2-propanol was removed by gentle streaming of N2. Aβ42 was then dissolved in 1 mL of 10 mmol/L HEPES (pH 7.4). This solution was incubated at 37°C with a micro stir bar at 500 rpm for 12 h and centrifuged at 14 000 g for 5 min to remove fibrils and other large aggregates. Protein concentration was determined using the Bio-Rad DC protein assay (Richmond, CA, USA). The presence of Aβ oligomers was confirmed by anti-Aβ antibodies, 6E10 and 4G8 following its separation on 10–20% Tris–Tricine gradient gels (Ma et al. 2006).
Cell lysate preparation from SH-SY5Y cells
SH-SY5Y cells were placed on ice, washed with cold PBS once, scraped with 1 mL of PBS, and transferred to a microfuge tube. Cells were centrifuged at 734.5 g for 5 min and pellets were dissolved in lysis buffer with a cocktail of protease and phosphatase inhibitors. After a brief sonication, the lysate was incubated at 4°C for 30 min, centrifuged at 16 000 g for 10 min, and the supernatants were collected.
Cell viability assays
The effect of Aβ42 oligomers on SH-SY5Y cell viability was determined by measuring the activity of lactate dehydrogenase released into the culture medium using a CytoTox 96® Non-Radioactive Cytotoxicity Assay. The assay was performed in accordance with the manufacturer’s protocol and absorption read at 490 nm.
For western immunoblotting, protein concentration was determined using Bio-Rad DC protein assay methods. Equal amounts of protein per sample were added to Laemmli loading buffer, and boiled for 3 min. Except for Aβ oligomer detection, 30 μg of protein per well was electrophoresed on 10% Tris–glycine gels and transferred to Immobilon-P PVDF membranes (Millipore, Bedford, MA, USA). For Aβ oligomer detection, 5 μg of Aβ preparation without boiling was electrophoresed on 10–20% Tris–Tricine gradient gels. Gels were stained with Coomassie blue to ensure equal protein loading. Membranes were blocked for 1 h at 25°C in 10% non-fat dried milk in PBS, followed by incubation overnight at 4°C with appropriate primary antibodies in PBS containing 0.05% Tween-20 (PBS-T) and 1.5% (w/v) albumin. pERK, actin, total ERK, and 6E10 were diluted to 1 : 2000; pCREB and NR2B were diluted to 1 : 1000. After being rinsed in PBS-T, the membranes were incubated with horseradish peroxidase-conjugated anti-mouse IgG (1 : 10 000) or anti-rabbit IgG (1 : 30 000) in PBS-T with 1.5% albumin for 1 h. Immunolabeled proteins were visualized by enhanced chemiluminescence detection reagents. Resulting films were scanned and quantified using densitometric software (Molecular Analyst II; Bio-Rad). To strip immunoblots, membranes were incubated for 30 min at 57°C in 100 mmol/L 2-mercaptoethanol, 2% sodium dodecyl sulfate, and 62.5 mmol/L Tris–HCl (pH 6.8). The membranes were rinsed 3 × 10 min in large volumes of PBS-T, followed by blocking and reprobing with appropriate antibodies, including anti-glial fibrillary acidic protein (GFAP) (1 : 1000), drebrin (1 : 500), synaptophysin (1 : 1000), synaptosomal-associated protein of 25 kDa (SNAP-25) (1 : 1000), and PSD-95 (1 : 1000) as previously described (Calon et al. 2004).
Interleukin-1β and tumor necrosis factor-α were analyzed by sandwich ELISA of TBS fractions as described (Lim et al. 2000).
Immunostaining and image analysis
Cryosectioned mouse brains were warmed to 25°C for 10 min then steamed for 1 h in citrate buffer. Sections were quenched in 0.6% hydrogen peroxide with methanol for 30 min at 25°C and washed three times with TBS-buffered solution (pH 7.4). Next sections were treated with 0.3% Triton X-100 in 0.1 mol/L TBS (pH 7.4). Sections were then blocked using 5% normal goat serum containing 3% bovine serum albumin/TBS for 1 h at 37°C followed by a 2 day incubation with pERK antibody at a dilution of 1 : 100 at 4°C. Vector biotinylated goat anti-rabbit antibody (1 : 1200) was used with 1.5% normal goat serum and 3% bovine serum albumin/TBS. Sections were incubated in secondary antibody followed by the ABC reagent for 1 h at 37°C. Slides were developed using peroxidase/DAB. Microglia was labeled with IB4 (biotinylated Griffonia Simplicifolia isolectin B4; Vector Labs) 1 : 50 overnight at 4°C and developed with avidin-peroxidase (ABC; Vector Labs) reagents and DAB. For immunofluorescent staining, pERK and pCREB were diluted to 1 : 100; NR2B was diluted to 1 : 50; microtubule-associated protein type II (MAP2) was diluted to 1 : 400.
For image analysis of IB4 or pERK, mid-region (−1.58 to −2.30 mm Bregma) sections were analyzed from a Nikon Eclipse E800M microscope (Valencia, CA, USA) with a Nikon microscope power unit video system (Irvine, CA, USA). The video signal was routed into a Macintosh via a Scion Corporation AG-5 averaging frame grabber, and these digitized images were analyzed with NIH-Image public domain software (Bethesda, MD, USA). Quantitative analysis was used to evaluate cell density. Ring analysis and regional analysis of IB4 labeled microglia with NIH image was performed as previously described (Frautschy et al. 2001).
Statistical analyses were performed with StatView 5.0 software (SAS Campus Drive, Cary, NC, USA). Differences among means were assessed by anova followed by Tukey–Kramer post hoc test.
APP transgene attenuates injury-induced ERK activation in vivo and IGF-induced ERK activation in SY5Y cells
To evaluate the possibility of an APP transgene-dependent dysregulation of ERK, we investigated whether APP transgene and Aβ oligomer influenced the pERK response to injury (associated with stereotaxic i.c.v. infusion surgery) at different ages. Results indicated that at 11–12 months of age, i.c.v. infusion significantly increased pERK in Tg− mice but not in Tg+ mice (Fig. 1a) when compared with groups of non-infused mice. We also analyzed pERK in animals at 16–18 months of age (Fig. 1b and c). In the cortex, i.c.v. infusion induced a threefold increase in pERK in Tg− mice but only a onefold increase in Tg+ mice (Fig. 1b). Intracerebroventricular infusion caused a four- to fivefold increase in pERK in hippocampal pyramidal neurons in Tg− mice (Fig. 1c). In contrast, in Tg+ mice, infusion did not increase pERK in the CA1 region of the hippocampus and the response was attenuated in the CA2 and CA3 regions (Fig. 1c). Thus, the APP transgene significantly attenuated the trauma-associated induction of pERK at both ages, consistent with the impairment in ERK regulation with chronic Aβ exposure. To address whether Aβ oligomers could also attenuate trophic factor induced ERK, we examined the ERK response to 100 ng/mL IGF-1, known to be elevated after injury in SY5Y cells. This dose of IGF-1 induced a 3.1-fold increase in ERK at 30 min, while pre-incubation with Aβ oligomers for 1.5 h resulted in an attenuated response (p <0.0003, a 2.4-fold increase in ERK, Fig. 2).
Anti-Aβ antibody blocked Aβ oligomer-induced decreases of activated ERK and CREB in Tg2576 mice
Our previous study found that acute infusion of anti-Aβ1–15 10G4 antibody into Tg2576 mice significantly reduced Aβ oligomers, including 12-mer, but did not reduce insoluble Aβ. Oligomer reductions significantly correlated with reduced tau phosphorylation by glycogen synthase kinase-3β (GSK3β), a major tau kinase (Ma et al. 2006). Aβ oligomer species were characterized in vivo and in vitro by silver staining and western blot analysis (Ma et al. 2006). In this study, we initially evaluated whether a loss of active ERK and CREB occurred in Tg+ mice; 10-month (n = 3 per group), 11- to 12-month (n = 4 for Tg+ and n = 8 for Tg−), 15-month (n = 3 per group), and 22-month-old Tg− (n = 15) or Tg+ (n = 16) mouse cortices were examined with pERK antibody by western blot. The 11- to 12- and 15-month-old Tg+ mice which were raised on PMI 5015 breeder chow showed a trend toward a decrease (p =0.10 and p =0.06) in pERK (Fig. 3b and c), whereas 22-month-old Tg+ mice showed a very large and significant decrease (p <0.01, Fig. 3d) in pERK compared with the age-matched Tg− mice. This suggested that while only a trend toward a decrease of cortical ERK activity could be seen at an early age (11–12 months) when plaques are emerging in Tg+ mice, a major ERK defect occurred as a function of APP transgene with age and a less protective diet. Similarly, a large and significant pCREB decrease (p <0.05, Fig. 3e) occurred in 22-month-old Tg+ mice compared with the age-matched Tg− mice on the same diet. pCREB was highly correlated with pERK (p <0.0001, R2 = 0.571, Fig. 3f).
As the APP transgene could suppress ERK signaling, we asked whether this could be Aβ-dependent by evaluating the levels of active ERK and CREB after infusion with anti-Aβ or control IgG2b antibody using 11- to 12-month-old Tg2576+ mice. The results indicated that, compared with non-specific IgG2b infused controls, IgG2b anti-Aβ-infused mice showed significantly increased levels of pERK2 (the 42 kDa band, p <0.05, Fig. 3g), and pCREB (p <0.05, Fig. 3h). There was a larger impact on pERK and pCREB in the temporal and entorhinal cortices (Figs 4 and 5) from anti-Aβ antibody treatment than frontal and parietal cortices, consistent with greater antibody perfusion of regions close to, and posterior to, the infused lateral ventricle. In addition to increase in intensity of pERK-ir cells after anti-Aß infusion, there was also an increase number of pERK-ir cells stained in the entorhinal cortex (not shown).Consistent with an anti-oligomer effect, frontal cortex showed no pERK change and also had no reduction in Aβ oligomer levels (not shown). These results suggest that immunoneutralizing Aβ oligomers can restore the diminished ERK response observed in the APP transgene positive mice in vivo.
An alternative interpretation might be that anti-Aβ antibody-induced microglial activation and inflammatory changes that caused increased ERK activation. However, anti-Aβ antibody infusion did not induce significant overall increases in microgliosis (IB4, image analysis of immunocytochemistry, Table 1) despite increases within plaques (Table 2). Regionally, anti-Aβ significantly reduced cortical and hippocampal microglial density except in the dentate gyrus. In contrast, there was a significant increase in the intensity of IB4 staining per cell, except in the dentate gyrus. Relative to control IgG2b, anti-Aβ (10G4) infusion significantly increased microglial density only within plaques but not outside plaques; similarly, there was a strong trend to increase microglial size within, but not outside plaques. While anti-Aβ infusion increased plaque-associated microglia, this effect was limited by the small number of plaques in the 11- to 12-month-old Tg2576. Microglial activation is transient and returns to baseline by 7 days in response to intracranial anti-Aβ antibody (Wilcock et al. 2004). Thus, global markers of inflammation including interleukin-1β and tumor necrosis factor-α (ELISA) and reactive astrocytes (GFAP and western) were not increased by anti-Aβ antibody, nor did anti-Aβ antibody cause damage indexed by westerns of synaptic markers (PSD-95, drebrin, synaptophysin, and SNAP-25) which were not significantly changed (Table 3).
Table 1. Overall impact of 10G4 and IgG2b on isolectin B4 (Iba4)
St Lac, Orien, Luc
#Treatment p <0.001, region p < 0.0002, Interaction p <0.001; ##Treatment effect p <0.01, region p <0.0001, interaction p <0.0015; ###Treatment NS, region p <0.001, Interaction (p <0.05); ####Treatment p <0.0005, region NS, interaction (p <0.05); ap <0.05 Treatment effect within region.
#Treatment effect NS, Ring effect p <0.0001, Interaction p <0.05; ##Treatment effect p <0.05, Ring effect p <0.0001, Interaction NS; ###Treatment effect NS, Ring effect p <0.00001, Interaction NS; ap <0.05 between treatments within ring.
Cell number# (per 1000 mm2)
5.748 ± 0.655
7.075 ± 0.991
7.506 ± 0.825
5.208 ± 0.679a
3.835 ± 0.546
2.792 ± 0.362
31.178 ± 2.829
25.158 ± 4.188
12.406 ± 1.325
7.074 ± 1.221a
3.963 ± 0.660
3.113 ± 0.555
Average cell size### (mm2)
80.275 ± 12.120
112.216 ± 41.65
12.588 ± 1.854
12.632 ± 2.106
8.373 ± 0.760
12.463 ± 2.025
Table 3. Effect of Aβ antibody infusion on IL-1β, TNFα, GFAP, and synaptic markers (mean ± SE)
Oligomer-specific antibody blocked Aβ oligomer-induced decreases of activated ERK and CREB in a human neuronal cell line
β-Amyloid peptide induces glial-derived cytokines and oxygen radicals that can regulate ERK signaling. To determine whether Aβ oligomers can directly modulate human neuronal ERK signaling, pERK and pCREB were evaluated in Aβ oligomer-treated human neuronal SH-SY5Y cells. Cells were pre-incubated with or without A11 antibody (2 ng/mL) for 1 h followed by treatment with 250 nmol/L of Aβ oligomer preparation for 1.5–24 h. (Fig. 6). Cells were collected and the levels of active ERK and CREB were assessed by immunoblot. Results (Fig. 6a and b) indicated that cells treated with Aβ oligomer alone had significantly decreased levels of pERK at 1.5 and 24 h (***p <0.001 and *p <0.05, respectively), and pCREB at 1.5 and 12 h (*p <0.05 and **p <0.01, respectively). To demonstrate that oligomers and not other forms of Aβ were responsible for ERK and CREB effects, an oligomer-specific antibody, A11, was used to immunoneutralize oligomeric Aβ. The levels of pERK at 1.5 and 24 h and pCREB at 12 h were significantly elevated in cells pre-treated with A11 antibody followed by Aβ42 oligomer compared with cells treated with Aβ42 oligomer alone (***p <0.001, *p <0.05, and **p <0.01, respectively, Fig. 6a and b). Total ERK and actin were unchanged confirming equal sample loading in each lane. Cell viability measured by lactate dehydrogenase was unchanged at the time of the decrease of pERK (1.5 and 24 h) or pCREB (1.5 and 12 h), suggesting that the observed effect was not simply a result of altered cell viability. A11 antibody also significantly protected from Aβ oligomer-induced delayed cytotoxicity which emerged at 24 h (p <0.01, Fig. 6c).
Oligomer-specific antibody prevented Aβ oligomer-induced decreases in activated ERK and CREB in primary hippocampal neurons
To further confirm the direct impact of Aβ oligomers on neuronal ERK and CREB, cultured hippocampal neurons were employed. Cells were pre-incubated with or without A11 antibody (2 ng/mL) for 1 h followed by treatment with 250 nmol/L of Aβ oligomers for 2 h. pERK and pCREB were visualized by immunofluorescence staining with anti pERK or anti pCREB (Fig. 7a and b). Aβ oligomers were visualized by staining with anti-Aβ antibodies, 6E10. Aβ oligomers specifically bound as puncta on neurons or processes (Fig. 7c) and significantly decreased the intensity of pERK (p <0.01) in the cytoplasm compared with controls and these losses were reversed by A11 antibody (Fig. 7a and d). In addition, Aβ oligomers induced a redistribution of pCREB from the nucleus, significantly decreasing the intensity of nuclear pCREB (p <0.01) when compared with control. These changes were also blocked by A11 antibody (Fig. 7b and e).
Anti-Aβ antibody blocked Aβ-induced decreases of the NMDAR subunit, NR2B
Activation of NMDARs can lead to long-term potentiation (LTP) and increased synaptic strength. The NMDAR family has a complex heteromeric composition comprised of variable NR1 and NR2 subunits. In fact, over-expression of only the NR2B subunit enhanced hippocampal LTP, prolonged NMDAR currents, and led to superior ability in learning and memory in various behavioural tasks (Tang et al. 1999). In contrast to NR2B, NR1 was not altered in Tg2576 (Calon et al. 2005). Because NR2B is directly coupled to ERK which is stimulated by NMDAR activation (Krapivinsky et al. 2003), we assessed the levels of NR2B. Our data revealed that anti-Aβ-infused mice had significantly increased levels of NR2B compared with controls (p <0.05, Fig. 8c). In addition, when mature 19-day-old primary hippocampal neurons were treated with 250 nmol/L of Aβ oligomers for 7 h, the Aβ oligomers were located as puncta on neurons or processes bearing a decreased intensity of NR2B when compared with untreated controls. A11 antibody prevented this Aβ oligomer-induced change (Fig. 8a and b).
We demonstrate that regardless of baseline ERK phosphorylation levels, ERK activation can be impaired by APP transgene, and this effect is mediated by Aβ oligomers. Previously age-dependent alterations in Tg2576 were evaluated in the hippocampus and baseline pERK was found to be increased as early as 4 months of age in the dentate gyrus prior to Aβ deposition (Dineley et al. 2001) and in the absence of detectable synaptic and neuronal loss (Irizarry et al. 1997). ERK hyperactivation has been linked to APP transgene-dependent up-regulation of the alpha7 nicotinic receptor and Aβ signaling to ERK (Dineley et al. 2001). We evaluated age-dependent changes in the cortex to see if parallel changes occurred. Unlike in the hippocampus, our results show that cortical pERK was not reduced or elevated at 12 months. Consistent with a previous report, we also observed a trend for increased baseline pERK in CA1 of the hippocampus (Dineley et al. 2001).
Infusion-induced pERK at both 11–12 months of age and at 16–18 months of age in wild-type (Tg−) mice, but these robust responses were either not present or blunted in APP Tg+ mice, demonstrating an APP transgene-mediated ERK dysregulation. To address the role of Aβ, we infused anti-Aβ antibodies. Infusion of antibodies against Aβ increased pERK and pCREB at both ages, essentially correcting the suppressed response to infusion/injury. Thus, the significant increase in ERK and CREB observed with anti-Aβ antibody treatment cannot be viewed simply as a correction of a pre-existing baseline deficit, but rather as a correction of a blunted ERK/CREB response to CNS infusion-induced factors. Trophic factors are one possibility because subtoxic levels of Aβ oligomers have been reported to blunt the ERK/CREB response to trophic factors including nerve growth factor (Chromy et al. 2003) and brain derived neurotrophic factor (Tong et al. 2004). IGF-1 is one example of a persistent cannulation-induced trophic factor (Garcia-Estrada et al. 1992). Our results showing that IGF induced ERK in SY5Y neuroblastoma is attenuated by Aβ oligomers further support the idea that Aβ oligomers may diminish ERK signaling. While we found the Tg+ mice had lost the ERK response to injury-induced stimulation, our results are consistent with the hypothesis that oligomers in APP Tg+ mice reduce the ERK response to injury-induced trophic factors.
In contrast to 12 month mice, significant cortical ERK and CREB deficits develop in older Tg2576 mice. These were further exacerbated by an omega-3 fatty acid depleted diet (not shown) that also increased selected synaptic deficits (Calon et al. 2005). Our in vivo and in vitro immunoneutralization data with anti-Aβ and anti-oligomer antibodies data suggest that these deficits are due to Aβ oligomers. Preparations enriched in Aβ oligomers, including trimer, 12-mer, and 24-mer, significantly down-regulated the level of active ERK (1.5 and 24 h) and active CREB (1.5 and 12 h). The conformation-specific anti-oligomer antibody, A11, ameliorated this Aβ-induced inhibition of active ERK and CREB in human SH-SY5Y cells and rat primary hippocampal neurons. As A11 antibody fails to detect monomer, low molecular weight oligomers, or fibrils (Kayed et al. 2003), active ERK and CREB losses may be attributed to one or more high molecular weight Aβ oligomer species.
Our findings of Aβ-dependent ERK suppression are distinct from data obtained with chronic incubation with aged or fibrillar Aβ which caused persistent ERK activation in hippocampal slice cultures (Bell et al. 2004) and aged hippocampal neurons (Anderson and Ferreira 2004). However, they are consistent with experiments using freshly prepared Aβ42 containing predominantly hexameric oligomer species that resulted in an initial ERK activation followed by ERK down-regulation at 16 h (Bell et al. 2004). Aβ or Aβ25–35 induced suppression of ERK or its downstream intermediate CREB have also been observed in neuroblastoma cells (Daniels et al. 2001), endothelial cells (Magrane et al. 2006), hippocampal preparations (Xie 2004), and in rats after injection of Aβ25–35 (Jin et al. 2005). In vitro, Aβ effects on ERK activation may reflect dose, aggregation status and treatment duration. Because we used 250–500 nmol/L of pre-aggregated Aβ42 with a predominant 12-mer (dodecamer) species to treat cells, we may have obtained a more rapid ERK down-regulation, consistent with the hypothesis that the Aβ 12-mer is a very potent inhibitor of mechanisms underlying cognitive function.
Large (> 50%) pCREB deficits have been reported in hippocampus of AD patients (Yamamoto-Sasaki et al. 1999). Although neither pERK nor pCREB show early hippocampal deficits in Tg2576 mice (Dineley et al. 2001), pCREB deficits occur in other regions. For example, we observed pCREB deficits in 16-month-old Tg2576 mouse entorhinal cortex where it was restored by treatment with curcumin (Cole et al. 2003), a compound that blocks Aβ oligomer and fibril formation (Yang et al. 2005) and reduces cognitive deficits induced by chronic soluble Aβ infusion (Frautschy et al. 2001). Defects in CREB signaling are likely Aβ-dependent. They are relevant to cognitive function because in bigenic mutant APP × presenlin 1 transgenic mice, increasing cAMP/protein kinase A (PKA)/CREB signaling with rolipram ameliorated deficits in LTP and cognitive function (Gong et al. 2004). PKA activity has been reported to be reduced by Aβ aggregates in vitro and in bigenic APP × presenlin 1 mice secondary to a 30–50% decrease in Uch-L1 hydrolase, but 9–12 month Tg2576 (used in our experiments) have only about a 15% decrease in Uch-L1 and no reported change in PKA activity or protein (Gong et al. 2006). Whether involving ERK or PKA or both, data from three very different compounds in three model systems show correction of Aβ-induced cognitive deficits involving concomitant increases in pCREB. Given the known role of CREB in plasticity, small molecules capable of increasing pCREB may be a useful downstream therapeutic approach.
Aberrantly increased or decreased, but dysregulated ERK signaling is likely to impact human cognition. ERK signaling events have been studied extensively in recent years for involvement in synaptic plasticity and memory function (English and Sweatt 1997). ERK alterations are involved in a human mental retardation syndrome, neurofibromatosis 1, in which defects in the ras/ERK signaling cascade contribute to hippocampal-dependent learning and memory deficits (Costa et al. 2002). Normally, ERK participates in regulating local dendritic protein synthesis (Ying et al. 2002), processing molecular information in dendrites (Watanabe et al. 2002), and stabilizing structural changes in dendritic spines (Goldin and Segal 2003).
Extracellular-signal regulated kinase activation occurs in neurons in response to excitatory glutamatergic signaling (Kurino et al. 1995; Xia et al. 1996). Glutamate receptors including NMDARs activate multiple biochemical pathways that transduce post-synaptic signals (Sheng et al. 2002). A reduction in specific NMDAR subunits, notably NR2B, occurs in AD patients and correlates with cognitive deficits (Greenamyre et al. 1987; Mishizen-Eberz et al. 2004). In Tg2576 mice, omega-3 fatty acid deficiency aggravates APP transgene-dependent decreases in NR2B, the developmentally regulated brain protein, drebrin, and PSD-95, all indices of a defect in excitatory neurons (Calon et al. 2004). In the present study, we found that anti-Aβ-infused Tg2576 mice had a significantly increased level of NR2B. In primary hippocampal neuron culture, Aβ oligomers specifically bound to neurons and neurites as puncta associated with decreased NR2B, consistent with a selective effect on NMDARs trafficking by Aβ (Snyder et al. 2005).
Considerable evidence indicates that immunotherapeutic approaches against Aβ reverse cognitive deficits (Morgan et al. 2000; Klyubin et al. 2005) and promote the rapid recovery of amyloid-associated neuritic dystrophy in APP mutant transgenic mice (Brendza et al. 2005). However, the mechanisms to explain why Aβ antibodies acutely improve learning and memory deficits even without clearance of Aβ-related pathology are not well understood. This study provides initial evidence that Aβ antibody treatment of 11–12 month Tg2576 mice has rapid effects on Aβ oligomer-modulated signal transduction, increasing NR2B and downstream active ERK and CREB, all critical components of signal transduction pathways involved in learning and memory. While insoluble deposits were not significantly reduced in our acute study, they may not play a central role in cognitive deficits. Some AD transgenic mouse models show significant cognitive dysfunction and synaptic damage but have no insoluble Aβ deposits (Kumar-Singh et al. 2000; Mucke et al. 2000). Insoluble Aβ deposits do not predict the degree of dementia in AD patients (McLean et al. 1999) and evidence has emerged suggesting that soluble Aβ oligomer may be a key toxic form of Aβ implicated in AD pathogenesis. For example, soluble oligomeric forms of Aβ caused selective synaptic protein loss and cognitive decline in Aβ oligomer-infused rats (Frautschy et al. 2001) and inhibited LTP and specifically disrupted cognitive function (Cleary et al. 2005). In particular, Aβ 12-mer has been detected as a highly stable, soluble extracellular aggregate in Tg2576 mice that is sufficient to induce cognitive deficits (Lesne et al. 2006). Our previous study demonstrated that an acute treatment with anti-Aβ antibody not only neutralized the level of ∼56 kDa oligomer (12-mer) of Aβ, but also counteracted the effect of Aβ oligomer-induced activation of the GSK3β/ptau signaling cascade without a significant reduction of insoluble Aβ in Tg2576 mice (Ma et al. 2006). Aβ vaccine or anti-Aβ antibodies are still among the most promising treatments for AD. A clinical trial report from one of the groups that participated in the ELAN/Wyeth-Ayerst AN1792 (QS-21) Phase 2A multicenter trial found benefits in patients who both received a primary and a booster immunization of Aβ42 and generated anti-Aβ plaque immunoreactivity (tissue amyloid plaque immunoreactivity assay). Compared with patients who are without such antibodies, the responders, including several with inflammatory side effects appeared to show significantly slower rates of decline of cognitive decline (Hoek et al. 2003).
In summary, this study shows that soluble Aβ oligomers can blunt the ERK and CREB response in vivo and provides direct evidence that passive immunization with anti-Aβ antibodies can increase ERK/CREB activation and NMDAR (NR2B) in Tg2576 mice and in oligomer-treated neuronal cultures. These data also strongly support therapeutic strategies targeting Aβ oligomer by highly specific anti-Aβ oligomer antibodies or drugs to limit dysregulation of multiple memory-related signaling pathways, including oligomer sensitive NMDAR/ERK/CREB, Akt/GSK3β/ptau (Ma et al. 2006), and PAK/LIMK/cofilin/drebrin (Zhao et al. 2006).
We thank Drs Charles G. Glabe and Rakez Kayed for kindly providing A11 antibody and thank Dr Fusheng Yang, Pingping Chen, Beverly Hudspeth for assistance. This work was supported by NIA NS43946 (GMC), NIA AG021975 (SAF), and AG022080 (MEH-W).