Brain-derived neurotrophic factor (BDNF) and its high-affinity full-length (FL) receptor, TrkB-FL, play a central role in the nervous system by providing trophic support to neurons and regulating synaptic plasticity and memory. TrkB and BDNF signaling are impaired in Alzheimer's disease (AD), a neurodegenerative disease involving accumulation of amyloid-β (Aβ) peptide. We recently showed that Aβ leads to a decrease of TrkB-FL receptor and to an increase of truncated TrkB receptors by an unknown mechanism. In the present study, we found that (1) Aβ selectively increases mRNA levels for the truncated TrkB isoforms without affecting TrkB-FL mRNA levels, (2) Aβ induces a calpain-mediated cleavage on TrkB-FL receptors, downstream of Shc-binding site, originating a new truncated TrkB receptor (TrkB-T′) and an intracellular fragment (TrkB-ICD), which is also detected in postmortem human brain samples, (3) Aβ impairs BDNF function in a calpain-dependent way, as assessed by the inability of BDNF to modulate neurotransmitter (GABA and glutamate) release from hippocampal nerve terminals, and long-term potentiation in hippocampal slices. It is concluded that Aβ-induced calpain activation leads to TrkB cleavage and impairment of BDNF neuromodulatory actions.

Introduction

Brain-derived neurotrophic factor (BDNF) is a neurotrophin that promotes neuronal survival, differentiation, and synaptic plasticity through activation of its full-length receptor, TrkB-FL. However, besides encoding for this receptor, the TrkB gene (NTRK2) also encodes for truncated isoforms (Klein et al. 1990), which may act as negative modulators of TrkB-FL signaling (Eide et al. 1996; Dorsey et al. 2006). Both decreases in the ratio between FL and truncated (Tc) receptors and reduced BDNF signaling have been detected in several neurodegenerative disorders. Particularly, hippocampal and cortical postmortem samples from Alzheimer's disease (AD) patients revealed a decrease in both BDNF and TrkB-FL and an increase in TrkB-Tc levels (Phillips et al. 1991; Connor et al. 1997; Allen et al. 1999; Ferrer et al. 1999; Holsinger et al. 2000). These changes are thought to be involved in spatial memory impairments and, accordingly, the activation or overexpression of TrkB-FL has been associated with spatial memory improvements (Blurton-Jones et al. 2009; Devi and Ohno 2012; Kemppainen et al. 2012).

AD is characterized not only by the accumulation of intracellular neurofibrillary tangles made of hyperphosphorylated tau proteins but also of extracellular plaques composed by amyloid-β peptides (Aβ). Aβ plaques are largely composed by Aβ40 and Aβ42, but also by Aβ fragments including the Aβ25–35 (Kubo et al. 2002), which has been proposed to be the active region of the full-length Aβ peptide responsible for its neurotoxic effects (Pike et al. 1995). Mechanisms underlying the neurotoxic actions of Aβ peptides are not fully understood, but the existing data suggest that oxidative stress, perturbation of calcium homeostasis, mitochondrial dysfunction, synaptic loss, and caspases and calpains activation are strongly involved (Holscher 1998, 2005; Selkoe 2002; Wei et al. 2008). Calpains are Ca2+-dependent proteases that play a physiologic role by the cleavage of several substrates, changing their function or localization. Abnormal activation of calpains and down-regulation of its endogenous inhibitor (calpastatin) have been linked to AD (Saito et al. 1993; Grynspan et al. 1997; Adamec et al. 2002). In addition, calpain overactivation contributes to tau hyperphosphorylation, a hallmark of AD, through the activation of cyclin-dependent kinase 5 (CDK5), followed by the cleavage of its regulatory protein—p35 (Lee et al. 2000; Noble et al. 2003; Cruz and Tsai 2004). Moreover, calpain also contributes to the formation and accumulation of Aβ peptides, and its inhibition prevents neurodegeneration and restores normal synaptic function and spatial memory in AD animal models (Trinchese et al. 2008; Granic et al. 2010; Medeiros et al. 2012).

Recently, we observed that Aβ induces a decrease in TrkB-FL receptors and an increase in truncated TrkB receptors on primary neuronal cultures, which is independent of the presence of glial cells (Kemppainen et al. 2012), suggestive of a relationship between the presence of Aβ and alterations in neuronal TrkB receptors. In the present study, we characterized the molecular alterations and the mechanisms activated by Aβ to affect TrkB receptors, as well as the functional consequences of such mechanisms. We uncovered that Aβ selectively increases mRNA levels of truncated TrkB-T1/T2 receptors and promotes a calpain-mediated cleavage of TrkB-FL, whereby producing a new truncated receptor and an intracellular fragment, the last one being also present in human brain samples. Importantly, Aβ-induced calpain-mediated TrkB cleavage impacts upon BDNF signaling since BDNF neuromodulatory actions were hampered by Aβ but rescued by calpain inhibition.

Materials and Methods

Animals and Brain Areas Used

Sprague–Dawley and Wistar rats were purchased from Harlan Interfauna Iberica, SL, and were housed in a temperature- (21 ± 1°C) and humidity-controlled (55% ± 10%) room with a 12:12 h light/dark cycle with food and water ad libitum. All animals were handled according to the current Portuguese Laws and to the European Union Directive (86/609/EEC) on the protection of Animals used for Experimental and other scientific purposes. All efforts were made to minimize animal suffering. Rats were deeply anesthetized with halothane before decapitation and tissue preparation. For functional studies, we used the hippocampus, which is a brain area severely affected in AD, and where the effects of BDNF are extensively characterized. Since Aβ-induced TrkB alterations are similar in cortical and hippocampal cultures (Kemppainen et al. 2012), we used cortical cultures for the molecular studies, to increase the culture yield and reduce the number of animals.

Human Brain Sample

Frontal cortex from a control case was obtained from the Lille Neurobank, France (Male, 41 years old, postmortem interval: 11 h) after scientific committee agreement.

Materials

Unless stated otherwise all reagents were purchased from Sigma. Culture reagents and supplements were from Gibco. Recombinant human BDNF was a gift from Regeneron Pharm. BDNF was used in a final concentration of 20 ng/mL (corresponding to ∼0.75 nm in cell cultures and in LTP experiments) and of 30 ng/mL (corresponding to ∼1.1 nm in neurotransmitter release assays). Rat recombinant m-calpain was from Calbiochem and the N-terminal His6-tagged recombinant human TrkB active (aa.455-end) was from Millipore. MG132, ALLN, MDL28170 and Pepstatin A were from Tocris Bioscience. Aβ25–35 peptide and zVAD(OMe)-FMK were purchased from Bachem, and Aβ1–42 peptide was purchased from rPeptide. [3H]GABA (4-amino-n-[2,3–3H]butyric acid, specific activity 92.0 Ci/mmol) and [3H]glutamic acid (l-[3,4–3H] glutamic acid, specific activity 49.6 Ci/mmol) were purchased from PerkinElmer Life Sciences.

Amyloid-β Peptides

Most of the experiments were performed using Aβ25–35 (from Bachem) and in selected key experiments the Aβ1–42 (from R-peptide) and the inverted Aβ35–25 peptide (from Bachem) were also used. Stock solutions of Aβ25–35, Aβ1–42, and Aβ35–25 peptides were performed in MilliQ water to a final concentration of 1 mg/mL. Stock solutions of Aβ25–35 and Aβ1–42 contain mainly protofibrillar and fibrillar amyloid structures as confirmed by atomic force microscopy and Thioflavin T binding as previously shown in Kemppainen et al. (2012).

The concentration of Aβ42 present in interstitial fluid (ISF) of human AD brain parenchyma is not known and difficult to predict. Nevertheless, it is estimated that Aβ42 is present in ISF at a concentration range of nanomolar (see Karran et al. 2011). Aβ toxicity may also differ according to the duration of Aβ exposure, the species (monomeric, oligomeric, or fibrillar ones), and type of Aβ peptide (e.g., 25–35, 42). It is, therefore, difficult to use conditions that mimic exactly what happens in human AD patients. The concentrations of Aβ used in this work were 25 µm for Aβ25–35 and 20 µm for Aβ1–42, the same as those used previously by us (Kemppainen et al. 2012) and similar to those used by others (Arancibia et al. 2008). It is worthwhile to note that, although the concentration of Aβ is relatively high, the exposure period to the peptide was much smaller (24 h for the cultures, 3 h in the acute slice preparations) than what happens for human AD patients (several years). Importantly, as a control, we used the inverted Aβ35–25 peptide, also at the concentration of 25 µm.

Primary Neuronal Cultures and Drug Treatments

Neurons were isolated from fetuses of 18-day pregnant females. The fetuses were collected in Hanks' balanced salt solution (HBSS-1), and brains were rapidly removed. The cerebral cortex were isolated and mechanically fragmented. Further tissue digestion was performed with 0.025% (wt/vol) trypsin solution in HBSS without Ca2+ and Mg2+ (HBSS-2) for 15 min at 37°C. After trypsinization, cells were washed and resuspended in Neurobasal medium supplemented with 0.5 mml-glutamine, 25 mm glutamic acid, 2% B-27, and 25 U/mL penicillin/streptomycin. Cells were plated at 7 × 104 cells/cm2 on poly-d-lysine (10 µg/mL)-coated dishes and maintained at 37°C in a humidified atmosphere of 5% CO2. Incubations with Aβ peptides were performed at 7 or 14 DIV for 24 h, as before (Kemppainen et al. 2012). In the experiments where protease inhibitors were used, the inhibitors were incubated 20 min before Aβ treatment.

Western Blot

Cells were washed with ice-cold phosphate-buffered saline (137 mm NaCl, 2.7 mm KCl, 8 mm Na2HPO4.2H2O, and 1.5 mm KH2PO4, pH 7.4) and lysed with 1% NP-40 lysis buffer containing (in mm): 50 Tris–HCl (pH 7.5), 150 NaCl, 5 ethylenediamine tetra-acetic acid (EDTA), 2 dithiothreitol (DTT), and protease inhibitors (Roche). Cell lysates were clarified by centrifugation (16 000g, 10 min), and the amount of protein in the supernatant was determined by Bio-Rad DC reagent. Each sample (40 µg of total protein) was separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) and transferred onto PVDF membranes (GE Healthcare). Membranes were stained with Ponceau S solution to check for protein transference efficacy. After blocking with a 5% nonfat dry milk solution in TBS-Tween (20 mm Tris base, 137 mm NaCl, and 0.1% Tween-20), membranes were incubated with the primary (overnight at 4°C) and secondary antibodies (1 h at room temperature). Finally, immunoreactivity was visualized using ECL chemiluminescence detection system (Amersham-ECL Western Blotting Detection Reagents from GE Healthcare), and bands intensities were quantified by digital densitometry (ImageJ 1.45 software). The intensity of α-tubulin or Ponceau S bands were used as loading control.

The pan-TrkB mouse monoclonal antibody (1:1500), raised against the extracellular domain of human TrkB (aa. 156–322), was from BD Bioscience. The C-terminal of Trk-FL rabbit polyclonal antibody (1:2000), raised against the C-terminus (C-14), and the αII-Spectrin (C-3) mouse monoclonal antibody (1:2500), raised against human spectrin (aa. 2368–2472), were from Santa Cruz, Inc. The phospho-TrkA (Tyr 490) rabbit polyclonal antibody (1:1500), which detects TrkB receptor when phosphorylated on the corresponding residue (Tyr515), was from Cell Signaling Technology. Phospho-TrkB (Tyr 816) antibody (1:1500) specifically detects TrkB when phosphorylated on Tyr816. The α-tubulin rabbit polyclonal antibody (1:5000) was from Abcam. The IgG-horseradish peroxidase-conjugated secondary antibodies used were goat anti-mouse and goat anti-rabbit (Santa Cruz). Mouse anti-rabbit IgG light chain specific (Jackson ImmunoResearch Laboratories) was used in Supplementary Fig. 2C.

N-Sequencing

Five micrograms of TrkB active (Millipore) were incubated with purified m-calpain (2.5 U) and CaCl2 (2 mm) in a final volume of 30 µL for 30 min at 25°C. The mixture was resolved by SDS–PAGE with 2 mm of thioglycolic acid in upper running buffer and transferred to a PVDF membrane. After Ponceau S staining, the band of interest (TrkB-ICD) was cut with a sharp clean blade. The data were provided by the protein sequencing service of Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa. Note that the proline residue is not detected by N-sequencing.

Calpain In-Vitro Digestion

In calpain digestion assays, the cultured cells or brain tissue were homogenized on ice in 1% NP-40 buffer containing (in mm): 50 Tris–HCl (pH 7.5), 150 NaCl, 0.1 EDTA, 2 DTT, 1 phenylmethylsulfonyl fluoride, and Aprotinin 5 µg/mL. The homogenates were clarified by centrifugation (16 000g, 10 min), and protein concentration was determined. In exogenous calpain digestion assays, the purified rat m-calpain (Calbiochem) was incubated for 30 min at 25°C in a 100 µL final volume of lysis buffer containing 100 µg of homogenate protein and 2 mm of CaCl2 (unless stated otherwise). In endogenous calpain digestion assays, 5 mm of CaCl2 and/or MDL28170 were added to the homogenates for 4 h at 25°C. In calpain digestion assays (exogenous or endogenous), each condition has the same amount of protein and total volume. For endogenous calpain activation in synaptosomes, CaCl2 and/or MDL28170 (20 µm) were added to the intact isolated synaptosomes suspended in KHR buffer, for 30 min at 37°C. All reactions were stopped by boiling the samples at 95°C in the presence of the denaturing SDS-sample buffer.

Immunoprecipitation

Trk-FL receptors were immunoprecipitated with 2 µg of C-terminal Trk-FL (sc14) antibody in 500 µL of neuronal cultures lysates (∼1 mg total protein). After overnight incubation at 4°C, 30 µL of packed G-protein agarose beads were added for 24 h at 4°C, and then, the tube was centrifuged and the supernatant (wash-flow) was collected. The remaining pellet of beads was washed 5 times with lysis buffer and resuspended in 100 µL of calpain lysis buffer containing purified m-calpain and CaCl2 (30 min at 25°C, as described earlier). The reactions were boiled at 95°C in the presence of denaturing SDS-sample buffer.

RNA Extraction and qPCR

RNA isolation and qPCR were performed as previously described (Aroeira et al. 2011). Briefly, total RNA was extracted from rat neuronal cultures (GE Healthcare RNAspin Mini RNA Isolation Kit). First-strand cDNA were synthesized from 1 µg of total RNA (in 20 µL) according to the manufacturer's recommendations (SuperScript First Strand Synthesis Systems for RT-PCR from Invitrogen). cDNA was amplified in Rotor-Gene 6000 real-time rotary analyzer thermocycler (Corbett Life Science) in the presence of SYBR Green Master Mix (Applied Biosystems) and each specific gene primer (0.2 µm for TrkB-FL and TrkB-T2 and 0.5 µm for TrkB-T1). Primer specificity was confirmed by melting curves (Supplementary Fig. 1A). The threshold cycle (Ct) and the melting curves required for the relative quantification (Pfaffl 2001) were acquired with Rotor-Gene 6000 Software 1.7 (Corbett Life Science). β-actin was used as reference internal standard. Replica reactions were always performed.

The primers used were: 5′-GTGATGCTGCTTCTGCTCAA-3′ and 5′-CCTCCGAAGAAGACGGAGTG-3′ for TrkB-FL; 5′-TAAGATCCCCCTGGATGGGTAG-3′ and 5′-AAGCAGCACTTCCTGGGATA-3′ for TrkB-T1; 5′-CGGGAGCATCTCTCGGTCT-3′ and 5′-TCCACTTAAGAAGCAAAATAAGC-3′ for TrkB-T2; 5′-AGCCATGTACGTAGCCATCC-3′ and 5′-CTCTCAGCTGTGGTGGTGAA-3′ for β-actin. The primers for TrkB-T2 were designed using the OligoAnalyzer 3.1 tool, provided by Integrated DNA Technologies. The TrkB-T2 mRNA sequence from Rattus norvegicus was obtained from the GenBank sequence database of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). The primers were synthesized by Invitrogen.

Freshly Prepared Hippocampal Slices

Male Wistar rats (8–10 weeks old) were deeply anesthetized with halothane before decapitation. The brain was quickly removed into ice-cold continuously oxygenated (O2/CO2: 95%/5%) artificial cerebrospinal fluid (aCSF) (124 mm NaCl, 3 mm KCl, 1.2 mm NaH2PO4, 25 mm NaHCO3, 2 mm CaCl2, 1 mm MgSO4, and 10 mm glucose, pH 7.40), and the hippocampi were dissected out. The hippocampal slices were cut perpendicularly to the long axis of the hippocampus (400 µm thick) and were allowed to recover functionally and energetically for at least 1 h in a resting chamber filled with continuously oxygenated aCSF, at room temperature (22–25°C). After recovering, slices were incubated for 3 h (the minimum time required to observe changes in TrkB receptor cleavage [Kemppainen et al. 2012]) with oxygenated aCSF (control), or with aCSF containing Aβ25–35 peptide (25 μm) or with aCSF containing Aβ25–35 and MDL28170 (calpain inhibitor, 20 µm) or with Aβ1–42 (20 μm) or with the inverted Aβ35–25. After this incubation period, the slices were used for electrophysiology recordings or for synaptosomal isolation to evaluate GABA and glutamate release.

Isolation of Synaptosomes

Hippocampal slices were homogenized in ice-cold isosmotic sucrose solution (0.32 m, containing 1 mm EDTA, 1 mg/mL bovine serum albumin, and 10 mm HEPES, pH 7.4) and centrifuged at 3000g for 10 min; the supernatant was centrifuged again at 14 000g for 12 min. The whole procedure was conducted at 4°C. The pellet was resuspended in 45% Percoll in KHR solution consisting of (in mm) NaCl 140, EDTA 1, HEPES 10, KCl 5, and glucose 5 and was centrifuged at 14 000g for 2 min. The synaptosomal fraction corresponds to the top buoyant layer and was collected from the tube. Percoll was removed by 2 washes with a KHR solution; synaptosomes were then kept on ice and used within 3 h.

[3H] Neurotransmiter Release from Hippocampal Synaptosomes

The [3H]GABA release experiments were performed as previously (Sousa et al. 2011). For each experiment, synaptosomes were prepared from ∼60 hippocampal slices (30 per condition) from 6 hippocampi of 3 animals. Synaptosomes (protein concentration 1–2 mg/mL) were resuspended in 2 mL of oxygenated Krebs medium (in mm: NaCl 125, KCl 3, NaH2PO4 1, glucose 10 NaHCO3 25, CaCl2 1.5 and MgSO4 1.2) and allowed to equilibrate for 5 min at 37°C. From this time onward, all solutions applied to the synaptosomes were kept at 37°C and continuously gassed with O2/CO2 (95%/5%). Aminooxyacetic acid (AOAA, 0.1 mm) was present in all solutions up the end of the experiments to prevent GABA catabolism by inhibition of GABA transaminase. The synaptosomes were loaded for 20 min at 37°C, with [3H]GABA (1.5 μCi/mL, 18.5 nm), together with 0.625 μm unlabeled GABA to decrease specific activity of the [3H]GABA solutions to 2.3 μCi/nmol and equally layered onto perfusion chambers over Whatman GF/C filters (flow rate, 0.8 mL/min; chamber volume, 90 μL).

The [3H]glutamate release assays were performed as routinely in our laboratory (Sousa et al. 2011). All procedures were similar to [3H]GABA release experiments with the necessary modifications. Synaptosomes were loaded with 0.2 µm [3H]glutamate (specific activity was 30–60 Ci/mmol) for 5 min and equally layered onto perfusion chambers over Whatman GF/C filters (flow rate, 0.6 mL/min; chamber volume, 90 µL).

After a 20-min washout period, the effluent was collected for 40 min in 2-min intervals. The GABA or glutamate release from synaptosomes was stimulated during 2 min with a high-K+ solution (15 mm, isomolar substitution of Na+ with K+ in the perfusion buffer) at the 5th (first stimulation period [S1]) and 29th (second stimulation period [S2]) minute after starting sample collection. BDNF (30 ng/mL) was added to the superfusion medium at the 9th minute, therefore before S2, and remained in the bath up to the end of the experiments, and its effect was quantified as percentage changes of the S2/S1 ratio compared with the S2/S1 ratio in the absence of BDNF in the same synaptosomal batch and under similar drug conditions. Thus, BDNF effect upon S2/S1 ratio was determined from synaptosomes prepared from slices incubated without any drug, incubated with Aβ or incubated with both Aβ and MDL28170.

Calculation of Drug Effects on GABA and Glutamate Release

At the end of each experiment, aliquots (500 µL) of each sample as well as the filters from each superfusion chamber were analyzed by liquid scintillation counting. The fractional release was expressed in terms of the percentage of total radioactivity present in the preparation at the beginning of the collection of each sample. The amount of radioactivity released by each pulse of K+ (S1 and S2) was calculated by integration of the area of the peak upon subtraction of the estimated basal tritium release. In each experiment, 2 synaptosome-loaded chambers were used as control chambers, the others being used as test chambers. In the test chambers, the test drug was added to the perfusion solution before S2 and the S2/S1 ratios in control and test conditions were calculated. The effect of the drug on the K+-evoked tritium release was expressed as percentage of change of the S2/S1 ratios in test conditions compared with the S2/S1 ratios in control conditions, in the same experiments (i.e., with the same pool of synaptosomes).

Ex-Vivo Electrophysiology Recordings

Long-term potentiation (LTP) induction and quantification were performed as described previously (Diogenes et al. 2011). Briefly, the hippocampal slices from 8- to 10-week-old Wistar rats were transferred to a recording chamber continuously superfused with oxygenated aCSF at 32°C (flow rate of 3 mL/min in open system). The stimulation pulses were delivered every 10 s alternately to 2 independent pathways through electrodes placed on Shaffer collateral/commissural fibers in stratum radiatum, and the fEPSPs were recorded in stratum radiatum of CA1 area. Control experiments were performed to confirm that LTP magnitude was similar in both pathways (see Supplementary Fig. 5A).

LTP was induced by θ-burst protocol consisting of 4 trains of 100 Hz, 4 stimuli, separated by 200 ms. We used θ-burst stimulation to induce LTP, since this pattern of stimulation is considered closer to what occurs physiologically in the hippocampus during episodes of learning and memory in living animals (Albensi et al. 2007). Furthermore, the facilitatory action of BDNF upon LTP is mostly seen under θ-burst stimulation (Chen et al. 1999). In addition, we previously showed that the effect of BDNF upon CA1 LTP is more evident under weak (as the used in this work) than under strong θ-burst stimulation paradigms (Fontinha et al. 2008). Therefore, we selected the optimal stimulation paradigm to observe an effect of BDNF upon LTP, so that we could evaluate the influence of Aβ upon the effect of BDNF.

One hour after LTP induction in 1 of the pathways, BDNF (20 ng/mL) was added to the superfusion solution and LTP was induced in the second pathway, no less than 20 min after BDNF perfusion. Whenever an increase on the slope of fEPSP was detected in the presence of BDNF, the intensity of stimulation was adjusted before LTP induction for similar values recorded before BDNF application. LTP was quantified as % change in the average slope of the fEPSP taken from 46 to 60 min after LTP induction in relation to the average slope of the fEPSP measured during the 14 min before the induction of LTP. The effect of BDNF upon LTP was evaluated by comparing the magnitude of LTP in the first pathway in the absence of BDNF (control pathway), with the magnitude of LTP in the second pathway in the presence of BDNF (test pathway). The independence of the 2 pathways was tested in the end of experiments by studying the pair-pulse facilitation across both pathways, <10% facilitation being usually observed.

Input/Output Curves

Input/output curves were performed after a stable baseline of at least 15 min. The stimulus delivered to the slice was decreased until no fEPSPs evoked and subsequently increased by steps of 20 μA. Data from 3 consecutive fEPSPs were collected for each stimulation intensity. The range of all input delivered to the slice was typically from 100 μA to a supramaximum stimulation amplitude of 360 μA. The input/output curve was plotted as the relationship of fEPSP slope versus stimulus intensity.

Statistical Analysis

The data are expressed as mean ± SEM of the n number of independent experiments. The significance of differences between the means of 2 conditions was evaluated by Student's t-test. To perform multiple comparisons between the means of >2 conditions, a one-way ANOVA followed by a Bonferroni post-test was performed. To perform comparisons on LTP magnitude in the presence or absence of BDNF between different slice treatments, a two-way ANOVA followed by a Bonferroni post-test was performed. Values of P < 0.05 were considered to represent statistically significant differences.

Results

Aβ Increases Levels of Truncated TrkB Receptors and Promotes TrkB-FL Cleavage

To evaluate the effects of Aβ upon TrkB receptors expression, we determined the mRNA levels of the main TrkB isoforms produced by alternative splicing (FL, T1, and T2). qPCR data showed that neuronal cultures incubated with Aβ25–35 (25 µm) for 24 h displayed a significant increase of truncated TrkB-T1 (45% ± 19%, n = 8, P < 0.05, Fig. 1A) and truncated TrkB-T2 (58% ± 17%, n = 8, P < 0.05, Fig. 1A) mRNA levels as compared with non-treated control cultures. Conversely, no significant change of TrkB-FL mRNA levels was detected upon Aβ25–35 incubation (Fig. 1A, n = 8). As shown in Supplementary Figure 1B, the TrkB-FL and TrkB-T1 are the main TrkB isoforms present in the cultures, whereas TrkB-T2 was detected at lower expression levels. Thus, only TrkB-T1 will be mentioned as the main spliced truncated isoform.

Figure 1.

Aβ peptide affects TrkB receptors by selective up-regulation of truncated TrkB mRNA and by cleavage of TrkB-FL receptors. (A) Analysis of mRNA levels by relative qPCR of TrkB-FL and truncated isoforms (TrkB-T1 and TrkB-T2) on DIV8 cortical cultures treated with (black bars) or without (white bars-CTR) Aβ25–35 (25 µm) for 24 h. β-actin was used as an internal loading control. *P < 0.05 comparing with control (CTR) of the respective isoform (n = 8, student's t-test). (B) In left panel is a representative western blot of DIV8 neuronal cultures showing TrkB-FL receptor levels and the TrkB cleavage fragment (TrkB-ICD: ∼32 kDa) after 24 h of Aβ25–35 incubation. The primary antibody used recognizes the C-terminal of Trk-FL. Right panels show the average band intensity of TrkB-FL (upper histogram) and TrkB-ICD (lower histogram). **P < 0.01 comparing with control (n = 10, student's t-test). (C) Western blot using a pan-TrkB antibody (extracellular TrkB epitope), which recognizes simultaneously the FL (∼145 kDa) and the truncated TrkB species: TrkB-T′ (TrkB-FL cleavage product, ∼100 kDa) and TrkB-T1 (natural truncated TrkB originated by alternative splicing, ∼90 kDa). These bands were detected in neuronal cultures extracts prepared after the exposure to Aβ25–35 (25 µm) for 3, 8, 24, and 48 h or from control cultures (CTR) as indicated above each lane. All values presented in (A) and (B) are mean ± SEM.

Figure 1.

Aβ peptide affects TrkB receptors by selective up-regulation of truncated TrkB mRNA and by cleavage of TrkB-FL receptors. (A) Analysis of mRNA levels by relative qPCR of TrkB-FL and truncated isoforms (TrkB-T1 and TrkB-T2) on DIV8 cortical cultures treated with (black bars) or without (white bars-CTR) Aβ25–35 (25 µm) for 24 h. β-actin was used as an internal loading control. *P < 0.05 comparing with control (CTR) of the respective isoform (n = 8, student's t-test). (B) In left panel is a representative western blot of DIV8 neuronal cultures showing TrkB-FL receptor levels and the TrkB cleavage fragment (TrkB-ICD: ∼32 kDa) after 24 h of Aβ25–35 incubation. The primary antibody used recognizes the C-terminal of Trk-FL. Right panels show the average band intensity of TrkB-FL (upper histogram) and TrkB-ICD (lower histogram). **P < 0.01 comparing with control (n = 10, student's t-test). (C) Western blot using a pan-TrkB antibody (extracellular TrkB epitope), which recognizes simultaneously the FL (∼145 kDa) and the truncated TrkB species: TrkB-T′ (TrkB-FL cleavage product, ∼100 kDa) and TrkB-T1 (natural truncated TrkB originated by alternative splicing, ∼90 kDa). These bands were detected in neuronal cultures extracts prepared after the exposure to Aβ25–35 (25 µm) for 3, 8, 24, and 48 h or from control cultures (CTR) as indicated above each lane. All values presented in (A) and (B) are mean ± SEM.

Although the above-mentioned results clearly show that TrkB-FL mRNA levels were not significantly affected by Aβ25–35, a strong decrease of 40% ± 5% in TrkB-FL protein levels was observed in Aβ25–35-treated cells when compared with control (Fig. 1B, n = 10, P < 0.01), as we previously observed for both Aβ25–35 and Aβ1–42 (Kemppainen et al. 2012). Thus, in order to assess whether Aβ could also promote TrkB-FL cleavage, an antibody recognizing the intracellular C-terminal of TrkB-FL was used to detect a possible product of such cleavage. The results show that the decrease on TrkB-FL receptors, in Aβ25–35-treated cultures, is concomitant with the formation of a ∼32-kDa band (Fig. 1B, n = 10, P < 0.01 compared with control), indicating that Aβ induces a cleavage of TrkB-FL receptor, whereby it generates an intracellular domain (ICD) fragment (designated for now on as TrkB-ICD). Moreover, in cells treated with the Aβ1–42 (20 µm), there is also an increase in the formation of TrkB-ICD (Fig. 2C).

Figure 2.

Aβ-induced cleavage of TrkB-FL is mediated by the calcium-dependent calpain proteases. (A) Western blot of DIV8 neuronal cultures showing a concentration-dependent inhibition of Aβ25–35-induced TrkB-FL cleavage and TrkB fragment production by E64-d (a general thiol proteases inhibitor). (B) Upper image: representative western blot of DIV8 neuronal cultures showing the impact of several protease inhibitors on the Aβ-induced TrkB-FL cleavage and TrkB fragment production. The protease inhibitors tested were as follows: zVAD-FMK 20 μM (pan-caspase inhibitor); MG132 2 μm, ALLN 20 μm, MDL28170 20 μm (inhibitors of calpain-like activity), and Pepstatin A 1 μm (aspartyl protease inhibitor). All bands represented in the image are from the same gel, and the order of the first 2 lanes was rearranged. Lower panel: average immunoreactive band intensity of TrkB-FL, TrkB-Tc, and TrkB-ICD bands (upper, center, and lower histogram, respectively). The order of the histogram bars is the same as the above lanes of the western blot (**P < 0.01 compared with CTR, #P < 0.05 compared with Aβ25–35, n = 5, one-way ANOVA with Bonferroni's multiple comparison test). (C) Left image: representative western blot showing the effect of 24 h of Aβ25–35 (25 μm) and Aβ1–42 (20 μm) incubation on DIV8 neuronal cultures upon brain αII-Spectrin levels and the formation of the calpain-specific spectrin breakdown products (SBDPs 145/150), caspase-3-specific SBDP (120), and TrkB-ICD fragment. Right histogram: analysis of calpain-specific SBDPs (145/150) immunoreactive band intensity of control (white bar) and Aβ25–35 treatment (black bar), ***P < 0.001 comparing with control (n = 4, student's t-Test). (D) Left image: representative western blot showing the production of TrkB-ICD upon 5-mm CaCl2 incubation on cell lysates for 4 h at 25°C. Right histogram: analysis of TrkB-ICD immunoreactive band intensity (**P < 0.01, n = 3, ANOVA). (E) Purified m-calpain concentration-curve and consequent cleavage of rat TrkB with the production of TrkB-T′ and TrkB-ICD in neuronal cultures lysates (left) and in adult rat brain homogenates (right). (F) Western blot of postmortem human cortical sample, showing endogenous levels of TrkB and TrkB-ICD (in control condition) and depletion of TrkB-FL with concomitant formation of TrkB-ICD when both exogenous m-calpain and calcium were added in the absence of calpain inhibitor MDL28170. Ponceau S staining was used for loading control. All values presented in panels BD are mean ± SEM.

Figure 2.

Aβ-induced cleavage of TrkB-FL is mediated by the calcium-dependent calpain proteases. (A) Western blot of DIV8 neuronal cultures showing a concentration-dependent inhibition of Aβ25–35-induced TrkB-FL cleavage and TrkB fragment production by E64-d (a general thiol proteases inhibitor). (B) Upper image: representative western blot of DIV8 neuronal cultures showing the impact of several protease inhibitors on the Aβ-induced TrkB-FL cleavage and TrkB fragment production. The protease inhibitors tested were as follows: zVAD-FMK 20 μM (pan-caspase inhibitor); MG132 2 μm, ALLN 20 μm, MDL28170 20 μm (inhibitors of calpain-like activity), and Pepstatin A 1 μm (aspartyl protease inhibitor). All bands represented in the image are from the same gel, and the order of the first 2 lanes was rearranged. Lower panel: average immunoreactive band intensity of TrkB-FL, TrkB-Tc, and TrkB-ICD bands (upper, center, and lower histogram, respectively). The order of the histogram bars is the same as the above lanes of the western blot (**P < 0.01 compared with CTR, #P < 0.05 compared with Aβ25–35, n = 5, one-way ANOVA with Bonferroni's multiple comparison test). (C) Left image: representative western blot showing the effect of 24 h of Aβ25–35 (25 μm) and Aβ1–42 (20 μm) incubation on DIV8 neuronal cultures upon brain αII-Spectrin levels and the formation of the calpain-specific spectrin breakdown products (SBDPs 145/150), caspase-3-specific SBDP (120), and TrkB-ICD fragment. Right histogram: analysis of calpain-specific SBDPs (145/150) immunoreactive band intensity of control (white bar) and Aβ25–35 treatment (black bar), ***P < 0.001 comparing with control (n = 4, student's t-Test). (D) Left image: representative western blot showing the production of TrkB-ICD upon 5-mm CaCl2 incubation on cell lysates for 4 h at 25°C. Right histogram: analysis of TrkB-ICD immunoreactive band intensity (**P < 0.01, n = 3, ANOVA). (E) Purified m-calpain concentration-curve and consequent cleavage of rat TrkB with the production of TrkB-T′ and TrkB-ICD in neuronal cultures lysates (left) and in adult rat brain homogenates (right). (F) Western blot of postmortem human cortical sample, showing endogenous levels of TrkB and TrkB-ICD (in control condition) and depletion of TrkB-FL with concomitant formation of TrkB-ICD when both exogenous m-calpain and calcium were added in the absence of calpain inhibitor MDL28170. Ponceau S staining was used for loading control. All values presented in panels BD are mean ± SEM.

Given that the cytosolic domain of rat TrkB-FL (starting at Lys454 until Gly821) has a predicted molecular weight of 41.6 kDa, and since TrkB-ICD fragment migrates in SDS–PAGE with a relative molecular weight of ∼32 kDa, the cleavage site might be located ∼10 kDa downstream of the transmembrane domain of the receptor. We thus anticipated that the Aβ-induced cleavage would lead to the generation of a new membrane-bound truncated TrkB receptor ∼10 kDa heavier than the natural truncated TrkB-T1 (which lacks the whole ICD). To directly evaluate this possibility, we used a pan-TrkB antibody that recognizes an extracellular epitope and we increased the separation in SDS–PAGE electrophoresis, which allowed us to identify 2 distinct truncated TrkB bands: 1 broad band at ∼90 kDa corresponding to the natural truncated TrkB-T1 receptor, and another broad band around ∼100-kDa band corresponding to the new truncated receptor produced by the cleavage of TrkB-FL (henceforth designated as TrkB-T′, Fig. 1C). The TrkB-T′ levels were very low in control neuronal cultures (Figs 1C and 2E), being also negligible in control rat brain homogenates (Fig. 2E), indicating that this fragment is only formed under conditions that trigger robust cleavage of TrkB-FL receptor.

Taken together, the data show that Aβ induces a selective up-regulation of truncated TrkB-T1 and T2 transcripts, whereas it simultaneously promotes TrkB-FL protein cleavage, thus producing a new truncated receptor (TrkB-T′) and an intracellular fragment containing the C-terminal of the receptor (TrkB-ICD). The natural truncated receptors produced by alternative splicing (TrkB-T1 and -T2) and the cleavage-generated truncated receptor (TrkB-T′) all contribute to the total pool of truncated TrkB receptors, which will be referred in this work as TrkB-Tc.

Calpain Mediates the Cleavage of TrkB-FL Induced by Aβ

The next series of experiments were designed to identify the enzyme involved in the cleavage of TrkB-FL by Aβ. The cell-permeable thiol proteases inhibitor, E-64d, caused a concentration-dependent inhibition of the TrkB-FL cleavage induced by Aβ25–35, with a maximal effect achieved at the concentration of 100 µm (Fig. 2A). To identify which thiol proteases were involved in the cleavage, several thiol inhibitors were tested, including the inhibitors of proteases with calpain-like activity, N-acetyl-Leu-Leu-norleucinal (ALLN 20 µm), N-acetyl-Leu-Leu-methional (MG132 2 µm), MDL28170 (20 µm), and the cell-permeable pan-caspases inhibitor (z-VAD-FMK 20 µm). In addition, a potent aspartyl protease inhibitor, pepstatin A (1 µm), was tested. Neither the caspase inhibitor zVAD-FMK nor pepstatin A mitigated Aβ-induced cleavage of TrkB-FL receptors (Fig. 2B). Conversely, the inhibitors of calpain-like activity ALLN, MG132, and MDL28170 significantly prevented Aβ-induced cleavage of TrkB-FL and the subsequent formation of TrkB-T′ and TrkB-ICD fragments (Fig. 2B, n = 5, P < 0.01). In contrary to what was observed for protein levels, the inhibition of calpains by MDL28170 did not affect the changes induced by Aβ25–35 on mRNA levels of TrkB receptors (Supplementary Fig. 1C, n = 4).

In parallel, we observed that the exposure of neuronal cultures to Aβ25–35 resulted in a strong activation of calpain, as confirmed by the 6-fold increase (Fig. 2C, n = 4, P < 0.001) in the formation of the calpain-specific αII-Spectrin breakdown products (SBDP 150/145 kDa), a standard assay for monitoring calpain activity. The Aβ1–42 peptide (20 µm) also induced robust calpain activation on neuronal cultures (see Fig. 2C). In contrast, the inverted peptide Aβ35–25 (25 µm) did not induce calpain activation and TrkB cleavage on neuronal cultures (Supplementary Fig. 2A).

Since calpains are calcium-dependent proteases, we next evaluated whether the activation of endogenous calpains by calcium would induce TrkB-FL cleavage by itself, in the absence of Aβ. Therefore, cell lysates of neuronal cultures were incubated with 5 mm of CaCl2 for 4 h at 25°C and, in these conditions, the characteristic TrkB-ICD band was detected, an effect fully blocked by the calpain inhibitor MDL28170 (Fig. 2D, n = 3, P < 0.05). This calcium-induced calpain-mediated cleavage of TrkB-FL was also detected in isolated nerve terminals (synaptosomes) prepared from adult rat hippocampus (Supplementary Fig. 2B). In addition, the cleavage of TrkB-FL and subsequent production of TrkB cleavage fragments was observed following the incubation of neuronal cell lysates or cortical homogenates from adult rat with purified recombinant rat m-calpain in the presence of 2-mm calcium (∼100-kDa TrkB-T′ and ∼32-kDa TrkB-ICD; Fig. 2E). To confirm the specificity of the fragments detected, the TrkB-FL receptors were immunopurified from neuronal cultures and incubated with purified rat m-calpain. The results showed that rat m-calpain cleaved the immunopurified receptors producing the characteristic TrkB-T′ and TrkB-ICD fragments, allowing to conclude that the fragments detected arise specifically from TrkB-FL receptors (Supplementary Fig. 2C).

Human TrkB-FL and rat TrkB-FL share the same amino acid sequence in the region of the calpain-cleavage site (Fig. 3B). Thus, it is expected that human TrkB-FL could also be cleaved by human calpains. Indeed, basal levels of TrkB-ICD were detected in a parietal cortex homogenate from a human control case (Fig. 2F). Addition of purified m-calpain completely cleaved human TrkB-FL, further enhancing the levels of TrkB-ICD fragment (Fig. 2F). These data clearly showed that human TrkB-FL is also prone to be cleaved by calpains, leading to the production of TrkB-ICD.

Figure 3.

Calpain cleaves TrkB downstream of Shc-binding site (Tyr515). (A) Ponceau S staining of a PVDF membrane after transfer from SDS–PAGE. Purified cytosolic domain of human TrkB (TrkB active aa.450-end, Millipore) was digested with purified m-calpain, producing TrkB-ICD. TrkB-ICD band was cut (dashed square) and submitted for N-terminal sequencing (Edman degradation). CaCl2 (2 mm) was present in all conditions. (B) Multiple alignment of TrkA, TrkB, and TrkC protein sequences for different species. The Shc-binding motif, the Trk kinase domain, and the calpain-cleavage site identified by N-sequencing are identified in the sequence. Protein sequences were obtained in UniProtKD, and the multiple alignments were performed using the Clustal Omega tool. The proline residue (P) present in the TrkB sequence (*) was not detected by N-sequencing. (C) Schematic representation of mature rat TrkB-FL, TrkB-T1, T2, and human TrkB-T-Shc isoforms showing the relevant domains and amino acid residues positions. Calpain-cleavage site (aa. 520) is represented based on N-sequencing data. Y515 and Y816 represent the phospho-tyrosine residues able to recruit Shc and PLCγ, respectively. Protein sequences obtained from UniProtKD (accession number: Q63604 for Rat and Q16620 for Human). (D) Left: representative western blot showing phosphorylated Tyr515 of TrkB-FL (upper image), pan-TrkB (middle image), and α-tubulin (lower image) on DIV8 neurons incubated firstly with (or without) 24 h of Aβ25–35 (25 μm) and with (or without) 10 min of BDNF (20 ng/mL). Right panel: levels of phosphorylated TrkB-FL (Tyr515) normalized for α-tubulin (upper histogram) and ratio between phosphorylated and total TrkB-FL levels (lower graph) for control (white bars) and Aβ25–35 treatment (black bars). Values are mean ± SEM (**P < 0.01, n = 4, student's t-Test, compared with CTR). (E) Western blot showing phosphorylated Tyr515 of TrkB-FL, pan-TrkB, TrkB-ICD, and α-tubulin on a DIV 15 cortical culture incubated firstly with (or without) Aβ25–35 (25 μm, 24 h) and then incubated with (or without) BDNF (20 ng/mL) for 10 min.

Figure 3.

Calpain cleaves TrkB downstream of Shc-binding site (Tyr515). (A) Ponceau S staining of a PVDF membrane after transfer from SDS–PAGE. Purified cytosolic domain of human TrkB (TrkB active aa.450-end, Millipore) was digested with purified m-calpain, producing TrkB-ICD. TrkB-ICD band was cut (dashed square) and submitted for N-terminal sequencing (Edman degradation). CaCl2 (2 mm) was present in all conditions. (B) Multiple alignment of TrkA, TrkB, and TrkC protein sequences for different species. The Shc-binding motif, the Trk kinase domain, and the calpain-cleavage site identified by N-sequencing are identified in the sequence. Protein sequences were obtained in UniProtKD, and the multiple alignments were performed using the Clustal Omega tool. The proline residue (P) present in the TrkB sequence (*) was not detected by N-sequencing. (C) Schematic representation of mature rat TrkB-FL, TrkB-T1, T2, and human TrkB-T-Shc isoforms showing the relevant domains and amino acid residues positions. Calpain-cleavage site (aa. 520) is represented based on N-sequencing data. Y515 and Y816 represent the phospho-tyrosine residues able to recruit Shc and PLCγ, respectively. Protein sequences obtained from UniProtKD (accession number: Q63604 for Rat and Q16620 for Human). (D) Left: representative western blot showing phosphorylated Tyr515 of TrkB-FL (upper image), pan-TrkB (middle image), and α-tubulin (lower image) on DIV8 neurons incubated firstly with (or without) 24 h of Aβ25–35 (25 μm) and with (or without) 10 min of BDNF (20 ng/mL). Right panel: levels of phosphorylated TrkB-FL (Tyr515) normalized for α-tubulin (upper histogram) and ratio between phosphorylated and total TrkB-FL levels (lower graph) for control (white bars) and Aβ25–35 treatment (black bars). Values are mean ± SEM (**P < 0.01, n = 4, student's t-Test, compared with CTR). (E) Western blot showing phosphorylated Tyr515 of TrkB-FL, pan-TrkB, TrkB-ICD, and α-tubulin on a DIV 15 cortical culture incubated firstly with (or without) Aβ25–35 (25 μm, 24 h) and then incubated with (or without) BDNF (20 ng/mL) for 10 min.

Altogether, the present work supports that Aβ leads to TrkB-FL cleavage through the activation of calpains.

Calpain Cleavage Site of TrkB-FL is Located Downstream of the Shc-Binding Site

Given the molecular weight (∼32 kDa) of the TrkB-ICD band detected by western blot, we hypothesized that the calpain-cleavage site would be located close to the Shc-binding site (Tyr515). To clarify this possibility, 5 µg of recombinant cytosolic domain of TrkB-FL (Human TrkB active, aa.455-end, Millipore) were digested by purified m-calpain. Following membrane staining after SDS–PAGE electrophoresis, we observed that the recombinant cytosolic domain of human TrkB-FL (∼42 kDa) was cleaved by m-calpain producing the same characteristic ∼32-kDa TrkB-ICD fragment band (Fig. 3A), as detected in neuronal cultures exposed to Aβ (Fig. 1B). The TrkB-ICD band was then cut (as depicted in Fig. 3A) and analyzed by N-terminal sequencing (Edman degradation) in order to identify the first 5 N-terminal amino acids, hence revealing the calpain-cleavage site position. The 5 N-terminal amino acids detected were Ser-Gln-Leu-Lys-Asp (S-Q-L-K-D), which allows to conclude that the TrkB-FL is cleaved between the Asn(N)520 and Ser(S)521 residues, considering the rat TrkB sequence (Fig. 3B). This cleavage site is located between the Shc-binding site (Tyr515) and the TrkB kinase domain (Fig. 3B,C). Therefore, these data indicate that the truncated TrkB-T′ receptor contains the Shc-binding site (Tyr515), whereas the TrkB-ICD fragment contains the complete tyrosine kinase domain of TrkB-FL receptor (Ile537-Leu806), as well as the C-terminal tail of TrkB (Gln807-Gly821), since the fragment is recognized by the antibody specific for the C-terminal tail of Trk-FL, as shown above (Fig. 1B).

Upon BDNF binding to TrkB-FL receptor, both Tyr515 and Tyr816 residues of the receptor are phosphorylated, allowing the binding of Shc adaptor protein and Phospholipase C-γ (PLCγ), respectively, with subsequent activation of signaling cascades (see Kaplan and Miller 2000). By using specific antibodies against phosphorylated Tyr515 and Tyr816 of TrkB, we evaluated whether TrkB-T′ or TrkB-ICD fragments could undergo phosphorylation on Tyr515 or Tyr816, respectively. In a first attempt, neurons were incubated with Aβ for 24 h to produce the TrkB fragments, and then, BDNF (20 ng/mL) was briefly applied (10 min) to induce TrkB phosphorylation. Whereas BDNF incubation induced robust phosphorylation in Tyr515 of TrkB-FL, such phosphorylation was not detected in the truncated TrkB-T′ fragment (Fig. 3D). The levels of phosphorylated (Tyr515) TrkB-FL upon BDNF exposure were 40% ± 9% lower in the Aβ-treated cultures than in control cultures (Fig. 3D, n = 4, P < 0.01), a reduction similar to that observed in total levels of TrkB-FL induced by Aβ (Fig. 1B). Thus, the ratio between the levels of total TrkB-FL and BDNF-induced phosphorylated TrkB-FL does not differ when comparing control cultures with Aβ-treated cultures (Fig. 3D, n = 4), suggesting that TrkB-FL phosphorylation efficacy remain similar regardless Aβ treatment. A similar result was also obtain in more mature cortical cultures with 15 DIV (Fig. 3E).

We further tested whether TrkB-ICD fragment could be phosphorylated upon BDNF incubation on the C-terminal located Tyr816. Similarly to the Tyr515, BDNF incubation promoted a robust phosphorylation of Tyr816 on the TrkB-FL, but not on TrkB-ICD fragment (Supplementary Fig. 3). Together, the above data show that the calpain-cleavage site of TrkB-FL is located downstream of the Shc-binding site and that calpain-generated TrkB-T′ and TrkB-ICD fragments are not phosphorylated.

Calpain Mediates Detrimental Effects of Aβ upon BDNF Actions

The influence of Aβ upon the modulatory action of BDNF on glutamate and GABA release was evaluated on synaptosomes, which were prepared from rat hippocampal slices pretreated with Aβ25–35 (25 µm) or Aβ1–42 (20 µm) for 3 h. The hippocampal synaptosomes were loaded with [3H]GABA or [3H]glutamate as previously described (Sousa et al. 2011), and neurotransmitter release was evoked twice (S1 and S2) by perfusion with 15 mm KCl for 2 min. In [3H]GABA release assays, the S2/S1 ratio was, in control conditions, 1.06 ± 0.03 (n = 7, Supplementary Fig. 4B), and it was significantly decreased to 0.82 ± 0.07 (n = 7, P < 0.05, Supplementary Fig. 4B) when BDNF (20 ng/mL) was added before S2, corresponding to a decrease of 26% ± 6% (n = 7, Fig. 4E and Supplementary Fig. 4B, P < 0.05) in the evoked release of GABA. In [3H]glutamate release assays, the S2/S1 ratio in control conditions was 0.75 ± 0.03 (n = 5, Supplementary Fig. 4A) and it was increased up to 0.93 ± 0.04 (n = 5, P < 0.05, Supplementary Fig. 4A) when BDNF (20 ng/mL) was added before S2, corresponding to an enhancement of 27% ± 7% (n = 5, Fig. 4F and Supplementary Fig. 4A, P < 0.05) in the evoked release of glutamate. BDNF-induced inhibition of GABA and facilitation of glutamate release from hippocampal synaptosomes was expected based on our previous studies (Canas et al. 2004). When synaptosomes were prepared from hippocampal slices that had been exposed for 3 h to Aβ25–35 (25 µm) or to Aβ1–42 (20 µm), the S2/S1 ratios in GABA release assays (S2/S1Aβ25–35: 1.04 ± 0.06, n = 13; S2/S1Aβ1–42: 1.16 ± 0.02, n = 3, Fig. 4A,E and Supplementary Fig. 4D) or glutamate release assays (S2/S1: 0.85 ± 0.08, n = 10; S2/S1Aβ1–42: 0.85 ± 0.03, n = 3, Fig. 4B,F and Supplementary Fig. 4C) were not significantly altered (P >0.05) as compared with control conditions. However, in synaptosomes prepared from Aβ25–35- and Aβ1–42-treated slices, BDNF lost its ability to decrease GABA release (S2/S1Aβ25–35+BDNF 0.96 ± 0.06, n = 13; S2/S1Aβ1–42+BDNF: 0.95 ± 0.17, n = 3 Fig. 4A,E and Supplementary Fig. 4D) and to increase glutamate release (S2/S1Aβ25–35+BDNF: 0.78 ± 0.07, n = 10; S2/S1Aβ1–42+BDNF: 0.86 ± 0.07, n = 3 Fig. 4B,F and Supplementary Fig. 4C). These results strongly suggest that Aβ causes a functional impairment of BDNF modulatory actions upon glutamate and GABA release in the hippocampus. In order to determine the contribution of calpains towards the Aβ-dependent impairment of BDNF modulation of neurotransmitters release, hippocampal slices were incubated simultaneously with both Aβ25–35 (25 µm) and MDL28170 (20 µm) for 3 h. In the absence of BDNF, MDL28170 did not impact S2/S1 ratios for both GABA (S2/S1Aβ25–35+MDL: 1.13 ± 0.05, n = 4, Fig. 4C,E) and glutamate release (S2/S1Aβ25–35+MDL: 0.78 ± 0.07, n = 4, Fig. 4D,F). On the contrary, in synaptosomes prepared from hippocampal slices incubated under similar conditions (Aβ25–35 and MDL28170), the addition of BDNF before S2 affected the S2/S1 ratio of GABA (S2/S1Aβ25–35+MDL+BDNF: 0.78 ± 0.15, P < 0.05 vs. S2/S1Aβ25–35+MDL, n = 4, Fig. 4C,E) and glutamate (S2/S1Aβ25–35+MDL+BDNF 0.96 ± 0.08, P < 0.05, vs. S2/S1Aβ25–35+MDL, n = 4, Fig. 4D,F) release similar to what had been observed when BDNF was added alone (Fig. 4D,F and Supplementary Fig. 4A,B). The incubation of hippocampal slices with MDL28170 alone for 3 h did not affect BDNF actions upon glutamate and GABA release (for glutamate: S2/S1MDL+BDNF: 0.95 ± 0.01, n = 3, Supplementary Fig. 4E; for GABA: S2/S1MDL+BDNF: 0.69 ± 0.28, n = 3, Supplementary Fig. 4F). These results strongly suggest that the impairment caused by Aβ upon the modulatory action of BDNF on neurotransmitter release is rescued by calpain inhibition.

Figure 4.

Aβ inhibits BDNF effect upon GABA and glutamate release from hippocampal synaptosomes in a calpain-dependent way. Fractional release of [3H] GABA (A and C) and [3H] glutamate (B and D) evoked by two 15-mm K+ stimuli of 2-min duration, at 5–7 min (S1) and 29–31 min (S2). BDNF (30 ng/mL) was added at 9 min and remained in the perfusion solution until the end of the experiments (closed circles), as indicated by the horizontal bar. Control curves in the presence of Aβ25–35 (25 μm) or Aβ25–35 (25 μm) and MDL28170 (20 μm), performed in parallel with the same synaptosomal batch, are represented by open circles. Modulation of BDNF (30 ng/mL) effect upon fractional release of [3H] GABA by Aβ25–35 (25 μm) (A) or by simultaneous treatment with Aβ25–35 (25 μm) and MDL28170 (20 μm) (C). Modulation of BDNF (30 ng/mL) effect upon fractional release of [3H]glutamate by Aβ25–35 (25 μm) (B) or by simultaneous treatment with Aβ25–35 (25 μm) and MDL28170 (20 μm) (D). E and F, S2/S1 ratios (%), calculated in each experiment from the fractional release curves, as described in Materials and Methods. BDNF (30 ng/mL) was tested in synaptosomes prepared from hippocampal slices treated or non-treated with Aβ25–35 (25 μm) or Aβ1–42 (20 μm) or treated simultaneously with Aβ25–35 (25 μm) and MDL28170 (20 μm), as indicated below each bar. In each experiment, the S2/S1 ratio obtained while BDNF was present during S2 was normalized, taking as 100% the S2/S1 ratio obtained in parallel chambers under the same drug conditions but in the absence of BDNF. The decline in the baseline in the glutamate release experiments results from glutamate incorporation in metabolic pathways, such as TCA cycle, resulting in a decrease in the availability for vesicular incorporation and release (Bak et al. 2006). Data are represented as mean ± SEM of 5–10 independent experiments. *P < 0.05, compared with 100%, except when otherwise indicated (one-way ANOVA followed by Bonferroni's multiple comparison test).

Figure 4.

Aβ inhibits BDNF effect upon GABA and glutamate release from hippocampal synaptosomes in a calpain-dependent way. Fractional release of [3H] GABA (A and C) and [3H] glutamate (B and D) evoked by two 15-mm K+ stimuli of 2-min duration, at 5–7 min (S1) and 29–31 min (S2). BDNF (30 ng/mL) was added at 9 min and remained in the perfusion solution until the end of the experiments (closed circles), as indicated by the horizontal bar. Control curves in the presence of Aβ25–35 (25 μm) or Aβ25–35 (25 μm) and MDL28170 (20 μm), performed in parallel with the same synaptosomal batch, are represented by open circles. Modulation of BDNF (30 ng/mL) effect upon fractional release of [3H] GABA by Aβ25–35 (25 μm) (A) or by simultaneous treatment with Aβ25–35 (25 μm) and MDL28170 (20 μm) (C). Modulation of BDNF (30 ng/mL) effect upon fractional release of [3H]glutamate by Aβ25–35 (25 μm) (B) or by simultaneous treatment with Aβ25–35 (25 μm) and MDL28170 (20 μm) (D). E and F, S2/S1 ratios (%), calculated in each experiment from the fractional release curves, as described in Materials and Methods. BDNF (30 ng/mL) was tested in synaptosomes prepared from hippocampal slices treated or non-treated with Aβ25–35 (25 μm) or Aβ1–42 (20 μm) or treated simultaneously with Aβ25–35 (25 μm) and MDL28170 (20 μm), as indicated below each bar. In each experiment, the S2/S1 ratio obtained while BDNF was present during S2 was normalized, taking as 100% the S2/S1 ratio obtained in parallel chambers under the same drug conditions but in the absence of BDNF. The decline in the baseline in the glutamate release experiments results from glutamate incorporation in metabolic pathways, such as TCA cycle, resulting in a decrease in the availability for vesicular incorporation and release (Bak et al. 2006). Data are represented as mean ± SEM of 5–10 independent experiments. *P < 0.05, compared with 100%, except when otherwise indicated (one-way ANOVA followed by Bonferroni's multiple comparison test).

BDNF has a well-documented ability to increase LTP on hippocampal CA1 area through TrkB-FL activation (Korte et al. 1995; Figurov et al. 1996). To evaluate the impact of Aβ upon BDNF effects on LTP, hippocampal slices were exposed for 3 h to oxygenated aCSF with or without Aβ25–35 (25 µm) or Aβ1–40 (20 µm) or even the inverted Aβ35–25 (25 µm) peptide. As mentioned in Materials and Methods section, the experiments were conducted using 2 independent stimulation pathways, being each pathway used as control or test in alternate days, in order to compare LTP magnitude in the absence and in the presence of BDNF, within the same slice. The LTP was firstly induced by θ-burst stimulation in 1 pathway and its magnitude quantified 46-60 min after LTP induction. BDNF was then added to the perfusing aCSF and allowed to equilibrate for at least 20 min before inducing LTP in the second pathway.

As expected (Fontinha et al. 2008), the θ-burst stimulus applied in the presence of BDNF (20 ng/mL) induced a robust LTP (LTPBDNF: 40.0% ± 1.7% increase in fEPSP slope), which was significantly higher (P < 0.01) than that obtained in the absence of BDNF (LTPCTR: 22.1% ± 3.9% increase in fEPSP slope; n = 11, Fig. 5A,E). Pretreatment of hippocampal slices with Aβ for 3 h did not affect LTP magnitude when compared with untreated slices (LTPAβ25–35: 23.5% ± 3.5%, n = 10 or LTPAβ1–42: 17.6% ± 7%, n = 5 vs. LTPCTR: 22.1% ± 3.9%, n = 11, Fig. 5E). However, in Aβ-treated slices, BDNF (20 ng/mL) failed to enhance LTP magnitude (LTPAβ25–35: 23.5% ± 3.5% vs. LTPAβ25–35+BDNF: 24.3% ± 4.7%, n = 10, P >0.05, Fig. 5B,E and LTPAβ1–42: 17.6% ± 7% vs. LTPAβ1–42+BDNF: 21.4% ± 4.0%, n = 5, P >0.05, Fig. 5C,E). In slices treated with 25 µm of inverted Aβ35–25 (control peptide), the facilitation of BDNF upon LTP was not lost (n = 4, P < 0.05; Supplementary Fig. 5B). To evaluate whether Aβ peptides could affect basal synaptic efficiency, input/output curves were performed and no significant differences were detected between control slices, and Aβ25–35- or Aβ1–42-treated slices (n = 4, Fig. 5F).

Figure 5.

Aβ decreases the facilitatory effect of BDNF upon θ-burst-induced LTP in a calpain-dependent way. Panels AD show the averaged time courses changes in field excitatory post-synaptic potential (fEPSP) slope induced by a θ-burst stimulation in the absence (open circle) or in the presence of BDNF of 20 ng/mL (filled circle) in the second stimulation pathway in rat hippocampal slices without (A, n = 11) or with a pre-exposure for 3 h to aCSF solution containing 25 μm25–35 (B, n = 10), 20 μm1–42 (C, n = 5), or 25 μm25–35 in the presence of 20 μm MDL28170 (D, n = 6). The traces from representative experiments are shown below panels AD; each trace is the average of 8 consecutive responses obtained before (1 and 3) and 46–60 min after (2 and 4) LTP induction. The traces are composed by the stimulus artifact, followed by the pre-synaptic volley and the fEPSP. The traces (1 and 2) and traces (3 and 4) were obtained in the absence and presence of BDNF, respectively. (E) LTP magnitude (change in fEPSP slope at 46–60 min) induced by θ-burst stimulation in relation to pre-θ-burst values (0%) for each group of pretreated slices (control, Aβ25–35, Aβ25–35 + MDL28170, and Aβ1–42). *P < 0.05; **P < 0.01, one-way ANOVA followed by Bonferroni's multiple comparison test. Values are mean ± SEM. (F) Input/output curves corresponding to fEPSP slope evoked by various stimulation intensities (100–360 μA) in non-treated hippocampal slices or treated for 3 h with Aβ25–35 (25 μm) or Aβ1–42 (20 μm) (n = 4).

Figure 5.

Aβ decreases the facilitatory effect of BDNF upon θ-burst-induced LTP in a calpain-dependent way. Panels AD show the averaged time courses changes in field excitatory post-synaptic potential (fEPSP) slope induced by a θ-burst stimulation in the absence (open circle) or in the presence of BDNF of 20 ng/mL (filled circle) in the second stimulation pathway in rat hippocampal slices without (A, n = 11) or with a pre-exposure for 3 h to aCSF solution containing 25 μm25–35 (B, n = 10), 20 μm1–42 (C, n = 5), or 25 μm25–35 in the presence of 20 μm MDL28170 (D, n = 6). The traces from representative experiments are shown below panels AD; each trace is the average of 8 consecutive responses obtained before (1 and 3) and 46–60 min after (2 and 4) LTP induction. The traces are composed by the stimulus artifact, followed by the pre-synaptic volley and the fEPSP. The traces (1 and 2) and traces (3 and 4) were obtained in the absence and presence of BDNF, respectively. (E) LTP magnitude (change in fEPSP slope at 46–60 min) induced by θ-burst stimulation in relation to pre-θ-burst values (0%) for each group of pretreated slices (control, Aβ25–35, Aβ25–35 + MDL28170, and Aβ1–42). *P < 0.05; **P < 0.01, one-way ANOVA followed by Bonferroni's multiple comparison test. Values are mean ± SEM. (F) Input/output curves corresponding to fEPSP slope evoked by various stimulation intensities (100–360 μA) in non-treated hippocampal slices or treated for 3 h with Aβ25–35 (25 μm) or Aβ1–42 (20 μm) (n = 4).

To explore whether calpains played a role in the Aβ-induced loss of BDNF effect upon LTP, hippocampal slices were pretreated simultaneously with the Aβ25–35 and the calpain inhibitor MDL28170 (20 µm), for 3 h. As shown in Fig. 5D, pretreatment with MDL28170 rescued the facilitatory effect of BDNF upon LTP (LTPAβ25–35+MDL: 15.6% ± 3.9% vs. LTPAβ25–35+MDL+BDNF: 32.5% ± 3.3%, n = 6, P < 0.05, Fig. 5D,E). MDL28170 by itself did not significantly affect LTP magnitude in slices treated with Aβ25–35 (Fig. 5D).

Taken together, these data demonstrate that Aβ severely hampers BDNF action on hippocampal LTP and neurotransmitter (GABA and glutamate) release and that these impairments are dependent on calpain activation. These functional results correlate with the results obtained in neuronal cultures treated with Aβ showing a calpain-mediated cleavage of the TrkB-FL BDNF receptor. Therefore, the data strongly suggest that Aβ impairs BDNF/TrkB-mediated actions in hippocampal slices through a mechanism that involves calpain activation.

Discussion

In the present work, we demonstrate that Aβ promotes a calpain-mediated cleavage of TrkB-FL receptor and impairs, in a calpain-dependent manner, BDNF modulation of neurotransmitter release, and synaptic plasticity. Moreover, we found that, in primary cortical cultures, Aβ significantly increases mRNA levels of truncated TrkB isoforms without affecting mRNA levels of TrkB-FL, in a mechanism independent of calpains.

The alterations at the transcriptional level, namely the increase in mRNA levels of truncated TrkB isoforms, are in line with previous data showing a correlation between amyloid-load and up-regulation of TrkB-T1 mRNA levels in cortical regions of a transgenic AD mice model without affecting TrkB-FL mRNA levels (Kemppainen et al. 2012). A selective up-regulation of truncated TrkB mRNA levels (TrkB-T1 and TrkB-T-Shc), without changes in TrkB-FL mRNA levels, was also reported in the hippocampus of AD postmortem human brain (Wong et al. 2012). As we now clearly show, the influence of Aβ upon TrkB-FL occurs at the post-translational level, rather than at the transcriptional level since Aβ strongly reduces TrkB-FL protein levels. Remarkably this occurs through calpain-mediated cleavage of TrkB-FL receptor protein, leading to a truncated receptor with a different molecular weight than the known isoforms of the truncated TrkB receptors, which we named as TrkB-T′. Interestingly, Aβ did not affect the proportion of BDNF-induced phosphorylated TrkB-FL over total TrkB-FL levels, in accordance with previous data showing that sub-lethal Aβ concentrations do not interfere with BDNF-induced phosphorylation of TrkB-FL (Tong et al. 2004). Thus, Aβ may impair BDNF signaling through a decrease in the TrkB-FL/TrkB-Tc ratio (Tc referring to all isoforms of truncated receptors), rather than by affecting the phosphorylation of the remaining TrkB-FL. In addition, Aβ could also affect downstream mediators of TrkB signaling, such as the docking proteins of TrkB, as already described for sub-lethal concentrations of Aβ (Tong et al. 2004).

By performing a detailed characterization of TrkB-FL cleavage, we could show that calpain cleavage of TrkB-FL occurs between the Asn520 and Ser521 residues, producing 2 TrkB cleavage products: (1) the new truncated TrkB receptor (TrkB-T′), which is heavier than the natural truncated TrkB-T1 and -T2 splicing products and (2) a fragment of ∼32 kDa, which corresponds to the ICD of TrkB-FL (TrkB-ICD). Moreover, we found that the calpain-cleavage site of TrkB-FL is located downstream of the Shc-binding site (Tyr515), indicating that the new truncated TrkB-T′ generated by calpain contains the Shc-binding site (Tyr515). Thus, the TrkB-T′ is only 16 amino acid residues shorter than the described truncated TrkB-T-Shc isoform, a neuron-specific alternative splicing product of human TrkB gene (Stoilov et al. 2002). Although TrkB-T-Shc function is poorly understood, it cannot be tyrosine phosphorylated (Stoilov et al. 2002), as the now described TrkB-T′, and it could act as a negative regulator of BDNF function. Our results suggest that TrkB-T′ may not act as a negative modulator of BDNF signaling since, in spite of its presence, the efficiency of TrkB-FL phosphorylation by BDNF, assessed as the ratio between pTrkB-FL/TrkB-FL, was not appreciably affected.

The second Aβ-induced TrkB-FL cleavage product, the TrkB-ICD, corresponds to the remaining ICD of TrkB-FL upstream of the cleavage site (Ser521-Gly821). The theoretical molecular weight of TrkB-ICD is 34.6 kDa, which is similar to the relative molecular weight observed in SDS–PAGE (∼32 kDa). TrkB-ICD was also detected in post-mortem human brain samples showing that human endogenous calpains could also cleave human TrkB-FL receptor. However, we cannot exclude the possibility that the presence of basal levels of TrkB-ICD detected in the human brain sample could be exacerbated due to calpain activation during the postmortem period. Nevertheless, in freshly prepared rat cortex homogenates, and therefore without a significant postmortem delay, it was also possible to detect small amounts of TrkB-ICD, suggesting that TrkB-FL cleavage could also occur in the alive healthy brain.

It is known that some members of the receptor tyrosine kinase family can undergo proteolytic cleavage by caspases, metalloproteases or secretases, producing ICD fragments that may possess a biological function (see Ancot et al. 2009), as it is the case of the pro-apoptotic fragment that results from caspase-mediated TrkC cleavage upon NT-3 deprivation (Tauszig-Delamasure et al. 2007). Calpain-mediated proteolytic cleavage usually occurs between 2 domains of the substrate, releasing big stable fragments that can also have biological activity (see Goll et al. 2003). As an example, the calpain-generated fragment p25 from p35 cleavage constitutively activates the cyclin-dependent kinase 5 (CDK5), contributing to tau hyperphosphorylation, morphological degeneration, and neuronal death (Patrick et al. 1999). Indeed, overexpression of p25 fragment in mice forebrain is sufficient to recapitulate the major hallmarks of AD, including hippocampal neuronal loss, aggregation of hyperphosphorylated tau, accumulation of Aβ, and impairments on synaptic plasticity and cognition (Cruz et al. 2003; Fischer et al. 2005). Therefore, one may propose that calpains are a key element of a vicious cycle, since its activation leads to β-amyloid generation (Cruz and Tsai 2004), which in turn leads to enhanced calpain activity with subsequent impaired signaling of key neurotrophic molecules such as BDNF (present work). Whether the resulting fragment, TrkB-ICD, also contributes to exacerbate neuronal damage, awaits further investigation.

The present work clearly demonstrates that Aβ impairs the facilitatory effects of BDNF upon glutamate release and the inhibitory effect of BDNF upon GABA release from isolated nerve terminals (synaptosomes). GABAergic and Glutamatergic hippocampal synaptosomes represent each one ∼40% of total synaptophysin-positive nerve terminals, <5% of nerve terminals being cholinergic (Canas et al. 2014). Interestingly, in spite of the greater vulnerability of the glutamatergic terminals to Aβ toxicity, as compared with the GABAergic ones (Bell et al. 2003; Canas et al. 2014), the effect of BDNF upon GABA release was also impaired by Aβ. Moreover, Aβ impairs BDNF-mediated effects upon LTP. In all cases, it was possible to rescue the effect of BDNF by adding a calpain inhibitor (MDL28170), indicating that the Aβ-induced loss of function of BDNF at the synapses is mediated by calpains. The finding that the calpain inhibitor prevents the molecular changes in TrkB receptors as well as prevents the loss of synaptic BDNF modulatory action strongly suggests that a single mechanism underlies both phenomena, highlighting the functional consequence of the Aβ-induced cleavage of TrkB receptors.

Interestingly, Aβ did not affect LTP magnitude (without BDNF), or basal synaptic transmission, as evaluated by input/output curves. This provides evidence that Aβ may impair BDNF signaling, even before impairment of synaptic transmission and plasticity, suggesting that loss of neuromodulation by neurotrophins is a very early sign of synaptic impairments induced by Aβ. Nevertheless, there are evidences that Aβ peptides could impair LTP (e.g., Lambert et al. 1998). In our experimental conditions, we did not detect any significant change in LTP magnitude induced by a very-weak θ-burst in hippocampal slices exposed to Aβ. This absence of Aβ effect upon LTP, already seen by others (Smith et al. 2009), could be due to several factors such as the stimulation protocol, the Aβ preparation, the developmental age, or genetic background of the animals used (Smith et al. 2009). Interestingly, Smith et al. (2009) showed that soluble oligomeric Aβ1–42 significantly blocked hippocampal LTP when induced by high-frequency stimulation but not by θ-burst, the type of stimulation used in this work.

Our findings provide a possible biochemical mechanism for previous observations that TrkB-FL receptors are decreased in pathological situations, including AD (Allen et al. 1999; Ferrer et al. 1999), where calpains are found overactivated (Saito et al. 1993). Calpain-dependent down-regulation of TrkB-FL protein also occurs after acute insults, such as excitotoxicity and ischemia (Gomes et al. 2012; Vidaurre et al. 2012). Whether calpains can also cleave other Trk family members, such as TrkA or TrkC, is yet unknown. However, by comparing the sequences of Trk receptors, we can predict that TrkA and TrkC are probably not cleaved by calpains, since they both lack the calpain-cleavage site present in TrkB, which is conserved within species. Calpain overactivation has been associated with several neuropathological conditions, including prion-like diseases, muscular dystrophies, Huntington's disease, Parkinson's disease, AD, multiple sclerosis, ischemia, stroke, and brain trauma. Calpain inhibition is therefore a promising therapeutic strategy with demonstrated efficacy in animal models. However, translation to clinical trials waits for the development of selective inhibitors of calpains to be used in humans (see Saez et al. 2006).

In summary, we highlighted the mechanisms responsible for Aβ-induced TrkB receptor dysregulation. Namely, we found that Aβ selectively increases the mRNA levels of truncated TrkB-T1 and T2 receptors, and it induces a calpain-mediated cleavage of TrkB-FL protein. The cleavage of TrkB-FL occurs between Asn520 and Ser521 and produces a new truncated receptor, containing the Shc-binding site (TrkB-T′), and a new intracellular cleavage product (TrkB-ICD), containing the complete kinase domain of TrkB-FL. At a functional level, Aβ severely impairs BDNF effects upon GABA and glutamate release and upon synaptic plasticity, in a calpain-dependent way. Taken together, the data demonstrate that calpain overactivation induced by Aβ severely impairs BDNF/TrkB-FL signaling, affecting the synaptic actions of BDNF. By detailing the mechanisms involved in the endogenous dysregulation of TrkB receptors induced by Aβ, as well as the early functional consequences of this dysregulation, this work reinforces the rational for the use of calpain inhibitors as a therapeutic tool in AD.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/.

Funding

This work was supported by Fundação para a Ciência e a Tecnologia (FCT) Grants SFRH/BD/62828/2009 (to A.J.S.) and SFRH/BPD/81627/2011 (to S.H.V.), EU (COST B-30 concerted action), Gabinete de Apoio à Investigação Científica, Tecnológica e Inovação (GAPIC)—15th Programme for Education and Science (to S.P and S.L), and Bayer grant (to A.P.C). D.B. and V.B.S. are supported by the LabEx (excellence laboratory) DISTALZ (Development of Innovative Strategies for a Transdisciplinary approach to Alzheimer's disease), Inserm, CNRS, DN2M, FEDER, France Alzheimer, Région Nord/Pas-de-Calais, LECMA, ANR (ADORATAU), and FUI MEDIALZ. E.C are supported by ERC AdG 322742-iPlasticity, Academy of Finland CoE Program and Sigrid Jusus foundation.

Funding

This work was supported by Fundação para a Ciência e a Tecnologia (FCT) Grants SFRH/BD/62828/2009 (to A.J.S.) and SFRH/BPD/81627/2011 (to S.H.V.), EU (COST B-30 concerted action), Gabinete de Apoio à Investigação Científica, Tecnológica e Inovação (GAPIC)—15th Programme for Education and Science (to S.P and S.L), and Bayer grant (to A.P.C). D.B. and V.B.S. are supported by the LabEx (excellence laboratory) DISTALZ (Development of Innovative Strategies for a Transdisciplinary approach to Alzheimer's disease), Inserm, CNRS, DN2M, FEDER, France Alzheimer, Région Nord/Pas-de-Calais, LECMA, ANR (ADORATAU), and FUI MEDIALZ. E.C are supported by ERC AdG 322742-iPlasticity, Academy of Finland CoE Program and Sigrid Jusus foundation.

Notes

We thank Regeneron for the gift of BDNF, W.W. Andersen (University of Bristol, Bristol, UK) for the gift of the data analysis (WinLTP) software, and Lille Neurobank, France, for providing the brain sample. PCR primers for β-actin were kindly provided by Dr Tiago Outeiro, Instituto de Medicina Molecular Lisbon, Portugal. The animal housing facilities of the Institute of Physiology of the Faculty of Medicine of the University of Lisbon are also acknowledged. Conflict of Interest: None declared.

References

Adamec
E
Mohan
P
Vonsattel
JP
Nixon
RA
2002
.
Calpain activation in neurodegenerative diseases: confocal immunofluorescence study with antibodies specifically recognizing the active form of calpain 2
.
Acta Neuropathol
 .
104
:
92
104
.
Albensi
BC
Oliver
DR
Toupin
J
Odero
G
2007
.
Electrical stimulation protocols for hippocampal synaptic plasticity and neuronal hyper-excitability: are they effective or relevant?
Exp Neurol
 .
204
:
1
13
.
Allen
SJ
Wilcock
GK
Dawbarn
D
1999
.
Profound and selective loss of catalytic TrkB immunoreactivity in Alzheimer's disease
.
Biochem Biophys Res Commun
 .
264
:
648
651
.
Ancot
F
Foveau
B
Lefebvre
J
Leroy
C
Tulasne
D
2009
.
Proteolytic cleavages give receptor tyrosine kinases the gift of ubiquity
.
Oncogene
 .
28
:
2185
2195
.
Arancibia
S
Silhol
M
Mouliere
F
Meffre
J
Hollinger
I
Maurice
T
Tapia-Arancibia
L
2008
.
Protective effect of BDNF against beta-amyloid induced neurotoxicity in vitro and in vivo in rats
.
Neurobiol Dis
 .
31
:
316
326
.
Aroeira
RI
Ribeiro
JA
Sebastiao
AM
Valente
CA
2011
.
Age-related changes of glycine receptor at the rat hippocampus: from the embryo to the adult
.
J Neurochem
 .
118
:
339
353
.
Bak
LK
Schousboe
A
Waagepetersen
HS
2006
.
The glutamate/GABA-glutamine cycle: aspects of transport, neurotransmitter homeostasis and ammonia transfer
.
J Neurochem
 .
98
:
641
653
.
Bell
KF
de Kort
GJ
Steggerda
S
Shigemoto
R
Ribeiro-da-Silva
A
Cuello
AC
2003
.
Structural involvement of the glutamatergic presynaptic boutons in a transgenic mouse model expressing early onset amyloid pathology
.
Neurosci Lett
 .
353
:
143
147
.
Blurton-Jones
M
Kitazawa
M
Martinez-Coria
H
Castello
NA
Muller
FJ
Loring
JF
Yamasaki
TR
Poon
WW
Green
KN
LaFerla
FM
2009
.
Neural stem cells improve cognition via BDNF in a transgenic model of Alzheimer disease
.
Proc Natl Acad Sci USA
 .
106
:
13594
13599
.
Canas
N
Pereira
IT
Ribeiro
JA
Sebastiao
AM
2004
.
Brain-derived neurotrophic factor facilitates glutamate and inhibits GABA release from hippocampal synaptosomes through different mechanisms
.
Brain Res
 .
1016
:
72
78
.
Canas
PM
Simoes
AP
Rodrigues
RJ
Cunha
RA
2014
.
Predominant loss of glutamatergic terminal markers in a beta-amyloid peptide model of Alzheimer's disease
.
Neuropharmacology
 .
76
Pt A
:
51
56
.
Chen
G
Kolbeck
R
Barde
YA
Bonhoeffer
T
Kossel
A
1999
.
Relative contribution of endogenous neurotrophins in hippocampal long-term potentiation
.
J Neurosci
 .
19
:
7983
7990
.
Connor
B
Young
D
Yan
Q
Faull
RL
Synek
B
Dragunow
M
1997
.
Brain-derived neurotrophic factor is reduced in Alzheimer's disease
.
Brain Res Mol Brain Res
 .
49
:
71
81
.
Cruz
JC
Tsai
LH
2004
.
Cdk5 deregulation in the pathogenesis of Alzheimer's disease
.
Trends Mol Med
 .
10
:
452
458
.
Cruz
JC
Tseng
HC
Goldman
JA
Shih
H
Tsai
LH
2003
.
Aberrant Cdk5 activation by p25 triggers pathological events leading to neurodegeneration and neurofibrillary tangles
.
Neuron
 .
40
:
471
483
.
Devi
L
Ohno
M
2012
.
7,8-dihydroxyflavone, a small-molecule TrkB agonist, reverses memory deficits and BACE1 elevation in a mouse model of Alzheimer's disease
.
Neuropsychopharmacology
 .
37
:
434
444
.
Diogenes
MJ
Costenla
AR
Lopes
LV
Jeronimo-Santos
A
Sousa
VC
Fontinha
BM
Ribeiro
JA
Sebastiao
AM
2011
.
Enhancement of LTP in aged rats is dependent on endogenous BDNF
.
Neuropsychopharmacology
 .
36
:
1823
1836
.
Dorsey
SG
Renn
CL
Carim-Todd
L
Barrick
CA
Bambrick
L
Krueger
BK
Ward
CW
Tessarollo
L
2006
.
In vivo restoration of physiological levels of truncated TrkB.T1 receptor rescues neuronal cell death in a trisomic mouse model
.
Neuron
 .
51
:
21
28
.
Eide
FF
Vining
ER
Eide
BL
Zang
K
Wang
XY
Reichardt
LF
1996
.
Naturally occurring truncated trkB receptors have dominant inhibitory effects on brain-derived neurotrophic factor signaling
.
J Neurosci
 .
16
:
3123
3129
.
Ferrer
I
Marin
C
Rey
MJ
Ribalta
T
Goutan
E
Blanco
R
Tolosa
E
Marti
E
1999
.
BDNF and full-length and truncated TrkB expression in Alzheimer disease. Implications in therapeutic strategies
.
J Neuropathol Exp Neurol
 .
58
:
729
739
.
Figurov
A
Pozzo-Miller
LD
Olafsson
P
Wang
T
Lu
B
1996
.
Regulation of synaptic responses to high-frequency stimulation and LTP by neurotrophins in the hippocampus
.
Nature
 .
381
:
706
709
.
Fischer
A
Sananbenesi
F
Pang
PT
Lu
B
Tsai
LH
2005
.
Opposing roles of transient and prolonged expression of p25 in synaptic plasticity and hippocampus-dependent memory
.
Neuron
 .
48
:
825
838
.
Fontinha
BM
Diogenes
MJ
Ribeiro
JA
Sebastiao
AM
2008
.
Enhancement of long-term potentiation by brain-derived neurotrophic factor requires adenosine A2A receptor activation by endogenous adenosine
.
Neuropharmacology
 .
54
:
924
933
.
Goll
DE
Thompson
VF
Li
H
Wei
W
Cong
J
2003
.
The calpain system
.
Physiol Rev
 .
83
:
731
801
.
Gomes
JR
Costa
JT
Melo
CV
Felizzi
F
Monteiro
P
Pinto
MJ
Inacio
AR
Wieloch
T
Almeida
RD
Graos
M
et al
2012
.
Excitotoxicity downregulates TrkB.FL signaling and upregulates the neuroprotective truncated TrkB receptors in cultured hippocampal and striatal neurons
.
J Neurosci
 .
32
:
4610
4622
.
Granic
I
Nyakas
C
Luiten
PG
Eisel
UL
Halmy
LG
Gross
G
Schoemaker
H
Moller
A
Nimmrich
V
2010
.
Calpain inhibition prevents amyloid-beta-induced neurodegeneration and associated behavioral dysfunction in rats
.
Neuropharmacology
 .
59
:
334
342
.
Grynspan
F
Griffin
WR
Cataldo
A
Katayama
S
Nixon
RA
1997
.
Active site-directed antibodies identify calpain II as an early-appearing and pervasive component of neurofibrillary pathology in Alzheimer's disease
.
Brain Res
 .
763
:
145
158
.
Holscher
C
1998
.
Possible causes of Alzheimer's disease: amyloid fragments, free radicals, and calcium homeostasis
.
Neurobiol Dis
 .
5
:
129
141
.
Holscher
C
2005
.
Development of beta-amyloid-induced neurodegeneration in Alzheimer's disease and novel neuroprotective strategies
.
Rev Neurosci
 .
16
:
181
212
.
Holsinger
RM
Schnarr
J
Henry
P
Castelo
VT
Fahnestock
M
2000
.
Quantitation of BDNF mRNA in human parietal cortex by competitive reverse transcription-polymerase chain reaction: decreased levels in Alzheimer's disease
.
Brain Res Mol Brain Res
 .
76
:
347
354
.
Kaplan
DR
Miller
FD
2000
.
Neurotrophin signal transduction in the nervous system
.
Curr Opin Neurobiol
 .
10
:
381
391
.
Karran
E
Mercken
M
De Strooper
B
2011
.
The amyloid cascade hypothesis for Alzheimer's disease: an appraisal for the development of therapeutics
.
Nat Rev Drug Discov
 .
10
:
698
712
.
Kemppainen
S
Rantamaki
T
Jeronimo-Santos
A
Lavasseur
G
Autio
H
Karpova
N
Karkkainen
E
Staven
S
Miranda
HV
Outeiro
TF
et al
2012
.
Impaired TrkB receptor signaling contributes to memory impairment in APP/PS1 mice
.
Neurobiol Aging
 .
33
:
1122 e1123–1139
.
Klein
R
Conway
D
Parada
LF
Barbacid
M
1990
.
The trkB tyrosine protein kinase gene codes for a second neurogenic receptor that lacks the catalytic kinase domain
.
Cell
 .
61
:
647
656
.
Korte
M
Carroll
P
Wolf
E
Brem
G
Thoenen
H
Bonhoeffer
T
1995
.
Hippocampal long-term potentiation is impaired in mice lacking brain-derived neurotrophic factor
.
Proc Natl Acad Sci USA
 .
92
:
8856
8860
.
Kubo
T
Nishimura
S
Kumagae
Y
Kaneko
I
2002
.
In vivo conversion of racemized beta-amyloid ([D-Ser 26]A beta 1–40) to truncated and toxic fragments ([D-Ser 26]A beta 25–35/40) and fragment presence in the brains of Alzheimer's patients
.
J Neurosci Res
 .
70
:
474
483
.
Lambert
MP
Barlow
AK
Chromy
BA
Edwards
C
Freed
R
Liosatos
M
Morgan
TE
Rozovsky
I
Trommer
B
Viola
KL
et al
1998
.
Diffusible, nonfibrillar ligands derived from Abeta1–42 are potent central nervous system neurotoxins
.
Proc Natl Acad Sci USA
 .
95
:
6448
6453
.
Lee
MS
Kwon
YT
Li
M
Peng
J
Friedlander
RM
Tsai
LH
2000
.
Neurotoxicity induces cleavage of p35 to p25 by calpain
.
Nature
 .
405
:
360
364
.
Medeiros
R
Kitazawa
M
Chabrier
MA
Cheng
D
Baglietto-Vargas
D
Kling
A
Moeller
A
Green
KN
LaFerla
FM
2012
.
Calpain inhibitor A-705253 mitigates Alzheimer's disease-like pathology and cognitive decline in aged 3xTgAD mice
.
Am J Pathol
 .
181
:
616
625
.
Noble
W
Olm
V
Takata
K
Casey
E
Mary
O
Meyerson
J
Gaynor
K
LaFrancois
J
Wang
L
Kondo
T
et al
2003
.
Cdk5 is a key factor in tau aggregation and tangle formation in vivo
.
Neuron
 .
38
:
555
565
.
Patrick
GN
Zukerberg
L
Nikolic
M
de la Monte
S
Dikkes
P
Tsai
LH
1999
.
Conversion of p35 to p25 deregulates Cdk5 activity and promotes neurodegeneration
.
Nature
 .
402
:
615
622
.
Pfaffl
MW
2001
.
A new mathematical model for relative quantification in real-time RT-PCR
.
Nucleic Acids Res
 .
29
:
e45
.
Phillips
HS
Hains
JM
Armanini
M
Laramee
GR
Johnson
SA
Winslow
JW
1991
.
BDNF mRNA is decreased in the hippocampus of individuals with Alzheimer's disease
.
Neuron
 .
7
:
695
702
.
Pike
CJ
Walencewicz-Wasserman
AJ
Kosmoski
J
Cribbs
DH
Glabe
CG
Cotman
CW
1995
.
Structure-activity analyses of beta-amyloid peptides: contributions of the beta 25–35 region to aggregation and neurotoxicity
.
J Neurochem
 .
64
:
253
265
.
Saez
ME
Ramirez-Lorca
R
Moron
FJ
Ruiz
A
2006
.
The therapeutic potential of the calpain family: new aspects
.
Drug Discov Today
 .
11
:
917
923
.
Saito
K
Elce
JS
Hamos
JE
Nixon
RA
1993
.
Widespread activation of calcium-activated neutral proteinase (calpain) in the brain in Alzheimer disease: a potential molecular basis for neuronal degeneration
.
Proc Natl Acad Sci USA
 .
90
:
2628
2632
.
Selkoe
DJ
2002
.
Alzheimer's disease is a synaptic failure
.
Science
 .
298
:
789
791
.
Smith
JP
Lal
V
Bowser
D
Cappai
R
Masters
CL
Ciccotosto
GD
2009
.
Stimulus pattern dependence of the Alzheimer's disease amyloid-beta 42 peptide's inhibition of long term potentiation in mouse hippocampal slices
.
Brain Res
 .
1269
:
176
184
.
Sousa
VC
Assaife-Lopes
N
Ribeiro
JA
Pratt
JA
Brett
RR
Sebastiao
AM
2011
.
Regulation of hippocampal cannabinoid CB1 receptor actions by adenosine A1 receptors and chronic caffeine administration: implications for the effects of Delta9-tetrahydrocannabinol on spatial memory
.
Neuropsychopharmacology
 .
36
:
472
487
.
Stoilov
P
Castren
E
Stamm
S
2002
.
Analysis of the human TrkB gene genomic organization reveals novel TrkB isoforms, unusual gene length, and splicing mechanism
.
Biochem Biophys Res Commun
 .
290
:
1054
1065
.
Tauszig-Delamasure
S
Yu
LY
Cabrera
JR
Bouzas-Rodriguez
J
Mermet-Bouvier
C
Guix
C
Bordeaux
MC
Arumae
U
Mehlen
P
2007
.
The TrkC receptor induces apoptosis when the dependence receptor notion meets the neurotrophin paradigm
.
Proc Natl Acad Sci USA
 .
104
:
13361
13366
.
Tong
L
Balazs
R
Thornton
PL
Cotman
CW
2004
.
Beta-amyloid peptide at sublethal concentrations downregulates brain-derived neurotrophic factor functions in cultured cortical neurons
.
J Neurosci
 .
24
:
6799
6809
.
Trinchese
F
Fa
M
Liu
S
Zhang
H
Hidalgo
A
Schmidt
SD
Yamaguchi
H
Yoshii
N
Mathews
PM
Nixon
RA
et al
2008
.
Inhibition of calpains improves memory and synaptic transmission in a mouse model of Alzheimer disease
.
J Clin Invest
 .
118
:
2796
2807
.
Vidaurre
OG
Gascon
S
Deogracias
R
Sobrado
M
Cuadrado
E
Montaner
J
Rodriguez-Pena
A
Diaz-Guerra
M
2012
.
Imbalance of neurotrophin receptor isoforms TrkB-FL/TrkB-T1 induces neuronal death in excitotoxicity
.
Cell Death Dis
 .
3
:
e256
.
Wei
Z
Song
MS
MacTavish
D
Jhamandas
JH
Kar
S
2008
.
Role of calpain and caspase in beta-amyloid-induced cell death in rat primary septal cultured neurons
.
Neuropharmacology
 .
54
:
721
733
.
Wong
J
Higgins
M
Halliday
G
Garner
B
2012
.
Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheimer's disease
.
Neuroscience
 .
210
:
363
374
.