Abstract
Acid-sensing ion channels (ASICs) are proton-gated cation channels found in peripheral and central nervous system neurons. The ASIC1a subtype, which has high Ca2+ permeability, is activated by ischemia-induced acidosis and contributes to the neuronal loss that accompanies ischemic stroke. Our laboratory has shown that activation of σ receptors depresses ion channel activity and [Ca2+]i dysregulation during ischemia, which enhances neuronal survival. Whole-cell patch-clamp electrophysiology and fluorometric Ca2+ imaging were used to determine whether σ receptors regulate the function of ASIC in cultured rat cortical neurons. Bath application of the selective ASIC1a blocker, psalmotoxin1, decreased proton-evoked [Ca2+]i transients and peak membrane currents, suggesting the presence of homomeric ASIC1a channels. The pan-selective σ-1/σ-2 receptor agonists, 1,3-di-o-tolyl-guanidine (100 μM) and opipramol (10 μM), reversibly decreased acid-induced elevations in [Ca2+]i and membrane currents. Pharmacological experiments using σ receptor-subtype-specific agonists demonstrated that σ-1, but not σ-2, receptors inhibit ASIC1a-induced Ca2+ elevations. These results were confirmed using the irreversible σ receptor antagonist metaphit (50 μM) and the selective σ-1 antagonist BD1063 (10 nM), which obtunded the inhibitory effects of the σ-1 agonist, carbetapentane. Activation of ASIC1a was shown to stimulate downstream Ca2+ influx pathways, specifically N-methyl-d-aspartate and (±)-α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid/kainate receptors and voltage-gated Ca2+ channels. These subsequent Ca2+ influxes were also inhibited upon activation of σ-1 receptors. These findings demonstrate that σ-1 receptor stimulation inhibits ASIC1a-mediated membrane currents and consequent intracellular Ca2+ accumulation. The ability to control ionic imbalances and Ca2+ dysregulation evoked by ASIC1a activation makes σ receptors an attractive target for ischemic stroke therapy.
Acid-sensing ion channels are a class of ligand-gated channels that are members of the degenerin/epithelial sodium channel superfamily and are expressed in both peripheral and central nervous system neurons (Waldmann et al., 1997). Thus far, four genes (ASIC1–ASIC4) and two splice variants of ASIC1 and ASIC2 (a and b) have been cloned (Wemmie et al., 2006) that encode protein subunits that form functional proton-gated homomultimeric or heteromultimeric channels (Wemmie et al., 2006). The pH of half-maximal activation and the tissue expression patterns differ between each channel subtype.
One of the most common ASIC subtypes in the central nervous system (CNS) contains the ASIC1a subunit, which can form homomultimeric or heteromultimeric channels with ASIC2a (Askwith et al., 2004). These channels are activated by pH ≤ 7 and have a pH of half-maximal activation of ∼6.0 to 6.5 (Waldmann et al., 1997; Hesselager et al., 2004). ASIC1a homomultimeric channels differ from other ASIC subtypes in that they are highly permeable to both Na+ and Ca2+ ions (Waldmann et al., 1997; Yermolaieva et al., 2004). This ASIC subtype has been implicated in a number of physiological processes such as synaptic plasticity, fear conditioning, and learning and memory (Wemmie et al., 2002, 2003, 2004). ASIC1a has also been shown to be activated after cerebral ischemia and has been unequivocally linked to neuronal cell death (Xiong et al., 2004; Gao et al., 2005; Pignataro et al., 2007). Transgenic mice deficient in ASIC1a have reduced infarct size in response to middle cerebral artery occlusion relative to wild-type mice (Xiong et al., 2004). Moreover, pharmacological inhibition of ASIC1a with either amiloride or psalmotoxin1, which is selective for homomultimeric ASIC1a channels (Diochot et al., 2007), diminishes ischemic brain injury (Xiong et al., 2004). Several studies have suggested that Ca2+ influx through these channels is a key mechanism leading to neurodegeneration (Xiong et al., 2004; Yermolaieva et al., 2004).
Despite efforts to determine the function of ASIC1a and the role of these channels in ischemia, little is known about endogenous mechanisms that control ASIC1a activity. Thus far, only the NMDA receptor, acting via a calcium/calmodulin-dependent protein kinase II, has been shown to modulate ASIC1a (Gao et al., 2005). Activation of NMDA receptors enhances ASIC1a-mediated currents, which consequently exacerbates acidotoxicity during ischemia (Gao et al., 2005).
σ Receptor activation has been shown to modulate multiple cell membrane ion channels in neurons. σ-1 Receptors regulate ionotropic glutamate receptors and voltage-gated K+ channels, whereas σ-2 receptors modulate voltage-gated Ca2+ channels (Hayashi et al., 1995; Aydar et al., 2002; Zhang and Cuevas, 2002, 2005). The inhibition of ionotropic glutamatergic receptors by σ receptors prevents elevations in [Ca2+]i associated with glutamate-induced excitotoxicity (Klette et al., 1995). All of these voltage-gated ion channels and NMDA receptors have been shown to contribute to the demise of neurons during an ischemic insult. Our laboratory has recently shown that σ-1 receptors inhibit Ca2+ dysregulation evoked by ischemia and that activation of σ receptors is neuroprotective at delayed time points in a rat model of ischemic stroke (Ajmo et al., 2006; Katnik et al., 2006). Interstitial pH in the brain remains low several hours after an ischemic event (Nedergaard et al., 1991), and pharmacological blockade of ASIC1a by amiloride or psalmotoxin1 administered even 5 h after middle cerebral artery occlusion has been shown to diminish stroke injury (Simon, 2006). These observations raise the possibility that σ receptors may regulate ASIC1a function and ASIC1a-induced intracellular calcium transients and provide neuroprotection when stimulated at delayed time points after an ischemic insult.
Experiments were conducted to determine the effects of σ receptors on ASIC-mediated membrane currents and transient [Ca2+]i elevations. It was determined that σ receptor agonists inhibit acidosis-induced increases in [Ca2+]i and peak membrane currents in cells expressing homomeric ASIC1a channels. Pharmacological studies demonstrated that the σ-1 receptor subtype was responsible for these effects. Moreover, acidosis was also shown to activate downstream Ca2+ influx pathways [e.g., NMDA and AMPA/kainate receptors and voltage-gated Ca2+ channels (VGCC)], and activation of σ-1 receptors also diminished Ca2+ entry via these channels. Therefore, σ-1 receptors couple to ASIC1a channels to inhibit channel function and also block ASIC1a-induced [Ca2+]i dysregulation. Our findings suggest that σ-1 receptors represent potential targets for improving outcome of stroke injury and expanding the therapeutic window for ischemic stroke treatment.
Materials and Methods
Primary Rat Cortical Neuron Preparation. Primary cortical neurons from embryonic (E18) rats were cultured as previously described by our laboratory (Katnik et al., 2006). All procedures were done in accordance with the regulations of the University of South Florida Institutional Animal Care and Use Committee. Cells were used between 10 and 21 days in culture.
Calcium Imaging Measurements. The effects of acidosis on intracellular Ca2+ concentrations were examined in isolated cortical neurons using fluorescent imaging techniques. Cytosolic free Ca2+ was measured using the Ca2+-sensitive dye, fura-2, as we have described previously (Katnik et al., 2006). Cells, plated on poly-l-lysine-coated coverslips, were incubated for 1 h at room temperature in Neurobasal (Invitrogen, Carlsbad, CA) medium supplemented with B27 (Invitrogen) and 0.5 mM l-glutamine or in physiological saline solution (PSS) consisting of: 140 mM NaCl, 5.4 mM KCl, 1.3 mM CaCl2, 1.0 mM MgCl2, 20 mM glucose, and 25 mM HEPES (pH to 7.4 with NaOH), 330 ± 10 mOsm. Both solutions contained 3 μM acetoxymethyl ester fura-2 and 0.3% dimethyl sulfoxide. The coverslips were washed in PSS (fura-2 free) before experiments were carried out.
Electrophysiology Recordings. Neurons plated on glass coverslips (as described above) were transferred to a recording chamber mounted on a Zeiss Axiovert 200 and visualized at 400×. Membrane currents were amplified using an Axopatch 200B (Molecular Devices, Sunnyvale, CA), filtered at 1 kHz, digitized at 5 kHz with a Digidata 1322A (Molecular Devices), and acquired using Clampex 8 (Axon Instruments Inc.). Electrical access was achieved using the amphotericin B perforated-patch method to preserve intracellular integrity of neurons (Rae et al., 1991). An amphotericin B stock solution (60 mg/ml in dimethyl sulfoxide) was made daily, kept on ice, light protected, and diluted to 240 μg/ml (0.4% dimethyl sulfoxide) in control pipette solution immediately before patch-clamp experiments. The pipette solution consisted of the following: 75 mM K2SO4, 55 mM KCl, 5 mM MgSO4, and 25 mM HEPES (titrated to pH 7.2 with N-methyl-d-glucamine, 300 ± 5 mOsm). Patch electrodes were pulled from thin-walled borosilicate glass (World Precision Instruments, Inc., Sarasota, FL) using a Sutter Instruments P-87 pipette puller (Sutter Instrument Company, Novato, CA) and had resistances of 1.0 to 1.5 MΩ. Access resistance (Rs) was monitored throughout experiments for stable values ≤ 20 MΩ and were always compensated at 40% (lag, 10 μs). All cells were voltage-clamped at -70 mV.
Solutions and Reagents. The control bath solution for all experiments was PSS. In one series of experiments requiring high extracellular K+, an additional 34.6 mM KCl was isosmotically substituted for NaCl. All drugs were applied in PSS (or high-K+ PSS) using a rapid application system identical to that described previously (Cuevas and Berg, 1998). ASIC activation was induced by applying PSS with a pH of 6.0 (±drug) to specifically target ASIC1a (Askwith et al., 2004). Individual cells were exposed to no more than four low pH applications, and no rundown of the responses was observed with this protocol. All chemicals used in this investigation were of analytic grade. The following drugs were used: DTG, opipramol, ibogaine, metaphit, nifedipine, AP5, and PB28 (Sigma-Aldrich, St. Louis, MO); carbetapentane, BD1063, CNQX, and PRE-084 (Tocris Bioscience, Ellisville, MO); dextromethorphan (MP Biomedicals, Irvine, CA); psalmotoxin1 (Spider Pharm, Yarnelle, AZ); tetrodotoxin and thapsigargin (Alomone Labs, Jerusalem, Israel); amiloride (Alexis Corporation, Lausen, Switzerland); cadmium (Thermo Fisher Scientific, Waltham, MA); and fura-2 acetoxymethyl ester (Invitrogen).
Data Analysis. Analyses of measured intracellular Ca2+ and membrane current responses were conducted using Clampfit 9 (Axon Instruments Inc.). Fluorescence intensities were recorded from fura-2-loaded neuronal cell bodies. Time-lapse imaging data files collected with SlideBook 4.02 (Intelligent Imaging Innovations, Inc., Denver, CO) were converted to a text format and imported into Clampfit for subsequent analysis. Statistical analysis was conducted using SigmaPlot 9 and SigmaStat 3 software (Systat Software, Inc., San Jose, CA). Statistical differences were determined using paired and unpaired Student's t tests for within- and between-group experiments, respectively, and were considered significant if p < 0.05. For multiple group comparisons, either a one- or a two-way analysis of variance, with or without repeat measures, were used, as appropriate. When significant differences were determined with an analysis of variance, post hoc analysis was conducted using a Tukey test to determine differences between individual groups. For the generation of concentration response curves, data were best fit using a singlesite Langmuir-Hill equation.
Results
ASIC subtypes are distinguishable by pH sensitivity and ion selectivity, with homomultimeric ASIC1a being the only subtype that is Ca2+-permeable (Yermolaieva et al., 2004). Experiments were carried out to determine the effects of ASIC activation on intracellular Ca2+ transients and to identify the specific ASIC subtype(s) affecting [Ca2+]i in our rat cortical neuron model. Figure 1A shows representative traces of [Ca2+]i as a function of time recorded from a single neuron during acidosis (pH 6.0) in the absence (Control) and presence of amiloride (100 μM) and after a 10-min washout of drug (Wash). The general ASIC inhibitor, amiloride, reversibly blocked ASIC-mediated increases in [Ca2+]i by 88 ± 1% (Fig. 1B). Psalmotoxin1 (PcTx1) from the venom of the tarantula Psalmopoeus cambridgei has been shown to be a selective blocker of homomultimeric ASIC1a channels (Diochot et al., 2007). Figure 1C shows representative traces of [Ca2+]i as a function of time recorded from a neuron before (Control), after a 10 to 20-min preincubation in bath applied PcTx1 (500 ng/ml venom protein) and after washout of the toxin (Wash). In identical experiments, PcTx1 produced a statistically significant reversible decrease in ASIC1a-mediated elevations in [Ca2+]i (57 ± 2%) (Fig. 1D). The effects produced by this concentration of PcTx1 are consistent with results obtained for ASIC1a responses in mouse cortical neurons (Xiong et al., 2004). These data indicate that, in cultured cortical neurons from embryonic rats, acidosis results in elevations in [Ca2+]i that are mediated via the activation of homomultimeric ASIC1a channels, as reported for cultured cortical neurons from embryonic mice (Xiong et al., 2004).
Activation of σ receptors has been shown to inhibit numerous plasma membrane ion channels in neurons (Hayashi et al., 1995; Zhang and Cuevas, 2002, 2005). Therefore, experiments were carried out to study the effects of σ receptor activation on ASIC1a function using the pan-selective σ receptor agonist, DTG. Figure 2A shows representative traces of [Ca2+]i recorded from a single neuron during acidosis in the absence (Control) and presence of 100 μM DTG, a concentration previously shown to effectively regulate other ion channels (Zhang and Cuevas, 2002, 2005) and after 10-min washout of the drug (Wash). DTG rapidly and reversibly inhibited the low pH-induced transient increases in [Ca2+]i. In identical experiments, 100 μM DTG produced a statistically significant decrease (46 ± 3%) in ASIC1a-mediated elevations in [Ca2+]i (Fig. 2B). Opipramol (10 μM), another pan-selective σ-1/σ-2 agonist, also inhibited ASIC1a-mediated increases in [Ca2+]i by 57 ± 1% (Fig. 2B). The concentration-response relationship for DTG inhibition of ASIC1a was determined to confirm that the actions of DTG on ASIC1a are consistent with σ receptor activation. Increasing DTG concentrations resulted in further depression of ASIC1a-mediated increases in [Ca2+]i (Fig. 2C). A plot of the concentration-response relationship obtained from measurements made in multiple cells is shown in Fig. 2D. The data were best fit using a single-site Langmuir-Hill equation with an IC50 value of 109 μM and a Hill coefficient of 0.9. These values are consistent with the effects of DTG being mediated via activation of σ receptors (Zhang and Cuevas, 2002, 2005; Katnik et al., 2006). Taken together, these results suggest that σ receptors depress acid-induced increases in [Ca2+]i and modulate ASIC1a function.
σ Receptor subtype-selective agonists were used to identify the specific σ receptor subtype(s) mediating these effects. Representative traces of [Ca2+]i as a function of time recorded from two cells during acidosis in the absence (Control) and presence of the σ-1 selective agonists carbetapentane (CBP) and dextromethorphan (DEX) at the indicated concentrations are shown in Fig. 3, A and B, respectively. σ-1 Selective agonists blocked the low pH-induced elevations in [Ca2+]i in a concentration-dependent and reversible manner. Concentration-response plots for mean changes in peak [Ca2+]i recorded in identical experiments using the σ-1-selective ligands CBP, DEX, and PRE-084 are shown in Fig. 3C. The data were best fit using the Langmuir-Hill equation, and values obtained for IC50 and the Hill coefficients were 13.8 and 0.7 μM (CBP), 22 and 0.8 μM (DEX), and 13.7 and 0.6 μM (PRE-084), respectively. Presented for comparison is the best fit to the data obtained for CBP inhibition, via σ-1 receptors, of chemical ischemia-induced increases in [Ca2+]i in cortical neurons (dotted line, IC50 = 18.7 μM; Hill coefficient, 0.8) (Katnik et al., 2006). This curve superimposes on the responses to CBP observed in the current study.
To provide further evidence that σ-1 receptors modulate ASIC1a-induced Ca2+ elevations, experiments were conducted using the irreversible σ antagonist, metaphit, and a selective σ-1 antagonist, BD1063, in combination with the σ-1 agonist, CBP. Cells were exposed to acidosis in the absence and presence of CBP (30 μM), with or without preincubation in metaphit (50 μM; 30 min to 1 h, 23°C). CBP decreased the acid-induced elevations in [Ca2+]i in control cells, and this inhibition was lessened by preincubation with metaphit (Fig. 4A). Figure 4B shows results from several similar experiments determining percentage inhibition of acid-induced increases in [Ca2+]i observed in the presence of CBP in control neurons (Control) and neurons preincubated with metaphit (MET). Although CBP decreased the elevations in [Ca2+]i evoked by acidosis in control cells by 52 ± 2%, the σ-1 receptor agonist only reduced the response by 30 ± 3% in cells preincubated in the irreversible σ receptor antagonist. This >40% decrease in the effects of CBP after metaphit preincubation was statistically significant (p < 0.001). The selective σ-1 antagonist, BD1063, showed more pronounced effects. Figure 4C shows representative traces of [Ca2+]i as a function of time recorded from two neurons during acidosis in absence (Control) and presence of CBP (30 μM), without (PSS) and with coapplication in BD1063 (10 nM). In identical experiments, the σ agonist CBP completely loses its ability to block the low pH-induced elevations in [Ca2+]i in the presence of the σ antagonist BD1063 compared with a 40% block in control cells (Fig. 4D). Inhibition of σ-1 receptors by BD1063 significantly blocked the effects of CBP on ASIC1a-mediated [Ca2+]i elevations compared with control cells (p < 0.001). These data confirm the effects of σ-1 agonists on ASIC1a-mediated increases in [Ca2+]i are the result of these compounds acting on σ-1 receptors.
Further experiments were carried out using the σ-2-selective ligands ibogaine and PB28 to determine whether the σ-2 receptor subtype also affects ASIC1a function. Figure 5A shows representative traces of [Ca2+]i as a function of time recorded from two cells during acidosis in the absence (Control) and presence of ibogaine (IBO; left traces) and PB28 (right traces) at the indicated concentrations. The σ-2 ligands inhibited acidosis-evoked increases in mean peak changes in [Ca2+]i in a concentration-dependent manner (Fig. 5B). Best fits to the data demonstrated that application of ibogaine and PB28 resulted in inhibition of ASIC1a-mediated increases in [Ca2+]i with IC50 values of 69 and 11 μM and Hill coefficients of 0.86 and 0.85, respectively. For comparison, the best fit to the data obtained for ibogaine inhibition of ICa-induced increases in [Ca2+]i is presented, which has been shown to be mediated by σ-2 receptors (dashed line; IC50 = 31 μM; Hill coefficient, 1.1) (Zhang and Cuevas, 2002). Unlike the similar concentration-response relationship observed for CBP inhibition of ASIC1a and ischemia responses, there is a discrepancy between the ibogaine block of responses mediated by ASIC1a and by VGCCs. To determine whether the effects of these σ-2 ligands were mediated by activation of σ-2 receptors, experiments using metaphit were carried out. PB28 (20 μM) produced an inhibition of ASIC1a-mediated increases in [Ca2+]i in both control (PSS) and metaphit (+MET)-treated cells (Fig. 5C). PB28 blocked ASIC1a-mediated increases in [Ca2+]i by 67 ± 10 and 60 ± 2% in the absence and presence of metaphit preincubation, respectively (Fig. 5D). The ∼10% reduction in the effects of PB28 produced by metaphit was not statistically significant (p = 0.64). Furthermore, BD1063 (10 nM) failed to block the effects of PB28 (20 μM) on ASIC1a-induced elevations in [Ca2+]i (data not shown). These results demonstrate that the effects of PB28 on ASIC1a-mediated increases in [Ca2+]i are not mediated by σ-2 receptors because the inhibition of these responses to acidosis by PB28 is metaphit-insensitive and occurs at concentrations inconsistent with σ-2 receptor activation.
The activation of σ receptors has been shown previously to directly affect Ca2+ release from intracellular stores (Cassano et al., 2006). Thus, experiments were conducted to resolve whether σ receptor activation reduces acid-induced increases in [Ca2+]i, in part via the inhibition of calcium-induced calcium release from the endoplasmic reticulum, triggered by Ca2+ influx through the plasma membrane. For these experiments, thapsigargin was used to block the sarcoplasmic/endoplasmic reticulum Ca2+-ATPase, which results in depletion of both ryanodine- and IP3-sensitive stores. Figure 6A shows representative traces of [Ca2+]i as a function of time recorded from a neuron during acidosis in the absence (Control) and presence of 100 μM DTG, whereas Fig. 6B shows traces in the absence and presence of DTG after preincubation (1 h, 23°C) in 10 μM thapsigargin (THAP and THAP + DTG, respectively). Thapsigargin alone did not decrease the elevations in [Ca2+]i produced by acidosis, and DTG depressed the increases in [Ca2+]i under both conditions (±thapsigargin preincubation). Analysis of the data collected in identical experiments indicates that preincubation with thapsigargin does not significantly alter the effects of DTG on acid-mediated increases in [Ca2+]i (Fig. 6C). Thus, DTG does not decrease the low pH-induced elevations in [Ca2+]i by affecting release of calcium from intracellular stores. The fact that depletion of intracellular stores fails to depress increases in [Ca2+]i evoked by acidosis suggests that Ca2+ influx through the plasma membrane accounts for most, if not all, of the increases in [Ca2+]i evoked by ASIC1a activation.
Simultaneous Ca2+ fluorometry and whole-cell patch-clamp recordings were performed to study ASIC1a-mediated membrane currents and to determine how much of the observed Ca2+ influx is due to Ca2+ entry through ASIC1a channels. Cells were voltage-clamped at -70 mV to minimize NMDA receptor and VGCC activation and, thus, isolate ASIC1a currents. Figure 7A shows that ASIC1a stimulation by low pH solution (pH 6.0) resulted in a small intracellular Ca2+ transient measured in a patched cell ([Ca2+]i V-Clamp). In contrast, a second cell ([Ca2+]i Control) in the same field of view, which was not voltage clamped, had a significantly larger increase in [Ca2+]i. Acidosis also resulted in a large inward current in the patched cell (Fig. 7A, inset). The cumulative results from several similar experiments demonstrate that ASIC1a activation of neurons voltage clamped at -70 mV resulted in elevations in [Ca2+]i that were an order of magnitude smaller than changes evoked in cells that were not voltage-clamped (Fig. 7B). These data confirm ASIC1a activation results in minimal Ca2+ influx through the ASIC1a channel itself and that the majority of the acid-induced [Ca2+]i increases are mediated by downstream Ca2+ influx pathways.
It was noted that in cortical neurons held under current-clamp mode, application of protons evokes a rapid depolarization that, unlike the case in voltage-clamped neurons, was associated with pronounced elevations in [Ca2+]i (data not shown). Thus, ASIC1a channels are probably promoting significant Ca2+ influx into the neurons by depolarizing the cells. To mimic this change in membrane potential evoked upon ASIC1a activation, cells were exposed to high-K+ (40 mM) extracellular solution. Figure 7C shows representative [Ca2+]i traces recorded in response to high- application in the absence (Control) and presence of CBP (100 μM, +CBP). Depolarizing the neurons in this manner evoked robust [Ca2+]i elevations that were blocked by addition of CBP. In identical experiments, CBP reduced the high -evoked increases in [Ca2+]i by 83 ± 1%, and this decrease was statistically significant (p < 0.001) (Fig. 7D). Thus, σ receptor activation inhibits Ca2+ channels downstream of ASIC1a in addition to the proton-gated channels themselves.
To determine the specific ion channels contributing to the ASIC1a-induced [Ca2+]i influxes, several inhibitors of plasma membrane ion channels were used. Activation of ASIC1a depolarizes these neurons, which could stimulate action potential firing and, consequently, synaptic transmission, both of which may elevate [Ca2+]i. Thus, tetrodotoxin (TTX; 500 nM) was used to inhibit voltage-gated Na+ channels to prevent the genesis of action potentials. Figure 8A shows that application of TTX inhibited ∼10% of the acid-induced increases in [Ca2+]i but did not significantly affect CBP (50 μM) modulation of the Ca2+ responses, compared with the PSS group. Because TTX had no effect on the σ-1/ASIC1a interaction, all subsequent experiments included 500 nM TTX in the bath solutions to prevent spontaneous action potentials from contributing to the measured increases in [Ca2+]i. Inhibition of NMDA receptors by AP5 (100 μM) significantly reduced the acid-induced elevations in [Ca2+]i, and coapplication with CBP (50 μM) resulted in a further decrease in [Ca2+]i that was statistically significant (Fig. 8A). These results show that ∼30% of ASIC1a-induced [Ca2+]i increases are dependent on NMDA receptor activation. Furthermore, because the effects of AP5 and CBP are less than additive, activation of σ-1 receptors depresses ASIC1a-evoked increases in [Ca2+]i, in part by blocking this NMDA receptor-dependent component. Blockade of L-type VGCCs by nifedipine (10 μM) significantly inhibited (>75%) acid-induced [Ca2+]i elevations, and CBP (50 μM) continued to provide further blockade of the residual increases in [Ca2+]i (Fig. 8A). Cadmium (100 μM), a broad-spectrum blocker of plasma membrane calcium channels, inhibited ∼90% of the acid-induced increases in [Ca2+]i, and CBP (50 μM) had no effect on the remaining Ca2+ influx (Fig. 8A). Thus, voltage-gated Ca2+ channels either directly (influx through the channel) or indirectly (facilitating glutamate release) account for most of the [Ca2+]i increases resulting from ASIC1a activation, and σ-1 receptor activation provides an inhibition similar to that observed with Cd2+.
Further experiments were conducted to determine whether AMPA/kainate receptors are involved in the [Ca2+]i elevations elicited upon ASIC1a activation. Figure 8B shows relative changes in [Ca2+]i in the absence (Control) and presence of the AMPA/kainate receptor blocker CNQX (10 μM) and the effects of CBP at two different concentrations (50 and 300 μM). Maximal blockade of AMPA/kainate receptors alone produced ∼40% reduction of the elevations of [Ca2+]i, and 50 μM CBP did not provide additional, statistically significant block. However, increasing the CBP concentration to 300 μM did produce an additional block of the Ca2+ response on top of the effects of CNQX. Figure 8B shows that coapplication of 300 μM CBP and CNQX blocked >80% of the acid-induced increases in [Ca2+]i, and this result was statistically significant compared with CNQX alone. Thus, [Ca2+]i elevations elicited upon ASIC1a activation involve Ca2+ influx through both ASIC1a and other Ca2+-permeable plasma membrane ion channels expressed in cortical neurons. More importantly, it is also evident that activation of σ-1 receptors results in depression of Ca2+ influx through all of these sources, either by inhibiting ASIC1a or the downstream channels themselves.
Two ASIC blockers, amiloride and PcTx1, were used to confirm that the acid-activated inward currents observed in our cortical neuron model were mediated by ASIC1 channels. Figure 9A shows representative traces of membrane currents as a function of time recorded from a single cell in the absence (Control) and presence of amiloride (100 μM, top traces, i), and from a different neuron in the absence (Control) and presence of PcTx1 (500 ng/ml venom, bottom traces, ii). Figure 9B summarizes the normalized peak proton-gated whole-cell currents recorded from several cells in the absence (Control) and presence of amiloride and PcTx1. Amiloride produced a 79 ± 6% inhibition of acid-activated currents, whereas 500 ng/ml PcTx1 containing venom produced a 42 ± 11% reduction in this current. Both reductions were significantly different from control (p < 0.01). Experiments were also carried out to confirm that ASIC1a currents were not affected by the channel blockers used in the imaging studies described above. Figure 9C shows relative peak proton-gated currents, normalized to control (PSS), during acidosis in the absence (PSS) and presence of the indicated drugs. These data demonstrate that TTX, AP5, CNQX, nifedipine, or cadmium do not have any direct effects on ASIC1a channels.
σ Receptor activation has been shown to modulate multiple cell membrane ion channels in neurons that are activated after ASIC1a stimulation (Hayashi et al., 1995; Zhang and Cuevas, 2002). This fact raises the possibility that σ-1 receptors only couple to the secondary events that result from ASIC1a activation and not to ASIC1a itself. To investigate whether σ receptors affect ASIC1a function, whole-cell patch-clamp recordings were performed in the presence of σ agonists. Figure 10A shows representative membrane current traces as a function of time recorded from a cell in the absence (Control) and presence of the pan-selective σ agonist DTG (100 μM). Results from several similar experiments show that activation of σ receptors with DTG inhibited ASIC1a-mediated currents by 53 ± 9%, and this decrease was statistically significant (Fig. 10B). To determine whether the effects of DTG on ASIC1a currents were mediated by σ-1 or σ-2 receptors, the effects of the σ-1 selective agonist CBP, on whole-cell currents were measured. Figure 10C shows representative membrane current traces as a function of time recorded in the absence (Control) and presence of CBP (50 μM). After multiple experiments, analysis of the measured current densities showed that control cells had statistically significant larger current densities than CBP-treated cells (Fig. 10D). In the presence of 50 μM CBP, ASIC1a-mediated currents were decreased by 30 ± 5% relative to control. In contrast, PB28 (20 μM) failed to inhibit ASIC1a-mediated membrane currents, suggesting σ-2 receptors do not regulate ASIC1a function (data not shown). Therefore, these results demonstrate that σ-1 receptors functionally couple to ASIC1a channels in cortical neurons and to targets downstream of ASIC1a activation.
Discussion
The principal finding reported here is that activation of σ receptors depresses membrane currents and elevations in [Ca2+]i mediated by ASIC1a channels in cortical neurons. The pharmacological properties of the receptor involved are consistent with the effects being specifically mediated by the σ-1 receptor subtype. Furthermore, most of the elevations in [Ca2+]i triggered by acidosis are the result of Ca2+ channels opening downstream of ASIC1a activation. Stimulation of σ-1 receptors effectively suppressed these secondary Ca2+ fluxes both by inhibiting ASIC1a and the other channels directly.
ASIC are regulated by various factors such as pH, membrane distention, and arachidonic acid and, therefore, function as signal integrators in the CNS (Allen and Attwell, 2002; Lopez, 2002). All of these factors elicit or potentiate ASIC-mediated responses. Information on endogenous mechanisms that inhibit ASIC function is lacking. It has been shown that NMDA receptors modulate ASIC1a function via the activation of a calcium/calmodulin-dependent protein kinase II signaling cascade, but activation of this pathway results in an increase in currents through ASIC1a (Gao et al., 2005). Thus, our finding that activation of σ receptors depresses ASIC1a-mediated responses is novel. Our conclusion that the responses observed are mediated specifically by ASIC1a is supported by the inhibition produced with the selective ASIC1a channel blocker, PcTx1 (Diochot et al., 2007), and that cultured cortical neurons from embryonic mice deficient in the ASIC1a subunit fail to show increases in [Ca2+]i or membrane currents at the proton concentrations used here (Xiong et al., 2004). ASIC2a and ASIC2b subunits are also expressed in the CNS, but homomeric ASIC2a channels are activated below pH 5.5, and ASIC2b does not generate currents in response to low pH (Lingueglia et al., 1997). Furthermore, neither homomeric ASIC2a nor heteromultimeric ASIC1a/ASIC2a channels conduct Ca2+ and thus could not account for the changes in [Ca2+]i observed here (Yermolaieva et al., 2004).
Results from Ca2+ imaging experiments suggest that it is specifically the σ-1 receptor subtype that modulates neuronal responses to ASIC1a activation. Studies have shown that the affinity of carbetapentane for σ-1 receptors is >50-fold greater than for σ-2 receptors (Rothman et al., 1991; Vilner and Bowen, 2000). The calculated IC50 for carbetapentane inhibition of ischemia-evoked increases in [Ca2+]i via σ-1 receptor activation is 18.7 μM (Katnik et al., 2006), which is comparable with the 13.8 μM IC50 for CBP inhibition of ASIC1a-induced [Ca2+]i increases. Carbetapentane also inhibits epileptiform activity in rat hippocampal slices via σ-1 receptors with an IC50 value of 38 μM (Thurgur and Church, 1998). Likewise, we show that the σ-1 agonists dextromethorphan (IC50 = 22 μM) and PRE-084 (IC50 = 13.7 μM), both of which have >100-fold greater affinities for σ-1 than σ-2 receptors, block ASIC1a-mediated responses at concentrations consistent with those reported in the literature. Dextromethorphan inhibits spreading depression in rat neocortical brain slices with an IC50 ∼ 30 μM (Anderson and Andrew, 2002), whereas PRE-084 protects human retinal cells against oxidative stress with an IC50 ∼ 10 μM (Bucolo et al., 2006). The fact that IC50 values determined here for carbetapentane, dextromethorphan, and PRE-084 are in the low micromolar range suggests that it is unlikely these agonists are affecting ASIC1a activity via σ-2 receptors because high micromolar to millimolar concentrations of these compounds are required to stimulate σ-2 receptors. Moreover, σ-2 selective agonists failed to inhibit ASIC1a-mediated responses at concentrations consistent with σ-2-specific effects.
The strongest evidence that σ-1 receptor activation modulates ASIC1a comes from experiments using the σ antagonists, metaphit and BD1063. Metaphit has been shown to bind irreversibly to σ-1 receptors with an IC50 value of 50 μM (Wu et al., 2003). Preincubation in metaphit blocks σ-1 receptor-mediated modulation of voltage-gated K+ channels in intracardiac neurons and depression of ischemia-induced elevations in [Ca2+]i in cortical neurons (Zhang and Cuevas, 2005; Katnik et al., 2006). Preincubation of cortical neurons in 50 μM metaphit antagonized CBP inhibition of ASIC1a by ∼40%. BD1063 has been shown to have a higher affinity for σ-1 than σ-2 receptors and attenuates the dystonia produced by DTG in rats in a dose-dependent manner, suggesting this ligand acts as an antagonist at σ sites (Matsumoto et al., 1995). Here, we show that CBP is unable to block acid-induced increases in [Ca2+]i when coapplied with BD1063. In addition, we found that metaphit fails to inhibit the effects of the σ-2 agonist, PB28, on ASIC1a-mediated responses. Taken together, these data show that increases in [Ca2+]i in response to ASIC1a activation are modulated only by σ-1 receptors.
Several studies have suggested that Ca2+ influx through ASIC1a channels is a key mechanism leading to cell death (Xiong et al., 2004; Yermolaieva et al., 2004). Depletion of Ca2+ from intracellular stores indicates that most, if not all, of the acid-induced increases in [Ca2+]i is due to plasma membrane influx. However, our results show that multiple ion channels downstream of ASIC1a activation contribute to acidosis-induced elevations in [Ca2+]i, including NMDA and AMPA/kainate receptors and VGCCs. The activation of NMDA and AMPA/kainate receptors after ASIC1a stimulation was observed even when neuronal conduction was inhibited with tetrodotoxin. This observation suggests a presynaptic localization of ASIC1a, whereby activation of the channel by protons results in synaptic transmission and subsequent activation of postsynaptic glutamatergic receptors. Consistent with this hypothesis, ASIC1a has been found to regulate neurotransmitter release probability in mouse hippocampal neurons (Cho and Askwith, 2008).
σ Receptors have been identified in both presynaptic and postsynaptic sites (Gonzalez-Alvear and Werling, 1995; Alonso et al., 2000) and thus may modulate channels in both regions. In the presence of specific inhibitors of ionotropic glutamate receptors, activation of σ-1 receptors with CBP further decreased proton-evoked increases in [Ca2+]i, but the effects of CBP and the glutamate channel inhibitors were less than additive. Thus, σ-1 receptors also inhibit Ca2+ entry via NMDA and AMPA/kainate receptors directly by inhibiting these channels and indirectly by depressing ASIC1a activation. Application of the L-type VGCC inhibitor, nifedipine, and the broad-spectrum Ca2+ channel inhibitor, cadmium, blocked ASIC1a-induced increases in [Ca2+]i by >70 and >90%, respectively. This observation indicates that most of the increases in [Ca2+]i produced upon ASIC1a activation are dependent on Ca2+ influx through VGCCs. Coapplication of CBP with nifedipine, but not with Cd2+, resulted in further reduction in the proton-evoked increases in [Ca2+]i. The conclusion that Ca2+ influx through ASIC1a channels itself contributed only a small fraction to the total observed [Ca2+]i increases was confirmed with simultaneous Ca2+ fluorometry and whole-cell patch-clamp recordings. Cells voltageclamped at -70 mV, which prevents NMDA receptor and VGCC activation, demonstrated minimal acid-evoked elevations in [Ca2+]i. Taken together, our results show that the increases in [Ca2+]i evoked by ASIC1a activation are the result of synaptic transmission and subsequent opening of multiple Ca2+ channels and that stimulation of σ-1 receptors down-regulates all of these events. However, the fact that activation of σ-1 receptors depressed ASIC1a-mediated currents in cells voltage-clamped at -70 mV indicates that σ-1 receptors are functionally coupled to ASIC1a and that the depression in acid-evoked increases in [Ca2+]i is not exclusively the result of σ-1 receptors blocking channels downstream of ASIC1a.
The finding that σ-1 receptors can inhibit ASIC1a channels has significant physiological and pathophysiological implications. It has been proposed that ASIC1a activation may facilitate neurotransmission by compensating for the decrease in excitatory neurotransmission caused by direct inhibition of postsynaptic Na+ and Ca2+ channels by protons that are released during exocytosis (Krishtal et al., 1987; Zha et al., 2006). Furthermore, the expression levels of ASIC1a have direct effects on the density of dendritic spines in hippocampal neurons (Zha et al., 2006). Thus, σ-1 receptors may influence cell-to-cell signaling in the CNS by affecting ASIC1a activity. One of the consequences of ASIC1a overexpression in mice is enhanced fear conditioning (Wemmie et al., 2004), whereas stimulation of σ-1 receptors is known to ameliorate conditioned fear stress (Kamei et al., 1997). These observations, coupled with our current report, suggest that σ-1 receptor activation may produce anxiolytic effects via the inhibition of ASIC1a channels.
The inhibition of ASIC1a by σ-1 receptors is a potential component of the neuroprotective properties of σ receptors because activation of ASIC1a has been shown to contribute to stroke injury (Xiong et al., 2004). More importantly, inhibition of ASIC1a has been shown to be neuroprotective at delayed time points after ischemic stroke (Simon, 2006). Thus, σ-1 receptor-mediated inhibition of ASIC1a may contribute to the enhanced neuronal survival after σ receptor activation 24 h poststroke in rats (Ajmo et al., 2006). Furthermore, our data suggest that activation of ASIC1a stimulates the activity of NMDA and AMPA/kainate receptors and VGCCs, all of which have been linked to ischemia-induced brain injury. Thus, σ-1 receptor activation may provide further neuroprotection by reducing the activity of these channels, which occurs subsequent to ASIC1a stimulation. Consistent with this pleiotropic effect of σ-1 receptors is our observation that σ-1 receptor activation suppressed extracellular high-K+-induced increases in [Ca2+]i, which would also activate these downstream effectors. In conclusion, σ-1 receptors inhibit ASIC1a channel function and blunt acidosis-evoked ionic fluxes and increases in [Ca2+]i. Thus, σ-1 receptors can be targeted for therapeutic intervention in pathophysiological conditions involving ASIC1a activation.
Acknowledgments
We thank Nivia Cuevas for comments on a draft of the manuscript.
Footnotes
-
This study was supported by an American Heart Association Florida/Puerto Rico Affiliate Grant-In-Aid Award and by a University of South Florida Signature Program in Neuroscience Award (to J.C.).
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.108.143974.
-
ABBREVIATIONS: ASIC, acid-sensing ion channel; CNS, central nervous system; NMDA, N-methyl-d-aspartate; AMPA, (±)-α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid; VGCC, voltage-gated Ca2+ channel; PSS, physiological saline solution; DTG, 1,3-di-o-tolyl-guanidine; nifedipine, 1,4-dihydro-2,6-dimethyl-4-(2-nitrophenyl)-3,5-pyridinedicarboxylic acid dimethyl ester; AP5, d-2-amino-5-phosphonovaleric acid; PB28, 1-cyclohexyl-4-(3-(5-methoxy-1,2,3,4-tetrahydronaphthalen-1-yl)-n-propyl)piperazine dihydrochloride; BD1063, 1-[2-(3,4-dichlorophenyl)-ethyl]-4-methylpiperazine dihydrochloride; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; PRE-084, 2-(4-morpholinethyl) 1-phenylcyclohexane-carboxylate hydrochloride; PcTx1, psalmotoxin1; CBP, carbetapentane citrate; DEX, dextromethorphan hydrobromide; MET, metaphit; IBO, ibogaine; THAP, thapsigargin; TTX, tetrodotoxin.
- Received July 23, 2008.
- Accepted August 21, 2008.
- The American Society for Pharmacology and Experimental Therapeutics