Supraoptic nucleus (SON) neurons secrete oxytocin or vasopressin in response to various physiological stimuli (e.g., lactation/suckling, dehydration). Released near fenestrated capillaries of the neurohypophysis, these peptides enter the blood and travel to peripheral target organs. The pervasive neuromodulator adenosine, acting at A1 receptors, is an important inhibitory regulator of magnocellular neuroendocrine cell activity. Another high-affinity adenosine receptor exists in this system, however. We examined the physiological effects of adenosine A2A receptor activation and determined its localization among various cell types within the SON. In whole cell patch-clamp recordings from rat brain slices, application of the selective adenosine A2A receptor agonist CGS-21680 caused membrane depolarizations in SON neurons, often leading to increased firing activity. Membrane potential changes were persistent (>10 min) and could be blocked by the selective A2A receptor antagonist ZM-241385, or GDP-β-S, the latter suggesting postsynaptic sites of action. However, ±-α-methyl-(4-carboxyphenyl)glycine or TTX also blocked CGS-21680 effects, indicating secondary actions on postsynaptic neurons. In voltage-clamp mode, application of CGS-21680 caused a slight increase (∼8%) in high-frequency clusters of excitatory postsynaptic currents. With the use of specific antibodies, adenosine A2A receptors were immunocytochemically localized to both the magnocellular neurons and astrocytes of the SON. Ecto-5′nucleotidase, an enzyme involved in the metabolism of ATP to adenosine, was also localized to astrocytes of the SON. These results demonstrate that adenosine acting at A2A receptors can enhance the excitability of SON neurons and modulate transmitter release from glutamatergic afferents projecting to the nucleus. We suggest that adenosine A2A receptors may function in neuroendocrine regulation through both direct neuronal mechanisms and via actions involving glia.
magnocellular neurons of the supraoptic nucleus (SON) and paraventricular nucleus synthesize either oxytocin (OT) or vasopressin (VP), peptide hormones that participate in a variety of physiological phenomena including lactation, parturition, and water balance. OT and VP released within the neural lobe enter the general circulation and ultimately activate receptors on their target organs. These hormones, however, can be released from anywhere along the SON neuronal plasma membrane (38), and peptide release profiles from magnocellular dendrites do not necessarily parallel those observed from the neural lobe (22). Therefore, knowledge concerning endogenous regulators of magnocellular excitability contributes to the understanding of neuronal, and in this case, hormonal regulation.
Adenosine is a ubiquitous neuromodulator that tends to exert primarily inhibitory actions on synaptic transmission and transmitter release. With four known receptors, it is through adenosine's activation of A1 receptors that its suppressive effects are mostly executed. In addition to A1 receptors, adenosine A2A receptors have also been shown to be present in the SON (19, 39). The staining pattern reported within the SON in those studies, however, left unclear its exact location on various cellular elements, e.g., axon terminals, glia, postsynaptic neurons, and/or blood vessels. The A2A receptor's function in vasodilation and its presence on endothelial cells are well documented. In this regard, it is noteworthy that these magnocellular nuclei have the highest capillary densities found in the brain (31).
That A2A receptor activation can cause glutamate release from astrocytes has been reported (21). Evidence for astrocyte-neuron signaling in the SON has long suggested that the glial cells in this nucleus behave proactively in the regulation of neuronal excitability (2, 15, 34). Astrocytes of the SON are uniquely organized, with most having their somata in the ventral glial lamina, often sending long processes coursing dorsally to permeate the entire nucleus. Glutamate release from astrocytes has been shown to be capable of modulating neuronal activity (33), and the possible presence of A2A receptors on its astrocytes would suggest that this could be the case in the SON.
As previous studies have shown, adenosine A1 receptor activation hyperpolarizes the membrane and causes inhibitory effects on synaptic transmission (30, 35). Here we investigate the physiological effects and anatomical location of adenosine A2A receptors in the SON. We report that A2A receptor activation causes a small membrane depolarization that can lead to neuronal firing changes. We also report an A2A receptor-mediated increase in the frequency of spontaneous postsynaptic currents, and the immunolocalization of adenosine A2A receptors on astrocytes and neurons of the SON.
MATERIALS AND METHODS
All salts used in the perifusion media and whole cell recording solution were purchased from Fisher Scientific (Fair Lawn, NJ). Adenosine, the A2A agonist 2-p-(2-carboxyethyl)phenethyl-amino-5′-N-ethylcarboxamidoadenosine (CGS-21680), TTX, tetraethyl ammonium chloride (TEA), EGTA, HEPES, ±-α-methyl-(4-carboxyphenyl)glycine (MCPG), and MOPS were purchased from RBI/Sigma (St. Louis, MO). The A2A receptor antagonist ZM 241385 was purchased from Tocris (Ellisville, MO).
Slice preparation and procedures.
Slices were prepared similarly to those previously described (35), and procedures were in accordance with the University of California, Riverside, animal handling and use guidelines. Briefly, adult male Sprague-Dawley rats (45–70 days old) were decapitated, and the brains were rapidly removed and placed in ice-cold oxygenated artificial cerebrospinal fluid (aCSF) consisting of (in mM) 126 NaCl, 3 KCl, 1.25 NaH2PO4, 1.3 MgSO4, 10 d-glucose, 26 NaHCO3, 2.4 CaCl2, 5 MOPS, pH 7.4. The combination of this organic buffer and NaHCO3 has been found in earlier studies to better stabilize the pH over prolonged recording sessions than does the use of the bicarbonate buffer alone. Brains were then placed ventral side up, blocked for slicing, and glued to a specimen holder of a Vibratome. Gassed (95% O2/5% CO2), ice-cold aCSF bathed the brain block and coronal hypothalamic slices containing the SON (300–400 μm thick) were cut and placed in a bath of gassed aCSF for slice bisection and cropping. One hemislice was then placed in a recording chamber held at 34°C, the others in a holding chamber at room temperature.
Whole cell recording.
Methods were similar to those described previously (35). Briefly, patch electrodes (borosilicate, tip diameter: 1–2 μm) were pulled using a multistage pipette puller (model P-97; Sutter Instruments) and filled with a recording solution consisting of (in mM): 130 K+-gluconate, 8 KCl, 2 MgCl2, 10 HEPES, 0.4 EGTA, 2 K2-ATP, 1 Na2-GTP, and Alexa Fluor 488 (Molecular Probes Eugene, OR). The final pH and osmolality of the recording solution was 7.3 (adjusted with KOH) and 298 ± 3 mOsm/kg (adjusted with H2O). The final DC resistance of the pipettes was 3–4 MΩ.
Patch pipettes were guided to cells visualized under near-infrared differential interference contrast video microscopy (Leica DMLFSA equipped with a Dage IR-1000 camera), and GΩ seals were obtained. Brief suction resulted in the establishment of whole cell configuration, and the bridge circuitry of the amplifier (Axoclamp 2-B; Axon Instruments) was immediately engaged and optimized.
Drugs were applied by bath perifusion at a flow rate of 1.5 ml/min. At the close of each experiment, images of the Alexa Fluor 488-filled neuron were collected with a charged-coupled device camera using the microscope under epifluorescence illumination for morphological verification that the recorded cell was magnocellular. Images were stored on a personal computer, and only SON neurons having an observed diameter > 12 μM, were used in the study.
Adult male rats were deeply anesthetized with pentobarbital sodium and transcardially perfused with 4% paraformaldehyde. Coronal sections (thickness = 50 μm) were cut using a Vibratome and stored in 24-well culture plates. All immunocytochemical reactions were done in the wells. Sections were washed in 0.1 M PBS and treated with 0.3% Triton X-100 for 30 min. Following a PBS wash, sections were incubated in a 0.3% gelatin blocking buffer for 30 min. Primary antibodies used were against the adenosine A2A receptor (Alpha Diagnostics, San Antonio, TX), the OT-neuron identifying protein neurophysin I (NP I; Santa Cruz Biotechnology, Santa Cruz, CA), ecto-5′nucleotidase (Santa Cruz Biotechnology), and glial fibrillary acidic protein (GFAP; Sigma, St. Louis, MO and Santa Cruz Biotechnology). Incubation in primary antibody solutions was done at room temperature for 3 h with gentle agitation on an orbital shaker. All antibodies were diluted in blocking buffer and were combined as outlined in Table 1. Following four PBS washes, the sections were again incubated in blocking buffer for 30 min, followed by secondary antibody solutions (Table 1) for 2 h at room temperature. All secondary antibodies were obtained from Molecular Probes (Eugene, OR) and made in donkey. Control experiments with the absence of primary antibody were done in parallel with those including primary antibodies. Images of the sections were collected on a confocal microscope (Leica SP2) under sequential scan mode. The pinhole was set to 1 airy unit corresponding to an optical slice thickness of <300 nm, and images are presented as z-series projections.
All electrophysiological data were digitized at 10 kHz, filtered at 5 kHz, and collected for off-line analysis using pClamp 8 software (Axon Instruments). Statistical analysis using SigmaStat software consisted of the paired or unpaired Student's t-tests or their nonparametric equivalents. Minimal statistical significance was taken as P < 0.05.
A2A receptor activation depolarizes SON neurons.
As shown in Fig. 1, application of the selective adenosine A2A receptor agonist CGS-21680 caused a slight, but lasting depolarization of the membrane. This occurred over all concentrations tested (20 nM-1 μM) and often led to increased firing activity (Fig. 1, A and B). The depolarization could be blocked by delivery of 1 mM GDP-β-S into the postsynaptic cell through the patch electrode (n = 4) or by extracellular application of the selective A2A receptor antagonist ZM-241385 (n = 5, 1 μM). To obtain membrane potential values, recordings were low-pass filtered at 1 Hz using an 8-pole Bessel filter algorithm. This resulted in the removal of action potentials and fast synaptic potentials, leaving only the slower membrane processes intact. The average membrane potential occurring 3 min into CGS-21680 application (measured over 1 min) was compared with control values taken just before the application. A summary of changes in membrane potential observed under several conditions is presented in Fig. 1C. As shown in our previous work and in contrast to activating only A2A receptors, application of adenosine 50–100 μM caused a strong hyperpolarization (−3.64 mV, P < .01, n = 8). This adenosine-induced hyperpolarization prevailed even in the presence of selective A2A receptor activation (Fig. 1D). Additionally, unlike the inhibitory effects of adenosine that were repeatable in a given cell, only the first application of CGS-21680 induced the depolarization, which is consistent with the A2A receptor's rapid desensitizing characteristics (1, 24, 32). Therefore, each cell was exposed to only a single application of CGS-21680.
A2A receptor activation slightly increases EPSC clustering.
To investigate the effects of adenosine A2A receptor activation on afferent activity, voltage-clamp recordings were obtained from SON neurons. Clamping the membrane at either −35 or −60 mV reveals inhibitory postsynaptic currents (IPSCs) or excitatory postsynaptic currents (EPSCs), respectively. With the addition of 20 mM KCl, the number of events observed increased dramatically (from ∼5 to ∼20 Hz). The further addition of CGS-21680 (500 nM) did not cause a change in EPSC amplitude (−18.7 ± 1.4 vs. −17.0 ± 3.5 pA, P < 0.42, n = 6), nor was there a significant further increase in EPSC frequency (17.9 ± 7.7 vs. 20.7 ± 7.5 Hz, n = 6). However, activation of A2A receptors slightly, but significantly, increased the number of EPSC clusters (by +7.7 ± 2% over control period, P < 0.02, n = 6), where a “cluster” was defined as a minimum of five consecutive events occurring with a maximum interevent interval of 50 ms (Fig. 2). The same analysis applied to five cells voltage clamped at −35 mV, which revealed IPSCs, did not reveal an effect of A2A receptor activation on frequency, amplitude, or clustering of IPSCs.
Adenosine A2A receptor activation has been associated with increases in intracellular calcium. To test for an A2A receptor-mediated enhancement of voltage-gated calcium currents, cells were treated with 0.6 μM TTX + 5 mM TEA and given current pulses to evoke Ca2+ spikes. CGS-21680 did not change the amplitude or duration of evoked calcium spikes, nor did it change individual components of the action potential (Fig. 3). Several components were measured, including the spike peak, width, AHP amplitude, and descending slope. The effect of CGS-21680 on these action potential components for 6 cells is presented in Fig. 3B.
Action potentials and metabotropic glutamate receptors participate in the A2A-mediated depolarization.
Recent studies have suggested that metabotropic glutamate receptors are functionally linked to adenosine A2A receptors (6, 9). To investigate involvement of metabotropic glutamate receptors and TTX-sensitive channels in the adenosine A2A receptor-mediated depolarization, we bath-applied either the nonselective metabotropic glutamate receptor antagonist MCPG (200 μM, n = 5) or TTX (0.6 μM, n = 6) before CGS-21680 (500 nM). Consistent with metabotropic glutamate receptors playing a permissive role for A2A receptor activation (26), MCPG prevented CGS-21680 from depolarizing the SON neurons (Fig. 4). An earlier study in this system using TTX found no membrane depolarization in response to application of the natural ligand adenosine (17). Consistent with those results, no A2A-mediated depolarization was observed in the presence of TTX, suggesting an involvement of action potentials/Na+ channels in the observed depolarization.
Localization of adenosine A2A receptors and ecto-5′nucleotidase in the SON.
Immunocytochemical experiments using antibodies to the adenosine A2A receptor, GFAP, and NP I were performed. The A2A receptor antibody used was a rabbit polyclonal antibody made against the canine A2A receptor. The reactivity of this antibody to rat adenosine A2A receptors has previously been described (25), and it recognizes an intracellular portion of the protein's COOH terminus. In agreement with a previous report (39), A2A receptor immunoreactivity is apparent in the SON (Fig. 5A). Large puncta were seen along large-diameter blood vessels (diameter >12 μm) entering the brain. Within the SON parenchyma, the heaviest labeling occurred in the ventral glial lamina, followed by the somatic zone, with lighter A2A receptor immunoreactivity in the dendritic zone. Furthermore, ventral to the astrocytes of the ventral glial lamina, meningeal cells also expressed A2A receptor immunoreactivity. Dual immunolabeling of the astrocyte intermediate filament protein GFAP together with the adenosine A2A receptor revealed colocalization in both astrocytic somata and along the GFAP-positive processes that extend toward the interior of the nucleus (Fig. 5B). Acquisition of a confocal z-series on the immunoprocessed sections revealed the presence of A2A receptor immunoreactivity completely surrounded by GFAP labeling (Fig. 5D). Light A2A receptor immunoreactivity was also seen in magnocellular somata both positive and negative for neurophysin I immunoreactivity (Fig. 5C), consistent with the physiological data that the receptor is expressed by both OT and VP neurons.
There are two main sources of endogenous extracellular adenosine, one from bidirectional nucleoside transporters and the other from metabolism of ATP. In the SON, the nucleoside transporter inhibitor dilazep produced effects similar to adenosine A1 receptor activation, suggesting there exists a constant supply of extracellular adenosine and that nucleoside transporters are important in maintaining extracellular adenosine concentrations (35). Dendritic release of OT and VP from large dense-core vesicles, however, is a well-described phenomenon in this system, and ATP is also highly concentrated in secretory vesicles. Because we were interested in knowing whether breakdown of ATP might be contributing to extracellular adenosine levels, we investigated the presence of ecto-5′nucleotidase, an enzyme involved in the metabolism of ATP to adenosine. Strong immunoreactivity to the enzyme was observed in the SON (Fig. 6). Interestingly, some of the strongest immunoreactivity for ecto-5′nucleotidase colocalized with long GFAP-positive processes, suggesting an astrocytic source of this enzyme and ascribing another function to this dynamic cell type in this region (Fig. 6C).
Adenosine is a ubiquitous neuromodulator that exerts strong effects on several cell types of the magnocellular hypothalamic neuroendocrine system. Within this system, adenosine acting on A1 receptors induces astrocyte stellation (40), reduces Ca2+ currents (29), reduces PSC frequency (30), and hyperpolarizes SON neurons resulting in reduced spike firing (35). Although the presence of adenosine A2A receptors in this system has been known for several years before this report, direct effects of activating A2A receptors on SON neuronal physiology and their specific locations had not been investigated. The present evidence suggests that activation of SON adenosine A2A receptors results in a depolarization of SON neuronal membrane potential and slightly increases high-frequency excitatory postsynaptic currents. We also present evidence suggesting the presence of A2A receptors both on SON neurons and astrocytes and for astrocytic expression of ecto-5′nucleotidase.
A2A receptor modulation of PSCs.
Evidence for adenosine A2A receptor influence on transmitter release has been reviewed (8, 41). Studies have suggested that A2A receptor activation can result in enhanced levels of several neurotransmitters, including glutamate and GABA. It has been suggested that these effects may not be due to activation of terminally located A2A receptors but result from disinhibition of axon terminals (8).
Conversely, a study monitoring A2A receptor-mediated modulation of GABA release from synaptosomes concluded that activation of this receptor enhances GABA release (3). It has been shown that A2A receptor activation leads to an increase in Ca2+ levels within terminals (10), likely through an enhancement of N- and P-type Ca2+ currents (11). In this study, we did not address the specific location of A2A receptors on the glutamatergic afferents. It is interesting that in the SON, activation of adenosine A2A receptors did not significantly alter the overall frequency of PSCs; however, it did increase the occurrence of high-frequency EPSC clusters. High-frequency EPSCs are more apt to summate and, in turn, may bring the membrane potential above threshold. An enhancement of high-frequency EPSCs, but not IPSCs, by adenosine A2A receptor activation is consistent with 1) an adenosine A2A receptor-induced disinhibition of glutamatergic afferents and 2) a direct enhancement of N- and P/Q-type channels on glutamatergic neurons. Either process could produce a net A2A-mediated excitatory effect ultimately leading to increased firing activity.
A2A receptor functions in the SON.
Spontaneous firing is a characteristic of SON neurons both in vivo and in vitro. OT cells tend to fire intermittently or continuously until they become activated, at which point they can then undergo periodic high-frequency bursts (37). VP cells, on the other hand, fire phasically upon their activation by dehydration. These cells undergo periods of spike firing that are followed by quiescent periods, and this phasic cycle repeats. It has been suggested that both cell types regulate their own firing activity through autocrine or retrograde transmission via dendritic release of OT or VP. Dendritically released OT tends to reduce inhibitory synaptic transmission to OT neurons. Interestingly, this effect can be mimicked by exogenous application of either OT or adenosine and is blocked by antagonists for both OT and adenosine A1 receptors (4). These results suggest that retrogradely transmitted adenosine (or its precursor) is released along with OT from the dendrites of OT neurons.
ATP is known to be colocalized with a number of neurotransmitters and peptides (16, 20) in neurons that send projections to the SON (13). Along with its capacity to activate P2 receptors, ATP is also readily metabolized to adenosine in the extracellular environment (7, 42). Here we present evidence that the SON itself is endowed with sources of ecto-5′nucleotidase, one of the enzymes involved in the metabolism of ATP in the extracellular space. Glial processes were seen to be particularly immunoreactive for the enzyme. As the glia of the SON are especially influential on SON neuronal activity, these results suggest that they may also function to regulate extracellular levels of ATP and adenosine. SON astrocytes retract from between adjacent neurons and dendrites during activated states (14). This would tend to reduce their effectiveness in metabolizing ATP to adenosine under these conditions, and promote prolongation of ATP action.
Finding an adenosinergic influence on SON cells was not surprising. Adenosine is fast becoming known as a potent inhibitory neurotransmitter. However, that adenosine tends to so tightly regulate SON neuronal firing activity and synaptic input in this system is of considerable interest. Also, that both excitatory A2A receptors and inhibitory A1 receptors are capable of altering the physiological activity of SON neurons in an opposing manner suggests that adenosine may play an integral role in neuroendocrine function. Whereas activation of A1 receptors leads to a sustained hyperpolarization (35), activation of A2A receptors leads to a depolarization. Here, however, we show that under conditions in which both receptors are activated, the inhibitory influence of the A1 receptors can override the excitatory effect of A2A receptor activation.
A2A receptors and SON neuronal membrane depolarization.
A2A immunoreactivity on SON neurons combined with the ability of GDP-β-S to block the effects of the A2A agonist, CGS-21680 on membrane potential may suggest the observed excitatory effects on neurons were due to activation of postsynaptic A2A receptors. However, this effect could be blocked by both TTX and MCPG, suggesting an involvement of action potentials and metabotropic glutamate receptors. Basal activation of metabotropic glutamate receptors has been shown to be necessary for observing effects of A2A receptor activation (26). Blockade of the A2A receptor-mediated depolarization is consistent with this occurring in the SON. TTX blocks the propagation of action potentials and thereby reduces the likelihood and amount of glutamate released. MCPG blocks the direct action of glutamate on metabotropic receptors.
There are several potential sources of glutamate release in the mHNS, possibly including the SON neurons themselves (36). Additionally, there may be a glial-induced effect via A2A receptor activation on astrocytes. It has been shown that astrocytes can release glutamate nonsynaptically in response to A2A receptor activation (27, 28) and that astrocytes can contain vesicular glutamate transporters (VGLUTs) (23). Furthermore, the astrocytes of the ventral glial lamina play an integral role in SON neuronal excitability (12, 14) and are immunopositive for VGLUT type 3 (36) and taurine (5, 18). It is established that the astrocytes of the SON can release taurine in response to hypoosmotic changes. They may similarly release glutamate in response to activating stimuli. If A2A receptors are involved in this release, then the immunoreactivity seen along glial processes coursing through the SON suggests a nonsynaptic mechanism for intranuclear glutamate release, leading to excitation of those neurons near the release sites.
In this study, we report physiological effects of adenosine A2A receptor activation on SON neurons and their specific localization among SON astrocytes and neurons. Adenosine A2A receptor activation causes a depolarization of the neuronal membrane potential that can lead to increased firing activity and an increase in high-frequency EPSC clusters. These effects contrast with those seen following activation of adenosine A1 receptors, suggesting this widespread neuromodulator can play both an activating and an inhibiting role in the SON.
In conclusion, our data suggest that activation of adenosine A2A receptors results in the depolarization of magnocellular neurons of the SON and in a slight enhancement of high-frequency EPSC clustering. Metabotropic glutamate receptor activation plays a permissive role in this depolarization. Also, the previous anatomical findings of Rosin et al. (39) were extended by immunolocalization of the adenosine A2A receptor to both neurons and astrocytes of the SON. We further explored the source of endogenous adenosine, showing that the presence of an enzyme participating in its production, ecto-5′nucleotidase, was found throughout the SON and was particularly enriched within GFAP-labeled astrocytes of the ventral glial lamina. Together, these results suggest ATP-derived adenosine may activate postsynaptic A2A receptors leading to membrane depolarization and subsequent selectively enhanced excitability.
This study was supported by National Institute of Neurological Disorders and Stroke Grant NS-009140
We are grateful to Drs. Vladimir Parpura and Glenn Stanley for helpful input on a previous draft of the manuscript. We also thank Reno Reyes for technical assistance and the Center for Plant Cell Biology (UCR) for facilitating use of the Leica confocal microscope.
Present address of T. A. Ponzio: Laboratory of Neurochemistry, National Institutes of Health, Bldg. 49/5C68, Bethesda, MD 20892.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society