BDNF prevents NMDA-induced toxicity in models of Huntington's disease: the effects are genotype specific and adenosine A2A receptor is involved

Authors

  • Alberto Martire,

    Corresponding author
    • Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author
  • Rita Pepponi,

    1. Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author
  • Maria Rosaria Domenici,

    1. Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author
  • Antonella Ferrante,

    1. Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author
  • Valentina Chiodi,

    1. Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author
  • Patrizia Popoli

    1. Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy
    Search for more papers by this author

Address correspondence and reprint requests to Alberto Martire, Section of Central Nervous System Pharmacology, Department of Therapeutic Research and Medicines Evaluation, Istituto Superiore di Sanità, Rome, Italy. E-mail: alberto.martire@iss.it

Abstract

NMDA receptor-mediated excitotoxicity is thought to play a pivotal role in the pathogenesis of Huntington's disease (HD). The neurotrophin brain-derived neurotrophic factor (BDNF), which is also highly involved in HD and whose effects are modulated by adenosine A2ARs, influences the activity and expression of striatal NMDA receptors. In electrophysiology experiments, we investigated the role of BDNF toward NMDA-induced effects in HD models, and the possible involvement of A2ARs. In corticostriatal slices from wild-type mice and age-matched symptomatic R6/2 mice (a model of HD), NMDA application (75 μM) induced a transient or a permanent (i.e., toxic) reduction of field potential amplitude, respectively. BDNF (10 ng/mL) potentiated NMDA effects in wild-type, while it protected from NMDA toxicity in R6/2 mice. Both effects of BDNF were prevented by A2AR blockade. The protective effect of BDNF against NMDA-induced toxicity was reproduced in a cellular model of HD. These findings may have very important implications for the neuroprotective potential of BDNF and A2AR ligands in HD.

Abbreviations used
A2AR

adenosine A2A receptor

ACSF

artificial CSF

AMPA

alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate

BDNF

brain-derived neurotrophic factor

FP

field potential

GABA

gamma-aminobutyric acid

min

minute

MSNs

medium spiny neurons

NMDA

N-methyl-d-aspartic acid

Huntington's disease (HD) is an inherited neurodegenerative disorder caused by a mutation in the exon 1 of the huntingtin (htt) gene (The Huntington's Disease Collaborative Research Group's 1993), namely an abnormal expansion of a CAG codon, resulting in devastating cognitive disturbance, motor impairment, and premature death of affected individuals (Vonsattel and DiFiglia 1998).

The expanded CAG trinucleotide encodes a long polyglutamine (poliQ) tract, which forms aggregates in nucleus and cytoplasm of neurons both in patients and in experimental models of the disease (Reddy et al. 1999). As the neurodegeneration mainly involves the striatum and the cerebral cortex (Vonsattel et al. 1985), the brain areas of the paradigmatic corticostriatal glutamatergic pathway, N-methyl-D-aspartic acid receptor (NMDAR)-mediated excitotoxicity is thought to play a pivotal role in the selective neuronal death that occurs in GABA-ergic medium-sized spiny neurons (MSNs) of the striatum. In agreement, an increased sensitivity to NMDA-induced toxicity has been consistently reported in the striatum of HD mice (Levine et al. 1999; Cepeda et al. 2001; Zeron et al. 2002; Martire et al. 2007). Functional NMDARs are tetrameric structures composed of two NR1 and at least two NR2 (A-D) subunits (Dingledine et al. 1999). The NR2A and NR2B subunits of NMDAR mediate cell survival and cell death, respectively (Liu et al. 2007); indeed, the over-expression of NR2B subunits potentiates striatal neurodegeneration in HD mice (Heng et al. 2009). In the striatum of R6/2 mice, a transgenic model of HD, we found a reduction in the NR2A/NR2B ratio (an index of vulnerability to excitotoxic cell death, Ali and Levine 2006), which correlated with an increased susceptibility to NMDA-induced toxicity (Martire et al. 2010), and which was influenced by adenosine A2A receptor (A2AR) stimulation (Ferrante et al. 2010) and blockade (Martire et al. 2010).

Brain-derived neurotrophic factor (BDNF), a glycoprotein belonging to neurotrophins, is very important for the survival of striatal neurons and for the activity of corticostriatal synapses (Cattaneo et al. 2005), where it controls glutamate release and allows striatal neurons to survive to excitotoxin-induced neurodegeneration (Bemelmans et al. 1999). An impairment in the synthesis and striatal transport of BDNF is considered a major determinant in the pathogenesis of HD (Zuccato et al. 2001; Zuccato and Cattaneo 2007). Interestingly, BDNF has been reported to influence NMDAR signaling in models of HD: a huge remodeling of post-synaptic density and a specific reduction of synaptic alpha CaMKII have been observed in double transgenic R6/1:BDNF(±) mice (Torres-Peraza et al. 2008). Moreover, BDNF functions and NMDA-mediated toxicity in the striatum are both strongly modulated by A2ARs (Martire et al. 2007; Potenza et al. 2007).

The aim of this study was to further investigate the role of BDNF toward NMDA-induced effects in HD models, and the possible involvement of A2ARs.

Here we report that 1) BDNF oppositely modulates NMDA-mediated toxicity in the striatum of transgenic HD (R6/2) versus wild-type (WT) mice, respectively, showing protective and worsening effects on synaptic transmission in the two genotypes; 2) BDNF effects are always counteracted by TrkB or A2AR blockade, confirming the permissive role of A2AR in BDNF signaling; and 3) a similar protective effect against excitotoxicity is reproduced in a cellular model of HD (ST14A/Q120 cell line).

These findings may have very important implications for the neuroprotective potential of both BDNF and A2AR ligands in HD.

Materials and methods

Animals

A colony of R6/2 (Mangiarini et al. 1996) and littermate WT mice was maintained at Charles River Laboratories (Calco, Italy). Male and female genotyped mice, usually not younger than 4.5 weeks of age, were delivered and housed in our animal facilities until the end of the experiments. In addition, 6- to 8-week-old C57BL/6 mice were used for patch-clamp experiments. All studies were conducted in accordance with the principles and procedures outlined in the EU (European Community Guidelines for Animal Care, DL 116/92, application of the European Communities Council Directive, 86/609/EEC), FELASA, and ARRIVE guidelines. The animals were kept under standardized temperature, humidity, and lighting conditions and had free access to water and food. All efforts were made to reduce the number of animals used and to minimize their suffering.

Electrophysiology

Corticostriatal slices were prepared according to the method described by Tebano et al. (2009). Transgenic R6/2 mice in a late symptomatic phase (12 weeks) and age-matched WT were used. Animals were killed by decapitation under ether anesthesia, the brain was removed quickly from the skull, and coronal slices (300 μm thick) including the neostriatum and the neocortex were cut with a vibratome. Slices were maintained at 22–24°C in artificial cerebrospinal fluid (ACSF) containing (mM) 126 NaCl, 3.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2 CaCl2, 25 NaHCO3, 11 glucose, saturated with 95% O2 and 5% CO2 (pH 7.3). After incubation in ACSF for at least 1 h, a single slice was transferred to a submerged recording chamber and continuously superfused at 32–33°C with ACSF at rate of 2.7–3 mL/min. Tested drugs were added to this superfusion solution.

Whole cell voltage-clamp recordings were made using a PC-501A patch clamp amplifier (Warner Instruments LLC; Hamden, CT, USA) and excitatory and inhibitory post-synaptic currents (EPSC and IPSC, respectively) were recorded. Data were filtered at 1 kHz, digitized at 10 kHz, stored on a computer and analyzed off-line using WinWCP software (J Dempster, Strathclyde University, Glasgow, Scotland). MSN in dorsolateral striatum were visualized using a 60× water-immersion objective and infrared differential interference optics video microscopy, and identified by small somata and basic membrane properties (input resistance, membrane capacitance, and time constant) (Cepeda et al. 1998). Passive membrane properties of MSN were determined in voltage-clamp mode by applying a depolarized step voltage command (10 mV) and using the cell test function integrated in WinWCP program. Series resistance was < 25 MΩ and checked periodically. If the series resistance changed > 30% by the end of the experiment, the cell was discarded. For EPSC recordings, borosilicate glass (1.5 mm outer diameter, WPI Instruments, Sarasota, FL, USA) patch electrodes (4–6 MΩ) contained the following solution (in mM): 130 Cs-methanesulfonate, 10 CsCl, 4 NaCl, 1 MgCl2, 5 MgATP, 5 EGTA, 10 HEPES, 0.5 GTP, 10 phosphocreatine, 0.1 leupeptin, and 4 QX-314 (pH 7.25–7.3, 280–290 mOsm). For IPSC recordings, patch electrodes were filled with (in mM) 128 K-gluconate, 10 HEPES, 2 MgCl2, 5 KCl, 0.2 CaCl2, 2 EGTA, 5 D-glucose, 20 creatinphosphate, 5 QX-314. QX- 314 was used to internally block voltage-dependent Na+ current. Currents were evoked by extracellular stimulation (stimulus duration: 0.01 ms; frequency: 0.05 Hz) with bipolar platinum/iridium concentric electrode (FHC) placed in the white matter, at a holding potential of −70 mV in ACSF containing 10 μM bicuculline for EPSC and with the stimulating electrode place in the striatum, at a holding potential of −40 mV, in ACSF containing CNQX (10 μM) and APV (50 μM) for IPSC recordings.

Extracellular field potentials (FPs) were recorded in the dorsomedial striatum with a glass microelectrode filled with 2-M NaCl solution (pipette resistance 2–5 MΩ) on stimulation of the white matter between the cortex and the striatum with a bipolar twisted NiCr-insulated electrode (50 μm o.d.). Each pulse was delivered every 20 s with a duration 100 μs and an intensity chosen to cause 60% of maximal response. Three consecutive responses were averaged and recorded using a DAM-80 AC differential amplifier (WPI Instruments, Sarasota, FL, USA), acquired and analyzed using LTP software (Anderson and Collingridge 2001). The data were expressed as mean ± SEM from n experiments (one slice tested in each experiment). Slices were obtained always from at least two animals for each set of experiments. To allow comparisons between different experiments, in each experiment, the FP amplitudes were normalized, taking as 100% the average of the values obtained over the 10-min period immediately before the test compound was applied, and considered as the basal value. The effects of the drugs were expressed as percentage variation with respect to basal values. In all experiments, NMDA (75 μM) was applied to the slices over 5 min and a reduction of at least 90% of basal FP amplitude, during and after NMDA application, was defined as FP disappearance. The washout period lasted at least 40 min.

Input/Output (I/O) plots

For each slice, single stimuli were delivered every 20 s (square pulses of 100 μs duration at a frequency of 0.05 Hz) and three consecutive responses were averaged. Once the response was stable, the minimum stimulus intensity necessary to evoke an observable response was measured. An I/O curve was then obtained by recording averaged responses at 20-μA increments, starting at the threshold stimulation intensity (~20 μA) and ending with a plateau at a maximum of 180 μA. Each point on the I/O curve was obtained by averaging responses over at least 5 min of recording.

Q15 and Q120 ST14A cell cultures

ST14A cell lines were provided by the Coriell Biological Material Repository by the High Q Foundation and CHDi, Inc. These cells express a human huntingtin N-terminal portion (residues 1–548). One cell line (ST14A/Q15) expresses normal huntingtin with a 15-glutamine repeat region, and the second cell line (ST14A/Q120) expresses mutant huntingtin with a 120-glutamine repeat region. ST14A cells were developed from embryonic day 14 rat striatal primordia by retroviral transduction of the temperature-sensitive SV40 large T antigen (Cattaneo and Conti 1998). These cells have typical features of the medium-sized spiny neurons that are affected by Huntington's disease, and are well described in the literature (Cattaneo and Conti 1998; Ehrlich et al. 2001). Normally, the cells were grown at 33°C under 5% CO2 in high-glucose Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 2 mM l-glutamine, 100 units/mL penicillin, and 0.1 mg/mL streptomycin.

ST14A cell differentiation induced by hormonal factors

For our studies, we used ST14A/Q15 and ST14A/Q120 cells differentiated as follows. Cells were first grown under normal conditions (described above) until the desired numbers of cells were produced. To induce differentiation, cells were transferred into high-glucose Dulbecco's modified Eagle's medium containing 2 mM l-glutamine, 100 units/mL penicillin, 0.1 mg/mL streptomycin, 1xN2 supplement (Invitrogen, Carlsbad, CA, USA), 5 μM forskolin (Sigma, St. Louis, MO, USA), 250 μM isobutylmethylxanthine (Sigma), 200 nM 12-O-tetradecanoylphorbol-13- acetate (Sigma), 10 μM dopamine (Sigma), and 10 ng/mL recombinant human acidic fibroblast growth factor (Peprotech). Cells were incubated at 33°C under 5% CO2 for 48 or 72 h until the experiments were done (Ehrlich et al. 2001).

Morphological changes were evaluated by immunofluorescence of the neuronal marker ‘Microtubule Associated Protein’ (MAP2). In particular, cells were washed three times in phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde at 20–22°C; after washing they were permeabilized in 0.2% Triton X- 100 for 5 min at 20–22°C; after extensive washing, cells were blocked in a solution of 2.5% horse serum in PBS for 20 min at 20–22°C; they were incubated with MAP2 (1 : 200; Millipore, Milan, Italy) for 1 h at 37°C; they were washed three times with PBS and incubated with AlexaFluor 555 (1 : 200, Life Technologies, Milan, Italy) for 1 h at 37°C; after washing, nuclei were counter-stained with Hoechst 33258 and mounted onto slides.

To obtain a similar level of differentiation in WT (ST14A/Q15) and HD (ST14A/Q120) cells, we checked the morphological appearance of cells by immunocytochemistry in many different attempts, finding comparable changes in ST14A/Q15 cells differentiated for 72 h and in ST14A/Q120 cells differentiated for 48 h. Then, cells were treated with NMDA at different concentrations, respectively, 10 mM and 3 mM in Q15 and in Q120 cells, for the same time (10 min), to obtain a comparable toxic effect in the two genotypes. BDNF (100 ng/mL) pre-treatment was performed 2 h before in the differentiating medium and NMDA treatment was performed for 10 min in Locke's solution containing (in mM) 134 NaCl, 2.5 KCl, 4 NaHCO3, 5 Hepes, 2.3 CaCl2, 5.5 d-glucose, 50 μM glycine, pH 7.4. After this time, differentiated cells were returned to the differentiating medium for 24 h until the cell survival was estimated by cell counting. To obtain the related bright-field pictures, a Petri dish containing a neuronal culture was positioned on the stage of an inverted microscope (Axiovert 35, Zeiss, Hamburg, Germany) equipped with a x10 (NA0.3) Neofluar objective and a ccd Sensicam (PCO, Kelheim, Germany) 12bit gray scale camera, with a resolution of 1280 × 1024 pixels.

Drugs

K252a, ZM241385, and CGS21680 were obtained from Tocris Biosciences (Bristol, UK); NMDA and BDNF were from Sigma-Aldrich (Milan, Italy). All pharmacological agents were diluted directly in the superfusion medium from stock solutions prepared in distilled water, NaOH 0.1 N (for NMDA), or in dimethylsulfoxide. Stock solutions were made to obtain concentrations of dimethylsulfoxide lower than 0.001% in the superfusing ACSF.

Statistics

Results from electrophysiology experiments were expressed as mean ± SEM from n slices. Statistical analysis of the data was performed using Mann–Whitney U-test. Results from cell cultures were expressed as percentage of control, which was considered as 100% and represent mean ± SEM values of three to four experiments. Statistical analysis of the data was performed using student's t-test. Statistical analyses and curve fittings were obtained by using Graphpad Prism software (GraphPad Software, San Diego, CA, USA). A p-value of < 0.05 was considered to indicate a significant difference.

Results

Electrophysiology

Despite the well-known ability of exogenous BDNF to induce a long-lasting enhancement of basal synaptic transmission in hippocampal slices from adult rats or mice (Diógenes et al. 2007; Tebano et al. 2008; Ji et al. 2010), the same neurotrophin failed to affect the FP amplitude in corticostriatal slices from WT and R6/2 mice (Fig. 1a, b).

Figure 1.

Brain-derived neurotrophic factor (BDNF) does not affect basal striatal synaptic transmission in corticostriatal slices. The application of exogenous BDNF to corticostriatal slices from wild-type (a) and R6/2 (b) mice did not affect the field potential (FP) amplitude. Insert shows FPs recorded before (left) and at the end (right) of the application of BDNF. Each trace is the average of three successive FPs (artifacts of stimulation have been truncated). Calibration bars: 0.5 mV, 10 ms. This lack of efficacy was further confirmed in whole-cell patch-clamp recordings of evoked excitatory and inhibitory post-synaptic currents (EPSC) and inhibitory post-synaptic currents (IPSC) in medium spiny neurons of the dorsal striatum in 6- to 8-week-old C57BL/6 mice (c, d). BDNF did not influence excitatory currents in the striatum and has only a minimal effect on inhibitory current, namely a slight reduction of IPSC (d). Insert shows EPSC and IPSC recorded before (left) and at the end (right) of the application of BDNF. Calibration bars: 20 pA, 50 ms.

The lack of effect on excitatory synaptic transmission in the dorsal striatum was further confirmed in whole-cell patch-clamp recordings. To determine whether excitatory currents could be regulated by BDNF, we performed whole-cell recordings of evoked EPSC in MSNs of the dorsal striatum. We recorded EPSC after stimulation of the glutamatergic afferents coming from the cortex by placing the stimulating electrode in the white matter between the cortex and the striatum. EPSC were recorded in the presence of bicuculline to block GABAA-mediated currents. We found that BDNF application (10 ng/mL) did not change EPSC amplitude (n = 4, Fig. 1c). At the end of the experiments, CNQX was bath applied to definitely demonstrate that the current was alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate mediated. Even by increasing the concentration up to 20 ng/mL, BDNF failed to modify excitatory synaptic transmission (data not shown). In addition, we evaluated the effect of BDNF on inhibitory synaptic currents in the dorsal striatum. After intrastriatal stimulation, we recorded GABA-mediated IPSCs in MSNs in the presence of APV (50 μM) plus CNQX (10 μM). Ten minutes of baseline recordings preceded BDNF bath application. After 30 min, BDNF (10 ng/mL) induced a slight reduction of IPSC (90 ± 4.4%, n = 3, Fig. 1d). At the end of each experiment, bicuculline 20 μM was applied and the IPSC reduced progressively until the disappearance to demonstrate the full dependence of IPSC on GABAA receptor activation.

In large agreement with previous results from our group (Domenici et al. 2007; Martire et al. 2007, 2010), a higher vulnerability to NMDA in the striatum of R6/2 versus WT mice has been found. In electrophysiology experiments, frankly symptomatic R6/2 mice (12–13 weeks) and age-matched controls were used. In corticostriatal slices from WT mice, the application of 75 μM NMDA for 5 min induced the disappearance of the FP followed by an almost complete recovery after 40 min of washout (75.06 ± 7.35% of basal, n = 9, Fig. 2a). Conversely, in R6/2 mice, the same treatment induced a toxic effect, as revealed by the only partial recovery at washout, which indicates a permanent impairment of synaptic activity (41.23 ± 5.53% of basal, n = 11, Fig. 2c). When coapplied with NMDA in slices from WT mice, BDNF, ineffective when tested at 2 ng/mL, at a higher concentration, namely 10 ng/mL, rendered the effects of NMDA frankly toxic, as revealed by the significant reduction of FP recovery versus NMDA alone (47.44 ± 8.76% of basal, n = 6, *p < 0.05 vs. NMDA alone according to Mann–Whitney U test, Fig. 2a, b). A comparable result was obtained by using BDNF 20 ng/mL, Fig. 2b. Interestingly, BDNF (10 ng/mL) exerted the opposite effect (namely it significantly attenuated NMDA toxicity), in slices from R6/2 mice (73.09 ± 5.11% of basal, n = 4, *p < 0.05 vs. NMDA alone according to Mann–Whitney U test, Fig. 2c, d). In our experimental conditions, we did not observe a clear dose dependency and BDNF, at 2, 10, and 20 ng/mL, affected NMDA toxicity in a similar way (Fig. 2d). Moreover, the opposite effect of BDNF was unrelated to changes in alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) receptor-dependent basal synaptic transmission, as the I/O curve remained similar in WT and R6/2 mice (Fig. 2e). The effect elicited by BDNF in WT mice was prevented by the inhibitor of Trk receptors autophosphorylation, (200 nM) (79.95 ± 6.95% of basal, n = 3, °p < 0.05 vs. BDNF+NMDA according to Mann–Whitney U test, Fig. 2b) and by the A2AR antagonist ZM241385 (100 nM) (69.90 ± 4.76% of basal, n = 3, #p < 0.05 vs. BDNF+NMDA according to Mann–Whitney U test, Fig. 2b). Also the effect elicited by BDNF in R6/2 mice was prevented by the Trk tyrosine kinase inhibitor K252a (200 nM) (52.25 ± 2.74% of basal, n = 4, °p < 0.05 vs. BDNF+NMDA according to Mann–Whitney U test, Fig. 2d), but also by the A2AR antagonist ZM 241385 (100 nM) (39.02 ± 9.91% of basal, n = 4, #p < 0.05 vs. BDNF+NMDA according to Mann–Whitney U test, Fig. 2d). Neither K252a nor ZM241385 alone influenced the basal FP amplitude or NMDA-induced effects (data not shown).

Figure 2.

Brain-derived neurotrophic factor (BDNF)-induced opposite effect on synaptic transmission recovery in wild-type (WT) and R6/2 mice is TrkB and A2AR dependent. In WT mice, the application of 75 μM NMDA for 5 min induced the disappearance of the field potential (FP) followed by an almost complete recovery after 40 min of washout (n = 9, a). BDNF rendered the effect of NMDA frankly toxic, as revealed by the significant reduction of FP recovery versus NMDA alone (10 and 20 ng/mL, 30 min, n = 6 and n = 4, respectively, *p < 0.05 vs. NMDA alone, a, b), while BDNF was ineffective when tested at 2 ng/mL. Conversely, in R6/2 mice, the NMDA treatment induced a toxic effect, as revealed by the only partial recovery at washout, which indicates a permanent impairment of synaptic activity (n = 11, c). BDNF significantly attenuated NMDA toxicity in slices from R6/2 mice (2, 10, and 20 ng/mL, 30 min, n = 4, n = 4 and n = 3, respectively, *p < 0.05 vs. NMDA alone, c, d). Input/output plots of FP amplitude values are not different in 12- to 13-month-old R6/2 mice as compared with aged-matched WT animals. n = 6–3/group (e). The effect elicited by BDNF 10 ng/mL in WT mice was prevented by the Trk tyrosine kinase inhibitor K252a (200 nM) (n = 3, °p < 0.05 vs. BDNF+NMDA, b) and by the A2AR antagonist ZM241385 (100 nM) (n = 3, #p < 0.05 vs. BDNF+NMDA, b). Similarly, the effect of BDNF 10 ng/mL in R6/2 mice was blocked by K252a (200 nM) (n = 4, °p < 0.05 vs. BDNF+NMDA, d) but also by the A2AR antagonist ZM 241385 (100 nM) (n = 4, #p < 0.05 vs. BDNF+NMDA, d).

The above results showed that in corticostriatal synaptic transmission, an endogenous adenosine tone on A2ARs is requested to allow TrkB-mediated BDNF effects; in the light of these data, and considering their similarity with the opposite effect of A2AR stimulation on the NMDA toxicity in WT and R6/2 mice (Martire et al. 2007), we wondered if the shift from the pro-toxic to the protective effects of CGS21680 could be mediated by Trk receptors. To investigate this aspect, we performed some specific electrophysiology experiments by using the paradigm in which the NMDA toxicity in symptomatic R6/2 mice was reduced by the coapplication of CGS21680 (10 min of slices perfusion with CGS21680 30 nM followed by coapplication with NMDA 75 μM, Fig. 3a); once we were able to reproduce this condition, we evaluated the effect of the Trk inhibitor K252a (200 nM, 20 min of pre-treatment) on the ‘protective’ effect of CGS21680. However, K252a failed to block the effect of CGS21680 on NMDA toxicity (Fig. 3a, b), suggesting that the protective effects of the A2AR agonist do not depend on the activation of Trk receptors.

Figure 3.

The protective effect of the A2AR agonist CGS21680 on NMDA toxicity in R6/2 mice is not mediated by TrkB receptor transactivation. The effect of brain-derived neurotrophic factor (BDNF) on synaptic transmission in wild-type (WT) mice does not depend on the degree of NMDA toxicity. We tested the Trk inhibitor K252a (200 nM, 20 min of pre-treatment) on the ‘protective’ effect of CGS21680; K252a failed to block the effect of CGS21680 on NMDA toxicity (a, b) suggesting that the permissive role is exerted only in one direction, namely by A2A on TrkB receptors. In WT slices, the application of 100 μM NMDA induced a toxic effect as revealed by the only partial recovery at washout (51.49 ± 9.55% of basal, n = 5, c, d). BDNF 10 ng/mL was not able to potentiate nor to reduce such a toxic effect of NMDA in WT slices (58.47 ± 8.78% of basal, n = 5, c, d).*p < 0.05 vs NMDA alone according to Mann–Whitney U test.

To verify whether the protective effect of BDNF could depend on the degree of NMDA toxicity, we reproduced a toxic effect in WT slices by the application of 100 μM NMDA. Although the effect of NMDA was frankly toxic, as revealed by the only partial recovery at washout (51.49 ± 9.55% of basal, n = 5, Fig. 3c), BDNF still failed to be protective in WT slices (58.47 ± 8.78% of basal, n = 5, Fig. 3c, d).

BDNF protects cells against mutant huntingtin toxicity

We then used ST14A/Q120 cells, an in vitro model of HD, to test whether BDNF can rescue them from NMDA-induced toxicity. In a series of preliminary experiments, we established that ST14A/Q120 have to be differentiated for 48 h to obtain a morphological change comparable to ST14A/Q15 differentiated for 72 h (see Fig. 4). Cells were treated with NMDA at 10 mM and 3 mM in Q15 and in Q120 cells, respectively, for 10 min, to obtain a comparable toxic effect in the two genotypes (65.50 ± 6.50% of control, n = 3, and 64.50 ± 2.01% of control, n = 6, respectively; in both cases *p < 0.05 vs. control according to student's t-test; Fig. 5a, b). BDNF (100 ng/mL) pre-treatment had a significant protective effect on NMDA toxicity only in differentiated Q120 cells (97.00 ± 8.50% of control, n = 3, °p < 0.05 vs. NMDA alone according to student's t-test, Fig. 5b, e).

Figure 4.

Differentiation in Q15 and in Q120 cells. Differentiation was induced using hormonal factors as described in 'Materials and methods'. Pictures were taken 72 and 48 h after induction of differentiation in Q15 and Q120 cells, respectively. Cells developed numerous neurite-like appearing processes. Of note, there was a marked variability between neurons so that not all developed an identical morphology in both genotypes.

Figure 5.

Protective effect of brain-derived neurotrophic factor (BDNF) against mutant huntingtin toxicity in differentiated Q120 ST14A cells. In ST14/Q120 cells, the treatment with 3 mM NMDA induced a toxic effect (64.5 ± 2% of control, n = 6, *p < 0.05 vs. control, b) as evaluated using a trypan blue exclusion method. BDNF pre-treatment was able to reduce the toxic effect of NMDA (97 ± 9% of control, n = 3, °p < 0.05 vs. NMDA alone, b). Such an effect was blocked by K252a (200 nM) (n = 3, #p < 0.05 vs. BDNF+NMDA, c) but also by the A2AR antagonist ZM 241385 (100 nM) (n = 3, #p < 0.05 vs. BDNF+NMDA, c). BDNF was effective also at 50 ng/mL (n = 3, °p < 0.05 vs. NMDA alone, d), while a lower concentration (10 ng/mL) failed to be still protective against NMDA (d). The efficacy of BDNF in Huntington's disease cells is also clear, given the appearance of treated cells (e). In ST14A/Q15, the same toxic effect was obtained using 10 mM NMDA (65.50 ± 6.50% of the control, n = 3, *p < 0.05 vs. NMDA alone, a). In this case, the application of BDNF before the NMDA treatment was not able to significantly reduce the toxic effect of the drug (78.00 ± 3.00% of the control, n = 3, #< 0.05 vs. control but not vs. NMDA alone, a).

Such an effect was blocked by K252a (200 nM) and ZM 241385 (100 nM), confirming the involvement of both TrkB and A2A receptor (Fig. 5c). BDNF was effective also at 50 ng/mL, while a lower concentration (10 ng/mL) failed to protect against NMDA (Fig. 5d). It is interesting that BDNF had no beneficial effect on Q15 cells (78.00 ± 3.00% of control, n = 3, #p < 0.05 vs. control but not vs. NMDA alone, according to student's t-test, Fig. 5a). These data suggest that the protective effects of BDNF are specifically observed in presence of mutant huntingtin.

Discussion

BDNF, which exerts an important role on the preservation of medium-sized spiny neurons (Gavalda et al. 2004) and is highly involved in the pathogenesis of HD (Zuccato et al. 2001; Gauthier et al. 2004; del Toro et al. 2006), influences phosphorylation, activity, expression, and trafficking of NMDA receptors in different brain areas, including the striatum of HD mice (Caldeira et al. 2007; Crozier et al. 2008; Torres-Peraza et al. 2008). Moreover, the turnover of BDNF in the striatum of symptomatic R6/2 mice is reduced, as demonstrated by the decrease in pro-BDNF levels associated with no changes in the mature form of BDNF (Traficante et al. 2007). Furthermore, the activity of BDNF is strictly linked to the functionality of A2ARs, which exert both a facilitatory (Diógenes et al. 2004; Pousinha et al. 2006) and a permissive (Tebano et al. 2008) role on the neurotrophin effects. It is reported in literature that A2AR density is significantly decreased in HD since the early stage of the disease (Glass et al. 2000).

In agreement with previous reports showing a decrease in A2AR mRNA and A2AR density in the striatum of symptomatic R6/2 mice (Cha et al. 1999; Tarditi et al. 2006), in this study, we further demonstrate the down-regulation of the A2AR protein (Figure S3). In addition, an aberrant A2A receptor signaling has been found both in an in vitro model of the disease and in peripheral circulating cells from HD patients (Varani et al. 2001, 2003).

Thus, the role of BDNF toward NMDA-induced effects in HD mice, and the possible involvement of A2ARs, was worth considering.

In WT mice, BDNF clearly potentiated NMDA effects, worsening the synaptic recovery. This finding is in line with the view that BDNF generally facilitates NMDA-dependent effects and that physiological BDNF release may be necessary to allow a normal NMDA receptor transmission and plasticity (Pattwell et al. 2012). On the contrary, BNDF displayed a frankly protective effect in R6/2 mice, as revealed by the significant reduction of NMDA-mediated toxicity.

The different effect elicited by BDNF in the two genotypes was not related to changes in AMPAR-dependent basal synaptic transmission, as suggested by the similar I/O plot found in WT and R6/2 mice. Furthermore, the different outcome could not be ascribed to a different synaptic effect of BDNF on its own, as the neurotrophin failed to affect the FP amplitude in corticostriatal slices from both WT and R6/2 mice. This lack of efficacy was further confirmed in whole-cell patch-clamp recordings, which showed that BDNF did not influence excitatory currents in the striatum and has only a minimal effect on inhibitory current.

Alternatively, the attainment of a neuroprotective effect in the HD striatum might depend on the dysregulation of TrkB receptor isoforms that occurs under pathological conditions; in particular, excitotoxicity has been reported to down-regulate TrkB.FL protein levels and to inhibit their signaling activity in parallel with the up-regulation of TrkB.T isoforms, a shift that may favor the neuroprotective effects of BDNF (Gomes et al. 2012).

As already published by our group (Martire et al. 2010), the full-length form of TrkB (TrkB.FL) is not changed in the striatum of R6/2 mice with respect to WT littermates, and is also comparable in ST14A Q15 and Q120 cells (Figure S4). Even if it cannot be excluded that the truncated form of TrkB (TrkB.T) could influence in some way the signaling of TrkB.FL, in our experiments, the effects exerted by BDNF in both R6/2 and WT mice, as in Q120 cells, were blocked by K252a, which selectively inhibits the tyrosine kinase activity present only in TrkB.FL (Tapley et al. 1992).

However, further investigations about the role of TrkB.T and p75, the lower affinity BDNF receptor, are worth performing in future studies.

The main defect of K252a is that it is not selective for TrkB receptors, but is a membrane-permeable inhibitor of all Trk receptors (high-affinity neurotrophin receptors) tyrosine kinases; undesirable effects on other tyrosine kinases, at least at the concentration used in our experiments (200 nM), can be ruled out and thus K252a is still an effective blocking agent against the biological actions of BDNF and other neurotrophins (Knüsel and Hefti 1992; Tapley et al. 1992).

Moreover, in almost all our experiments, K252a has been used to block the effects induced by exogenous BDNF applied to slices or cells. Those effects were always fully blocked by K252a and then they can be reasonably considered as mediated by TrkB signaling. Unfortunately, a more selective chemical agent against TrkB receptor is not yet available.

In ST14A/Q120 cells, an in vitro model of HD, BDNF had a significant protective effect on NMDA toxicity, while it was ineffective on control (Q15) cells. Such an effect was blocked by both K252a and ZM 241385, showing the involvement of both TrkB and A2ARs. The levels of this latter receptor were found reduced in Q120 cells (Figure S4 and Varani et al. 2001).

Despite a comparable level of NR1 subunit in Q15 and Q120 cells (Figure S4), the profound changes that NMDA receptors undergo in models of HD might contribute to the shift in the effects of BDNF. For instance, we recently reported a reduction in the NR2A/NR2B ratio in the striatum of R6/2 mice (Martire et al. 2010), and it is well accepted that changes in subunit composition result in NMDA receptors with very different functional characteristics (Cull-Candy et al. 2001). Interestingly, in the present electrophysiological experiments, the opposite influence of BDNF toward NMDA-induced synaptic effects strongly resembles the similar pro-toxic and neuroprotective actions induced by the A2AR agonist CGS21680, respectively, in WT and R6/2 genotypes (Martire et al. 2007). This leads to suppose that A2ARs are involved in the dual modulatory role of the neurotrophin. Indeed, consistently with the well-known role of A2ARs in regulating BDNF levels and effects (Diógenes et al. 2004; Minghetti et al. 2007; Tebano et al. 2008), the opposite effect of exogenous BDNF on striatal NMDA toxicity was blocked by the A2AR antagonist ZM241385.

Moreover, considering that the stimulation of A2ARs can lead to TrkB receptor transactivation (Lee and Chao 2001), we wondered if the previously discovered genotype-dependent shift from the pro-toxic to the protective effects of CGS21680 (Martire et al. 2007) could be mediated by Trk receptors. This was apparently not the case, given that the Trk tyrosine kinase inhibitor K252a failed to block the beneficial effect of CGS21680 on NMDA toxicity in R6/2 mice. Thus, while BNDF effects require the activation of A2ARs, the effects of A2ARs on NMDA do not depend on Trk receptors functioning.

As the protective effects of BDNF were observed in conditions in which NMDA exerted toxic effects, we wanted to verify whether BDNF could become protective also in WT striata in presence of a toxic concentration of NMDA. This was clearly not the case, as in control striata, BDNF did not exert any protection toward a degree of NMDA toxicity comparable with that achieved in R6/2 mice. At the same time, however, BDNF lost its potentiating effects in a condition of stronger NMDAR stimulation. This seems in line with the above-reported observations that BDNF facilitates a “normal” NMDA receptor transmission.

In conclusion, our data show that BDNF exerts neuroprotective effects toward NMDA-dependent toxicity in experimental models of HD, that such effects of BDNF seem specifically related to the pathological genotype, and that they require endogenous A2AR activation.

Considering the pivotal role of NMDA receptors, BDNF, and A2ARs in the pathogenesis of HD, these data have obvious (potential) implications for therapeutics.

Acknowledgements

This work has been developed and supported through ISS intramural research grant N. 524/2011. We acknowledge Maria Teresa Tebano for her valuable comments on the manuscript. The authors declare no conflict of interest.

Ancillary