The anterior cingulate cortex (ACC), a limbic region associated with pain-related working memory and memory acquisition, receives a dense cholinergic innervation. To further understand the role of acetylcholine in ACC, we characterized the firing properties of pyramidal neurons following muscarinic receptor activation. Using whole-cell patch clamp recordings in acute brain slices, we report long-lasting nonsynaptic plateau potentials and persistent firing induced by carbachol (CCh) in pyramidal neurons in layers II/III of rat ACC. CCh responses were abolished by the muscarinic receptor antagonist atropine or by inhibitors of G proteins and phospholipase C. Inhibiting L-type calcium channels with nifedipine, removing extracellular calcium or chelating intracellular calcium with BAPTA also abolished plateau potentials and persistent firing. Blockade of nonselective cation channels with flufenamic acid, 2-aminoethyl diphenylborinate or SKF-96365 suppressed CCh responses and voltage-clamp recordings of CCh-sensitive currents revealed a transient receptor potential canonical-like cationic conductance. The group I metabotropic glutamate receptor (mGluR) agonist (S)-3,5-dihydroxyphenylglycine hydrate induced plateau potentials and persistent firing that were mediated by mGluR5. Our data demonstrate that receptor-operated channels drive calcium-dependent plateau potentials and persistent firing in layers II/III of ACC. Therefore, acetylcholine- and glutamate-evoked persistent activity in ACC may play a mnemonic role by allowing transient storage of information during pain processing.
The anterior cingulate cortex (ACC), a component of the limbic system, plays an important role in multiple high-level brain functions including attention, sensory perception, motor control, emotion, working memory, and long-term memory (Vogt et al. 1992; Vogt and Gabriel 1993; Devinsky et al. 1995; Rainville 2002; Sewards TV and Sewards MA 2002). This region specifies the emotional aspects of pain, controls the motor responses to noxious stimuli, and is involved in learning associated with the prediction and avoidance of noxious stimuli (Devinsky et al. 1995; Sewards TV and Sewards MA 2002; Malin et al. 2007). Functional imaging studies in humans have demonstrated that ACC is the most consistently activated structure during pain stimulation (Apkarian et al. 2005). Electrophysiological studies have shown that neurons in ACC respond to noxious stimulation (Sikes and Vogt 1992; Yamamura et al. 1996; Hutchison et al. 1999), are activated during pain anticipation or pain avoidance behavior (Koyama et al. 1998, 2000; Kuo and Yen 2005) and display enhanced response and synaptic long-term potentiation in rats with digit amputation (Wei and Zhuo 2001; Zhuo 2006).
ACC receives a dense cholinergic innervation from medial septal nucleus and the vertical limb of the diagonal band of Broca (Bigl et al. 1982; Butcher and Woolf 1986; Woolf 1991). Infusion of the muscarinic receptor agonist oxotremorine into ACC enhances inhibitory avoidance memory retention (Malin et al. 2007) while injection of scopolamine, a muscarinic receptor antagonist, disrupts the acquisition of nociceptive memory (Henzi et al. 1990; Ortega-Legaspi et al. 2003). These studies demonstrated the crucial role of cholinergic inputs in pain-related learning and memory processes. ACC, as the central part of the medial pain pathway, receives its inputs mainly from the midline and intralaminar thalamic nuclei as well as other cortical and subcortical areas (Vogt and Gabriel 1993; Vogt 2005; Zhuo 2006). Numerous electrophysiological studies indicate that the medial thalamus is the primary source of glutamatergic inputs to cingulate neurons (Greengard et al. 1991; Gigg et al. 1992; Gemmell and O'Mara 2002). ACC integrates information from pain pathway and other areas of the brain and participates in pain perception, pain modulation, and pain-related memory processes (Vogt and Gabriel 1993; Bush et al. 2000; Zhuo 2006).
It has been shown that during discriminative avoidance learning in the rabbit, a discriminative neuronal activity (i.e., greater neuronal discharges) in ACC is associated with presentation of the positive conditional stimulus rather than the negative conditional stimulus. This training-induced activity in superficial layers of the ACC is relatively “flexible” and is believed to represent substrates for the labile form of memory termed working memory (Orona and Gabriel 1983). Activation of cholinergic receptors and metabotropic glutamate receptors (mGluRs) enhances various types of memory encoding (Hasselmo 2006; Anwyl 2009). For example, activation of both types of receptors increases the strength of afferent inputs (Chrobak and Buzsaki 1994; Buesa et al. 2006) and synaptic long-term potentiation (Adams et al. 2004; Wu et al. 2008). Furthermore, activation of both cholinergic receptors and mGluRs induces intrinsic mechanisms for persistent firing, a cellular mechanism for working memory observed in entorhinal cortex (Klink and Alonso 1997; Egorov et al. 2002; Reboreda et al. 2007; Tahvildari et al. 2008; Yoshida et al. 2008; Zhang et al. 2010), prefrontal cortex (Haj-Dahmane and Andrade 1998; Sidiropoulou et al. 2009; Yan et al. 2009), and amygdala (Egorov et al. 2006; Faber et al. 2006). It has been shown that persistent firing is associated with the delay phase in delayed match or nonmatch to sample task performance (Suzuki et al. 1997; Young et al. 1997). By holding the input information necessary to guide goal-directed behaviors, persistent activity provides a robust cellular mechanism for working memory (Hasselmo 1999; Hasselmo and Stern 2006). However, the molecular components underlying these intrinsic neuronal properties are still unknown. To further understand the cellular basis of learning and memory in ACC, we investigated the intrinsic firing properties of its projection neurons under cholinergic or glutamatergic modulation. Using electrophysiological recordings and pharmacology, we report in the present study that, in layers II/III pyramidal neurons of ACC, activation of muscarinic receptors or mGluR5 induces robust calcium-dependent plateau potentials and persistent firing that require a phospholipase C (PLC)-coupled transient receptor potential canonical (TRPC)-like cation conductance. Such neurotransmitter-evoked persistent firing may contribute to the mechanisms of pain-related memory by enabling the transient storage of information during noxious stimuli.
Materials and Methods
Preparation of Acute Brain Slices
All experimental procedures were approved by the McGill University Animal Care Committee and were in compliance with the guidelines of the Canadian Council on Animal Care. Acute brain slices were obtained from adult Long-Evans rats (20–23 days old) (Charles River Canada). Rats were anesthetized with ketamine:xylazine cocktail (60:5 mg/kg) and transcardially perfused with ice-cold choline chloride–based artificial cerebrospinal fluid (cutting solution) consisting of (in mM): 110 choline-Cl, 1.25 NaH2PO4, 25 NaHCO3, 7 MgCl2, 0.5 CaCl2, 2.5 KCl, 7 glucose, 3 pyruvic acid, and 1.3 ascorbic acid, bubbled with carbogen (O2 95%, CO2 5%). Coronal rat brain slices (300 μm) were prepared from the forebrain. Briefly, after transcardial perfusion, rat brain was exposed dorsally and was cut vertically between cerebral cortex and cerebellum. Rat brain was quickly removed from cranial cavity and immersed in ice-cold cutting solution for 1–2 min. Coronal slices (Bregma 1.6 ∼ −0.26 mm, Paxinos and Watson 1998), obtained using a vibratome Leica VT1000S in the same solution, were separated into 2 halves by cutting the corpus callosum and then transferred to a normal extracellular solution (see below) to settle down at room temperature for at least 1 h before recording.
Brain slices were placed in a recording chamber mounted on the stage of an upright microscope Axioskop (Zeiss) equipped with ×63 water immersion objective and differential contrast optics. A near-infrared charged-coupled device camera (Sony XC-75) was used to visualize the neurons. Brain slices were stabilized using a U-shaped stainless steel anchor with Lycra threads at 1.5-mm spacing (Warner Instruments). Layers II/III principal neurons with typical pyramidal shape were chosen for electrophysiological recording. Brain slices were perfused by gravity at a speed of 0.5–1 mL/min. The temperature of perfusion solution was maintained at 32–33 °C using a TC-324B temperature controller (Warner Instruments). Patch pipettes (5–7 MΩ) were pulled on a Brown Flaming puller (P-97, Sutter Instruments) using borosilicate glass electrodes. Tight seals (∼5 GΩ) were obtained by applying constant negative pressure. Electrical signals were amplified using an Axopatch 200B amplifier (Axon Instruments, Molecular Devices), low-pass filtered at 10 kHz, digitized at 10 kHz via a Digidata 1322A interface (Axon Instruments), and stored on a Pentium computer using pClamp 18.104.22.168 software (Axon Instruments) for off-line analysis. Series resistance (Rs), on average 16–18 MΩ, was estimated online by canceling the fast component of whole-cell capacitive transients evoked by −10 mV voltage steps with the amplifier's compensation section (with the low-pass filter set at 10 kHz). Rs was compensated by 40% with the amplifier's built-in compensation section. In current-clamp recordings, the holding current was around 0 pA or slightly adjusted to obtain a membrane potential of –60 mV unless otherwise indicated.
Drugs and Solutions
All drugs were purchased from Sigma, except 2-aminoethyl diphenylborinate (2-APB) and 6-methyl-2-(phenylethynyl)pyridine (MPEP) from Tocris Bioscience, ET-18-OCH3 from BIOMOL International, and tetrodotoxin (TTX) from Research Biochemicals International. Atropine, BAPTA, carbachol (CCh), SKF-96365, (S)-3,5-dihydroxyphenylglycine hydrate (DHPG), and TTX were dissolved in water; 2-APB, flufenamic acid (FFA), MPEP, nifedipine, U73122, and U73343 were dissolved in dimethyl sulfoxide (DMSO). ET-18-OCH3 was dissolved in anhydrous ethanol. All drugs were freshly diluted from stock to the desired concentrations. The final concentration of DMSO or ethanol did not exceed 0.1%. GTP-γ-S and GDP-β-S were dissolved in intracellular solution immediately before experiments.
Normal extracellular solution in current-clamp experiments contained (in mM): 125 NaCl, 2.5 KCl, 1.6 CaCl2, 2 MgCl2, 25 NaHCO3, 1.25 NaH2PO4, 3 pyruvic acid, 1.3 ascorbic acid, 10 glucose, 2 kynurenic acid, and 0.1 picrotoxin. Kynurenic acid and picrotoxin were used to block N-methyl-D-aspartate (NMDA), non-NMDA (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and Kainate) receptors, and GABAA receptor, respectively. pH was maintained at 7.4 by constant bubbling with carbogen (95% O2, 5% CO2). In some experiments, removal of extracellular Ca2+ ions was achieved by replacing CaCl2 (1.6 mM) with MgCl2 (2 mM) and ethyleneglycol-bis(2-aminoethylether)-N,N,N',N'-tetra acetic acid (EGTA) (1 mM) in the perfusion medium. Intracellular solution in current-clamp recording contained (in mM): 120 Kgluconate, 20 KCl, 2 MgCl2, 0.2 EGTA, 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 7 di-tris phosphocreatine, 4 Na2 ATP, 0.3 Tris-GTP. pH was adjusted to 7.3 with KOH.
Extracellular solution for voltage-clamp recording was the same as in current-clamp recording except that 1 μM TTX and 5 mM CsCl were added to block Na+ and K+ channels, respectively. Intracellular solution for voltage-clamp recording contained (in mM): 120 CsMeSO3 (methanesulfonate), 10 HEPES, 0.2 EGTA, 2 MgCl2, 20 CsCl, 7 di-tris phosphocreatine, 4 Na2ATP, 0.3 Tris-GTP, pH was titrated to 7.3 with CsOH.
Electrophysiological data were analyzed using Clampfit 9 (Axon Instruments) and Origin 6.0 (Microcal Software). Values are expressed as means ± standard error of the mean. In current clamp, after whole-cell configuration was reached, membrane voltage was held at −60 mV. Firing frequency was defined as the average spiking frequency within 20 s after the depolarizing current pulse (2-s duration, 50–150 pA). Plateau potential amplitude was defined as the difference between the mean membrane potential (mV) measured at baseline within 1 min (before the pulse) and the mean membrane potential measured during the steady-state phase of persistent firing (excluding action potentials and afterhyperpolarizing potentials). Membrane input resistance was estimated by the voltage deflection induced by a 100 pA negative current pulse to the neurons at holding potential of −60 mV in control conditions (without CCh application). The amplitude of afterhyperpolarization (AHP) was measured as the difference between the holding potential (∼ −60 mV) during the 100-ms period immediately before the onset of the depolarizing current and the membrane potential 500 ms after the offset of the depolarizing current pulse in control conditions. Two-sample paired t-tests were used to compare values obtained in the same neurons before and after drug administration. Two-sample independent t-tests were used for comparison between 2 independent groups. One-way analysis of variance with Bonferroni correction was used for comparison of multiple independent groups. Differences were considered statistically significant when P < 0.05.
ACC Pyramidal Neurons Display Plateau Potentials and Persistent Firing
Electrophysiological recordings were performed on synaptically isolated principal neurons (n = 91) in layers II/III of the dorsal area of caudal anterior cingulate cortex (dACC, at Bregma +1.6 to −0.26 mm) in Long-Evans male rats (21–23 days, n = 42) (Fig. 1A). The recorded neurons are visually identified as large pyramidal neurons. We divided them into 2 groups based on their firing properties. Group I neurons (∼75% of recorded neurons) displayed a resting membrane potential of −73.9 ± 0.9 mV, an input resistance of 167.2 ± 9.9 MΩ, and minimal AHP of −1.6 ± 0.3 mV following a short depolarizing current pulse injection (2 s, 50–150 pA) (see Table 1). Group II neurons (∼25% of recorded neurons) have a similar resting membrane potential of 72.3 ± 1.7 mV (n = 12, P = 0.3354, compared with group I). However, they displayed both higher membrane input resistance (223.2 ± 2.0 MΩ, n = 12, P = 0.0115, compared with group I) and larger AHP (7.2 ± 0.9 mV, n = 12, P = 0.0001, compared with group I) after depolarizing current pulse injection (see Table 1). Both groups of neurons responded to depolarizing current pulse injection with tonic firing which displayed slight spike frequency adaptation. In all neurons tested, the current pulse-induced tonic firing terminated at the end of the stimulus (Fig. 1B, left panel). In group I neurons, however, when CCh (10 μM) was added in the perfusion medium, a sustained depolarizing membrane potential and superimposed repetitive action potential firing persisted after the end of the stimulus (Fig. 1B, right panel). In the presence of CCh, this sustained plateau potential and persistent firing lasted for several minutes (tested up to 7 min) without self-termination. Typically persistent firing was terminated by applying a negative bias current hyperpolarizing the membrane below –60 mV (Fig. 1B, right panel). After termination, the CCh-evoked response could be reactivated by applying depolarizing current pulse at a later time without desensitization within a period of at least 1 h in the presence of CCh. The CCh-evoked responses were blocked by bath perfusion of the muscarinic receptor antagonist atropine (1 μM, n = 4) indicating that the effects of CCh depend on muscarinic receptor activation (Fig. 1C). In group II neurons, however, neither plateau potential nor persistent firing was induced in the presence of CCh (data not shown). Therefore, no further investigation was made in this group of neurons.
|RMP (mV)||Rin (MΩ)||AHP (mV)||Persistent firing||N (sampling)|
|Group I||−73.9 ± 0.9||167.2 ± 9.9||1.6 ± 0.3||Yes||25|
|Group II||−72.3 ± 1.7||223.2 ± 2.0**||7.2 ± 0.9**||No||12|
|RMP (mV)||Rin (MΩ)||AHP (mV)||Persistent firing||N (sampling)|
|Group I||−73.9 ± 0.9||167.2 ± 9.9||1.6 ± 0.3||Yes||25|
|Group II||−72.3 ± 1.7||223.2 ± 2.0**||7.2 ± 0.9**||No||12|
Note: RMP, resting membrane potential; Rin, input resistance; ** P < 0.01.
Persistent Firing Is Calcium-Dependent and Requires G Protein-Coupled PLC Activation
To test if extracellular Ca2+ ions are required for the expression of persistent firing of dACC pyramidal neurons, Ca2+ ions were replaced with Mg2+ (2 mM) in presence of EGTA (1 mM) in the recording solution. In these conditions, depolarizing current pulses failed to induce plateau potential and persistent firing in the presence of CCh (10 μM) (Fig. 2A, n = 4). However, after calcium wash-in with normal extracellular solution, persistent firing partially recovered with sustained afterdischarges lasting for several seconds after the end of the depolarizing current pulse (n = 4). To test if Ca2+ entry was critical, we added the L-type Ca2+ channel blocker nifedipine (100 μM) to the perfusate. Nifedipine effectively blocked both plateau potentials and persistent firing, indicating that calcium influx through L-type voltage-gated Ca2+ channels is required (Fig. 2B, n = 5). Furthermore, when we buffered intracellular Ca2+ ions using BAPTA (10 mM), the persistent activity was gradually suppressed, confirming that intracellular Ca2+ ions play a major permissive role for CCh-evoked persistent responses (Fig. 2C, n = 5).
To investigate if G proteins are involved in the muscarinic receptor-dependent persistent firing, an intracellular solution lacking GTP was used. Application of CCh (10 μM) in the bath induced persistent firing during the first 5–10 min of the recording (n = 4). However, persistent firing gradually disappeared within 20 min, reflecting the intracellular dilution of GTP (Fig. 3A,D). When GDP-β-S (1 mM), a nonhydrolyzable competitive inhibitor of G protein activation, was included in the patch pipette in the absence of GTP, persistent firing was completely abolished (Fig. 3B,D; n = 6) despite a significant inhibitory effect on the AHP. The AHP amplitude is −2.8 ± 0.29 mV before GDP-β-S and −0.06 ± 0.52 mV after GDP-β-S (n = 6, P = 0.0018). Interestingly, when the nonhydrolyzable GTP analog GTP-γ-S (1.5 mM), a G-protein activator, was added in the patch pipette, in the absence of CCh, it evoked a transient plateau potential with afterdischarges (Fig. 3C,D; n = 8) following depolarizing current pulse injection, within 10 min of the start of recording. This observation suggests that activation of G proteins is necessary to induce afterdepolarization and afterdischarges but is not sufficient to evoke long-lasting persistent firing.
To confirm the activation of the PLC pathway during CCh-evoked persistent firing, PLC blockers were applied in the bath solution. The extensively used PLC inhibitor U73122 (10 μM) blocked persistent firing (data not shown). However, its inactive enantiomer U73343 (10 μM) was also effective at blocking persistent firing (data not shown). It is possible that this effect of U73122 and U73343 relies on the inhibition of L-type Ca2+ channels reported previously (Macrez-Leprêtre et al. 1996). Therefore, we applied another selective PLC inhibitor, ET-18-OCH3, in the bath solution. ET-18-OCH3 (10 μM), which does not seem to block L-type Ca2+ channel (Lemmens et al. 2001), significantly suppressed plateau potential and persistent firing in layers II/III principal neurons of dACC (n = 5, Fig. 4) indicating the involvement of G protein-coupled PLC pathway in the CCh-evoked persistent firing.
Nonselective Cation Channels Mediate Plateau Potentials and Persistent Firing
According to a model of persistent firing proposed by Fransén et al. (2006), CCh-evoked plateau potentials and persistent firing could be due to a balance between hyperpolarizing outward currents mediated by multiple K+ conductances, including leak K+ currents, Ca2+-activated K+ currents, fast inactivating A currents and M-currents, and depolarizing inward currents mediated by cation conductances.
Therefore, we tested the effects of the generic K+ channel blocker Ba2+ on the persistent firing. As shown in Figure 5A, blocking K+ channels with Ba2+ (100 μM) did not induce persistent firing following depolarizing current pulse injection while, in the same cells, addition of CCh (10 μM) induced strong plateau potential and persistent firing (n = 5). Interestingly, in all cells tested the CCh-evoked persistent activity was enhanced by addition of Ba2+: the firing frequency was increased (Fig. 5A), a result consistent with the current model (Fransén et al. 2006). Furthermore, the stable sustained CCh-evoked persistent firing gradually developed into periodic oscillatory firing 20 min after CCh application in all cells tested (Fig. 5B).
Persistent Na+ current has been reported to mediate plateau potential in spinal motor neurons and spike afterdepolarization in medium spiny neurons of the nucleus accumbens in rats (Li and Bennett 2003; D'Ascenzo et al. 2009). Therefore, we tested if persistent Na+ current contributes to CCh-evoked persistent firing in dACC neurons. We observed that afterdepolarizations and plateau potentials were not sensitive to blockade of Na+ channels by TTX. As shown in Figure 5C, CCh-evoked plateau potentials that mediate persistent firing were left intact in the presence of TTX (1 μM, n = 4). In contrast, persistent firing was blocked by substitution of extracellular Na+ with equimolar concentration of choline (140 mM), suggesting that cation channels that mediate persistent firing use Na+ as major charge carrier (Fig. 5D, n= 3).
Next, we tested the effect of the nonselective cation channel blockers FFA, 2-APB, and SFK-96365 on persistent firing. FFA (100 μM, n = 5), 2-APB (100 μM, n = 6) as well as the TRPC channel blocker SKF-96365 (50 μM, n = 3) strongly suppressed CCh-evoked plateau potentials and persistent firing in dACC neurons (Fig. 6). To isolate the CCh-induced currents and assess their current–voltage relationship, we applied a ramp (+60 to −100 mV in 3 s) to dACC neurons in voltage-clamp recording before and after CCh (20 μM) application, in the presence of 5 mM CsCl and 1 μM TTX in the extracellular medium. Current subtraction revealed a CCh-induced inward current which exhibits a negative slope and a reversal potential near 0 mV, indicating a nonselective cation permeability for this channel (Fig. 7A; n = 13). Interestingly, this I-V phenotype resembles the one of TRPC4/5 subunit-containing channels expressed in heterologous expression system (Strübing et al. 2001; Plant and Schaefer 2005). The maximum inward current (Iin) measured at the membrane potential of −45.9 ± 2.0 mV is 61.2 ± 8.7 pA. We confirmed that this cholinergic cation current is inhibited by the TRPC channel blocker SKF-96365 (Fig. 7B, n = 4).
Activation of Group I mGluRs Induces Persistent Activity in ACC
Numerous reports have shown that mGluR activation induces slow afterdepolarization (e.g., Gee et al. 2003; Tozzi et al. 2003; Bengtson et al. 2004; Fowler et al. 2007) or persistent firing (Yoshida et al. 2008) in neurons of various central regions. Therefore, we tested the effect of activation of group I mGluRs on the firing properties of dACC neurons. Application of the group I mGluR agonist DHPG (20 μM) in perfusion medium induced plateau potential and persistent firing following depolarizing current pulse injection in 81% (21/26) of pyramidal neurons tested (Fig. 8A,C). Group I mGluRs include 2 subtypes of Gq-coupled glutamate receptors: mGluR1 and mGluR5. mGluR1 is mainly expressed in nonpyramidal neurons and almost all mGluR1-immunoreactive neurons in adult neocortex and hippocampus are γ-aminobutyric acidergic interneurons (Baude et al. 1993; Stinehelfer et al. 2000), whereas mGluR5 is expressed by both pyramidal and nonpyramidal neurons in adult neocortex and hippocampus (Baude et al. 1993; Luján et al. 1996; Stinehelfer et al. 2000). Therefore, we applied the mGluR5 antagonist MPEP to test the specific contribution of mGluR5. We observed that MPEP (50 μM) blocked DHPG-induced persistent firing in dACC principal neurons (Fig. 8A,C; n = 5–6). Furthermore, we observed that 10 min after washout of the DHPG effect, plateau potentials and persistent firing were evoked again by application of CCh (10 μM) (n = 5). Conversely, 10 min after washout of the CCh effect, plateau potentials and persistent firing could be evoked again by application of DHPG (20 μM). Interestingly, in neurons where saturating amounts of CCh (100 μM, empirical dose) induced strong persistent firing, application of DHPG (20 μM) to these same neurons did not enhance the persistent responses (Fig. 8B,C; n = 4). Therefore the effect of DHPG was occluded by application of CCh, demonstrating that mGluR5 and muscarinic receptors converge on the same intracellular pathway for induction of persistent firing.
In the present study, we characterized the firing properties of ACC neurons under metabotropic modulation and demonstrated that the majority of the layers II/III principal neurons in ACC exhibit cholinergic or group I mGluR-evoked persistent activity. Our results provide evidence that this metabotropic receptor-dependent persistent firing in ACC requires permissive intracellular calcium levels and the activation of a nonselective cation conductance likely mediated by receptor-operated and PLC-coupled TRPC-like channels. These results are consistent with previous data obtained in entorhinal cortex (Egorov et al. 2002; Shalinsky et al. 2002; Reboreda et al. 2007; Yoshida et al. 2008; Zhang et al. 2010), prefrontal cortex (Yan et al. 2009), postsubiculum (Yoshida and Hasselmo 2009), and amygdala (Egorov et al. 2006; Faber et al. 2006) showing the induction of persistent firing by metabotropic receptor activation and the involvement of TRPC-like channels in mediating this intrinsic neuronal property. The ionic mechanism revealed here also fits with the computational model of persistent firing in entorhinal cortex (Fransén et al. 2006) predicting that the stability of persistent firing depends on the balance between nonselective cation conductances and K+ conductances.
Our electrophysiological recordings were performed on large pyramidal neurons in dACC, a prominent nociceptive area according to animal electrophysiological studies and noninvasive human research (Vogt and Gabriel 1993; Vogt 2005). Interestingly, most of the nociception-specific neurons were found in layers II and III of this area (Sikes and Vogt 1992; Shyu et al. 2008) so the neurons showing persistent firing might be involved both in nociception and in pain anticipation that precedes the avoidance of noxious stimuli. The apical dendritic trees of layers II/III principal neurons in dACC extend into layer I where both extracingulate inputs from the midline and intralaminar thalamic nuclei (medial pain pathway) and cholinergic fibers from basal forebrain nuclei terminate. Numerous reports have shown that [H3]-pirenzepine binding sites for cholinergic M1 muscarinic receptors are mainly present in superficial layers of the ACC in rats (Messer et al. 1987), monkeys (Mash et al. 1988), and humans (Zavitsanou et al. 2004). This is consistent with our finding that CCh induces persistent firing in layers II/III neurons of the ACC through activation of PLC-coupled muscarinic receptors.
One of the key properties of persistent firing in the principal neurons of dACC is the Ca2+ dependence. This is demonstrated by exclusion of extracellular Ca2+ ions or by the blockade of calcium influx with L-type calcium channel blocker. The failure of CCh to induce persistent firing in these conditions suggests that an extracellular source of Ca2+ ions is required for the expression of persistent activity. Because chelation of intracellular Ca2+ ions with BAPTA also eliminated the CCh-evoked plateau potential and persistent firing, a Ca2+-dependent signaling pathway plays a critical role. Without intracellular calcium-dependent signaling, L-type calcium channel-mediated membrane depolarization alone is insufficient to induce persistent firing since chelating Ca2+ intracellularly by BAPTA without blocking Ca2+ inward currents did not induce persistent firing. It is worth noting that intracellular Ca2+ release from intracellular stores may not play a major role in persistent activity since it has been shown that intracellular application of the endoplasmic reticulum calcium ATPase inhibitors cyclopiazonic acid and thapsigargin failed to affect persistent firing in entorhinal cortex (Fransén et al. 2006).
Persistent firing in dACC is not induced by blockade of the many K+ conductances sensitive to extracellular Ba2+ ions. Ba2+ has been shown to block multiple K+ channels including channels that mediate M-currents, leak currents, fast inactivating A-type current, small (SK) and big (BK) conductance Ca2+-activated K+ currents, and inwardly rectifying K+ currents. Although blocking of these K+ channels alone is not sufficient to induce persistent firing, we do see an enhancement of CCh-evoked persistent responses after Ba2+ application. Blocking of K+ conductances by Ba2+ ions increased the firing frequency and changed the stable firing pattern into oscillatory firing in all neurons tested suggesting that K+ channels contribute to the stability of CCh-evoked persistent firing in normal conditions as simulated in a single-cell model of layer V pyramidal neurons of entorhinal cortex (Fransén et al. 2006). Furthermore, the slow persistent Na+ current reported to mediate plateau potentials in spinal cord and nucleus accumbens (Li and Bennett 2003; D'Ascenzo et al. 2009) and to be sensitive to TTX (Pace et al. 2007; Koizumi and Smith 2008; D'Ascenzo et al. 2009), may not play a major role in neurons of the ACC since the CCh-evoked plateau potential mediating persistent firing is left intact in the presence of TTX. Interestingly, persistent firing disappeared when extracellular Na+ was replaced by choline, indicating a requirement for a Na+-mediated cation conductance in the mechanisms of persistent firing in ACC.
The CCh- and DHPG-evoked persistent firing in the layers II/III principal neurons of dACC likely involves TRPC nonselective cation channels. This is supported by the following evidence: 1) TRPC channels are abundantly expressed in superficial layers of the ACC (Fowler et al. 2007), 2) both the persistent firing and the activation of TRPC (e.g., TRPC5) channels are calcium dependent (Okada et al. 1998; Ordaz et al. 2005; Gross et al. 2009), 3) both persistent firing and the activation of TRPC channels are mediated by the activation of Gq/11-proteins and the PLC pathway (Ramsey et al. 2006), 4) persistent firing in ACC is blocked by nonselective cation channel and TRPC channel blockers (FFA, 2-APB, and SKF-96365), and 5) I-V relationship analysis in voltage-clamp recording revealed a CCh-sensitive current with a negative slope and a near-zero reversal potential that resembles the current mediated by heteromeric TRPC4/5 subunit-containing channels (Plant and Schaefer 2005).
Results from the present study do not support a role for TRPM4 and/or TRPM5 channels in persistent firing because these calcium-activated channels are inhibited, but not activated, by the activation of muscarinic receptors (Nilius et al. 2006). Up to now, despite its physiological importance, the exact molecular identity of the channels mediating plateau potentials and long-lasting persistent firing is still elusive. Due to the absence of selective pharmacological ligands for the TRPC multigene family, a combination of multiple approaches will be needed for the final elucidation of this issue.
Interestingly, we also observed that mGluR5 activation induced plateau potentials and persistent firing in dACC layers II/III principal neurons. Indeed, similarly to what was reported in entorhinal cortex (Yoshida et al. 2008), this effect was blocked by the mGluR5-selective antagonist MPEP. Also in agreement with our observations, several reports have shown that the mGluR-mediated slow neuronal excitation and excitatory post-synaptic potentials involve the activation of TRPC-like cation conductances (e.g., Gee et al. 2003; Tozzi et al. 2003; Bengtson et al. 2004; Faber et al. 2006; Fowler et al. 2007). It is important to note that both muscarinic and mGluR5 activation induce persistent firing in the same group of neurons in dACC since neurons that showed CCh responses also responded to DHPG and vice versa. The fact that activation of mGluR5 induces persistent firing in ACC is also consistent with behavioral data showing that activation of mGluRs in ACC enhances escape response and pain-related fear memory in mice (e.g., Tang et al. 2005). Moreover, the DHPG effect was occluded by previous application of a saturating amount of the muscarinic receptor agonist CCh, clearly indicating that both metabotropic systems converge on the same downstream signaling pathway that drives persistent activity. A synergistic interaction of these 2 receptor systems has been reported for the modulation of plastic burst firing in neurons of the subiculum (Moore et al. 2009). Therefore, the dual cholinergic and glutamatergic innervation of ACC pyramidal neurons could provide a cellular mechanism for concerted action in mediating pain-related memory process. However, the exact nature of their interaction and the conditions under which this might occur will remain to be investigated.
In summary, the present study demonstrated that activation of muscarinic receptors or mGluR5 induces plateau potentials and long-lasting persistent firing in superficial layers of the dACC. These metabotropic receptor-mediated neuronal responses are calcium dependent and require the opening of PLC-linked TRPC-like nonselective cation conductances. During discriminative avoidance learning, the training-induced neuronal activity in ACC has been proposed to reflect the operation of a recency system, a component of working memory, that holds information during memory acquisition (Goldman-Rakic 1990), while posterior cingulate cortex, by receiving input from ACC, plays an important role in the maintenance and retention of memory (Gabriel 1993). We propose that the metabotropic cholinergic or glutamatergic long-lasting persistent firing activity in ACC provides a cellular mechanism for holding and storing information during discriminative learning and pain signal processing.
Canadian Institutes of Health Research (NRF 77566).
Z.Z. held an AstraZeneca-McGill Alan Edward Centre for Research on Pain postdoctoral fellowship during the preparation of this manuscript and holds now a Jeanne Timmins Costello postdoctoral fellowship. We thank our colleague Dr Ariel Ase (Montreal Neurological Institute) for helpful discussions. Conflict of Interest: None declared.