Besides a reduction of L-type Ca2+-currents (CaV1), muscarine and the peptidic M1-selective agonist, MT-1, reduced currents through CaV2.1 (P/Q) and CaV2.2 (N) Ca2+ channel types. This modulation was strongly blocked by the peptide MT-7, a specific muscarinic M1-type receptor antagonist but not significantly reduced by the peptide MT-3, a specific muscarinic M4-type receptor antagonist. Accordingly, MT-7, but not MT-3, blocked a muscarinic reduction of the afterhyperpolarizing potential (AHP) and decreased the GABAergic inhibitory postsynaptic currents (IPSCs) produced by axon collaterals that interconnect spiny neurons. Both these functions are known to be dependent on P/Q and N types Ca2+ channels. The action on the AHP had an important effect in increasing firing frequency. The action on the IPSCs was shown to be caused presynaptically as it coursed with an increase in the paired-pulse ratio. These results show: first, that muscarinic M1-type receptor activation is the main cholinergic mechanism that modulates Ca2+ entry through voltage-dependent Ca2+ channels in spiny neurons. Second, this muscarinic modulation produces a postsynaptic facilitation of discharge together with a presynaptic inhibition of the GABAergic control mediated by axon collaterals. Together, both effects would tend to recruit more spiny neurons for the same task.
Striatal acetylcholine (ACh) comes from intrinsic cholinergic interneurons (Bolam et al. 1984) and the neostriatum is a nucleus with one of the highest concentrations of both ACh and cholinergic receptors in the brain (Vilaro et al. 1991; Volpicelli and Levey 2004). ACh mediates important postsynaptic and presynaptic effects in the neostriatum (Calabresi et al. 2000; Zhou et al. 2003) and anticholinergic agents are important supplementary drugs for Parkinson's disease therapeutics, intended to restore a supposed imbalance between the dopaminergic and the cholinergic systems (Zhou et al. 2003).
Striatal spiny neurons originate the output pathway of the neostriatum to other basal ganglia nuclei. These spiny neurons predominantly express 2 types of muscarinic receptors: M1 and M4 (Bernard et al. 1992; Hersch and Levey 1995; Yan et al. 2001). M1-type receptors belong to the M1-class, which is preferentially coupled to the hydrolysis of phosphatidylinositol. Instead, M4-type receptors belong to the M2-class, whose activation reduces adenylate cyclase activity or inhibits voltage-gated Ca2+ channels (Caulfield and Birdsall 1998; Jones 1993). In many neurons, M2-class receptors, including the M4 subtype, regulate N- and/or P/Q-type Ca2+ channels through fast membrane-delimited pathways (Allen and Brown 1993; Bernheim et al. 1992; Cuevas and Adams 1997; Liu and Rittenhouse 2003; Murai et al. 1998; Shapiro et al. 1999; Yan and Surmeier 1996; Wanke et al. 1994). However, the activation of M1-class receptors may also regulate N- and P/Q-type Ca2+ channels through slow and diffusible signaling cascades (Bernheim et al. 1992; Liu and Rittenhouse 2003; Melliti et al. 2001; Shapiro et al. 1999). This work attempts to answer which of these mechanisms is the most important to regulate Ca2+ influx through voltage-gated Ca2+ channels in neostriatal projection neurons. To answer this question we used the most selective antagonists of the muscarinic receptors known to date: MT-3 and MT-7 (Karlsson et al. 2000; Potter et al. 2004).
It is of crucial importance to know which class of muscarinic receptor regulates Ca2+ entry through N- and P/Q-type Ca2+ channels in neostriatal projection neurons because these Ca2+ channels control vital cellular functions. For example, Ca2+ entry through N- and P/Q-type Ca2+ channels generate the afterhyperpolarizing potential (AHP) that makes up the interspike interval. In turn, the AHP regulates a basic neuronal function: the frequency response to a given stimulus, i.e., the input/output, or gain, function (Bargas et al. 1999; Perez-Garci et al. 2003; Pineda et al. 1992, 1995; Vilchis et al. 2000). In addition, N- and P/Q-type Ca2+ channels are responsible for triggering neurotransmitter release in most central synapses including those in the neostriatum (Catterall 1998; Momiyama and Koga 2001; Murakami et al. 2002; Reuter 1996; Rhee et al. 1999; Tecuapetla et al. 2005; Wheeler et al. 1994). Therefore there is a possibility that muscarinic actions have a global action on spiny cells performance arising from modulation of N- and P/Q-type voltage-gated Ca2+ channels. This work was previously reported in abstract form (Perez-Rosello et al. 2000, 2003).
Preparation of slices
Briefly, and as described elsewhere (Bargas et al. 1999), male adult Wistar rats (100–120 g) were used to record from acutely dissociated cells. Rats from our animal house were anesthetized and their brains quickly removed into ice-cold saline (4°C) containing (in mM): 126 NaCl, 3 KCl, 1 MgCl2, 2 CaCl2, 25 NaHCO3, and 11 glucose (pH 7.4 with NaOH, 298 mOsm/L with glucose; aerated with 95% CO2-5% O2). Parasagittal neostriatal slices (300 μm thick) were cut in 4°C saline using a vibratome (Ted Pella, Reading, CA). Slices were then transferred to room-temperature saline (23–25°C) and allowed to recover for 1 h.
Voltage-clamp recordings in dissociated cells
Neostriatal neurons were acutely dissociated using procedures similar to those previously described (Bargas et al. 1994, 1999; Vilchis et al. 2002). Briefly, slices were taken into a HEPES-buffered saline and the dorsal striatum was dissected. Striata were then placed in the same HEPES solution, now containing 1–1.5 mg/ml of pronase E–type XIV (Sigma, St. Louis, MO) at 32°C. After about 20 min, the tissue was removed into a low-Ca2+ HEPES saline. It was then rinsed and cells mechanically dissociated with Pasteur pipettes. The cell suspension (2 ml) was plated in a 35-mm petri dish mounted on the stage of an inverted microscope containing 1 ml of the recording saline (in mM): 0.001 tetrodotoxin (TTX), 140 NaCl, 3 KCl, 5 BaCl2, 2 MgCl2, 10 HEPES, and 10 glucose (pH 7.4 with NaOH; 298 mOsM/L with glucose). Note that TTX blocked inward sodium currents and the charge carrier (through calcium channels) Ba2+ is also a potent and unspecific K+-channel blocker. In addition, the intracellular solution did not contain K+ ions. This made us confident that the recorded inward current was mainly carried by Ba2+ ions passing through Ca2+ channels (Bargas et al. 1994). Further, the inward current isolated in this way was totally blocked by 400 μM Cd2+ (not shown) and as shown in the results, partially blocked by several known peptides known as specific Ca2+ current blockers.
After allowing the cells to settle, superfusion began at about 1 ml/min with saline of the same composition. Recordings were made only from medium-sized neurons (10–12 μm of soma diameter and about 6–7 pF of whole cell capacitance) that had only a few short (<50 μm) proximal dendrites. Recordings of whole cell currents used standard techniques: Electrodes were pulled from borosilicate glass (WPI, Sarasota, FL) in a Flaming–Brown puller (Sutter Instrument, Novato, CA) and fire polished before use. The internal saline contained (in mM): 180 N-methyl-d-glucamine (NMG), 40 HEPES, 4 MgCl2, 10 EGTA, 2 Na2ATP, 0.2 Na3GTP, and 0.1 leupeptin (pH = 7.2 with H2SO4, 265–270 mOsM/L). Electrode DC resistances were 3–7 MΩ in the bath. Recordings were obtained with an Axopatch-1D patch-clamp amplifier (Axon Instruments, Foster City, CA), and controlled and monitored with a PC clone running pClamp (version 5) with a 125-kHz DMA interface (Axon Instruments). After seal rupture, the series resistance (<15 MΩ) was compensated (70–80%) and monitored before and after drug application. Voltage control was assessed after compensation by examining the tail currents generated by strong depolarizations. Current–voltage relationships before and during drug blockade were evoked with both 20-ms voltage commands from −80 to 50 mV in 10-mV steps and with current responses to voltage ramps (0.7 mV/ms) from −80 to 50 mV. Because results from both methods coincided (Fig. 2), for the sake of clarity most figures illustrate only ramp responses. Drugs were applied with a gravity-fed system that positioned a glass capillary 200 μm from the recorded cell in the direction of flow superfusion. Solution changes were performed with a DC-controlled microvalve system (Lee, Essex, CT).
Slices obtained as above (see Preparation of slices), but sometimes from older animals (100–200 g), were also recorded in a submerged chamber and superfused with the same saline at 1 ml/min (34–36°C). Intracellular recordings were performed with microelectrodes filled with 3 M K-acetate (DC resistances: 80–120 MΩ) and the help of an active bridge electrometer (Neuro Data, Cygnus Technologies, DWG format, Philadelphia, PA) (Pineda et al. 1992). Records were digitized and saved on VHS tapes (40 kHz) and analyzed off-line in a PC clone. Stimulation consisted of intracellular injections of constant-current steps to evoke either the AHP after a single action potential (Pineda et al. 1992) or repetitive firing. Stimuli were of suprathreshold intensity and given at a holding potential of −55 to −60 mV by adjusting constant current. Bridge balance as well as recovery periods (without DC current) were monitored between sample records. After recording some neurons were injected with biocytin as previously described (Galarraga et al. 1999). All neurons identified in the present study were medium-sized spiny projection neurons.
Voltage-clamp recordings of inhibitory postsynaptic currents (IPSCs)
In this case, neostriatal slices from PD14 rats were transferred to a custom Plexiglas recording chamber and superfused with oxygenated saline (3–6 ml/min) as above. Individual neurons were visualized (×40 water-immersion objective) under differential interference contrast (DIC) enhanced visual guidance using infrared videomicroscopy in an upright microscope (Diaphot, Nikon, Melville, NY) adapted with a CCD camera (CCD-100, Dage-MTI, Michigan City, IN). Synaptic events were evoked with a bipolar concentric tungsten electrode (12 μm at the tip; FHC, Bowdoinham, ME) located at the globus pallidus (GP) to stimulate antidromically the axons of spiny cells (Guzman et al. 2003). Paired shock stimulation (45–50 ms of interstimulus interval; 0.2- to 0.4-ms duration; 1–4 V; at a frequency of 0.1 Hz) was delivered with a computer interface (see following text). Isolation units (Digitimer, Hertfordshire, UK) between the computer and the stimulating electrodes were used to adjust stimulus parameters during the experiment. The distance between recording and stimulating electrodes was about 1 mm. Synaptic responses in these conditions were of moderate amplitude and had amplitude variation, but without exhibiting failures in most cases. Traces shown are the average of approximately 2-min recordings (10 traces) taken once the amplitude had been stabilized in a given condition. A hyperpolarizing voltage command (15 mV) continuously monitored input conductance. Internal solution was (in mM): 72 KH2PO4, 36 KCl, 2 MgCl2, 10 HEPES, 1.1 EGTA, 0.2 Na2ATP, 0.2NA3GTP, 5 QX 314 (to prevent neuronal firing and enhance input resistance), and 0.5% biocytin (pH = 7.2, 275 mOsM/L). The bath solution contained 6-cyano-7-nitroquino-2,3-dione (CNQX) and d-2-amino-5-phosphonopentanoic acid (AP5) to block glutamatergic currents. Cells with resting potential more negative than −70 mV (at zero current), input resistance >200 MΩ, and holding current (in voltage-clamp mode) ≤0.02 nA to maintain a holding potential near the resting potential of the cell, were chosen. Whole cell recordings were made using a Axoclamp 2B/2A amplifiers (Axon Instruments). Whole cell access resistances were in the range 5–20 MΩ. Access resistance was continuously monitored and experiments abandoned if changes >20% were encountered. No cell capacitance, series resistance, or liquid junction potential (2 mV) compensations were made. All recordings were filtered at 1–3 kHz and digitized with an AT-MIO-6040E, a DAQ (NI-DAQ) board (National Instruments, Austin, TX) and a PC clone. Online data acquisition used custom programs made in the LabVIEW environment (National Instruments). The NI-DAQ board was used to save the data on binary files in the computer hard disk for further off-line analysis.
Drugs were dissolved in the bath saline from stock solutions made daily using a gravity-driven superfusion system. AP5, CNQX, muscarine, nitrendipine, QX-314, and TTX were all purchased from Sigma (St. Louis, MO). Muscarinic ligands were: muscarine (RBI, Natick, MA), the muscarinic peptides muscarinic toxin 1 (MT-1) and muscarinic toxin 3 (MT-3) from Alomone Labs (Jerusalem, Israel), and muscarinic toxin 7 (MT-7) from Peptides International (Louisville, KY). Nomenclature chosen to name these peptides in the present work follows that of commercial suppliers. Calcium channel antagonists: ω-conotoxin GVIA (ω-CgTx), ω-agatoxin TK (ω-AgaTK), and ω-conotoxin MVIIC (ω-CTx) were obtained from both Peptides International and Alomone Labs. Most drugs were dissolved in water to obtain stock solutions and added to the superfusate to give the final concentration. Nitrendipine was dissolved in dimethylsulfoxide (DMSO, 0.1%).
Digitized data were imported for analysis and graphing into commercial software (Origin v. 6. Microcal, Northampton, MA). Means ± SE of all ICa2+s, AHPs, and IPSCs are reported. However, free-distribution statistical tests were used to assess statistical significance of the differences between samples: Mann–Whitney U test or Wilcoxon's t-test.
Activation of muscarinic M1-type receptors reduces Ca2+ currents through both N- and P/Q-type Ca2+ channels in neostriatal neurons
Figure 1A (filled circles) illustrates the time course of action of the cholinergic muscarinic agonist, muscarine (1 μM), on whole cell Ca2+ currents recorded from neostriatal medium-sized neurons (Ba2+ is the charge carrier). Note that muscarine actions are reversible (Fig. 1A, filled circles and representative traces in Fig. 1B). Ca2+ currents from another neuron (Fig. 1A, open circles) were recorded without adding any drug to appreciate the typical rundown in these experiments: 4 ± 0.4% after 15-min recording (Fig. 1C, box plot; n = 35).
Figure 2A (control) shows a family of currents evoked with depolarizing voltage commands in control conditions (Fig. 2A, bottom; see methods for an explanation of details of the voltage protocol). Figure 2A (muscarine) illustrates a family of currents recorded in the same cell after the addition of muscarine (1 μM). Muscarine reduced both Ca2+ currents and their slow tail currents for equivalent voltages (vg., 40 mV) (Fig. 2A, inset). Figure 2B shows Ca2+ currents evoked in the same neuron, but this time with a depolarizing ramp command from −80 to 50 mV (indicated at the top; see methods). Ramp-evoked currents in the presence of muscarine were smaller than those evoked in control conditions. Figure 2C shows the current–voltage relationships (I–V plots) obtained from experiments depicted in Fig. 2, A and B. Open (control) and filled (muscarine) circles are measurements from the currents in Fig. 2A (at the arrows), whereas continuous lines are taken from currents illustrated in Fig. 2B, plotted as a function of ramp voltage. Note that ramp-evoked I–V plots can be seen as the “fit” of I–V plots obtained with step commands. Thus the next figures will show I–V plots obtained with ramp commands only. However, most experiments were done with both protocols.
Figure 2D illustrates the time course of action of the selective muscarinic M1 receptor agonist, the MT-1 peptide (48 nM) (Jerusalinsky and Harvey 1994), on Ca2+ current amplitude. The I–V plot in Fig. 2E shows that MT-1 mimics the action of muscarine on Ca2+ currents. Muscarine (1 μM) reduced Ca2+ currents 45 ± 3% (Fig. 1B; n = 9; P < 0.01, Wilcoxon's t-test). Larger muscarine concentrations (5–10 μM) did not produce additional blockade (not shown). MT-1 (48 nM) reduced these currents 42 ± 6% (Fig. 2E; n = 7; P < 0.03, Wilcoxon's t-test). There were no significant differences between the effects of both agonists (Fig. 1F; Mann–Whitney U test). The concentration of MT-1 used is larger than the affinity constant obtained by displacing pirenzepine from brain synaptosomal membranes (19 nM; Jerusalinsky et al. 1992), but below the concentration that affects M4 receptors working as an antagonist (Jerusalinsky and Harvey 1994; Kornisiuk et al. 2001). In addition muscarine had no further effect after this MT-1 concentration and larger concentrations of MT-1 had only minimal effects in one out of 3 tested cells (not shown). MT-1 action, at these concentrations, then supports the view that most muscarinic actions on Ca2+ currents (CaV1 and CaV2) are by the activation of M1-type receptors in spiny neurons.
To further test this view, we first isolated the Ca2+ current components that are not L-type by using saturating concentrations of a dihidropyridine blocker, nitrendipine (10 μM) (Bargas et al. 1994), to block L-type Ca2+ channels present in spiny cells (CaV1.2 and CaV1.3; Olson et al. 2001). Thereafter, we intended to quantify how much muscarinic modulation remained. It was seen that muscarinic modulation (1 μM muscarine), when L-type Ca2+ current was blocked, reduced the remaining current by 29 ± 1% (Fig. 3A; n = 14; P < 0.001, Wilcoxon's t-test). That is, nitrendipine only partially occluded muscarinic action (time course in Fig. 3B) and a substantial modulation remained in most cells (box plot in inset of Fig. 3B shows variability). Modulation with nitrendipine was significantly different from that obtained without nitrendipine (29 vs. 45%; see above; P < 0.001, Mann–Whitney U test). It is known that L-type Ca2+ channels are a main target for muscarinic modulation (Howe and Surmeier 1995). Noticeably, because nitrendipine reduces Ca2+ currents by about 23 ± 2% (n = 15, P < 0.002, Wilcoxon's t-test) (e.g., Bargas et al. 1994), and the experiments showed that blockage of L-channels occluded muscarinic modulation by about 36%, then, the result requires that modulation by muscarine be about 69% of available L-type Ca2+ current (Howe and Surmeier 1995). It is not known what is being targeted by the remaining modulation and which is the receptor involved. Thus we added other specific channel blockers together with nitrendipine: either ω-CgTx, which blocks N-type channels, or ω-AgaTK, which blocks P/Q-type channels (see methods and Bargas et al. 1994), the rationale being that, if one of these channels is targeted, its blockage would further occlude muscarinic actions.
Figure 3, C and D show that in the presence of both nitrendipine (10 μM) and ω-CgTx (1 μM) (that is without L- and N-type Ca2+ channels), muscarinic actions were further occluded. Muscarine was now able to reduce the remaining current by only 15 ± 2% (from 29 to 15%; n = 8; P < 0.001, Mann–Whitney U test). This suggested that N-type Ca2+ channels are also a target for muscarinic modulation. By itself, ω-CgTx blocks about 22 ± 2% (n = 12) of the whole cell Ca2+ current (see following text and Bargas et al. 1994).
Similar experiments were performed in the presence of both nitrendipine (10 μM) and ω-AgaTK (400 nM), a specific P/Q-type Ca2+ channel blocker. Again, further occlusion of muscarinic modulation was found, current reduction now being only 11 ± 1% (from 29 to 11%; n = 5; P < 0.004, Mann–Whitney U test). This result suggested that P/Q-type Ca2+ channels are a third target for muscarinic actions (Fig. 3, E and F). By itself, ω-AgaTK blocked about 30 ± 3% of the Ca2+ current, being the most abundant Ca2+ current component (Bargas et al. 1994; Vilchis et al. 2002). In addition, 3 experiments were done with the 3 Ca2+ channel blockers administered together: nitrendipìne (10 μM), ω-CgTx (1 μM), and ω-AgaTK (400 nM). In these experiments muscarine (1 μM) had no further effect whatsoever (data not shown). Therefore to explain the actions of muscarine with and without the different combinations of Ca2+ blockers, the 45% reduction ascribed to muscarinic modulation was distributed in the following way by solving a linear system of equations (see Vilchis et al. 2002): 45 = L(0.7) + P/Q(0.23) + N(0.27) + R, where L, P/Q, N, and R denote the percentage that each channel type contributes to the whole cell Ca2+ current (approximately): 23 + 30 + 22 = 75%, R-channels being about 20–25% (see above). These results then pointed to the following conclusions: first, they confirmed that L-type channels are preferentially targeted by muscarine (nearly 70% block) (Howe and Surmeier 1995) and, second, they also showed that an important part of the remaining modulation (50%) was shared by other channel types: 23% of P/Q and 27% of N. Bottom graph in Fig. 3G summarizes these data. To conclude, the experiments suggested that the activation of muscarinic receptors was targeting 3 types of Ca2+ channel: L, N, and P/Q.
Next we wanted to see which receptor modulates N- and P/Q-type Ca2+ channels because it is known that L-type Ca2+ channels are controlled by the M1-type receptor (Howe and Surmeier 1995). Figure 4A illustrates that a substantial amount of muscarinic modulation remained in the presence of MT-3 (10–20 nM), a specific antagonist of muscarinic M4-type receptors (Caulfield and Birdsall 1998; Potter et al. 2004). With the block of both M4-type receptors and L-type Ca2+ channels, Ca2+ currents were reduced by 23 ± 4% (n = 19, range: 12–38%), which is not significantly different from the reduction without MT-3 (29%; see above). In contrast, muscarinic modulation was substantially reduced in the presence of MT-7 (10–30 nM), a specific antagonist of muscarinic M1-type receptors (Caulfield and Birdsall 1998; Karlsson et al. 2000) (Fig. 4B). The remaining muscarinic modulation after MT-7 was 9 ± 1% (P < 0.005, Mann–Whitney U test). Cell-to-cell variability is illustrated in the box plot of Fig. 3(inset) . The histogram at the bottom (Fig. 4C) summarizes these experiments, which do not discard a slight modulation by M4-type receptors but clearly suggest that M1-type receptor activation mediates most muscarinic actions on N- and P/Q-type Ca2+ currents.
The above conclusion was further confirmed by the use of the selective M1-type receptor agonist MT-1 (48 nM). Figure 5 compares the action of the MT-1 peptide (48 nM) in the presence of L-type Ca2+ current blockage, first in the absence (top) and then in the presence (middle) of the MT-3 peptide (20 nM). Modulation was not significantly different: 20 ± 2% (n = 8; Fig. 5A) and 25 ± 4% (n = 6; Fig. 5B; P > 0.2, Mann–Whitney U test), indicating that M4-receptor block does not appreciably influence muscarinic effects. A histogram (Fig. 5C) summarizes these results. In contrast, MT-7, the specific M1-type receptor antagonist, reduced MT-1 action to 8 ± 0.2% (n = 3; data not shown).
However, it could still be argued that the action of M4-type receptor activation is hard to see because, in some cells, it is concentrated only on one type of Ca2+ channel, and that only when both M4-type receptors and a given Ca2+ channel are both blocked, an important reduction in muscarine actions could be seen. Thus muscarine was tested with the following blockers administered together: 1) the specific M4-type receptor antagonist MT-3 (20 nM) (M4-type receptors blocked), 2) nitrendipine to block L-type channels (L-type Ca2+ channels blocked), and 3) either N- or P/Q-type Ca2+ channel antagonists (N- or P/Q-type Ca2+ channels blocked). In these conditions, the experiments showed the same magnitude of muscarine effects. That is, blockage of M4-type receptors did not alter muscarinic actions on any one of the channels, N or P/Q, as illustrated in Fig. 6: 23 ± 4% block in 10 μM nitrendipine (Fig. 6A; n = 19; P > 0.19); 14 ± 2% in nitrendipine plus 1 μM ω-CgTx (Fig. 6B, n = 7; P > 0.4); and 13 ± 2% in nitrendipine plus 400 nM ω-AgaTK (Fig. 6C; n = 11; P > 0.8). No mean result was significantly different from the equivalent value obtained with available M4 receptors (Mann–Whitney U test). To summarize, all data obtained with the use of specific muscarinic receptor antagonists indicated that, besides modulating L-type Ca2+ channels, M1-type receptors are also in charge of modulating N- and P/Q-type Ca2+ channels. Given the variability of the responses, the action of M4-type receptor activation could not be completely ruled out, although it is expected to be slight, at best.
Impact of N–P/Q-channel modulation on neuronal firing
Although both N- and P/Q-type Ca2+ channels cooperate to produce the Ca2+ influx that activates the AHP in spiny neurons (Vilchis et al. 2000), it has been found that both channel types have to be inhibited at the same time to significantly reduce the AHP (Perez-Garci et al. 2003). Blockage of only one channel type, N or P/Q, has no impact on the AHP in most cases (Perez-Garci et al. 2003). Experiments described above show that muscarinic M1-receptor activation modulates both channel types. However, this modulation only partially reduces current through each channel type: 27% of N and 23% of P/Q (see above). Therefore experiments were carried out to see whether this modulation was enough to exert an impact on the AHP, i.e., if a partial reduction of Ca2+ entry through each channel was capable of decreasing the AHP. Figure 7A and B, show that MT-7 impedes the action of muscarine on the AHP (n = 3/3). In contrast, muscarine reduced the AHP in the presence of MT-3 (n = 4/4; Fig. 7, C and D) (see Pineda et al. 1995). The result then suggests that the AHP in spiny cells is controlled by M1 but not M4 muscarinic receptors.
During repetitive firing the AHP generates an important component of the interspike interval in spiny neurons. Accordingly, we next evaluated whether the action on the AHP had any impact on evoked discharge. Figure 8, A–D show that both muscarine (1 μM) and MT-1 (48 nM) increased evoked discharge in spiny cells (Pineda et al. 1995). An increase in firing frequency of 44 ± 6% with respect to the control (range = 23–79%; n = 10; P < 0.005, Wilcoxon's t-test) was obtained with muscarine and an increase of 26 ± 5% (range = 16–38%; n = 5; P < 0.05, Wilcoxon's t-test) was obtained with MT-1. Figure 8, E and F also show that posttrain (cumulative) AHP was reduced at the same time. Thus muscarinic actions on the AHP have a clear impact on cell excitability (note that a depolarizing holding potential reduces the impact of inward rectification on evoked firing).
Impact on γ-aminobutyric (GABA) release
A way to selectively stimulate the GABAergic terminals that medium spiny neurons have onto other spiny neurons has been described (Guzman et al. 2003; Tecuapetla et al. 2005) (Fig. 9A). These terminals use N- and P/Q-type Ca2+ channels to trigger GABA release and inhibit other spiny cells (Tecuapetla et al. 2005). Because the above results indicated that cholinergic activation of muscarinic M1 receptors regulates these same Ca2+ channels, a presynaptic regulation of GABA release by M1 receptors can be hypothesized. The presence of this modulation would indicate that M1 receptors are transported from the soma to the synaptic terminals of spiny cells. The paired-pulse protocol and the paired-pulse ratio (PPR) were used to find evidence of this presynaptic control (Barral et al. 1999; Guzman et al. 2003). Figure 9B (top) shows a pair of IPSCs evoked from the GP by antidromically exciting the passing axons from spiny cells (i.e., striofugal fibers). Note the slight synaptic facilitation of the second response in control conditions. The addition of muscarine (1 μM) (Fig. 9B, middle) to the bath saline reduced the IPSC amplitude from a mean of 58 ± 5.4 to 22 ± 4 pA (or 60% amplitude reduction in the first response; n = 6; P < 0.01, Wilcoxon's t-test). IPSC reduction coursed with an increase in PPR or synaptic facilitation in all cells tested (Fig. 9C): from 1.05 ± 0.13 in the control to 1.42 ± 0.16 during muscarine (n = 6; P < 0.05, Wilcoxon's t-test). A time course for this action is illustrated in Fig. 9D, which shows the appearance of failures after muscarine. It was concluded that there is a cholinergic muscarinic modulation of the synaptic terminals that interconnect medium spiny projection neurons.
Figure 10 shows that MT-7 completely blocks this muscarinic effect on synaptic terminals (n = 4; Fig. 10A), whereas MT-3 failed to block it (Fig. 10B) (n = 4). Muscarine still reduced the first IPSC by 52 ± 8% (n = 4) in the presence of MT-3. Consistently, MT-3 did not block the PPR change produced by muscarine (from 1.04 ± 0.01 to 1.34 ± 0.06). It was concluded that, as expected by the types of Ca2+ channel present on these terminals (Tecuapetla et al. 2005), M1 muscarinic receptors control the cholinergic modulation that in turn controls GABA release from medium spiny synaptic terminals at this age.
Are both N- and P/Q-type Ca2+ channels involved in this presynaptic effect? If this were the case, a partial occlusion should be seen after any of their respective blockers. Accordingly, the addition of muscarine in the presence of saturating concentrations of ω-CgTx (1 μM) did not completely block the muscarinic action on the IPSC (Fig. 11A). However, ω-CgTx partially occluded this action to about half (n = 3/3); note the absence of failures (cf. Fig. 9D). The experiment then showed that modulation of N-type Ca2+ channels is important for controlling transmitter release in the GABAergic terminals of spiny neurons at this age.
The specific P/Q-type channel blocker, ω-AgaTK, blocked virtually all evoked GABAergic responses when used in saturating concentrations (Tecuapetla et al. 2005). However, for the purposes of the present work, it was only necessary to show that a partial block of P/Q channels was sufficient to produce a partial occlusion of muscarinic actions to imply P/Q channels in the cholinergic actions. Thus the experiment shown in Fig. 10B used only subsaturating concentrations (50 nM) of ω-AgaTK. In the presence of these concentrations of ω-AgaTK, muscarine still reduced IPSCs by 48 ± 2% (n = 4; cf. 60% without AgaTK; see above), thus yielding again a partial occlusion. Note again the absence of failures in spite of a clear effect (cf. Fig. 9D). The increase in PPR was also partially occluded from 0.95 ± 0.13 to only 1.14 ± 0.21. To summarize, M1-receptor action on synaptic terminals of medium spiny cells modulates both N- and P/Q-channel types to regulate GABA release.
The present results demonstrate that cholinergic modulation of CaV2.1 (P/Q-type) and CaV2.2 (N-type) Ca2+ channels in neostriatal neurons is preferentially mediated by the activation of muscarinic M1-type receptors, suggesting that the signaling cascade that mediates this modulation is a slow diffusible pathway, probably linked to phospholipase C and protein kinase C (Caulfield and Birdsall 1998; Jones 1993; Perez-Rosello et al. 2004). M1-receptor actions on N- and P/Q-type Ca2+ channels importantly resulted in modulating the discharge frequency of spiny neurons (neostriatal output) and the surrounding inhibitory control between spiny cells (neostriatal lateral inhibition). Therefore ACh both facilitates firing frequency and inhibits GABAergic interconnections. These 2 actions combined should lead to an increase in the recruitment of projection neurons (Wickens and Oorschot 2000). In addition, it has been reported that muscarinic actions presynaptically inhibit glutamatergic afferents from the cortex (Barral et al. 1999; Calabresi et al. 2000; Hernandez-Echeagaray et al. 1998) and postsynaptically potentiate N-methyl-d-aspartate (NMDA)-dependent responses (Calabresi et al. 1998). Assuming that these actions may go on simultaneously, what then is the global action? Similar results in the cortex, fed to a network model, led to the hypothesis that ACh selects some incoming afferents, while simultaneously ensuring that the recipient neurons respond vigorously to these afferents. This may improve the encoding of new memory traces (Hasselmo and McGaughy 2004).
In addition to the results obtained with specific peptide antagonists (see results), our experiments were dedicated to see whether the modulation had a voltage-dependent component that yielded negative results (data not shown). Although M4-type receptors do not necessarily use a voltage-dependent pathway (Shapiro et al. 1999), when taken together, the available evidence suggests that M4-type (M2-class) receptors are not the main regulators of Ca2+ currents in neostriatal projection neurons (see Bernheim et al. 1992; Liu and Rittenhouse 2003; Shapiro et al. 1999). Thus the relatively high abundance of M4-type receptors (Potter et al. 2004) is still in need of an explanation (Bernard et al. 1992; Potter et al. 2004; Yan et al. 2001). M2-class receptors are the main regulators of Ca2+ channels and firing of cholinergic interneurons (Calabresi et al. 2000; Galarraga et al. 1999; Yan and Surmeier 1996), indicating that different signaling systems are used for projection neurons and interneurons.
Ca2+ current in spiny neurons and muscarinic receptors
A muscarinic reduction of Ca2+ action potentials had been described in neostriatal neurons (Misgeld et al. 1986). Although the present voltage-clamp experiments were done in dissociated cells without dendrites, recordings were obtained from 6–7 pF cells, which were the majority in the preparation, and were previously characterized as somata belonging to medium spiny neurons (Yan and Surmeier 1996). Moreover, percentages of Ca2+ current types, similar to those found here, have been shown to be expressed in both substance P and enkephalin-containing striatal neurons (Bargas et al. 1994). Finally, the impact of M1 muscarinic actions was also tested by recording neurons in slices, and many of these neurons were filled with biocytin and later seen as medium spiny. This made us reasonably sure that the results presented here can be attributed to medium spiny neostriatal projection neurons.
Up to now, precise identification of the main cholinergic receptor involved in controlling Ca2+ entry in spiny neurons had been hampered by the lack of specific receptor agonists and antagonists (Caulfield and Birdsall 1998; Jones 1993). The presence of both M1- and M4-type muscarinic receptors (Bernard et al. 1992; Yan and Surmeier 1996; Zhou et al. 2003) had obscured previous results because, for example, pirenzepine has a weak preference for M1- over M4-type receptors and both pirenzepine and the MT-3 peptide compete for the M4-type receptor (Olianas et al. 1996; Purkerson and Potter 1998). This prompted some investigators to classify pirenzepine as a selective M1/4 antagonist (see Caulfield and Birdsall 1998), not adequate to be used on cells with both receptor types.
Because of this uncertainty, we resorted to muscarinic ligands that are considered as the most selective antagonists found to date: the MT- 3 and MT- 7 peptides. MT-7 has a 4 orders-of-magnitude higher affinity for M1- over M4-type receptors. Conversely, MT-3 has a 100-fold higher affinity for M4 than for M1 receptors (Caulfield and Birdsall 1998; Karlsson et al. 2000; Potter et al. 2004). It was assumed that the combined use of both peptides on medium spiny cells would tend to decrease error. Besides, these peptides have been seen to affect the function, at the systems level, of several brain circuits, including the neostriatum (Cuevas and Adams 1997; Karlsson et al. 2000; Marino et al. 1998; Potter et al. 2004; Purkerson and Potter 1998).
It was under these conditions that a preferential muscarinic modulation of CaV2.1 (P/Q-type) and CaV2.2 (N-type) Ca2+ channels by M1-type receptors was observed. High levels of mRNA expression for M1 muscarinic receptor have been found in virtually all medium spiny neurons tested, whereas the messenger for M4 muscarinic receptors is only weakly present in subsets of these neurons (Bernard et al. 1992; Yan et al. 2001). In agreement with this, muscarinic inhibition of N- and/or P/Q-type Ca2+ currents was consistently seen, with the expected variability, in all neurons tested for the present study. This adds up to other probable M1-receptor actions on spiny cells (Calabresi et al. 1998; Galarraga et al. 1999; Howe and Surmeier 1995; Hsu et al. 1996; Pineda et al. 1995).
Impact of Ca2+ current modulation on the firing pattern
Ca2+ entry through N and P/Q channels, in particular, makes up the Ca2+ source that activates Ca2+-dependent K+ currents in neostriatal spiny neurons (Vilchis et al. 2000). These K+ currents are not activated by Ca2+ entry through L-type Ca2+ channels (Vilchis et al. 2000), which have a completely different function (Perez-Garci et al. 2003; see also Catterall 1998; Reuter 1996). In turn, Ca2+-dependent K+ currents generate the AHP (Bargas et al. 1999; Galarraga et al. 1989; Pineda et al. 1992) and are thus in position to regulate the interspike interval and the firing frequency. However, N and P/Q channels need to cooperate to tune the firing frequency because the complete blockade of only one of them, N or P/Q, is not sufficient to significantly change the firing frequency. Both have to be inhibited simultaneously to increase evoked discharge (Perez-Garci et al. 2003). However, the muscarinic modulation described in this work involves only a partial reduction of each current type (between 20 and 30% in each case). Then, one question is whether this partial modulation is sufficient to change the firing frequency. The present experiments show that partial muscarinic modulation of both channel types (N and P/Q), by M1-type receptors, was enough to significantly reduce the AHP and to cause an increase in firing frequency. The participation of the M4-type receptor in these actions was negligible. Moreover, MT-1 concentrations that selectively activate M1-type receptors mimicked muscarine actions. The almost complete blockage of these actions by MT-7 (M1 antagonist) and the absence of effects of the MT-3 peptide (M4 antagonist) indicated that the effects occurred, virtually exclusively, by the M1 modulation of N- and P/Q-type Ca2+ channels.
This may prove to be important for understanding striatal physiology because another important transmitter, dopamine, uses a similar signaling pathway by D2 receptors to modulate Ca2+ currents and to induce changes of firing frequency in these neurons (Hernandez-Lopez et al. 2000): both D2 and M1 receptors decrease Ca2+ flow through L- and other Ca2+-channel types. Unexpectedly, however, the changes these transmitters produce are opposite: D2 receptor activation decreases (Hernandez-Lopez et al. 2000), whereas M1 receptor activation increases firing frequency (Galarraga et al. 1999), as seen in the present work. In the case of D2 receptors, the effect on L-type current is predominant because the result of reducing this current is a decrease in firing (Hernandez-Lopez et al. 2000; Olson et al. 2001; Perez-Garci et al. 2003; see also Catterall 1998). Consequently, this cannot be the main effect of muscarinic M1 receptors because their activation increases firing. Therefore the main effect of producing changes in firing has to be the reduction of Ca2+ currents that activate the AHP and, possibly, some additional effects on other K+ currents. The present data then reveal an intriguing divergence of effects for receptors that supposedly use the same signaling cascade (Rakhilin et al. 2004).
M1 modulation of synaptic transmission
Finally, the present findings indicate that GABA release from axon collaterals of medium spiny neurons is also controlled by M1-receptor activation. This is a departure from the commonly used membrane-delimited pathway used in many synaptic terminals (Currie and Fox 2002). It is also a counterintuitive result because the Gq/11-PLC pathway may lead to IP3 formation and Ca2+ increase in the synaptic terminals with a consequent enhancement of synaptic transmission (Chameau et al. 2001; Yang et al. 2001). However, presynaptic inhibition has been reported for transmitters that use the Gq/11-PLC signaling pathway, including muscarinic agonists (e.g., Hernandez-Echeagaray et al. 1998; Kimura and Baughman 1997; Tan et al. 2003).
Muscarinic inhibition of GABA release in the neostriatum has previously been reported in synaptosomal preparations (Raiteri et al. 1990) and in in vivo microdialysis studies (Smolders et al. 1997). In addition, field-evoked GABAergic potentials have been reported to be inhibited by muscarinic agonists, an effect that was blocked by pirenzepine (Kimura and Baughman 1997; Momiyama and Koga 2001; Sugita et al. 1991). Nevertheless, the actions reported with the use of brain slice preparations have been conflicting: both muscarinic inhibition and stimulation of GABA release have been described (Harsing and Zigmond 1998; Kayadjanian et al. 1994; Kuriyama et al. 1984), probably arising from an unspecific augmentation of spiny neurons excitability that surpasses the presynaptic inhibition of GABA release. Alternatively, indirect effects are possible because several interneurons also express muscarinic receptors (Bernard et al. 1992; Galarraga et al. 1999; Hersch and Levey 1995; Yan and Surmeier 1996).
Thus to study muscarinic actions on GABAergic transmission, the present work made use of isolated inhibitory synaptic currents (IPSCs) from the recurrent axon collaterals that interconnect spiny neurons (see Guzman et al. 2003; Tecuapetla et al. 2005). In this defined set of afferents, muscarine produced a reduction of evoked IPSCs, accompanied by an increase in the paired-pulse ratio, suggesting presynaptic inhibition. These actions were virtually abolished by MT- 7 but not by MT-3. Therefore the actions were mediated by M1-type receptors. This outcome fits well with M1 actions on somatic Ca2+ currents (see above) because N- and P/Q-type Ca2+ channels control synaptic transmission in this and many other synapses (Momiyama and Koga 2001; Murakami et al. 2002; Reuter 1996; Rhee et al. 1999; Tecuapetla et al. 2003; Wheeler et al. 1994; Wu and Saggau 1994). However, because the channel types used at the synaptic terminals may change during development (Iwasaki et al. 2000; Momiyama 2002; Urbano et al. 2002), we asked whether muscarinic actions specially targeted a defined type of Ca2+ channel at the terminals. Our data indicated that both N- and P/Q-type Ca2+ channels were equally targeted at this stage.
In summary, most Ca2+ channel modulation in spiny neostriatal neurons falls into the M1-type receptor domain. The same is true for the inward rectification and input conductance (leak) of these cells (Figueroa et al. 2002; Galarraga et al. 1999). It is concluded that M1 receptors are a predominant factor in regulation of the excitability of neostriatal projection neurons. These results are expected to explain, in part, the important cholinergic actions of basal ganglia physiology found or inferred from clinical and systems studies.
This work was supported by the following grants: Dirección General de Asuntos del Personal Académico–Universidad Nacional Autónoma de México IN201603 and IN219304 to J. Bargas and IN200803 and IN205804 to E. Galarraga and by Consejo Nacional de Ciencia y Tecnología (Mexico) Grants 31839 to J. Bargas and 42636 to E. Galarraga.
The authors thank D. Tapia and A. Laville for technical support. A. Perez-Burgos collaborated in some final experiments.
↵✠ Deceased August 21, 2004.
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 © 2005 by the American Physiological Society