Introduction

Brain-derived neurotrophic factor (BDNF) is an endogenous glycoprotein belonging to neurotrophins, a family of signalling molecules that play a key role in regulating neuronal proliferation, differentiation and survival [1–3]. Among the neurotrophins, BDNF has the widest distribution in the central nervous system [4, 5]. In the adult hippocampus, BDNF is critically involved in the regulation of synaptic plasticity [6] and facilitates long-term potentiation (LTP) [7–10; see 11 for review]. Most biological effects of BDNF are mediated by the tyrosine kinase receptor TrkB. Both BDNF and its receptors are highly expressed in the hippocampus, and the activation state of the complex BDNF/TrkB appears critical for modulating synaptic efficacy [12] and the response to excitotoxic injury [13, 14].

In hippocampal neurones, adenosine has been reported to transactivate TrkB receptors, an effect involving the A2A receptor subtype (A2ARs) [15]. Furthermore, Diogenes et al. [16] reported that A2ARs facilitate the excitatory action of BDNF on hippocampal synaptic transmission.

Besides its involvement in hippocampal functions, BDNF is also very important for the survival of striatal neurons and the activity of corticostriatal synapses [17]. Impairment in BDNF function is though to play a major role in Huntington’s disease (HD), an inherited neurodegenerative disease caused by a mutation in the protein huntingtin and characterised by marked striatal degeneration (see [18] for review). Specifically, it has been shown that the activity of BDNF depends on the presence of normal huntingtin [19–21]. In vitro and in vivo data showed that wild-type huntingtin, but not the mutant protein, stimulates cortical BDNF production by acting at the level of Bdnf gene transcription [20, 21]. At the corticostriatal synapse, BDNF controls glutamate release and its exogenous administration allows striatal neurons to survive excitotoxin-induced neurodegeneration [22].

Mounting evidence also indicates an involvement of striatal A2A receptors in HD (see [23] and [24] for reviews). First, A2ARs are mainly localised on the neurons that degenerate early in HD [25]. Second, A2ARs and underlying signalling systems undergo profound changes in cellular and animal models of HD (see [24] for review). Thus, not only A2ARs seem to regulate BDNF functions in the brain, but both BDNF and A2ARs seem to be implicated in HD.

In this article we report and critically discuss some recent findings on the following issues obtained by our group: (1) the regulatory role exerted by hippocampal A2ARs on BDNF functions and levels as revealed by studies performed in A2AR KO mice, and (2) the effects of A2AR blockade on striatal BDNF levels in experimental models of HD or HD-like striatal degeneration.

In hippocampal slices from wild-type mice, the facilitatory effects exerted by BDNF on synaptic transmission require the endogenous activation of A2AR

In hippocampal slices from wild-type (WT) mice (400-μm thick, see [26] for a detailed methodological description), BDNF (10 ng/ml over 30 min) induced a long-lasting increase in the slope of field excitatory postsynaptic potential (fEPSP) recorded in the CA1 area (Fig. 1). This effect was abolished not only by the inhibitor of Trk phosphorylation K-252a, but also by the selective adenosine A2A antagonists ZM241385 and SCH58261 (Tebano et al., manuscript in preparation). The effects of A2ARs seem to involve the cyclic adenosine monophosphate/protein kinase A (cAMP-PKA) pathway, the main transduction mechanism operated by A2ARs, as the selective PKA inhibitor KT 5720 abolished the excitatory action of BDNF (Tebano et al., manuscript in preparation).

Fig. 1
figure 1

Brain-derived neurotrophic factor (BDNF) facilitates synaptic transmission in hippocampal slices from wild-type (WT) but not A2A receptor (R) knock-out (KO) mice. In hippocampal slices from WT mice, BDNF (10 ng/ml) induces an increase of the excitatory postsynaptic field potentials (fEPSP) slope, whereas a lower concentration (5 ng/ml) is ineffective. BDNF (5, 10 and 20 ng/ml) was totally ineffective in hippocampal slices from A2AR KO mice. *P < 0.05 vs. baseline (paired t test)

Full size image

The finding that BDNF enhances synaptic transmission in mice hippocampal slices confirms and strengthens previous data on the acute synaptic effects of BDNF at adult central synapses [17, 27]. Although the mechanisms responsible for such synaptic effects are not completely understood, they do not seem to be related to the neuroprotective ability of BDNF. That the above functions of BDNF recognise different molecular mechanisms is indicated by the fact that ligand-induced TrB translocation into lipid rafts is required for short-term synaptic modulation, but not neuronal survival, by BDNF [28]. Whatever the mechanisms, the blocking effect of ZM241385 and SCH58261 suggests that A2ARs have to be tonically activated by endogenous adenosine to manifest BDNF effects. To further explore the apparently “permissive” role of adenosine A2ARs on the synaptic effects of BDNF, A2AR KO mice were used.

BDNF is no longer able to facilitate synaptic transmission in hippocampal slices from A2A KO mice

A2AR KO mice (A2AR-/-) were generated, as previously described [29]. In hippocampal slices from A2AR KO mice, BDNF was no longer able to increase the fEPSP slope (Tebano et al., manuscript in preparation). As shown in Fig. 1, none of the tested concentrations of BDNF (5, 10 and 20 ng/ml) potentiated the synaptic response in slices from A2AR KO mice. Since the effectiveness of BDNF may depend on the proper expression of its receptors, we compared the levels of TrkB protein in WT and A2A KO mice by Western blotting. By using primary anti-TrkB antibody (Bioscience), we found that the levels of full-length TrkB isoforms were very similar in hippocampal extracts from WT and A2AR KO mice (Tebano et al., manuscript in preparation). Thus, the loss of BDNF activity observed in A2AR KO mice is not associated with a reduced density of TrkB receptors.

BDNF protein levels are significantly reduced in the brain of A2AR KO vs. WT mice

The BDNF protein levels were then measured on hippocampal extracts from A2AR KO and WT mice using the BDNF immunoassay system [enzyme-linked immunosorbent assay (ELISA) kit; Promega]. Interestingly, hippocampal BDNF levels were significantly lower in the A2AR KO mice compared with the WT littermates. As shown in Fig. 2, such an effect was not confined to the hippocampus, as a significant loss of BDNF levels was observed also in the striatum of A2A KO vs. WT mice.

Fig. 2
figure 2

Brain-derived neurotrophic factor (BDNF) levels are significantly reduced in the hippocampus and striatum of A2A receptor (R) knock-out (KO) mice. BDNF concentration was significantly reduced in extracts of hippocampal and striatal tissues prepared from A2AR mice compared with wild-type WT mice. The amount of BDNF (total) is expressed as pg/mg of wet tissue [means ± standard error of the mean (SEM) from five mice/group]. *P < 0.05 vs. WT

Full size image

This result indicates that the presence of A2ARs is required to maintain normal BDNF levels in the brain. Although the mechanism by which A2ARs regulate BDNF is yet undetermined, it is worth mentioning that the cAMP-PKA pathway (the main transduction system operated by A2ARs) has been implicated not only in “gating” BDNF functions [30], but also in modulating BDNF gene transcription [31] and release [32]. Furthermore, Kolarow et al. [33] showed that the PKA inhibitor Rp-cAMP significantly inhibited and delayed BDNF secretion in hippocampal neurons. Interestingly, the endogenous state of activation of A2ARs seems to be adequate to sustain a normal BDNF tone in WT mice, as no increase in the hippocampal levels of BDNF were observed in WT mice treated with a single (0.5 mg/kg i.p sacrifice 24 h later) or with repeated (0.5 mg/kg per day over 5 days) doses of the A2A agonist CGS 21680.

On their whole, the experiments performed in A2A KO mice indicate that A2ARs are essential for the tone and synaptic activity of BDNF. It should be noted, however, that the stimulation or the blockade of A2ARs can induce effects very different from those achieved after stimulation or blockade of the BDNF/TRkB system. For instance, whereas BDNF is thought to play an important role in memory processes [34, 35], adenosine A2ARs seem to negatively influence them. Indeed, an improvement in spatial memory has been reported in A2AR KO mice [36], whereas working memory was impaired in rats overexpressing A2ARs [37]. Another consideration that apparently weakens the importance of the A2AR/BDNF interaction is that A2ARs and TrkB Rs undergo opposite age-related changes in the hippocampus [38]. According to the authors’ interpretation, however, the relationship between age-related changes in the density of TrkB and A2AR receptors does play a role in the modulation of BDNF effects, thus even reinforcing the hypothesis of a close functional interplay between A2ARs and BDNF.

A2ARs regulate striatal BDNF levels in experimental models of HD or HD-like striatal degeneration

Two different models of HD were used: the quinolinic-acid (QA)-lesioned rats (pathogenetic model of HD-like striatal degeneration, [39]) and transgenic R6/2 mice (genetic HD model, [40]). For the method of QA-induced striatal lesions, see [41]. Briefly, QA (180 nmol/1 μl) was unilaterally injected in the striatum of anaesthetised rats, whereas vehicle [1 μl phosphate-buffered saline (PBS)] was injected in the other side. SCH58261 was administered at the doses of 0.01 and 1 mg/kg i.p. 20 min before QA lesion. The levels of BDNF were measured 24 h after surgery. As shown in Fig. 3, BDNF levels were significantly increased in the QA-lesioned vs. the control side. In animals pretreated with both doses of SCH58261 the rise in striatal BDNF was completely prevented. Worthy of note, SCH58261 markedly reduced BDNF levels in the control (vehicle-injected) side also (Fig. 3). As the early rise in BDNF levels most probably reflects a compensatory mechanism following QA injection, one can consider that SCH58261 may worsen QA effects by preventing BDNF increase. Previous studies have, however, demonstrated that this is not the case. We found, indeed, that whereas SCH58261 exerted very clear neuroprotective effects towards QA at 0.01 mg/kg i.p., it was no longer neuroprotective when administered at 1 mg/kg [39]. The finding that SCH58261 reduces BDNF levels irrespective of the dose (neuroprotective or non-neuroprotective) indicates that early changes in the striatal BDNF levels do not seem to correlate with the clinical and neuropathological outcome in the QA model.

Fig. 3
figure 3

The A2A receptor (R)-antagonist SCH58261 reduces striatal brain-derived neurotrophic factor (BDNF) levels in control and quinolinic-acid (QA)-lesioned rats. QA (180 nmol/1 μl) was unilaterally injected in the striatum of anaesthetised rats, whereas vehicle [1 μl phosphate-buffered saline (PBS)] was injected in the other side. SCH58261 was administered at the doses of 0.01 and 1 mg/kg i.p., 20 min before QA lesion. The levels of free mature BDNF were measured 24 h after surgery. BDNF levels were significantly increased in the QA-lesioned vs. the control side. In animals pretreated with both doses of SCH58261, the rise in striatal BDNF was completely prevented. N 5/group. °P < 0.05 vs. PBS. *P < 0.05 vs. corresponding vehicle

Full size image

R6/2 HD mice may represent a more suitable model to evaluate whether A2AR antagonists have a worsening influence through the reduction of striatal BDNF levels. Indeed, the administration of BDNF has been reported to have beneficial effects in transgenic models of HD [42, 43]. Considering that in the transgenic mice the neurodegenerative process progressively takes place over several weeks, to ascertain whether A2AR blockade could influence striatal BDNF levels, we decided to treat R6/2 mice chronically with SCH58261. The drug was administered at the dose of 0.01 mg/kg per day i.p. over 1 or 3 weeks starting from the fifth week of age. Another group of R6/2 mice was treated with vehicle over 3 weeks. BDNF levels were then assayed at the beginning of the symptomatic phase (8 weeks of age). Three groups of age-matched WT littermates were used as controls and treated in the same way described as above. As shown in Fig. 4, BDNF protein levels were not reduced, and showed instead a nonsignificant tendency to increase, in the striatum of early symptomatic R6/2 vs. WT mice. This finding was unexpected, as a significant decrease in BDNF mRNA has been recently reported in the striatum of R6/2 mice of the same age [43]. Although we have not explored the mechanisms responsible for such a discrepancy, a dissociation between the levels of BDNF protein and mRNA has already been observed in the brains of transgenic R6/1 HD mice [44]. Another reason for which one would expect a reduction in BDNF levels in the striatum of R6/2 mice is that a marked reduction in the expression of A2ARs (which, as mentioned above, are very important to maintain normal BDNF levels) was reported in these mice [45]. More recently, however, the A2AR-stimulated adenylyl cyclase, the A2AR density, and their affinity for the agonist CGS21680 were found unchanged in the striatum of clearly symptomatic R6/2 mice [46].

Fig. 4
figure 4

The A2A receptor (R)-antagonist SCH58261 reduces striatal brain-derived neurotrophic factor (BDNF) levels in R6/2 HD mice SCH58261 was administered at the dose of 0.01 mg/kg per day i.p. over 1 or 3 weeks starting from the fifth week of age. Free mature BDNF levels were then assayed at the beginning of the symptomatic phase (8 weeks of age). Three groups of age-matched wild-type (WT) littermates were used as controls and treated in the same way as described above. Both schedules (1 and 3 weeks) of SCH treatment significantly reduced BDNF levels in the striatum of either R6/2 and WT mice. N  4–7/group. *P < 0.05 vs. corresponding control group (two-tailed unpaired t test)

Full size image

As for the influence of A2AR blockade, both schedules (1 and 3 weeks) of SCH treatment significantly reduced BDNF levels in the striatum of either R6/2 and WT mice (Fig. 4). Again, however, such an effect does not necessarily imply a negative influence of SCH58261 on HD mice. Indeed, in recent experiments, we found that SCH treatment (0.01 mg/kg per day i.p. between the 5th and the 6th weeks of age) exerted some beneficial effects, namely, normalisation of emotional behaviour and restoration of a normal sensitivity to NMDA in corticostriatal slices (Domenici et al., manuscript in preparation). Thus, although the pharmacological blockade of A2ARs significantly reduces striatal BDNF levels in experimental models of HD or HD-like neurodegeneration, this does not seem to have a negative impact on the course of the disease. Of course, it is still possible that reducing BDNF levels at other time points of the pathological process has a much marked influence on the disease. This issue, as well as the evaluation of BDNF protein changes at different stages of the disease, will be the object of future studies.

Conclusions

Over the last several years, BDNF has emerged not only as a potent neuromodulator but also as a substance exerting fast excitatory actions in neurons, controlling resting membrane potential, neuronal excitability and synaptic transmission and participating in the induction of long-term changes in synaptic transmission [47]. In agreement with the recent evidence that A2ARs play a major role in regulating BDNF functions [16, 48], data reported in this article indicate that the presence and tonic activation of A2ARs are necessary to allow BDNF-induced potentiation of synaptic transmission. As for the mechanisms responsible for the permissive effects of A2ARs, data indicate an involvement of the cAMP-PKA pathway, the main transduction mechanism operated by A2ARs. The finding of reduced BDNF levels in the brain of A2AR KO mice indicates that A2ARs are critical not only to allow the synaptic effects of BDNF but also to maintain normal BDNF levels. Findings also point out that A2ARs tonically regulate BDNF levels in different areas of the brain, irrespective of their levels of expression. The tonic effect exerted by A2ARs in the regulation of BDNF levels is also evident in experimental models of HD (R6/2 mice) and of HD-like striatal degeneration (QA lesion in rats).

Given that BDNF delivery has beneficial effects in models of HD [42, 43], the possible functional consequences of reducing striatal BDNF levels in those models need to be clarified. In particular, further investigations on trophic factors and associated signal pathways are warranted before A2AR antagonists can be considered as a suitable therapeutic approach to HD [24].