The Journal of Neurophysiology Vol. 79 No. 1 January 1998, pp. 181-189
Copyright ©1998 by the American Physiological Society

EPSP Amplitude Modulation at the Rat Ia-Alpha Motoneuron Synapse: Effects of GABA_B Receptor Agonists and Antagonists

Kavita R. Peshori, William F. Collins III, and Lorne M. Mendell

Department of Neurobiology and Behavior, State University of New York at Stony Brook, New York 11794-5230

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Peshori, Kavita R., William F. Collins III, and Lorne M. Mendell. EPSP amplitude modulation at the rat Ia-alpha motoneuron synapse: effects of GABA_B receptor agonists and antagonists. J. Neurophysiol. 79: 181-189, 1998. The object of this study was to examine the relationship between excitatory postsynaptic potential (EPSP) amplitude, posttetanic potentiation, and EPSP amplitude modulation at synapses made by group Ia afferents on motoneurons in the rat. These relationships were evaluated in cells in untreated rats and in cells in rats treated with the -aminobutyric acid-B (GABA_B) receptor agonist baclofen and antagonist CGP-35348, which were used to manipulate Ca²⁺ entry into presynaptic terminals and consequently probability of transmitter release from them. There was no evidence for postsynaptic action of these drugs from measurement of their effects on motoneuron properties. During high-frequency stimulation (32 shock bursts at 167 Hz), EPSP amplitude either decreased (negative modulation) or increased (positive modulation) in response to successive stimuli at different connections. In untreated rats this frequency-dependent amplitude modulation behavior was inversely but weakly correlated with EPSP amplitude measured at low frequency. Intravenous (iv) administration of the GABA_B agonist, baclofen, produced a marked and progressive decrease in EPSP amplitude measured at low frequency coincident with a change in frequency-dependent EPSP amplitude modulation toward more positive values (synaptic facilitation). In contrast, an increase in EPSP amplitude occurred after iv administration of the GABA_B antagonist CGP-35348 that was accompanied by a negative shift in EPSP amplitude modulation during high-frequency stimulation. The negative shift in EPSP amplitude modulation (synaptic depression) after CGP-35348 application was much smaller than the positive shift induced by baclofen when normalized to the change in EPSP amplitude. Posttetanic potentiation decreased after baclofen but did not increase after CGP-35348. The relationship between modulation and EPSP amplitude was much steeper after GABA_B receptor manipulation in either direction than that observed in the population of motoneurons in untreated preparations. This suggests that in the rat differences in probability of release play at most a small role in determining EPSP amplitude across the motoneuron pool.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

A series of studies in the cat has illustrated that the successive excitatory postsynaptic potential (EPSPs) generated in motoneurons of different types during high-frequency stimulation of Ia afferent fibers differ in their amplitude modulation (Collins et al. 1984; Honig et al. 1983; Mendell et al. 1995). Large-amplitude EPSPs generated in motoneurons with a small value of rheobase, large input resistance, and afterhyperpolarization (AHP) of long duration tend to undergo a progressive decrease in amplitude during high-frequency stimulation (negative modulation, see METHODS). Conversely, small-amplitude EPSPs in motoneurons with a low input resistance, high rheobase and short-duration AHP increase in amplitude during high-frequency stimulation (positive modulation). Positive correlations have also been observed between posttetanic potentiation and EPSP amplitude after high-frequency stimulation (Davis et al. 1985, 1987). These observations suggest that the processes that underlie EPSP amplitude modulation and potentiation contribute to the distribution of Ia EPSP amplitudes across motoneurons (Mendell et al. 1990).

By analogy with peripheral systems (Del Castillo and Katz 1954; Frank 1973; Gertler and Robbins 1978; Kuno 1964), it was hypothesized that the differences in frequency-dependent behavior at these synapses resulted from differences in presynaptic properties, specifically in probability of transmitter release (Collins et al. 1984). A buildup of Ca²⁺ in the presynaptic terminals from repetitive stimulation (residual Ca²⁺) is considered to lead to an increase in the probability of transmitter release resulting in facilitation of the postsynaptic response during a high-frequency burst (Katz and Miledi 1968; Mallart and Martin 1967, 1968). At connections with high initial probability of transmitter release (large initial values of P) and consequently large EPSP amplitudes, depression is seen during high-frequency stimulation, despite the residual Ca²⁺ buildup, possibly because of the unavailability or depletion of adequate releasable transmitter (Betz 1970) or inability to release available transmitter on subsequent stimuli (Korn et al. 1981). At connections with low initial probability of release, EPSP amplitude increases during high-frequency stimulation. Thus in this work we have assumed that the behavior of EPSP amplitude during high-frequency stimulation is an index of the average probability of release at that connection.

The focus of the present study was to investigate the dependence of EPSP amplitude during high-frequency stimulation on probability of transmitter release more explicitly. The approach was to manipulate probability of release pharmacologically by using agonists and antagonists of the -aminobutyric acid (GABA_B) receptor. This receptor, which is known to be present on terminals of Ia fibers (Bowery et al. 1987; Price et al. 1984), acts to control entry of Ca²⁺ into sensory neurons (Deisz and Lux 1985; Dolphin and Scott 1987; Dunlap and Fischbach 1981). Previous studies suggested that the GABA_B agonist baclofen reduces monosynaptic EPSP amplitude (Edwards et al. 1989; Lev Tov et al. 1988; Peng and Frank 1989) by decreasing probability of transmitter release (Edwards et al. 1989; Peng and Frank 1989) because of reduction in Ca²⁺ entry into the presynaptic terminals (Curtis et al. 1997). The GABA_B antagonist CGP-35348 is also known to reduce the level of presynaptic inhibition in the spinal cord (Curtis 1998; Curtis and Lacey 1994). By using these agents, the aim in the present experiments was to examine in detail the relationship between EPSP amplitude, EPSP amplitude modulation, and posttetanic potentiation in individual cells. One question was whether activation and blockade of the GABA_B receptor yields the same functional relationship between EPSP amplitude modulation and amplitude. A second question was whether or not this relationship is similar to that observed across the population of connections. This would speak to the question of whether or not differences in EPSP amplitude in different motoneurons can be attributed to differences in probability of transmitter release. Preliminary results from these studies have been published in abstract form (Mendell et al. 1991; Peshori et al. 1991, 1993).

	METHODS

Abstract Introduction Methods Results Discussion References

Adult, male Sprague-Dawley rats (300-500 g) were anesthetized with a mixture of ketamine and xylazine (90 and 10 mg/kg, im, respectively) and anesthesia was maintained by a continuous infusion (0.4-1.6 ml/hr) of this mixture diluted in 10% dextrose saline. The rats were mechanically ventilated and heart rate, ECG, blood pressure, expired CO₂, and rectal temperature were monitored. Body temperature was maintained at 37°C by using a heat lamp and a heating pad. The spinal column was placed in a custom-made spinal clamp and a dorsal laminectomy was performed to expose segments T₁₂-S₂ of the spinal cord. The left medial gastrocnemius (MG) and lateral gastrocnemius-soleus (LGS) nerves were exposed and cut. The proximal cut ends were placed on bipolar platinum stimulating hook electrodes and covered with warm mineral oil. The dura overlying the spinal cord was removed and the cord was covered with mineral oil. A platinum ball electrode was placed on the dorsal columns and the stimulus intensities for the MG and LGS nerves were set at 2 times threshold for the dorsal column response to ensure maximal activation of group Ia afferent fibers. Incoming afferent volleys were monitored in this manner to ensure faithful conduction of afferent volleys during high-frequency stimulation.

Intracellular recordings were made by using glass microelectrodes filled with 3 M potassium acetate with tip resistance between 4 and 16 M. Impaled motoneurons were identified as MG or LGS by their response to antidromic stimulation of the MG or LGS muscle nerve and the latency was used to calculate axonal conduction velocity. The analysis was restricted to motoneurons with antidromic action potential amplitudes >75 mV. Motoneuron rheobase was measured as the minimum intracellular depolarizing current (50 ms duration) required to generate an action potential. Input resistance (R_n) was measured as the averaged (n = 64) amplitude of the voltage change in response to 1 nA of hyperpolarizing current (50-ms duration) injected intracellularly. Before measuring R_n, the averaged record was corrected for membrane potential Sag by using a procedure similar to that used by Zengel et al. (1985).

Composite Ia-EPSPs were elicited by stimulation of the heteronymous muscle nerve. Two stimulus paradigms were employed; first, low-frequency (18 Hz) trains and second, 32 shock bursts at high-frequency (167 Hz) delivered every 2 s. Successive triggered events were averaged in register (n = 256 for 18 Hz trains, n = 64 for 167-Hz bursts) and the amplitudes of the averaged EPSPs were measured (EPSP_{18 Hz}; and for the burst, EPSP₁, EPSP₂, ..., EPSP₃₂). A correction factor was added to the measurements of all but the first EPSP within a burst to compensate for the summation of each EPSP with the falling phase of the previous EPSP. The value of this correction factor was calculated as the magnitude of the voltage decay of the last (32nd) EPSP in the burst between times representing where the onset and peak of an imaginary 33rd EPSP would have occurred.

These two stimulus paradigms have been used in previous studies in cats (Collins et al. 1984, 1988; Davis et al. 1985, 1987; Koerber and Mendell 1991; Mendell et al. 1995) for several reasons. First, both frequencies are within the normal physiological range for Ia afferents (Prochazka et al. 1977; Loeb and Duysens 1979). Second, no systematic variability is observed in the amplitude of Ia-motoneuron EPSPs during 18 Hz stimulation (Collins et al. 1984). Thus EPSP_{18 Hz} is considered to be the control EPSP for comparison with EPSPs generated at higher frequencies of stimulation. Third, 32 shock bursts at 167 Hz is an optimal paradigm for revealing the heterogeneity in the response of Ia-motoneuron connections to repetitive stimulation. At frequencies >167 Hz, most Ia-motoneuron synapses tend to exhibit depression and conduction failure in afferent terminal branches can occur (Collins et al. 1984). In preliminary studies we found these criteria to be valid for the rat as well, and for the sake of consistency and easier comparison between species, we chose the same stimulus paradigms as those used in the cat.

To assess the change in EPSP amplitude during high-frequency stimulation, EPSP amplitude modulation, potentiation and facilitation/depression (F-D) were calculated by using the following formulae:

(1)

(2)

(3)

Modulation is defined as the change in EPSP amplitude during a high-frequency, 32 shock burst and was calculated as the ratio of the amplitude of the EPSPs at the end of the burst (average of 30th and 31st EPSPs) to the amplitude of the first EPSP within the burst (Eq. 1). A modulation value <1 represents a decrease in EPSP amplitude during the burst and is termed negative modulation. A modulation value >1 signifies positive modulation, i.e., an increase in EPSP amplitude during the burst. Because the interburst interval used in averaging bursts was 2 s, the initial EPSP in each burst (except the first burst) was often potentiated in amplitude because of the influence of the previous burst (Collins et al. 1984; Davis et al. 1985). This potentiation was quantified by calculating the ratio of the amplitude of the first averaged EPSP in a highfrequency burst to the amplitude of the averaged control EPSP_{18 Hz}(Eq. 2). The F-D ratio represents the amplitude of the EPSP at the end of the burst referred to the value measured at low-frequency stimulation (EPSP_{18 Hz}). Because the first EPSP in the burst (EPSP₁) was generally potentiated, EPSP₁ > EPSP_{18 Hz} [potentiation (Eq. 2) > 0] and so F-D was generally more positive than modulation (see Fig. 3).

View larger version (25K): [in this window] [in a new window]   FIG. 3. Modulation (Eq. 1; A), potentiation (Eq. 2; B), and facilitation-depression (Eq. 3; C) as functions of EPSP18 Hz amplitude in untreated rats. Each point represents data obtained from a single motoneuron. (- - -), ratio value = 1.0; (), linear regressions. r = correlation coefficient for the regression.

Pharmacological studies

After obtaining measurements of electrical and synaptic properties at a given Ia-motoneuron connection, either l-() baclofen (10 mg/kg in saline, Ciba-Geigy, Suffern, NY), (±) baclofen (10 mg/kg in phosphate buffered saline, Sigma Chemical, St. Louis, MO or Research Biochemicals International, Natick, MA), CGP-35348 (30-400 mg/kg in saline, Ciba-Geigy, Basel), or saline was injected intravenously while the intracellular penetration was maintained. A single injection was usually sufficient for baclofen effects which were not reversible over the intervals studied here (up to 4 h). In the case of CGP-35348 the initial small dose often did not have any effect or it wore off very quickly and supplementary injections were given. This accounts for the wide range of dosages for CGP-35348. The measurements were repeated immediately after drug administration and periodically throughout the duration of a stable penetration (up to 4 h). All pre- and postdrug comparisons were made with the paired-sample t-test.

	RESULTS

Abstract Introduction Methods Results Discussion References

Composite Ia-motoneuron EPSP amplitude

The amplitudes of composite EPSPs were measured in a total of 103 motoneurons in response to 18 Hz stimulation (Fig. 1A) and varied from 0.15 mV to 3.35 mV with a mean of 1.15 ± 0.68 (SD) mV. Motoneuron rheobase varied from 0.8 to 19.5 nA (mean = 8.1 ± 3.9 nA; n = 100) and motoneuron input resistance varied from 1.2 to 4.1 M (mean = 2.2 ± 0.6 M; n = 38). EPSP amplitude at 18 Hz was weakly but significantly correlated with both motoneuron rheobase (linear regression, r = 0.23, P < 0.02; Fig. 2A) and motoneuron input resistance (r = 0.43, P < 0.01; Fig. 2B).

View larger version (20K): [in this window] [in a new window]   FIG. 1. Examples of averaged, composite, heteronymous excitatory postsynaptic potential (EPSPs) recorded from a medial gastrocnemius (MG) motoneuron in response to stimulation of lateral gastrocnemius-soleus (LGS) muscle nerve at (A) low-frequency (18 Hz train; n = 256 sweeps) and (B) high-frequency (32 shock bursts at 167 Hz; 2 s interburst interval, n = 64 sweeps).

View larger version (27K): [in this window] [in a new window]   FIG. 2. Plots of relationships between (A) EPSP18 Hz amplitude and motoneuron rheobase and (B) EPSP18 Hz amplitude and motoneuron input resistance in untreated rats. Each point represents data obtained from a single motoneuron. Linear regressions are plotted as solid lines through data. r = correlation coefficient for regression.

Frequency dependence of EPSP amplitude

EPSP amplitude modulation during 32 shock bursts (Fig. 1B) was calculated using Eq. 1 and varied from 0.46 (negative modulation) to 1.56 (positive modulation) with a mean of 1.25 ± 0.2 (n = 103). As in the cat, large EPSPs tended to exhibit more negative modulation than small EPSPs. Plotting modulation against EPSP amplitude (Fig. 3A) revealed a significant negative correlation, but with a relatively small correlation coefficient (r = 0.25, P < 0.01). The slope of the linear regression is 0.08, suggesting a small dependence of amplitude modulation on EPSP amplitude. No correlation was detected between amplitude modulation and either motoneuron rheobase or motoneuron input resistance.

One of the complications in evaluating differences in amplitude modulation (Eq. 1) across the motoneuron pool is that the first EPSP in the burst is often potentiated with respect to the low frequency (18 Hz) EPSP amplitude (Davis et al. 1985, 1987). This is evident in the averaged records of the present study (Fig. 1) and occurs as a consequenceof the high-frequency activity of the preceding burst (2-sinterburst interval). Thus differences in EPSP amplitude modulation across the population of motoneurons could result from systematic differences either in potentiation of the first EPSP in the burst or in faciliation-depression (F-D) behavior during the burst, or both. The contribution of these factors to the dependence of amplitude modulation on EPSP amplitude was examined in more detail by evaluating the correlation of potentiation (Eq. 2) and F-D (Eq. 3) with EPSP amplitude. Potentiation of the first EPSP in the burst was found not to depend on EPSP amplitude(P > 0.78; Fig. 3B), unlike in the cat where a positive correlation between potentiation and EPSP amplitude was observed (Davis et al. 1985). In contrast, a significant negative correlation was detected between F-D and EPSP amplitude (r = 0.25; P < 0.01; Fig. 3C), which was nearly identical to that observed for modulation. Thus the dependence of modulation on EPSP amplitude at rat Ia-motoneuron synapses results from changes occurring during bursts of high-frequency stimulation (i.e., facilitation and depression), rather than changes between bursts (i.e., potentiation).

Manipulation of EPSP amplitude with baclofen and CGP-35348

A major aim of these experiments was to determine explicitly the effects of manipulating the probability of transmitter release pharmacologically on EPSP amplitude and EPSP amplitude modulation. In 12 motoneurons, baclofen was administered iv while the motoneuron penetration was maintained and EPSP amplitude, amplitude modulation, and potentiation were measured during the period of EPSP amplitude decline. An example of this is shown in Fig. 4, A and B. Note that during the 45 min over which the motoneuron penetration was maintained, EPSP amplitude decreased from 0.43 mV to 0.11 mV and EPSP amplitude modulation ratio increased concomitantly from 0.86 to 1.77. In addition, the potentiation ratio decreased from 1.4 to 1.18 (16%). Similar data were obtained at all connections studied in this way (Fig. 5). On average (n = 12), baclofen produced maximum changes in EPSP amplitude, modulation, and potentiation of 70 ± 14%, 67 ± 46%, and 23 ± 14%, respectively, although these are approximate values because they were obtained over different treatment intervals (i.e., until the motoneuron penetration was lost). The decreases in EPSP amplitude and potentiation and the positive shift in modulation could be detected within 2 min after application of baclofen (Fig. 6A) and persisted for as long as the motoneuron penetration was maintained (up to 2.8 h). These effects did not wear off over the several hours that the intracellular penetration was maintained.

View larger version (38K): [in this window] [in a new window]   FIG. 4. Ia-motoneuron EPSPs generated in a MG motoneuron in response to 18-Hz train stimulation and 167-Hz burst stimulation of LGS nerve before (A) and after (B) baclofen administration and before (C) and after (D) CGP-35348 administration. Note that baclofen administration produces a decrease in EPSP18 Hz amplitude and a positive shift in amplitude modulation during high-frequency stimulation, whereas the opposite occurs after CGP-35348.

View larger version (28K): [in this window] [in a new window]   FIG. 5. Histograms of EPSP18 Hz amplitude (A), modulation (B), and potentiation (C) in untreated rats (top) and after baclofen (middle) or CGP-35348 (bottom) administration. Control group (n = 103) was either untreated rats or ones that had been injected with saline. Values after baclofen (n = 32) were steady-state values from cells (n = 12) followed for several hours immediately after baclofen was administered and values obtained from other cells (n = 20) impaled up to 3.3 h after baclofen had been given. In case of CGP-35348 only 8 cells studied over course of administration of this drug are shown because unlike effects of baclofen, effects of CGP-35348 appeared to be short-lived. Histograms document smaller values of EPSP amplitude and potentiation and more positive values of modulation after baclofen and larger values of EPSP amplitude, slightly more negative (less positive) values of modulation and identical values of potentiation after CGP-35348.

View larger version (27K): [in this window] [in a new window]   FIG. 6. Plots of time course of changes in EPSP18 Hz amplitude, modulation, and potentiation after baclofen (A) and CGP-35348 (B) administration in two different motoneurons in different preparations. Note opposite effects of these agents on EPSP amplitude and modulation and nonreciprocal effects on potentiation.

Administration of the GABA_B antagonist CGP-35348 permitted evaluation of these same relationships when probability of release was increased (Fig. 4, C and D). Intravenous administration of CGP-35348 produced a rapid and marked increase in EPSP amplitude (50 ± 18%, n = 8). However only a small change in amplitude modulation was observed (13 ± 9%, n = 8; Figs. 4-6), even after repeated doses (see METHODS) unlike the finding after baclofen administration where modulation underwent substantial changes in conjunction with those in EPSP amplitude (Figs. 4-6). Thus in the example of Fig. 4 (C and D), EPSP amplitude increased from 1.90 mV to a maximum of 2.95 mV but modulation decreased only slightly (from 1.02 to 0.83). Unexpectedly, a small but significant decrease (11 ± 6%, n = 8) in potentiation of the initial EPSP (i.e., potentiation at 2 s) was observed after CPG-35348 administration (see DISCUSSION).

To better quantify the relationship between amplitude and modulation, we plotted the change in amplitude modulation as a function of the change in EPSP amplitude for individual motoneurons in baclofen- and CPG-35348-treated rats (Fig. 7). Each straight line represents the least squares fit of the points in an individual experiment obtained after the drug was administered (as shown in Fig. 6, as a function of time). For the 12 connections studied with baclofen, 2-14 measurements were made at individual connections after administration of baclofen. Because of the small number of measurements in some cases or because the rapid decrease in amplitude and increase in amplitude modulation caused many of the points to be clustered, the data at only six connections exhibited a significant correlation. However, all 12 are shown to demonstrate the similarity in the results. For CGP-35348 experiments, 5-23 measurements were made per cell and the plots revealed qualitatively similar results although the slope was substantially smaller than in the case of baclofen. Five of the eight experiments yielded a significant correlation. It should be noted that the EPSP amplitude before drug treatments was similar in all these experiments and became smaller after baclofen and larger after CGP-35348 with corresponding changes in the modulation values.

View larger version (19K): [in this window] [in a new window]   FIG. 7. Each line is linear regression fit to values of EPSP amplitude and EPSP amplitude modulation values for each of 12 connections studied for up to 2.8 h after baclofen administration (an example of which is displayed as a function of time in Fig. 6A) and 8 connections in rats treated with CGP-35348 (Fig. 6B). Note that slope of changes after CGP-35348 is less negative than after baclofen. Thin line is regression line for EPSPs evoked in 103 cells in untreated animals.

We have previously demonstrated a semilogarithmic relationship between modulation and EPSP amplitude for Ia EPSPs in cat medial gastrocnemius motoneurons (Collins et al. 1988). We replotted the data for the subset of experiments with significant correlations displayed in Fig. 7, using "ln amplitude" as abcissa (i.e., semilog plot; Fig. 8). Although the effect of CGP-35348 on modulation is much smaller than that seen with baclofen (Fig. 7), the slopes of the semilogarithmic regression lines were similar. In the case of baclofen-treated motoneurons, the slopes of the regression lines between modulation and ln EPSP amplitude ranged from 0.29 to 0.63 (mean 0.55 ± 0.14, n = 6). The slopes of the corresponding regression lines for motoneurons studied with CGP-35348 ranged from 0.22 to 0.42 (mean 0.36 ± 0.08, n = 5). However, the slope of the regression line for modulation as a function of EPSP amplitude across the population of medial gastrocnemius motoneurons was markedly different (0.08, n = 103).

View larger version (21K): [in this window] [in a new window]   FIG. 8. Plot of regression lines of modulation versus ln EPSP18 Hz amplitude for individual motoneurons after baclofen (dark arrows) or CGP-35348 (shaded arrows) administration. Each arrow is regression through several measurements of EPSP18 Hz amplitude and modulation acquired at different times after drug administration (as in Fig. 6) and arrow indicates direction of change. Only cells for which regression line was significant (P < 0.05) are displayed here. EPSPs from cells treated with baclofen or CGP-35348 exhibited similar slopes when plotted on semilog coordinates. Thin line is regression through similar data obtained from 103 motoneurons in untreated rats. Note that regression lines from baclofen- and CGP-35348-treated motoneurons have similar slopes which are markedly different from slope of regression line through population of motoneurons in untreated rats.

The shape of the EPSP was examined before and after administration of baclofen (n = 8) and CGP-35348 (n = 7). After baclofen administration, no change was observed in the shape of the EPSP in a majority of the cells (5/8). However, faster decay of the EPSP was detected in one cell and, at two others, the rise time of the EPSPs appeared to be faster. After CGP-35348 administration, there was no change in the time course or shape of the EPSPs in five of the seven cells examined. One of the remaining cells exhibited a faster EPSP decay whereas the other one exhibited a slower EPSP decay.

Peak potentiation

As pointed out above, application of baclofen or CGP-35348 generally reduced potentiation of the first EPSP in the burst evoked 2 s after the preceding burst. In previous work in the cat, we demonstrated that peak potentiation after the burst is reached 50-100 ms after the burst and it then undergoes a decrease (Davis et al. 1985, 1987). In a small subset of cells studied here, we examined potentiation 50-100 ms after the burst and found that it was generally greater than that seen 2 s after the burst. During baclofen administration, peak potentiation declined, whereas after CGP-35348 it was unaffected.

Effects of baclofen and CGP-35348 on motoneuron properties

To assess the possibility that baclofen or CGP-35348 may influence EPSP amplitude via a postsynaptic action, motoneuron input resistance and rheobase were measured before and after drug administration. In addition, repeated measurements were obtained from motoneurons in the absence of drug administration. In no case was a significant change (paired sample t-test) in motoneuron input resistance or rheobase detected after drug administration or with repeated measurements. Before and after baclofen treatment, mean motoneuron input resistances were 1.53 and 1.67 M, respectively (9.8%, P > 0.1, n = 10), and mean rheobase values were 8.66 and 8.63 nA, respectively (0.5%, P > 0.9, n = 8). After CGP-35348 treatment, an increase in mean input resistance from 1.70 to 1.95 M (13.5%) was noted, which was not statistically significant (P > 0.05, n = 9) and mean rheobase was unaffected (5.73-5.72 nA, 0.2%, P > 0.9, n = 9). Similar repeated measurements over approximately the same time period without drug administration revealed no significant change in mean input resistance (1.86-1.94 M, 4.3%, P > 0.6, n = 16) or in mean rheobase (7.07-6.76 nA, 4.4%, P > 0.4, n = 33). Activation of GABA_B receptors has been reported to hyperpolarize hippocampal cells via an increased K conductance (Dutar and Nicoll 1988; Gahwiler and Brown 1985). However, no consistent change in motoneuron membrane potential was observed after baclofen or CGP- 35348, either directly by measurements of membrane potential or indirectly by comparison of action-potential amplitude before and after application of these drugs. Together, these results are consistent with those from previous studies and indicate that the actions of baclofen in the Ia-motoneuron system are exclusively presynaptic, and suggest the same for CGP-35348, although more direct tests are required.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

In this study, Ia-motoneuron EPSP amplitude was altered by pharmacologically decreasing or increasing the probability of transmitter release and the resultant changes in EPSP amplitude modulation and potentiation were observed. A major question was whether or not the correlated changes in these synaptic variables measured in a single motoneuron would match those measured across the different motoneurons in the pool. Such an observation would strengthen the inference that probability of release of transmitter differs systematically in terminals on motoneurons across the pool according to the amplitude of the EPSP they produce.

Baclofen was used to diminish probability of transmitter release because previous studies indicate that the baclofen-induced decrease in EPSP amplitude at the Ia-motoneuron synapse results from an exclusively presynaptic action (Edwards et al. 1989; Lev-Tov et al. 1988; Peng and Frank 1989). This is a GABA_B receptor-mediated, G-protein linked, direct effect on voltage-sensitive Ca²⁺ channels (i.e., not through second messengers) leading to a decrease in Ca²⁺ entry into the presynaptic terminals (Dunlap and Fischbach 1981) and a decrease in the probability of transmitter release (Deisz and Lux 1985; Dolphin and Scott 1987). GABA_B receptors have been reported at the sites of presumed Ia-afferent terminals in receptor-binding experiments (Bowery et al. 1987; Price et al. 1984). In the present study, no significant change in motoneuron rheobase, input resistance, or membrane potential was observed after either baclofen or CGP-35348 administration (see also Lev-Tov et al. 1988), consistent with the lack of a postsynaptic effect, although the change in decay time of some of the EPSPs provides some challenge to the generality of that conclusion. Quantal analysis has revealed a reduction in quantal content rather than in quantal amplitude after baclofen administration (Edwards et al. 1989; Peng and Frank 1989), suggesting no desensitization of postsynaptic receptors mediating the Ia EPSP, although this may not be generally true (see Discussion section in Otis and Trussell 1996). Furthermore, baclofen has similar effects on proximally and distally generated EPSPs suggesting no selective effect on synaptic inputs to distal dendrites (Edwards et al. 1989). The lack of an effect of baclofen on terminal potentials (Lev-Tov et al. 1988; Peng and Frank 1989) and on the electrical component of mixed electrical/chemical synapses (Peng and Frank 1989; Shapovalov and Shiriaev 1982) argues against an effect on impulse invasion into the terminals. Antagonism of GABA_B receptors via CGP-35348 is also likely presynaptic, although direct effects on postsynaptic receptor sensitivity cannot yet be ruled out for this drug.

Baclofen administration resulted in a marked decrease in EPSP amplitude that was coincident with a shift from negative to positive (or an increase in positive) amplitude modulation, whereas CGP-35348 produced the opposite increase in EPSP amplitude and a negative shift in modulation. A likely explanation for the positive shift in amplitude modulation with baclofen is diminished Ca²⁺ entry into the terminals resulting in a decreased probability of transmitter release and consequently less transmitter depletion and thus less EPSP depression during repetitive stimulation (see INTRODUCTION). The reduced depression revealed the underlying tendency of synapses to exhibit facilitation resulting in more positive (or less negative) amplitude modulation during high-frequency Ia-afferent stimulation. A similar decrease in EPSP amplitude and positive shift in amplitude modulation has been observed in the in vitro neonatal rat spinal cord when extracellular Ca²⁺ concentration was reduced (Pinco and Lev-Tov 1993; Seebach and Mendell 1996).

The opposite effects on EPSP amplitude and amplitude modulation during high-frequency stimulation would be expected to occur after CGP-35348 administration because of an increase in Ca²⁺ current in Ia terminals (release from GABA_B-induced decrease in Ca²⁺ entry). The increased Ca²⁺ entry would elevate probability of transmitter release and residual Ca²⁺, but this enhanced release would also increase transmitter depletion, tending to mask any underlying facilitation. Thus the net effect expected is a negative shift in amplitude modulation. Substantial increases in EPSP amplitude occurred as expected; however, these were accompanied by only relatively small changes in EPSP amplitude modulation toward more negative values. Assuming that the magnitude and sign of modulation depends on the balance between residual Ca²⁺ and transmitter depletion, it appears that this balance measured as a function of the change in EPSP amplitude is affected differently by baclofen and by CGP-35348. These differences in the responses to theGABA_B receptor agonist and antagonist may relate to the EPSPs being in different portions of the amplitude spectrum, which was the result of treatments that altered Ca²⁺ influx in opposite directions. These findings are consistent with a nonlinear relationship between probability of transmitter release, EPSP amplitude and modulation, probably closer to a semilogarithmic relationship (Fig. 8), as has been found in the cat (Collins et al. 1988). The cellular basis for this nonlinearity is presently unknown, but it is important to recognize that transmitter release is probably not uniform at the different release sites of a given afferent (Redman 1990; Walmsley et al. 1988) and this could be a source of nonlinearity in these relationships.

Manipulation of GABA_B receptor activity led to a much steeper change in EPSP amplitude modulation as a function of EPSP amplitude than was observed when these variables were compared across the motoneuron pool. Because the GABA_B receptor controls Ca²⁺ channel activity, manipulation of its activity can be assumed to affect probability of release as confirmed by studies of quantal release variables. Thus the finding that manipulation of probability of release directly results in a much steeper slope for the relationship between EPSP amplitude modulation and EPSP amplitude than what is observed across the motoneuron pool suggests that differences in EPSP amplitude in different motoneurons are not the result of systematic differences in probability of transmitter release. Rather it seems likely in the rat that most of the differences in EPSP amplitude across the motoneuron pool are the result of other factors such as the convergence of afferents, number of boutons per afferent, and motoneuron input resistance.

In the cat a much stronger correlation has been described across the motoneuron pool between modulation and EPSP amplitude than was observed here. One possible reason for the differences between cat and rat may relate to differences in the composition of the motoneuron pool whose members have been suggested to retrogradely determine the transmitter release properties of the presynaptic terminals synapsing on them (Davis and Murphey 1994; Koerber and Mendell 1991; Mendell et al. 1994). In the cat, type S motoneurons make up about 25% of the medial gastrocnemius motoneuron pool (Burke 1967), whereas in the male rat used here the percentage of type S motoneurons has been estimated at about 10% (Gardiner 1993). Because type S motoneurons in the cat display the most consistent negative values of modulation (Mendell et al. 1995), we can infer that in the present experiments in the rat these were few in number. Furthermore, no values of modulation >55% were seen in the rat in contrast to values >75% seen regularly in the cat at connections on type F motoneurons with small EPSPs (Collins et al. 1988; Mendell et al. 1995). Both of these factors likely contribute to the shallow slope in the modulation versus amplitude relationship in the rat.

The uniform decrease in potentiation after baclofen is consistent with the mechanism of action of this agent, i.e., decreasing Ca²⁺ entry into the presynaptic terminals. This does not agree with previous reports of an increase in potentiation after baclofen (Lev-Tov et al. 1988). However, the changes in potentiation after baclofen in that experiment likely resulted from the removal of the profound depression that follows the 30-s, 500-Hz train used as the conditioning burst. The brief bursts used in the present experiments do not have this complication (see Davis et al. 1985, 1987). It is interesting that baclofen produces a decrease in potentiation after the burst but a tendency to exhibit more facilitation during the burst. This implies that different mechanisms account for the enlarged EPSPs during and after the burst. As discussed above, the behavior during the burst may be determined by the balance between transmitter depletion and facilitation resulting from residual Ca²⁺, whereas for at least 2 s after the burst the residual Ca²⁺ may be dominant. Unexpectedly, a decrease in potentiation was observed after CGP-35348 when enhanced Ca²⁺ entry might have been expected to increase it. We speculate that other mechanisms, e.g., interference with transmitter replenishment or a very efficient Ca²⁺ pump, may be operating under these conditions and that this may also account in part for the relatively small changes in modulation induced by CGP-35348.

The fact that CGP-35348 results in an increase in Ia-motoneuron EPSP amplitude in the absence of baclofen indicates the presence of tonic GABA_B-mediated presynaptic inhibition in this system, although not excluding a possible role for -aminobutyric acid-A (GABA_A) mediated presynaptic inhibition. Stuart and Redman (1992) concluded that GABA-mediated presynaptic inhibition at this synapse is mediated via GABA_A and not via GABA_B receptors. However it is important to note that our observations, like those of Peng and Frank (1989), reflect the level of tonic presynaptic inhibition of unknown origin, whereas Stuart and Redman (1992) induced presynaptic inhibition by a conditioning volley to the posterior biceps-semitendinosus nerve. More recently, the GABA_B receptor has been reported to be involved in the tonic component of the segmentally evoked presynaptic inhibition of the lumbar extensor monosynaptic reflexes in the cat spinal cord in contrast to the GABA_A receptor, which is involved in the short-latency component (Curtis and Lacey 1994). Whether or not this involves different pathways or is the result of distinct pharmacological properties of these receptors is not presently known (see Curtis 1998).

The composite EPSPs studied in these experiments resulted from activation of populations of afferents and so the reported values of amplitude, modulation, and potentiation are averages. We assume that the single fiber and composite EPSPs responded in the same manner to baclofen and CGP-35348. It is known that the modulation produced by different afferents in the same motoneuron exhibits significantly less variability than that observed in different motoneurons in response to stimulation of the same afferent (Koerber et al. 1991). Furthermore, single fiber and composite EPSPs exhibit similar values of modulation during the burst (Koerber and Mendell 1991) and potentiation after the burst (Davis et al. 1985, 1987). Thus the composite EPSP is likely to have properties similar to those of the single fiber EPSPs of which it is composed. The possibility that transmitter release from different sites of the same afferent differs(Redman 1990; Walmsley et al. 1988) also emphasizes the need to consider the present results as averages. However, it is clear that these averages were affected in a consistent manner by the various treatments carried out here.

In conclusion, a major objective of these experiments was to determine whether manipulation of probability of transmitter release at connections on a given motoneuron would cause correlated changes in EPSP amplitude and frequency-dependent amplitude modulation similar to those seen across the population of motoneurons. Qualitatively, the similarity of these relationships is indicative of a common process. However, the correlated changes in modulation and amplitude were much greater after pharmacological manipulation of EPSP amplitude via CGP-35348 and particularly after baclofen than at synapses of different amplitude across the motoneuron pool. Assuming that the pharmacological manipulations affected mainly probability of transmitter release, this suggests that differences in this factor play only a minor role in determining EPSP amplitude across the motoneuron pool in the rat. It remains to be seen whether these pharmacological manipulations would suggest a more prominent role for probability of transmitter release in the determination of EPSP amplitude as has been suggested to occur in the cat (Mendell et al. 1990).

	ACKNOWLEDGEMENTS

  We thank T. C. Cope for valuable comments on a draft of this manuscript.

  This work was supported by National Science Foundation Grant IBN9421695 to W. F. Collins and National Institutes of Neurological Disorders and Stroke Grants 2RO1 NS-16996 (Javitts Neuroscience Award), RO1 NS-32264, and PO1 NS-14899 to L. M. Mendell.

	FOOTNOTES

  Address reprint requests to L. M. Mendell.

  Received 2 July 1997; accepted in final form 4 September 1997.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• BETZ, W. J. Depression of transmitter release at the neuromuscular junction of the frog. J. Physiol. (Lond.) 206: 629-644, 1970.[Abstract/Free Full Text]
• BOWERY, N. G., HUDSON, A. L., PRICE, G. W. GABA_A and GABA_B receptor site distribution in the rat central nervous system. Neuroscience 20: 365-383, 1987.[Medline]
• BURKE, R. E. Motor unit types of cat triceps surae muscle. J. Physiol. (Lond.) 193: 141-160, 1967.[Medline]
• COLLINS, W. F. III, DAVIS, B. M., MENDELL, L. M. Modulation of EPSP amplitude during high frequency stimulation depends on the correlation between potentiation, depression and facilitation. Brain Res. 442: 161-165, 1988.[Medline]
• COLLINS, W. F. III, HONIG, M. G., MENDELL, L. M. Heterogeneity of group Ia synapses on homonymous -motoneurons as revealed by high frequency stimulation of Ia afferent fibers. J. Neurophysiol. 52: 980-993, 1984.[Abstract/Free Full Text]
• CURTIS, D. R. Two types of inhibition in the spinal cord. In: Presynaptic Inhibition and Neural Control, Edited by P. Rudomin, R. Romo, and L. M. Mendell. New York: Oxford Univ. Press, 1998, p. 150-161.
• CURTIS, D. R., GYNTHER, B. D., LACEY, G., BEATTIE, D. T. Baclofen-reduction of presynaptic calcium influx in the cat spinal cord in vivo. Exp. Brain Res. 113: 520-533, 1997.[Medline]
• CURTIS, D. R., LACEY, G. GABA_B receptor-mediated spinal inhibition. Neuroreport 5: 540-542, 1994.[Medline]
• DAVIS, B. M., COLLINS, W. F. III, MENDELL, L. M. Potentiation of transmission at Ia-motoneuron connections induced by repeated short bursts of afferent activity. J. Neurophysiol. 54: 1541-1552, 1985.[Abstract/Free Full Text]
• DAVIS, B. M., DRUZINSKY, R. E., MENDELL, L. M. Distribution of potentiation following short high frequency bursts to motoneurons of different rheobase. Exp. Brain Res. 65: 639-648, 1987.[Medline]
• DAVIS, G. W., MURPHEY, R. K. Long-term regulation of short-term transmitter release properties: retrograde signaling and synaptic development. Trends Neurosci. 17: 9-13, 1994.[Medline]
• DEISZ, R. A., LUX, H. D. -aminobutyric acid-induced depression of calcium currents of chick sensory neurons. Neurosci. Lett. 56: 205-210, 1985.[Medline]
• DEL CASTILLO, J., KATZ, B. Statistical factors involved in neuromuscular facilitation and depression. J. Physiol. (Lond.) 124: 574-585, 1954.
• DOLPHIN, A. C., SCOTT, R. H. Calcium channel currents and their inhibition by ()-Baclofen in rat sensory neuronesmodulation by guanine nucleotides. J. Physiol. (Lond.) 386: 1-17, 1987.[Abstract/Free Full Text]
• DUNLAP, K., FISCHBACH, G. D. Neurotransmitters decrease the calcium conductance activated by depolarization of embryonic sensory neurons. J. Physiol. (Lond.) 317: 519-535, 1981.[Abstract/Free Full Text]
• , 1988. DUTAR, P., NICOLL, R. A. A physiological role for GABA_B receptors in the central nervous system. Nature 332: 156-158, 1988.[Medline]
• EDWARDS, F. R., HARRISON, P. J., JACK, J.J.B., KULLMANN, D. M. Reduction by baclofen of monosynaptic EPSPs in lumbosacral motoneurones of the anaesthetized cat. J. Physiol. (Lond.) 416: 539-556, 1989.[Abstract/Free Full Text]
• FRANK, E. Matching of facilitation at the neuromuscular junction of the lobster: a possible case for influence of muscle on nerve. J. Physiol. (Lond.) 233: 635-658, 1973.[Abstract/Free Full Text]
• GAHWILER, B. H., BROWN, D. A. GABA_B-receptor-activated K⁺ current in voltage-clamped CA3 pyramidal cells in hippocampal cultures. Proc. Natl. Acad. Sci. USA. 82: 1558-1562, 1985.[Abstract/Free Full Text]
• GARDINER, P. Physiological properties of motoneurons innervating different muscle unit types in rat gastrocnemius. J. Neurophysiol. 69: 1160-1170, 1993.[Abstract/Free Full Text]
• GERTLER, R. A., ROBBINS, N. Differences in neuromuscular transmission in red and white muscles. Brain Res. 142: 160-164, 1978.[Medline]
• HONIG, M. G., COLLINS, W. F. III, MENDELL, L. M. -Motoneuron EPSPs exhibit different frequency sensitivities to single Ia-afferent fiber stimulation. J. Neurophysiol. 49: 886-901, 1983.[Free Full Text]
• KATZ, B., MILEDI, R. The role of calcium in neuromuscular facilitation. J. Physiol. (Lond.) 195: 481-492, 1968.[Abstract/Free Full Text]
• KOERBER, H. R., MENDELL, L. M. Modulation of functional synaptic transmission during high frequency stimulation: role of postsynaptic target. J. Neurophysiol. 65: 590-597, 1991.[Abstract/Free Full Text]
• KORN, H., TRILLER, A., MALLET, A., FABER, D. S. Fluctuating responses at a central synapse, n of binomial fit predicts number of stained presynaptic boutons. Science 213: 898-901, 1981.[Abstract/Free Full Text]
• KUNO, M. Quantal components of excitatory synaptic potentials in spinal motoneurones. J. Physiol. (Lond.) 175: 81-99, 1964.
• LEV-TOV, A., MYERS, D.E.R., BURKE, R. E. Activation of type B -amino butyric acid receptors in the intact mammalian spinal cord mimics the effects of reduced presynaptic Ca²⁺ influx. Proc. Natl. Acad. Sci. USA. 85: 5330-5334, 1988.[Abstract/Free Full Text]
• LOEB, G. E., DUYSENS, J. Activity patterns in individual hindlimb primary and secondary muscle spindle afferents during normal movements in unrestrained cats. J. Neurophysiol. 43: 968-985, 1979.
• MALLART, A., MARTIN, A. R. An analysis of facilitation of transmitter release at the frog neuromuscular junction. J. Physiol. (Lond.) 193: 679-694, 1967.[Medline]
• MALLART, A., MARTIN, A. R. The relation between quantum content and facilitation at the neuromuscular junction of the frog. J. Physiol. (Lond.) 196: 593-604, 1968.[Abstract/Free Full Text]
• MENDELL, L. M., COLLINS, W. F. III, KOERBER, H. R. How are synapses distributed on spinal motoneurons to permit orderly recruitment? In: The Segmental Motor System, edited by M. D. Binder, and L. M. Mendell . New York: Oxford Univ. Press, 1990, p. 328-340
• MENDELL, L. M., COLLINS, W. F. III, MUNSON, J. B. Retrograde determination of motoneuron properties and their synaptic input. J. Neurobiol. 25: 707-721, 1994.[Medline]
• MENDELL, L. M., PESHORI, K. R., PIERCE, J. P., KOERBER, H. R., COLLINS, W. F. III. Short-term plasticity at Ia connections on motoneurons: physiological and functional implications. In: Plasticity of Motoneuronal Connections, edited by and A. Wernig . Amsterdam: Elsevier, 1991, p. 311-322
• MENDELL, L. M., TAYLOR, J. S., JOHNSON, R. D., MUNSON, J. B. Rescue of motoneuron and muscle afferent function in cats by regeneration into skin. II. Ia-Motoneuron Synapse. J. Neurophysiol. 73: 662-673, 1995.[Abstract/Free Full Text]
• OTIS, T. S., TRUSSELL, L. O. Inhibition of transmitter release shortens the duration of the excitatory synaptic current at a calyceal synapse. J. Neurophysiol. 76: 3584-3588, 1996.[Abstract/Free Full Text]
• PENG, Y.-Y., FRANK, E. Activation of GABA_B receptors causes presynaptic inhibition at synapses between muscle spindle afferents and motoneurons in the spinal cord of the bullfrog. J. Neurosci. 9: 1502-1515, 1989.[Abstract]
• PESHORI, K. R., COLLINS, W. F. III, MENDELL, L. M. Change in EPSP amplitude modulation during high frequency stimulation is correlated with changes in EPSP amplitudea baclofen study. Soc. Neurosci. Abstr. 17: 647, 1991.
• PESHORI, K. R., COLLINS, W. F. III, MENDELL, L. M. Effects ofGABA_B manipulation on EPSP amplitude and frequency-dependent modulation at rat Ia-motoneuron synapses-evidence for tonic GABA_B activity. Soc. Neurosci. Abstr. 19: 538, 1993.
• PINCO, M., LEV-TOV, A. Modulation of monosynaptic excitation in the neonatal rat spinal cord. J. Neurophysiol. 70: 1151-1158, 1993.[Abstract/Free Full Text]
• PRICE, G. W., WILKIN, G. P., TURNBULL, M. J., BOWERY, N. G. Are baclofen-sensitive GABA_B receptors present on primary afferent terminals of the spinal cord? Nature 307: 71-74, 1984.[Medline]
• PROCHAZKA, A., WESTERMAN, R. A., ZICONE, S. P. Ia afferent activity during a variety of voluntary movements in the cat. J. Physiol. (Lond.) 268: 423-448, 1977.[Abstract/Free Full Text]
• REDMAN, S. J. Quantal analysis of synaptic potentials in neurons of the central nervous system. Physiol. Rev. 70: 165-198, 1990.[Free Full Text]
• SEEBACH, B. S., MENDELL, L. M. Maturation in properties of motoneurons and their segmental input in the neonatal rat. J. Neurophysiol. 76: 3875-3885, 1996.[Abstract/Free Full Text]
• SHAPOVALOV, A. I., SHIRIAEV, B. I. Selective modulation of chemical transmission at a dual-action synapse (with special reference to baclofen). Gen. Physiol. Biophys. 1: 423-433, 1982.
• STUART, G. J., REDMAN, S. J. The role of GABA_A and GABA_B receptors in presynaptic inhibition of Ia EPSPs in cat spinal motoneurones. J. Physiol. (Lond.) 447: 675-692, 1992.[Abstract/Free Full Text]
• WALMSLEY, B., EDWARDS, F. R., TRACEY, D. J. Non-uniform release probabilities underlie quantal synaptic transmission at a mammalian excitatory central synapse. J. Neurophysiol. 60: 889-908, 1988.[Abstract/Free Full Text]
• ZENGEL, J. E., REID, S. A., SYPERT, G. W., MUNSON, J. B. Membrane electrical properties and prediction of motor-unit type of medial gastrocnemius motoneurons in the cat. J. Neurophysiol. 53: 1323-1344, 1985.[Abstract/Free Full Text]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society