HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (5) Citing Articles via Google Scholar Google Scholar Articles by Moldavan, M. G. Articles by Allen, C. N. Search for Related Content PubMed PubMed Citation Articles by Moldavan, M. G. Articles by Allen, C. N.

Presynaptic GABA_B Receptors Regulate Retinohypothalamic Tract Synaptic Transmission by Inhibiting Voltage-Gated Ca²⁺ Channels

¹Center for Research on Occupational and Environmental Toxicology and ²Department of Physiology and Pharmacology, Oregon Health and Science University, Portland, Oregon

Submitted 31 August 2005; accepted in final form 25 February 2006

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Presynaptic GABA_B receptor activation inhibits glutamate release from retinohypothalamic tract (RHT) terminals in the suprachiasmatic nucleus (SCN). Voltage-clamp whole cell recordings from rat SCN neurons and optical recordings of Ca²⁺-sensitive fluorescent probes within RHT terminals were used to examine GABA_B-receptor modulation of RHT transmission. Baclofen inhibited evoked excitatory postsynaptic currents (EPSCs) in a concentration-dependent manner equally during the day and night. Blockers of N-, P/Q-, T-, and R-type voltage-dependent Ca²⁺ channels, but not L-type, reduced the EPSC amplitude by 66, 36, 32, and 18% of control, respectively. Joint application of multiple Ca²⁺ channel blockers inhibited the EPSCs less than that predicted, consistent with a model in which multiple Ca²⁺ channels overlap in the regulation of transmitter release. Presynaptic inhibition of EPSCs by baclofen was occluded by -conotoxin GVIA (72%), mibefradil (52%), and -agatoxin TK (15%), but not by SNX-482 or nimodipine. Baclofen reduced both evoked presynaptic Ca²⁺ influx and resting Ca²⁺ concentration in RHT terminals. Tertiapin did not alter the evoked EPSC and baclofen-induced inhibition, indicating that baclofen does not inhibit glutamate release by activation of Kir3 channels. Neither Ba²⁺ nor high extracellular K⁺ modified the baclofen-induced inhibition. 4-Aminopyridine (4-AP) significantly increased the EPSC amplitude and the charge transfer, and dramatically reduced the baclofen effect. These data indicate that baclofen inhibits glutamate release from RHT terminals by blocking N-, T-, and P/Q-type Ca²⁺ channels, and possibly by activation of 4-AP–sensitive K⁺ channels, but not by inhibition of R- and L-type Ca²⁺ channels or by Kir3 channel activation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Suprachiasmatic nucleus (SCN) neurons of mammals contain a biological clock that synchronizes physiological processes with changing environmental conditions. To maintain proper temporal coupling between the circadian clock and the environment, neural systems have evolved to sense changing conditions and communicate that information to the clock. The daily change in light intensity (day/night) is the most important of these environmental signals. The SCN receives light intensity information by the retinohypothalamic tract (RHT), a direct glutamatergic projection from retinal ganglion cells (RGCs) to the SCN (Berson et al. 2002; Card and Moore 1991; Castel et al. 1993; Liou et al. 1986)

Light’s effect on circadian clock timing can be modified by neurotransmitters activating presynaptic receptors, thereby reducing glutamate release from RHT terminals. For example, activation of 5-HT_1B receptors reduces c-fos expression and the magnitude of light-induced phase shifts (Pickard et al. 1996, 1999). Similarly, activation of -aminobutyric acid type B (GABA_B) receptors inhibits light-induced phase advances and phase delays (Colwell et al. 1993; Gillespie et al. 1997; Ralph and Menaker 1989). Baclofen, a GABA_B agonist, reduces the light-induced c-fos expression in the SCN that is often used as a physiologic measure of the activation of SCN neurons by light stimulation (Colwell et al. 1993; Gillespie et al. 1997; Pickard et al. 1996). Thus GABA_B receptors as well as 5-HT_1B receptors "act to prevent photic information from reaching the SCN" (Colwell et al. 1993; Pickard et al. 1996). In hypothalamic slices baclofen reduces the amplitude of excitatory postsynaptic currents (EPSCs) evoked by optic nerve stimulation without altering the intrinsic activity of postsynaptic glutamate receptors (Jiang et al. 1995). GABA_B-receptor activation also reduces the frequency but does not alter the distribution of miniature EPSC amplitudes in SCN (Jiang et al. 1995). Further, the GABA_B antagonist phaclofen potentates the field potential induced by optic nerve stimulation, consistent with a tonic GABA_B-receptor–mediated inhibition (Gannon et al. 1995). Some of the RGCs that possess G-protein–coupled GABA_B receptors may contribute axons to the RHT (Chen et al. 2004; Slaughter 1995). These data are consistent with the hypothesis that activation of presynaptic GABA_B receptors inhibits glutamate release from RHT terminals (Gannon et al. 1995; Jiang et al. 1995).

Currently, the specific mechanism(s) by which GABA_B receptors regulate neurotransmitter release at RHT terminals is not known. In the hippocampus, GABA_B-receptor activation inhibits voltage-dependent Ca²⁺ channels and activates K⁺ channels through a G(i/o)-type G-protein–mediated pathways (Andrade et al. 1986; Sodickson and Bean 1996). The GABA_B-receptor–mediated effect is attenuated by pertussis toxin uncoupling the G(i/o)-protein from the GABA_B receptor. Loading G-protein beta-gamma subunits (G) into the calyceal nerve terminal partially occluded the inhibitory effect of baclofen on presynaptic Ca²⁺ currents (Kajikawa et al. 2001). G directly inhibits Ca²⁺ channels, putting them into a "reluctant" state (Bertram et al. 2003; Zamponi and Snutch 1998). Activation of GABA_B receptors suppresses GABA release from the terminals of cultured SCN neurons through a G-protein–mediated inhibition of N- and P/Q-type Ca²⁺ channels (Chen and Van den Pol 1998). Also baclofen-induced inhibition of N-type Ca²⁺ channels was found in axons of neonatal rat optic nerve (Sun and Chiu 1999). In addition, baclofen activates fast inactivating A-type K⁺ channels and Ca²⁺-activated K⁺ channels (SK channels) (Bettler et al. 2004; Saint et al. 1990). We performed experiments to test the hypothesis that activation of presynaptic GABA_B receptors acts to decrease evoked glutamate release from RHT terminals by inhibiting voltage-dependent Ca²⁺ channels or by activating K⁺ channels.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Preparation of SCN brain slices

Male Sprague–Dawley rats (4–6 wk old) were housed in an environmental chamber (Percival Scientific, Perry, IA) maintained at 20–21°C on a 12 h light/12 h dark schedule. During the lights-on phase, rats were deeply anesthetized with halothane, their brains removed and submerged in an ice-cold Krebs solution consisting of (in mM): NaCl 126, KCl 2.5, NaH₂PO₄ 1.2, MgCl₂ 4.0, CaCl₂ 0.5, glucose 11, and NaHCO₃ 26, saturated with 95% O₂-5% CO₂ (pH 7.3–7.4, 301–303 mOsm). Coronal 250- to 300-µm-thick slices of the hypothalamus containing the SCN were cut with a vibrating-blade microtome (Leica VT 1000 S, Nussloch, Germany). The Institutional Animal Care and Use Committee of OHSU approved all experimental procedures involving animals and all efforts were made to minimize pain and the number of animals used.

Whole cell patch-clamp recording

Recordings were made at 28–30°C using the whole cell patch-clamp technique from 1 to 8 h after slicing (Fig. 1A). The recording solution was identical to the slicing solution, but contained (in mM): NaCl 130, NaHCO₃ 22, CaCl₂ 2.4, and MgCl₂ 1.2 (pH 7.3–7.4, 300–305 mOsm). Microelectrodes with resistances of 7–9 M were pulled from borosilicate glass (World Precision Instruments, Sarasota, FL) and filled with a solution containing (in mM): CsCH₃O₃S 105, CsCl 20, CaCl₂ 1, HEPES 10, EGTA 11, CsOH 25, ATP 3, GTP 0.3, and QX-314 5. QX-314 was included in the patch pipette solution to block voltage-activated Na⁺ channels and K⁺ channels. Cs⁺ was used to block K⁺ channels including GABA_B-activated K⁺ channels (Jiang et al. 1995). To prevent activation of GABA_A receptors, by GABA released from SCN GABAergic neurons after RHT stimulation, picrotoxin 50 µM was added to the external solution. Individual SCN neurons were visualized with infrared illumination and differential interference contrast optics using a Leica DMLFS microscope with video camera and display (Sony). On-line data collection and analysis were performed using an EPC-7 patch-clamp amplifier (HEKA Electronik, Lambrecht, Germany), a Macintosh G3 computer, and Pulse and PulseFit (HEKA). The records were filtered at 3 kHz and digitized at 10 kHz.

View larger version (15K): [in this window] [in a new window]   FIG. 1. Schema of recording and test agent application onto suprachiasmatic nucleus (SCN) neurons in the rat brain coronal slices. A: whole cell patch-clamp recording. B: Ca2+ probe injection into the optic chiasm for Ca2+ imaging.

Optic chiasm stimulation

EPSCs were evoked by electrical stimulation of the optic chiasm. A bipolar concentric tungsten electrode (FHC, Bowdoinham, ME) connected to a stimulus isolation unit (model SIU5B, Grass Medical Instruments, Quincy, MA) was placed in the optic chiasm and stimulated using a Grass S88 stimulator (Grass Medical Instruments). The pulse duration was 0.13–0.17 ms and the stimulation intensity was set 1.5–2 times higher than that needed to evoke a threshold response and usually varied between 8 and 40 V. EPSCs were elicited by electrical stimuli (square pulses) at 0.08 Hz.

Test agent application

All test agents were applied by perfusion of the recording chamber with artificial cerebrospinal fluid (ACSF) containing the final concentration of the compound. Chamber volume was about 400 µl; a complete change of the external solution took <30 s at a flow rate of 1.5–2 ml/min. (±)Baclofen (GABA_B receptor agonist, 0.03–30 µM), mibefradil (T-type Ca²⁺ channel blocker, 20 µM), nimodipine (L-type Ca²⁺ channel blocker, 10 µM), lidocaine N-ethylchloride (QX-314), D-(–)-2-amino-5-phosphonovaleric acid (D-APV, 50 µM), CNQX (10 µM), picrotoxin (50 µM), and DMSO were purchased from Sigma (St. Louis, MO). -Agatoxin TK (P/Q-type Ca²⁺ channel blocker, 100 or 500 nM), -conotoxin GVIA (N-type Ca²⁺ channel blocker 1 µM), SNX-482 (R-type Ca²⁺ channel blocker, 150 nM), tetrodotoxin (TTX, 750 nM), and tertiapin (a potent inhibitor of Kir3.1, Kir3.4, Kir1.1, and K_ACh inwardly rectifying K⁺ channels, 10 or 100 nM) were purchased from Alomone Labs (Jerusalem, Israel). 4-Aminopyridine (4-AP, 1–5 mM) was purchased from Sigma (St. Louis, MO) and Tocris Cookson (Ellisville, MO). Appropriate stocks were made and diluted with ACSF just before application. -Agatoxin TK, -conotoxin GVIA, SNX-482, tertiapin, mibefradil, and baclofen, in experiments with the Ca²⁺ channel blockers, were applied by perfusion through a micropipette with an internal diameter of 100 µm. ACSF applied through micropipette contained the final concentration of test agent. The micropipette was placed close to the slice surface upstream of the SCN and the ACSF containing the toxins flowed out of the perfusion micropipette in the same direction to the flow of ACSF in the recording chamber (Fig. 1A). The flow from micropipette completely covered the SCN. These Ca²⁺ channel blockers were applied for 20–40 min.

Optical measurements

Stock solutions (2.5 mM) of Fluo-4 acetoxymethyl ester (AM) (K_d 345 nM) or Fura-Red AM (K_d 140 nM; Molecular Probes, Eugene, OR) dissolved in DMSO (Mallinckrodt, Hazelwood, MO) were diluted with double-strength ACSF (containing in mM NaCl 240, KCl 5, HEPES 20, pH 7.4 without Mg²⁺, Ca²⁺, or PO₄) to a final concentration of 250 µM, and briefly triturated. Glass pipettes with tips of approximately 5 µm were filled with probe solution and pressure injected (Picospritzer II; General Valve, Fairfield, NJ) over 20–40 min into the optic chiasm approximately 250–300 µM away from the SCN. The pipette tip was positioned at an angle pointing away from the SCN. A suction pipette was placed adjacent to the injection site to remove any extraneous probe (Fig. 1B). In addition, the laminar flow (2 ml/min) of ACSF in the chamber was aimed away from the SCN to further reduce the possibility of depositing probe outside of the optic chiasm. Further, the likelihood of the probe getting into SCN neurons is low because rats of this age (>4 wk) have minimal uptake of AM probes into neurons while presynaptic terminals load well (Colwell 2000; Yuste 2000). After injection, 2–3 h were allowed for probe transport into the RHT terminals. For experiments with Fluo-4 AM, a small amount of Texas Red-dextran (Molecular Probes) was added to the solution to help visualize the time course of probe loading. Fluorescent images were obtained with an upright microscope (DM LFS; Leica) with a water immersion objective (HCX AP L63X/0.9W U-V-I; Leica). To capture fast Ca²⁺ transients, single-wavelength excitation was used. Excitation light (Fluo-4 480 nm; Fura-Red 440 and 488 nm) was from a monochronometer (Polychrome II or IV; Till Photonics, Martinsried, Germany), attenuated with a 0.6 or 1.3 neutral density filter (Chroma, Brattleboro, VT), passed through a band-pass filter (Fluo 4, HQ480/40x; Fura Red, D500/200, Chroma, Rockingham, VT) and reflected by a dichroic mirror (Fluo-4 Q505LP; Fura-Red 505dcxru; Chroma). The emitted light was filtered (Fluo-4, HQ535/50m; Fura-Red HQ645/75; Chroma). Optical recordings were made with either a cooled charge-coupled device (CCD, 16-bit level) camera (C6790 or ORCA ER; Hamamatsu Photonics, Hamamatsu, Japan) with binning 32x32 to minimize photo bleaching of the probe and maximize speed of data collection (3–49 ms/image), or a photodiode (Till Photonics S4753-02, Martinsried, Germany). Experiments using the CCD camera were controlled by digital imaging software (ARGUS HiSCA; Hamamatsu Photonics) and photodiode experiments by Pulse (HEKA). During the recording, slices were perfused in a 0.5 to 0.75 ml chamber at a temperature of 25°C. Electrical stimulation was by a concentric bipolar electrode placed in the optic chiasm about 200 µm from the SCN, providing 200 µs pulses at 0.08 Hz (PG4000 Digital Stimulator and SIU90; Neurodata Instrument, New York, NY). Optical data from photodiode and images from the CCD camera were converted to relative fluorescence intensity units (F), background was subtracted and corrected for signal loss over time, and the stimulation-induced presynaptic Ca²⁺ transients were calculated using the fluorescence ratio F/Fo, where Fo is the baseline fluorescence before stimulation. In some experiments, the resting estimated Ca²⁺ concentration (est [Ca²⁺]) was determined by linearizing the Fura-Red AM response to Ca²⁺ by using two excitation wavelengths, 440 and 488 nm, and the formula Est [Ca²⁺] = (R – Rmin)/(Rmax – R) x Sf/b x K_d (Grynkiewicz et al. 1985), where Rmin, Rmax, and Sf/b (free/bound fluorescence at 488 nM) were determined in vitro using Fura-Red tetrapotassium salt with either 10 mM Ca²⁺ for Rmax or 10 mM EGTA for Rmin; R is the ratio 440/488 nm corrected for background, and K_d was taken to be 140 nM (Molecular Probes).

Math and statistical analysis

To calculate the inhibitory effect of baclofen on each Ca²⁺ channel type we used the equation: Ic = {1 – [(Itb – It)/Ib]} x 100%, where Ic is the percentage occlusion of baclofen effect by a Ca²⁺ channel blocker, It is the percentage inhibition of the EPSC by the Ca²⁺ channel blocker, Ib is the percentage inhibition of the EPSC induced by baclofen alone, and Itb is the percentage inhibition of EPSC produced by baclofen and the Ca²⁺ channel blocker together.

The peak amplitude, charge transfer, and 10–90% rise time for EPSCs were averaged across neurons (five current traces were averaged for each neuron). EPSC amplitudes were measured as the difference between the peak EPSC current and the baseline current before the stimulus artifact. Data for each neuron represent an average of EPSC amplitudes from five sweeps over 1 min. Between neurons, these EPSC amplitudes were normalized and presented as the means ± SE. Charge transfer represented the area enclosed by the EPSC and was analyzed during a 0.15-s period. To compare experimental protocols, amplitude, charge transfer, and rise time of EPSC were expressed as a percentage of control. Igor Pro (Version 5.0, Wave Metrics, Lake Oswego, OR) was used for curve fitting and data analysis. ANOVA, two-tail paired t-test, and unpaired t-test were performed using StatView 5.0.1 (SAS Institute, Cary, NC) or Excel 11.1.1 (Microsoft, Redmond, WA). A confidence level of 95% was used to determine statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
GABA_B activation inhibits evoked EPSCs

Joint application of CNQX (10 µM) and APV (50 µM) with picrotoxin (50 µM) was used to confirm that the currents activated in SCN neurons by stimulation of the optic chiasm were EPSCs mediated by glutamate release. CNQX (10 µM) and APV (50 µM) applied together significantly decreased the EPSC amplitude to 11.0 ± 1.9 pA or 7.3 ± 1.5% of control (control amplitude 156 ± 16.8 pA, n = 5, P < 0.0001; Fig. 2). This reduces the EPSC amplitude and confirms that stimulation of the optic chiasm evokes EPSCs mediated by glutamate-receptor activation.

View larger version (13K): [in this window] [in a new window]   FIG. 2. Inhibition of evoked excitatory postsynaptic currents (EPSCs) by -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-D-aspartate (NMDA) receptor antagonists. CNQX (10 µM) and D-(–)-2-amino-5-phosphonovaleric acid (D-APV, 50 µM) were applied together in the presence of picrotoxin (50 µM). Each point represents the mean amplitude of 5 evoked EPSCs. Inset: examples of individual EPSCs.

View larger version (11K): [in this window] [in a new window]   FIG. 3. Reduction by baclofen of the amplitude of EPSCs evoked in SCN neurons by optic chiasm stimulation. A: example of the reduction of EPSC amplitude by baclofen (10 µM). Each point represents the means ± SE of the amplitudes of 5 EPSCs recorded from a neuron during a 1-min period. Optic chiasm was stimulated once each 12 s (0.08 Hz). B: individual EPSCs recorded at the times indicated by the numbers on the graph (A) shown above. , onset of optic chiasm stimulation. Vh = –60 mV. C: concentration dependency of the baclofen inhibition of EPSCs during the subjective day and night. Baclofen reversibly inhibited EPSCs in a concentration-dependent (0.03–30 µM) manner. IC50 of baclofen during the light and dark phases was not significantly different. Each point represents the means ± SE of the EPSC amplitudes from 4 to 9 neurons. Light phase (circles, n = 42 neurons); dark phase (squares, n = 75 neurons).

Because the relationship between the presynaptic Ca²⁺ concentration and glutamate release is unknown for RHT terminals, we examined the effects of changing the extracellular Ca²⁺ concentration ([Ca²⁺]_e) on the EPSC amplitude (Fig. 4). Decreasing the [Ca²⁺]_e from 2.4 to 1.2 mM significantly reduced the EPSC amplitude to 52.9 ± 3.8% of control (98.1 ± 5.0 pA, P < 0.0001, n = 8), whereas a reduction to 0.6 mM further significantly reduced the EPSC amplitude to 20.6 ± 3.5% of control (P < 0.0001, n = 7, Fig. 4A). The EPSC amplitude increased as [Ca²⁺]_e was raised from 0.6 to 2.4 mM, saturated when the extracellular [Ca²⁺]_e reached 2.4 mM, and remained stable through 3.5 mM [Ca²⁺]_e (Fig. 4B).

View larger version (20K): [in this window] [in a new window]   FIG. 4. Effect of external Ca2+ concentration ([Ca2+]e) and baclofen on evoked EPSCs and Ca2+ transient (F/Fo). A: example showing the change in the amplitude of EPSCs evoked by optic chiasm stimulation in artificial cerebrospinal fluid (ACSF) with varying [Ca2+]e. B: effect of varying [Ca2+]e on Ca2+ transients (Fluo-4 AM), EPSC amplitudes, and the inhibitory action of baclofen (10 µM). Magnitude of presynaptic inhibition induced by baclofen depends on the [Ca2+]e. EPSC and Ca2+ transient amplitudes recorded in 2.4 µM [Ca2+]e were used as the control value. C: relationship between EPSC amplitude and the magnitude of the Ca2+ transient. D: image of Fluo-4–loaded retinohypothalamic tract (RHT) terminals in the SCN; white bar = 10 µm. E: example showing the effect of baclofen on evoked RHT presynaptic Ca2+ transients (arrows) and resting RHT Ca2+, using Fluo-4 AM. F: representative experiment, using Fluo-4 AM, showing reduction of electrically evoked Ca2+ transient amplitude by baclofen. Each point represents a single evoked Ca2+ transient as a percentage of control.

GABA_B-receptor activation could inhibit EPSCs by reducing the action potential–evoked increase in the presynaptic terminal Ca²⁺ concentration required to trigger neurotransmitter release. Because of the nonlinearity of the relationship between Ca²⁺ entry and transmitter release a relatively small decrease in the evoked presynaptic Ca²⁺ transient might significantly reduce glutamate release. To test this hypothesis we measured the changes in RHT terminal Ca²⁺ probe fluorescence after optic chiasm stimulation. We first confirmed that the optical measurements were from RHT terminal Ca²⁺ and not from postsynaptic SCN neurons. Stimulation of the optic nerve produced a rapid increase in Ca²⁺ probe fluorescence (transients) that was fully blocked by TTX (750 nM). Simultaneous application of CNQX (10 µM), APV (50 µM), and picrotoxin (50 µM) did not change the RHT Ca²⁺ transients, demonstrating that activation of postsynaptic AMPA, NMDA, or GABA_A receptors did not confound the RHT-presynaptic Ca²⁺ signal (Gompf et al. 2005). Visual inspection of RHT terminals demonstrated probe fluorescence in terminal processes, but not in neuronal cell bodies (Gompf et al. 2005). These data strongly support the conclusion that the Ca²⁺ signal measured was presynaptic. Because presynaptic Ca²⁺ transients might reach the micromolar range, it is possible that high-affinity Ca²⁺ probes such as Fluo-4 or Fura-Red may become saturated. However, presynaptic Ca²⁺ transients could be enhanced by 4-AP (1 mM) or by stimulation of the optic chiasm with pulses trains (Gompf et al. 2005). Therefore changes in Ca²⁺ probe fluorescence after a single stimulating pulse were within the linear range for the Ca²⁺ concentration response of the probe.

Relationship between presynaptic Ca²⁺ entry and neurotransmitter release

To examine the relationship between the evoked EPSC amplitude and evoked presynaptic Ca²⁺ transients, the extracellular Ca²⁺ concentration ([Ca²⁺]_e) was varied from 0.6 to 3.5 mM while electrically stimulating the optic chiasm. Concentration–response curves for the [Ca²⁺]_e and either EPSC amplitude or Ca²⁺ transients demonstrated a saturating nonlinear relationship, with near saturation at about 2.4 mM [Ca²⁺]_e (Fig. 4B). At 0.6 mM [Ca²⁺]_e, presynaptic Ca²⁺ transient amplitudes in RHT terminals were reduced nearly 45% compared with a roughly 80% reduction of the EPSC amplitude (Fig. 4C). This relationship implies that small changes in the amplitude of RHT terminal Ca²⁺ transients equate to a much larger reduction in the EPSC amplitude.

Effect of baclofen on electrically induced RHT presynaptic Ca²⁺ transients

Recordings of evoked Ca²⁺ transients and resting Ca²⁺ concentration were made using Fluo-4 AM; loaded RHT terminals with either photometry or Ca²⁺ imaging techniques (six and eight experiments, respectively) showed similar results (Fig. 4D). Experiments were performed in the presence of CNQX, APV, and picrotoxin. Baclofen (30 µM) significantly attenuated (87.9 ± 1.3% of control, P < 0.0001) evoked Ca²⁺ transients and reduced the baseline Ca²⁺ probe fluorescence 1.38 ± 0.25% (P < 0.01, n = 14). Because of the need for fast recordings, we used a single-wavelength Ca²⁺ probe in the linear range to calculate F/Fo [i.e., (peak – baseline)/baseline]. This assumes that the baseline Ca²⁺ concentration does not change with drug exposure. However, we have observed that baclofen lowers resting RHT terminal Ca²⁺ (Fig. 4E), and thus the evoked Ca²⁺ transient could underestimate the ability of baclofen to lower presynaptic Ca²⁺. Because measured Ca²⁺ transients reflect only activated terminals, whereas resting Ca²⁺ reflects all the terminals, we were unable to combine the baclofen-induced reduction of resting Ca²⁺ with the reduction in the Ca²⁺ transient. Nonetheless, the reductions observed in EPSCs (roughly 80%) versus presynaptic calcium transients (nearly 12%) during baclofen application are consistent with the nonlinear relationship between Ca²⁺ entry into presynaptic terminals and neurotransmitter release (Fig. 4C).

Dependency of baclofen effect on extracellular Ca²⁺

Reducing [Ca²⁺]_e to 0.6 mM decreased the EPSC amplitude during baclofen (10 µM) application to 2.1 ± 0.2% of control, a value that was significantly lower (P < 0.0001) than that when baclofen was applied at 2.4 mM [Ca²⁺]_e (26.3 ± 3.0% reduction). Decreasing [Ca²⁺]_e to 1.2 mM during baclofen application reduced the evoked EPSC amplitude to 5.0 ± 1.1% (P < 0.0001) of control. The magnitude of baclofen’s inhibition was stable between 2.4 and 3.5 mM [Ca²⁺]_e, which produced maximal EPSCs and Ca²⁺ transients (Fig. 4B). These data indicate a strong correlation between the magnitude of baclofen’s inhibition, Ca²⁺ transient amplitude, EPSC amplitude, and [Ca²⁺]_e.

Role of presynaptic voltage-dependent Ca²⁺ channels in glutamate release from RHT terminals

Selective Ca²⁺ channel blockers were used to identify the voltage-dependent Ca²⁺ channels that regulate neurotransmitter release from RHT terminals. -Conotoxin GVIA (1 µM) significantly reduced the evoked EPSC amplitude (66.1 ± 3.3%), consistent with an important role for N-type Ca²⁺ channels in mediating glutamate release from RHT terminals (P < 0.0001, n = 40; Fig. 5, A and B). Application of -agatoxin TK (500 nM) produced a 36.1 ± 2.5% reduction (P < 0.0005, n = 49; Fig. 5, C and D), whereas a lower -agatoxin concentration (100 nM) produced a nonsignificant 20.3 ± 4.8% reduction (P = 0.5, n = 4). SNX-482 (150 nM) induced a significant 17.5 ± 3.0% reduction of the EPSC amplitude (P < 0.03, n = 15; Fig. 5, E and F), and nimodipine (10 µM) a nonsignificant reduction of 5.6 ± 2.7% (P = 0.53, n = 5; Fig. 5, G and H). These data show that R-type but not L-type Ca²⁺ channels participate in glutamate release from RHT terminals. Mibefradil (20 µM) produced a significant 31.5 ± 3.1% reduction of the EPSC amplitude (P < 0.0002, n = 5; Fig. 5, I and J), indicating that T-type Ca²⁺ channels contribute to the release of transmitter from RHT terminals.

View larger version (15K): [in this window] [in a new window]   FIG. 5. Inhibition of evoked EPSCs by voltage-dependent Ca2+ channel blockers. A and B: -conotoxin GVIA (1 µM). C and D: -agatoxin TK (500 nM). E and F: SNX-482 (150 nM). G and H: nimodipine (10 µM). I and J: mibefradil (20 µM). A, C, E, G, and I: examples of EPSCs recorded before, during, and after blocker application. B, D, F, H, and J: control EPSC amplitude plotted vs. the EPSC amplitude (test EPSC) in the presence of Ca2+ channel blocker. Each point represents an individual experiment. Dashed line, 0% inhibition; solid line, mean % inhibition of EPSC amplitude by the Ca2+ channel blocker.

View larger version (26K): [in this window] [in a new window]   FIG. 6. Overlap of the effect of Ca2+ channel blockers on the EPSC amplitude. A: Ca2+ channel blockers were applied together and the EPSC amplitude recorded. Magnitude of the reduction of the EPSC amplitude produced by the combination of blockers was compared with the calculated sum of their individual effects. Observed inhibition was generally less than that predicted by summing the effects of the individual blockers. Error bars represent SE; NS, nonsignificant, **P < 0.01, ***P < 0.001. For blockers applied together: n = 9 (agatoxin, SNX), n = 3 (conotoxin, SNX), n = 14 (agatoxin, conotoxin), n = 10 (agatoxin, conotoxin, nimodipine), n = 14 (agatoxin, conotoxin, nimodipine, SNX), n = 5 (agatoxin, conotoxin, nimodipine, SNX, mibefradil). B and C: changing the order of -conotoxin and -agatoxin application had no effect on the magnitude of EPSC inhibition.

View larger version (17K): [in this window] [in a new window]   FIG. 8. Occlusion of baclofen’s inhibition of EPSCs by Ca2+ channel blockers. A: occlusion baclofen’s effect by -conotoxin GVIA (1 µM), mibefradil (20 µM), -agatoxin TK (500 nM), and SNX-482 (150 nM). Data in histogram: normalized % of full effect of baclofen (0.3 and 10 µM, respectively). White squares: quantity of neurons. B: occlusion of baclofen’s effect by a cocktail of Ca2+ channel blockers: -conotoxin GVIA, -agatoxin TK, SNX-482, nimodipine, mibefradil, and effect of Cd2+. Baclofen 10 µM. Inset: EPSC recording: control, cocktail, baclofen, and cocktail; effect of Cd2+ on evoked EPSC in the same neuron.

Experiments were performed to determine whether GABA_B-receptor activation decreased RHT transmission by inhibiting voltage-dependent Ca²⁺ channels. The EPSC amplitudes recorded when baclofen was applied on a background of prior toxin exposure were compared with the EPSC amplitudes in the presence of only baclofen. Occlusion of the baclofen effect by one of the toxins would suggest that baclofen is acting by inhibiting the activity of that channel.

Baclofen (0.3 and 10 µM) evoked a significant reduction of the EPSC amplitude (26.6 or 73.7% respectively; Fig. 7, A and B). The maximal inhibitory effect occurred 3.3 ± 0.3 min (n = 13) after beginning the baclofen application. After washout, the EPSC amplitude recovered to 70–100% of control for 25–40 min. To determine whether there was an interaction between GABA_B receptors and N-, P/Q-, T-, R-, or L-type Ca²⁺ channels, various Ca²⁺ channel blockers were applied alone or together with baclofen (0.3 or 10 µM) (Fig. 7). The -conotoxin (1 µM) plus baclofen (0.3 or 10 µM) effect (73.7 ± 6.0%, n = 12 or 92.1 ± 1.7%, n = 7 reduction, respectively) was significantly greater than the effect of -conotoxin alone (Fig. 7, C and D). From experiments combining -conotoxin with baclofen at low concentration (0.3 µM) and a nearly maximal concentration (10 µM), we calculate 72.2 and 48.6% (respectively) of baclofen’s effect was mediated by inhibition of N-type Ca²⁺ channels (Fig. 8A) (see Math and statistical analysis). -Agatoxin (500 nM) and baclofen (0.3 or 10 µM) were applied together to study a possible interaction between GABA_B receptors and P/Q-type channels (Fig. 7, E and F), and reduced the EPSC amplitude 62.2 ± 5.5% (n = 12) or 96.1 ± 0.5% (n = 4), respectively. These data indicate that P/Q-type Ca²⁺ channels contribute 15.4% at low and 6.8% at high baclofen concentrations to baclofen’s inhibition of evoked EPSCs (Fig. 8A).

View larger version (17K): [in this window] [in a new window]   FIG. 7. Occlusion of baclofen effects on EPSC amplitude by Ca2+ channel blockers. A–K: histograms indicate the inhibition produced by baclofen alone (0.3 or 10 µM) and by application of Ca2+ channel blocker alone and together with baclofen. Left column: baclofen 0.3 µM. Right column: baclofen 10 µM. A and B: baclofen alone. C and D: baclofen and conotoxin (1 µM) (C, n = 12; D, n = 7). E and F: baclofen and agatoxin (500 nM) (E, n = 12; F, n = 4). G and H: baclofen and SNX-482 (150 nM) (G, n = 6; H, n = 5). I and J: baclofen and mibefradil (20 µM) (I, n = 8; J, n = 6). K: baclofen and nimodipine (10 µM) (n = 5). Current traces are individual EPSCs recorded in control, channel blocker, and channel blocker plus baclofen conditions. NS, nonsignificant; *P < 0.05, **P < 0.01, ***P < 0.001.

GABA_B-receptor–mediated inhibition of T-type Ca²⁺ channels was studied by applying mibefradil (20 µM) together with baclofen (0.3 or 10 µM). Mibefradil plus baclofen produced a 50.8 ± 5.3% (n = 8) or 82.4 ± 6.2% (n = 6) reduction, respectively, for 0.3 and 10 µM baclofen (Fig. 7, I and J). The component of baclofen’s effect occluded by mibefradil was estimated to be 51.9 and 23.9% in low and high concentrations of baclofen, respectively. Thus mibefradil has a greater effect at lower than higher baclofen concentrations, and indicates a strong effect of baclofen on T-type Ca²⁺ channels. In contrast, evoked EPSCs inhibited by application of baclofen (10 µM) together with nimodipine (10 µM) were not significantly different from baclofen alone (Fig. 7K).

Joint application of -agatoxin, -conotoxin, SNX-482, nimodipine, mibefradil, and baclofen (10 µM) inhibited EPSCs 97.5 ± 1.9% compared with 81.8 ± 1.1% inhibition evoked by these toxins alone (n = 3). Thus the blocker cocktail does not completely inhibit the evoked EPSCs. Because baclofen (10 µM) applied with this cocktail almost completely blocked the remaining EPSCs, we calculated that the residual inhibitory baclofen effect was 15.7% (Fig. 8B).

Finally, because the application of Ca²⁺ channel blockers alone or jointly cannot completely block glutamate release, it suggests the existence of Ca²⁺ currents resistant to the applied blockers. These remaining Ca²⁺ currents were almost completely inhibited by the addition of baclofen (10 µM). Cd²⁺ (100 µM) in the same experiments also completely blocked the EPSC (Fig. 8B).

Baclofen does not activate presynaptic K⁺ channels

GABA_B receptors couple to G(i/o)-type G-proteins and activate G-protein–activated inward rectifier K⁺ channels (Kir3) (Andrade et al. 1986; Sodickson and Bean 1996). We designed experiments to determine whether the GABA_B-receptor–mediated inhibition of glutamate release at RHT terminals could be attributable, in part, to an activation of presynaptic Kir3 channels. To block Kir3 channels we used a peptide toxin tertiapin, a potent blocker of Kir3.1, Kir3.4, Kir1.1, and K_ACh inwardly rectifying K⁺ channels and Ba²⁺ (BaCl₂).

Tertiapin (10 nM) application produced a small nonsignificant decrease in the EPSC amplitude to 95.9 ± 7.9% of control (control 70.4 ± 10.0 pA) and did not change the charge transfer (n = 7; Fig. 9, A, E, and F). Similarly, the EPSC amplitude was not significantly reduced (92.4 ± 5.7% of control, P = 0.32, n = 11) in the presence of a higher tertiapin (100 nM) concentration. Application of baclofen (10 µM) with tertiapin (10 nM) reduced the EPSC amplitude to 23.7 ± 6.7% and the charge transfer (n = 7; Fig. 9, A, E, and F), values that were not significantly different from baclofen alone (Fig. 9, E and F).

View larger version (28K): [in this window] [in a new window]   FIG. 9. Interaction between baclofen and the K+ channel blockers tertiapin, Ba2+, and 4-aminopyridine (4-AP). A–D: graphs indicate time course of the application of K+ channel blockers and baclofen (10 µM) and their effect on the evoked EPSC amplitude. Inset: EPSC recordings: control, K+ channel blockers applied alone, and with baclofen. A: effect of baclofen and tertiapin (10 nM). B: effect of baclofen and BaCl2 (1 mM). C and D: effect of baclofen and 4-AP (5 mM). Each point represents the means ± SE of the amplitudes of 5 EPSCs recorded during a 1-min period. E: summary of the effect of tertiapin (10 nM), BaCl2, and 4-AP alone and applied together with baclofen (10 µM) on the EPSC amplitude. F: effect of each K+ channel blocker applied alone and together with baclofen (10 µM) on the charge transfer (EPSC). Quantity of neurons: 4-AP (n = 9), Ba2+ (n = 5), tertiapin (n = 7), baclofen alone (n = 13). NS, nonsignificant; *P < 0.05, **P < 0.01, ***P < 0.001.

Application of 4-AP (5 mM) significantly increased the EPSC amplitude to 153 ± 14.6% of control (control 109.0 ± 9.4 pA), the charge transfer (n = 9; Fig. 9, C–F), and the amplitude of the evoked Ca²⁺ transients (Gompf et al. 2005). Baclofen added together with 4-AP reduced the EPSC amplitude to 75.8 ± 9.0%, a value that was significantly different from the effect of baclofen alone (n = 9; Fig. 9, C–E). After baclofen application the charge transfer remained 31% larger than in control (131.2 ± 24.2% of control) and more than fourfold larger than baclofen alone (30.0 ± 3.3% of control; Fig. 9F). These data show that application of 4-AP significantly decreases the magnitude of the baclofen-induced inhibition. After joint application of 4-AP and baclofen (10 µM), the EPSC amplitude remained larger than the control EPSC amplitude in two neurons (Fig. 9D).

4-AP and Ba²⁺ increased the charge transfer (225.0 ± 24.1 and 144.1 ± 18.0% of control, respectively). Charge transfer (131.2 ± 24.2% of control, P < 0.006) remained significantly higher than in control when baclofen was applied together with 4-AP. Ba²⁺ or tertiapin together with baclofen induced a charge transfer (29.9 ± 9.1 and 32.5 ± 6.1% of control, respectively) that was not significantly different (P = 0.99 and P = 0.72) from baclofen alone (30.0 ± 3.2%).

Unlike Ba²⁺, 4-AP increased the 10–90% rise time of the EPSC [108.8 ± 13.3% (P = 0.96, n = 5) and 174.4 ± 26.3% of control, respectively]. Application of baclofen (10 µM) with BaCl₂ or with 4-AP increased the EPSC rise time in both cases (129.6 ± 35.8 and 252.3 ± 34.8%, respectively). These data demonstrate that the baclofen effect on the rise time in the presence of Ba²⁺ was not significantly different from that of baclofen alone (135.9 ± 20.7%, P = 0.12). Thus Ba²⁺, unlike 4-AP, does not alter the EPSC rise time when applied alone or together with baclofen. Thus 4-AP and Ba²⁺ significantly increased the EPSC amplitude and charge transfer. Ba²⁺, unlike 4-AP, did not change the EPSC rise time in either the presence or the absence of baclofen. The charge transfer remained significantly larger than control when baclofen was applied together with 4-AP, in contrast to Ba²⁺ or tertiapin, which did not alter baclofen’s effect.

Effect of increasing the extracellular K⁺ on EPSC inhibition evoked by baclofen

To evaluate whether changes in the presynaptic membrane potential could alter the effect of baclofen on glutamate release, we varied the extracellular KCl concentration from 2.5 to 25 mM, and correspondingly decreased the NaCl concentration, to balance osmolality. Increasing the extracellular K⁺ ([K⁺]_e) depolarizes presynaptic terminals (Hoss and Labkovsky 1986). As expected, raising the [K⁺]_e resulted in an exponential increase in resting presynaptic Ca²⁺ concentration, and a biphasic concentration–response relationship for both evoked EPSCs and presynaptic Ca²⁺ transient amplitudes (Fig. 10D). The EPSC was completely blocked by [K⁺]_e >20 mM (Fig. 10A). These data suggest that strong depolarization of optic nerve terminals can completely block EPSCs by blocking action potential propagation. In contrast, raising [K⁺]_e to 7–12 mM increased the EPSC amplitude (Fig. 10, B, C, and D) to 139.5 ± 9.7% of control (control 78.6 ± 10.9 pA, n = 16). The EPSC amplitude inhibition induced by baclofen (10 µM) did not change significantly after increasing of [K⁺]_e (Fig. 10D). Thus the depolarization of RHT terminals evoked by increasing [K⁺]_e does not significantly alter baclofen’s inhibitory effect on EPSCs.

View larger version (21K): [in this window] [in a new window]   FIG. 10. Influence of extracellular K+ concentration ([K+]e) increasing on evoked EPSC amplitude, Ca2+ transients and baclofen effect. A: effect of 15 mM and 20 mM [K+]e on the evoked EPSC amplitude. B: effect of baclofen (10 µM) in 7 mM [K+]e. C: baclofen (10 µM) in 12 mM [K+]e. D: dependency of the baclofen effect on the amplitude of evoked EPSCs on the [K+]e. EPSC before (white circles) and after (black circles) baclofen application (n = 23). Changes in Ca2+ transient (white squares) (Fura-Red AM; ratio 440/488 nm), and resting Ca2+ concentration (black squares) (n = 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

We have shown that GABA_B-receptor activation reduces transmission at RHT synapses by inhibiting primarily N-, T-, and P/Q-type voltage-gated Ca²⁺ channels that are required for glutamate release. R-type voltage-gated Ca²⁺ channels also contribute to glutamate release but are not affected by activation of presynaptic GABA_B receptors. L-type Ca²⁺ channels and Kir3 channels contribute neither to glutamate release from RHT terminals nor to GABA_B presynaptic inhibition. The magnitude of the effect on RHT synaptic transmission was dependent on the GABA_B agonist concentration but was not dependent on the circadian time.

Role of voltage-dependent Ca²⁺ channels in glutamate release from RHT terminals

Selective Ca²⁺ channel blockers were used to determine the contribution of specific types of Ca²⁺ channels to RHT transmission. These toxins do not have postsynaptic effects on glutamatergic responses nor do they alter the number of stimulated afferents (Luebke et al. 1993; Mintz et al. 1995). We have shown that N-, P/Q-, T-, and R-type, but not L-type, Ca²⁺ channels contribute the Ca²⁺ required for glutamate release from RHT terminals. During low-frequency stimulation (0.08 Hz) the EPSC amplitudes were reduced nearly 70% by N-type channel inhibition, nearly 30% by either P/Q-type or T-type channel inhibition, and nearly 20% by R-type Ca²⁺ channel inhibition. Blocking L-type channels with nimodipine had no significant effects on RHT synaptic transmission. -Agatoxin produced almost twice the reduction of the evoked EPSC at 500 than at 100 nM, suggesting that Q-type (IC₅₀ = 90 nM) channels have the same importance in regulating release as P-type channels (IC₅₀ = 1–3 nM).

To study which types of Ca²⁺ channels are involved we used specific blockers (see above). However, mibefradil is not completely selective for T-type Ca²⁺ channels. Our observations on mibefradil effects on EPSCs can be explained in two different ways. First, mibefradil specifically blocks low-voltage–activated T-type Ca²⁺ channels (Lacinova 2004). In this case, we would conclude T-type Ca²⁺ channels contribute to mediating transmitter release from RHT terminals. Alternately, mibefradil blocks several Ca²⁺ channel types to reduce the evoked EPSC amplitude (Bezprozvanny and Tsien 1995; Liu et al. 1999; Viana et al. 1997). Our data were consistent with the first possibility because in the presence of N-, P/Q-, R-, and L-type Ca²⁺ channel blockers mibefradil blocked an additional roughly 10% of the EPSC amplitude. This interpretation must be made with caution because we cannot exclude the blocking effect of mibefradil on Na⁺ and K⁺ channels (Liu et al. 1999).

Multiple types of Ca²⁺ channels with overlapping functions mediate glutamate release at RHT terminals. Application of any two toxins reduced the EPSC amplitude less than would be expected from the sum of their individual ability to reduce EPSC amplitude (Fig. 6). The sum of the inhibition produced individually by N-, P/Q-, and T-type Ca²⁺ channels blockers was greater than inhibition produced by their joint application. For example, the overlap between the effect of N- and P/Q-type Ca²⁺ channels was estimated to be 33%, but there was no overlap with R-type Ca²⁺ channels. These data are consistent with the activity of multiple Ca²⁺ channels overlapping to initiate transmitter release. Previous studies have similarly demonstrated that blocking synaptic currents by selective antagonists of N- and P/Q currents sum to >100% (Mintz et al. 1995; Reid et al. 1998; Wheeler et al. 1996). A model was proposed in which "multiple types of Ca²⁺ channels exert synergistic control over individual release sites, and that the domains of several Ca²⁺ channels must overlap at each release site" (Mintz et al. 1995). The regulation of glutamate release from RHT terminals by multiple types of voltage-dependent Ca²⁺ channels is consistent with such a model.

Joint application of -conotoxin GVIA, -agatoxin TK, mibefradil, SNX-482, and nimodipine in saturating concentrations evoked an 87% decrease of the control EPSC amplitude and should completely block N-, P/Q-, T-, R-, and L-type Ca²⁺ channels. The component of the EPSC (13%) that was not inhibited by these blockers was inhibited by Cd²⁺ and might reflect the contribution of SNX-482–resistant R-type Ca²⁺ channels (Tottene et al. 2000). In axons of the rat optic nerve and RGCs nearly 40–60% of the evoked Ca²⁺ influx accounted for N-type Ca²⁺ channels, whereas P/Q-type Ca²⁺ channels make little, if any, contribution (Guenther et al. 1994; Sun and Chiu 1999). In RGCs, L-type Ca²⁺ channels mediate 25% of the Ca²⁺ current (Guenther et al. 1994). Immunocytochemical studies confirmed the presence of L-type Ca²⁺ channels in axons of adult rat optic nerve, although they contribute little to evoked Ca²⁺ influx (Brown et al. 2001; Sun and Chiu 1999). The insignificant contribution of L-type channels in transmitter release can be explained by localization of L-type channels far from release sites at these axonal terminals.

During electrophysiological recordings, we noted that there was a large variation of the magnitude of toxin effects (Fig. 5). For example, -conotoxin produced as much as 100% inhibition and as little as 20% inhibition. One explanation is that delivery of the toxin varied across the experiments but this is unlikely because baclofen application from the same micropipette showed a consistent effect. An alternate explanation is that different RHT fibers have different complements of Ca²⁺ channels (Reid et al. 2003). Those that rely on N-type channels for glutamate release would be more sensitive to block by -conotoxin than those that rely more on P/Q- or R-type channels. Our data suggest that the different types of Ca²⁺ channels can vary considerably in different RHT terminals (Fig. 5).

Presynaptic effect of baclofen on voltage-dependent Ca²⁺ channels

GABA_B-receptor activation reduces the release of glutamate from axon terminals of the RHT (Jiang et al. 1995). The magnitude of the effect is dependent on the GABA_B-agonist concentration but is not dependent on the circadian time (Fig. 3). The IC₅₀ for baclofen-evoked inhibition during the subjective day was not significantly different from that during the subjective night (IC₅₀ 1 µM) and was similar to those measured in the rat optic nerve and in the calyx of Held (Sun and Chiu 1999; Takahashi et al. 1998). Further the concentration–response curves recorded during the day or night were not significantly different (Fig. 3).

Activation of GABA_B receptors inhibited the release of glutamate from RHT terminals as measured by a reduction in the EPSC amplitude. Blocking specific types of voltage-dependent Ca²⁺ channels attenuated the magnitude of the GABA_B-agonist–induced reduction. The GABA_B-receptor regulation is not uniform across all types of Ca²⁺ channels. We estimate that depending on the baclofen concentration (0.3–10 µM) that was applied, nearly 49–72% of the presynaptic inhibition produced by GABA_B-receptor activation was mediated by block of N-type Ca²⁺ channels, nearly 24–52% mediated by block T-type Ca²⁺ channels, and nearly 7–15% mediated by block of P/Q-type Ca²⁺ channels. The effect of GABA_B-receptor activation was not mediated by R-type (nearly 1% inhibition) or L-type Ca²⁺ channels. Similarly, baclofen-induced inhibition was occluded by block of N-type Ca²⁺ channels in axons of neonatal rat optic nerve (Sun and Chiu 1999). Recently G-proteins were shown to bind more strongly to Cav2.2 ( 1B, N-type) than Cav2.1 ( 1A, P/Q-type) Ca²⁺ channels (Agler et al. 2003). This may explain why baclofen inhibits N-type more than P/Q-type Ca²⁺ channels. These effects of baclofen are mediated by G that directly inhibits Ca²⁺ channels, putting them into a reluctant state (Kajikawa et al. 2001; Zamponi and Snutch 1998).

Effect of baclofen on resting and evoked presynaptic Ca² transients

The relationship between evoked EPSC amplitude and [Ca²⁺]_e was sigmoidal, whereas the relationship between presynaptic Ca²⁺ transients and [Ca²⁺]_e approximated a hyperbolic response (Mintz et al. 1995). Reduction of Ca²⁺ influx in presynaptic terminals by baclofen has previously been reported in different brain structures (Barnes-Davies and Forsythe 1995; Dittman and Regehr 1996; Isaacson 1998; Wu and Saggau 1997). A similar effect of baclofen on presynaptic Ca²⁺ influx and EPSC amplitude (correspondingly about 50 and 90% reduction of initial value) was shown in presynaptic terminals of the calyx of Held (Sakaba and Neher 2003). The Ca²⁺ concentration in RHT terminals after optic chiasm stimulation was reduced by GABA_B-receptor activation, and glutamate release from RHT terminals was nonlinearly dependent on changes of the Ca²⁺ concentration in those terminals. We observed baclofen (30 µM) inducing roughly 12% reduction in presynaptic Ca²⁺ transients compared with nearly 80% reduction in EPSC amplitudes. Although this is consistent with the nonlinear relationship discussed above, the reduction of Ca²⁺ transients by baclofen was less than expected.

There are several possible reasons for this difference. First, the Ca²⁺ channels, which mediate transmitter release, are believed to be located near the vesicles to be released. However, we are recording the bulk Ca²⁺ in the terminals, which may underestimate the change in Ca²⁺ required for release. Second, nonlinearity of a single-wavelength fluorescent Ca²⁺ probe may underestimate Ca²⁺ responses. Directly quantifying the magnitude of the Ca²⁺ level with a dual-excitation probe conceivably would have eliminated this problem. However, single-wavelength excitation recordings provide the fast data acquisition needed for measurement of Ca²⁺ transients in presynaptic terminals. Finally, it is important to note that in our studies baclofen reduced resting Ca²⁺. This reduction is not included in determining the roughly 12% reduction of Ca²⁺ transients. Therefore this may result in an underestimation of the inhibition by baclofen. Previous studies examining the effect of baclofen on presynaptic Ca²⁺ influx in the hippocampus and cerebellum have not shown this reduction in resting Ca²⁺ (Dittman and Regehr 1996; Wu and Saggau 1997). These studies differed by examining different regions of the brain and by using different Ca²⁺ indicators. Overall, our data are consistent with a nonlinear relationship between presynaptic RHT Ca²⁺ concentration and glutamate release, such that GABA_B receptor activation producing a small reduction in the evoked presynaptic Ca²⁺ transient can have a large effect on inhibiting glutamate release.

Role of K⁺ channels in RHT transmission and in modulation of baclofen effect

Tertiapin and BaCl₂ were used to study whether the GABA_B agonist baclofen activates Kir3 channels on RHT presynaptic terminals. Tertiapin did not change the EPSC amplitude or charge transfer. Neither tertiapin nor Ba²⁺ blocked baclofen-induced presynaptic inhibition in RHT synapses. These data are in agreement with studies showing that Ba²⁺ does not block presynaptic inhibition induced by baclofen in excitatory (glutamatergic) synapses (Cui et al. 2000; Takahashi et al. 1998; Thompson and Gahwiler 1992). Similarly, baclofen does not directly increase a K⁺ conductance in the optic nerve (Sun and Chiu 1999). Moreover, G-protein–activated Kir current can be detected in the presynaptic terminal but it cannot be activated by baclofen (Takahashi et al. 1998). Baclofen-induced inhibition does not require activation of Kir3 channels on excitatory terminals because the presynaptic effect of GABA_B receptors was unaltered in mutant mice lacking the GIRK2 gene (Lüscher et al. 1997). Thus our data show that baclofen does not activate Kir3 channels in RHT axon terminals, unlike its effect on RGCs (Chen et al. 2004).

More than 60% of outward K⁺ currents in presynaptic terminals are sensitive to 4-AP and are very important for regulation of synaptic transmission (Forsythe 1994; Reiff and Guenther 1999). 4-AP potentiates EPSCs by broadening presynaptic action potentials and slowing action potential repolarization (Ishikawa et al. 2003; Sun and Chiu 1999). 4-AP application onto the optic nerve induced a delayed inward Ca²⁺ current that forms a second "hump" on the action potential (Sun and Chiu 1999). The "hump" was blocked partially by baclofen, although the Ca²⁺ transient remained unaffected (Sun and Chiu 1999). These data are intriguing because 4-AP application increased Ca²⁺ influx and the EPSC amplitude and dramatically decreased the baclofen-induced inhibition in RHT terminals (Fig. 9D) (Gompf et al. 2005; Liang et al. 2002). The 4-AP effect was so strong that after baclofen application charge transfer remained 31% larger than in control.

The mechanism of disinhibition of baclofen’s presynaptic effect by 4-AP was not clear. One explanation was that baclofen’s effect could be altered by depolarization of presynaptic terminal. It is known that depolarization of the nerve terminal relieves baclofen- and G-induced inhibition of presynaptic Ca²⁺ channels (Isaacson 1998; Kajikawa et al. 2001). Another explanation is that 4-AP blocks voltage activated K⁺ channels (Kv) Kv3.1/Kv2.1 and enhances Na⁺ and Ca²⁺ influx in presynaptic terminals (Ishikawa et al. 2003; Kirsch and Drewe 1993). Elevation of axonal [Na⁺]_i also induces Ca²⁺ influx into the axon (Verbny et al. 2002). Increasing the Ca²⁺ influx could partially relieve baclofen-induced inhibition (Fig. 4B). At the same time it is unclear why depolarization of RHT terminals evoked by Ba²⁺ application or by high [K⁺]_e did not reduce the baclofen effect. Thus we can state only that 4-AP–sensitive K⁺ currents were important for baclofen-induced inhibition because baclofen significantly increased the transient outward K⁺ current (I_A) and shifts the activation of I_A to more positive potentials (Saint et al. 1990; Takeda et al. 2004). The effects of GABA_B-receptor activation on I_A could be similar to other G-protein–coupled receptors (cannabinoid, opioidergic, and serotonergic) whose presynaptic inhibition used an opening 4-AP–sensitive K⁺ channel by G-proteins (Childers et al. 1993; Kishimoto et al. 2001). We conclude that activation of