0306-4522/91$3.00+ 0.00 Pergamon Press plc IBRO
NeuroscienceVol. 41, No. 1, pp. 6149, 1991 Printed in Great Britain
DORSAL SPINAL
ROOT POTENTIALS IN THE ISOLATED FROG CORD: AMINO ACID NEUROTRANSMITTERS AND MAGNESIUM IONS J. C. HACKMAN and R. A. DAMDOFF*
Department of Neurology, University of Miami School of Medicine and the Neurophysiology Laboratory, Veteran’s Administration Medical Center, Miami, FL 33101, U.S.A. Abstract-Sucrose gap techniques recorded dorsal root potentials evoked by supramaximal dorsal root stimulation in in oitro, hemisected frog spinal cords. In 0 mM Mg 2+ large (mean 13.0 mV), long lasting (mean 8.1 s) dorsal root potentials were recorded which consisted of two components: (1) an early component sensitive to picrotoxin, bicuculline, and low [Cl-], and presumably produced by activation of GABA, receptors; and (2) a long-duration second component enhanced and lengthened by picrotoxin, bicuculline and low [Cl-], and thought to result from increased intemeuron discharges resulting from depression of GABA-mediated pre- and postsynaptic inhibition. Both the early and late components were reduced by over 90% in amplitude and duration by 20mM Mg+ or by kynurenate and bicuculline. The early component of the dorsal root potential may depend mainly upon activation of non-N-methylp-aspartate receptors, but the late component requires both N-methyl-D-aspartate and non-iv-methyl-Daspartate receptors. Thus, the N-methyl-D-aspartate antagonist D-( -)-2-amino-S-phosphonovalerate caused only a modest reduction in the amplitude of the early dorsal root potential component while the non N-methyl-D-aspartate antagonist 6-cyano-7-nitroquinoxaline-2,3-dione caused a much more substantial reduction. Exposure of the spinal cord to a “physiological” concentration of M$+ (1.0 mM) greatly reduced the duration and somewhat reduced the amplitude of the dorsal root potential. The reduction of dorsal root potentials by 1.OmM M2+ appears to be caused by both pre- and postsynaptic factors. These include: (1) reduction of evoked transmitter release; the reduction of the dorsal root potential by 1.0 mM M+$+ was partly reversed by doubling [Ca*+],; (2) block of the N-methyl-o-aspartate receptor ibn channelq(3)
decrease of GABA-depolarization of primary tierent fiber terminals: GABA-denolarizations of terminals were depressed by l.OmM Mgr+; and (4) decreased release of K+ by’afferent volleys; exposure to 1.0 mM Mg+ reduced the increment in [K+], produced by repetitive afferent stimulation. D-( -)-2-amino-5-phosphonovalerate had only minimal effects on dorsal root potentials in 1.0 mM M$+ presumably because the N-methyl-o-aspartate receptor-ion channel complex is largely inactivated by the Mg’+. In contrast, kynurenate and 6-cyano-7-nitroquinoxaline-2.3-dione substantially diminished dorsal root potential amplitude and duration.
Afferent volleys produce a prolonged depolarization of the intraspinal portion of primary afferent fibersa process termed primary afferent depolarization (PAD).i7 PAD is electrotonically conducted from the intraspinal presynaptic terminal region out along afferent fibers where electrodes placed along the cut end of a dorsal root (DR) can record the change of potential as the dorsal root potential (DRP).’ There are two major hypotheses about the genesis of PAD. (1) PAD is a synaptic process produced by the action of a specific transmitter.25 Much pharmacological and chemical evidence has implicated GABA as the responsible transmitter.‘0s36 (2) Ionic changes external to the central terminals of DR fibers, in particular, an increase of [K+],, cause PAD.5
Studies using K+-selective microelectrodes have shown substantial increments of [K+], following afferent activity. 34*MThe two hypotheses are not mutually exclusive, for there is evidence that PAD consists of two components: an “early” phase caused by the synaptic release of GABA and a “late” phase produced by elevated [K+],.‘2,34,38*50 However, other mechanisms may be involved. For example, the increment in [K+], evoked by afferent volleys is estimated to account for l&40% of the potential change underlying the DRP,4,3*,44*” but data now indicate that activation of GABAergic or K+dependent processes may depend upon the stimulation paradigm used to activate afferent fiberq3’ and although the increase in [K+], is thought to be produced by intemeuronal discharges,” it is unclear whether the increase is evoked direclty as a result of an afferent volley, or indirectly by the action of a second transmitter released as a result of afferent activity. With regard to the GABA hypothesis, much weight has been placed on findings that the GABA antagonists picrotoxin and bicuculline reduce DRPs.“,~~ However, d.c. recordings from DRs of the
*To whom correspondence should be addressed. APV, D( -)-2-amino-S-phosphonovalerate; BMI, bicuculline methiodide; CNQX, 6-cyano-7nitroquinoxaline-2,3-dione; DR, dorsal root; DRP, dorsal root potential; 2-hydroxy-saclofen, 3-amino-2-(4chlorophenyl)-2-hydroxypropyl sulphonic acid; NMDA, N-methyl-D-aspartate; PAD, primary aBerent depolarization.
Abbreviations:
61
62
J. C. HACKMANand R. A. DAVIDOFF
isolated frog cord show that these agents reduce only the early phase of the DRP produced by stimulating DRs; the latter portion of the potential is actually enlarged and prolonged.4,3E~45,49 However, picrotoxin increases the K+ transients produced by DR stimulation,49r50 and accumulated interstitial K+ may account for the large, slow component of the DRP evoked after application of picrotoxin and bicuculline.34*49Alternatively, a second neurotransmitter (e.g. L-glutamate)4.22 resistant to picrotoxin and bicuculline may be involved in PAD. Interpretation of experimental data is further complicated by findings that both GABA and L-glutamate can increase extracellular K+ .22.‘5,47 There is also conflicting information about the roles played by N-methyl-Daspartate (NMDA) and non-NMDA excitatory amino acid receptors putatively activated by synaptically released L-glutamate.24.41 In addition to these physiological controversies, it must be noted that much of the data about PAD were obtained from experiments in amphibian spinal cord preparations bathed in Mg2+-free medium. Such preparations leave the NMDA receptor unblocked and in a presumably unphysiological state.’ DRPs recorded from such amphibian preparations are much larger and of longer duration than DRPs recorded from mammalian spinal cords. The reasons for the differences have never been investigated. The present experiments were performed to provide additional data about the pharmacology of the DRP. We used the isolated superfused amphibian spinal cord and d.c. sucrose gap recording from DRs. An account of some of these findings has been reported.27
The ninth DR was placed across a 3-mm sucrose gap 10 record the membrane potential of afferent terminals electrotonically conducted along the population of axons contained in the root. Differential d.c. recordings were obtained between the spinal cord bath and the distal end of the DR with calomel electrodes connected via agar--Ringer bridges to each recording site. The preparation was left ungrounded. Data were stored in an IBM AT computer. The integral [the area under the DRP or under the DR depolarization produced by application of GABA or by elevation of K + m the medium (mV ‘s)], the peak amplitude (mV). and the duration (s) of the DRP or other responses were determined by reprogrammed Asyst software; the potentials were plotted on a laser printer. In some cases three to six samples were averaged. In addition. signals were amplified and recorded with an oscilloscope and with a rectilinear pen recorder. With those recording conditions, an upward pen deflection represents depolarization of the intraspinal portion of afferent fibers. The fibers contained in the 10th DR were stimulated with supramaximal rectangular pulses (1 .Oms) delivered via bipolar silver/silver chloride electrodes. [K+l, was measured with double-barrel liquid eonexchanger microelectrodes. The electrode tips were posttioned in the intermediate portion of rhe gray matter 6O&lOOO~m from the cord dorsal surface. The sensitivity of the microelectrodes was tested with Ringer solution containing known concentrations of K+ before and after each experiment.
EXPERIMENTAL PROCEDURES Experiments were performed on adult frogs (Rmapipiem, Kons Scientific, Co., 3&55 g) chilled on ice until they were
thoroughly anesthetized. After decapitation and laminectomy, the lumbar cord with attached roots was quickly removed, hemisected sagittally (to reduce diffusional barriers), and one-half cord continuously superfused with HCO;-buffered Ringer solution [(in mM) NaCl, 114;CaCl,,
1.9; KCI, 2.0; NaHCO,, 10; glucose, 5.51.M$+ was either omitted from the medium to facilitate study of responses mediated by NMDA receptors,’ or added at a “physiological” concentration of l.OmM or at a concentration of 20mM to block synaptic transmission (see Results). The Ringer solution was kept at pH 7.2 rt 0.2 by gassing with 95% 0,/5% CO,. A Peltier cooling unit maintained its temperature at 18 k 1°C. Test substances were dissolved in Ringer solution and the pH and osmolarity adjusted as necessary. Electronically controlled solenoid valves were used for rapid switching between normal Ringer solution and superfusate solutions containing known concentrations of test substances.
RESULTS Large (up to 18 mV), stable, and long-lasting (up to 19) DRPs were evoked by DR volleys in the isolated, hemisected frog spinal cord maintained in medium devoid of Md+ ion$.“*” (Table 1). A typical sucrose gap recording of such a DRP elicited in response to supramaximal stimulation of an adjacent DR is illustrated in Fig. IA. When, as was customary in many laboratories,‘9 an a.c.-coupled amplifier system with a long time constant (I .O s) was used to record DR potential changes, the DRP was reduced in amplitude, considerably shortened, and followed by a prominent artifactual positive component (Fig. IA). The slow components of d.c.-recorded DRPs were very sensitive to the rate of stimulation and were reduced when the DR was stimulated at 5.0-30-s intervals (Fig. IB). For this reason, stimuli were delivered at I-min intervals in all experiments. Raising the temperature of the spinal cord by IO-C resulted in a significant shortening of the duration of the DRP (Fig. IC); amplitude was less affected. Synaptic transmission in the in vitro frog spinal cord was blocked by adding Mg*+ (20 mM).” Under such conditions, DRPs were attenuated to 6.9% of their amplitude and 7.4% of their duration in 0 mM Mg” (Fig. 1D, Table 1).‘8~4R
Table 1. Effects of magnesium ions on dorsal root potentials MS? OmM (64) MS+ 1.0 rnG (i3) Mg2+ 20 mM (4)
Amulitude (mV) .
Duration (s)
Area (mV.s)
13.0 + 0.5 10.1 f 0.5 0.9 + 0.2
8.1 f 0.5 3.3 + 0.4 0.620.1
23.7 + 2.9 5.9 f 0.8 0.4 f 0.2
All values are means + S.E.M. Numbers of preparations tested are in parentheses.
Frog dorsal root potentials
A dc.
record
O.C.record
l.OmM Mg2+
20mM Mg2+
H
Fig. I. Dorsal root potentials produced by single supramaximal dorsal root volleys. All records are sucrose gap recordings from the ninth DR of cords. AC are from cords superfused with Ringer solution devoid of Mg*+. Negativity is indicated by an upward deflection and signifies a depolarization of the intramedullary portion of afferent fibers. (A) DRPs recorded with d.c. and a.c. (time constant 1.0s) amplification. (B) DRPs recorded with d.c. amplification produced by stimuli delivered at 1.0~min (0.016 Hz) and 5-s (0.2 Hz) intervals. (C) DRP recorded with d.c. amplification at two different temneratures (12 and 22°C). (D) DRPs evoked in Ringer solution containing 0 mM M$’ ,‘1.OmM M$+ (30 min), and 20 mM Mg2+ (25 min). Time bar = 1 s (AC), 1.25 s (D). The early component of the dorsal root potential medium devoid of M&+
in
As demonstrated in previous studies involving amphibians,4,3*,4S exposure of the cord to medium containing bicuculline methiodide (50 PM) or picrotoxin (100 p M) progressively reduced the peak amplitude of the DRP by 15-30%. However, as shown in Fig. 2A, this reduction reflected a decrease in the size of the initial 150-200 ms of the DRP. Similar results were produced by bathing the cord in Ringer solution in which isethionate ions were substituted for 85% of
63
the chloride ions (Fig. 2B).3 The results are summarized in Table 2. These findings presumably reflect changes in GABA, sites, but GABA released as a result of afferent volleys may activate GABA, receptors as well as classical bicuculline-sensitive GABAA sites. Thus, when the GABA, receptor antagonist 3-amino-2-(4-chlorophenyl)-2-hydroxypropyl sulphonic acid (Zhydroxy-saclofen; 100 PM) was added to the superfusate, the area of the DRP was increased by 32 + 3% (n = 3), although there was no effect on the amplitude of the potential (100.3 f 2%). To determine which excitatory transmitter receptors might be involved in the synaptic generation of the DRP, we examined the effects of selective excitatory amino acid antagonists. Two antagonists were tested by bath application: the competitive NMDA receptor antagonist D( -)-Zamino-Sphosphonovalerate (APV; 10 PM) and the competitive non6-cyano-7-nitroantagonist NMDA receptor quinoxaline-2,3-dione (CNQX; 10 nM).29*s’ Both reduced the DRP, although the effects of the two antagonists on the early component differed. APV caused only a small (16%) reduction in the amplitude of the early component of the DRP, while CNQX caused a much more substantial reduction (45%) (Table 2, Fig. 3A,B). Both antagonists shortened the DRP and reduced the area of the potential (Table 2). The combination of APV and CNQX or kynurenate (2.0 mM), a non-selective NMDA and non-NMDA antagonist in the spinal cord,3o effectively and reversibly reduced both the amplitude (72%) and the duration (88%) of the DRP (Table 2, Fig. 3C). However, when 20 mM Mgr+ was subsequently added to the medium, the size of the residual potential was further reduced (Fig. 3C). From this finding one can infer that non-excitatory amino acid-mediated synaptic transmission may make a small contribution to the DRP. In addition, because bicuculline (50 PM, n = 3) reduced the amplitude (by more than 60%) and the duration (by more than 15%) of the small residual DRP remaining after exposure of the cord to kynurenate (2.0 mM) (Fig. 4D) it appears that there is a small GABAergic portion of the DRP
9 control
LOWcrlo
Fig. 2. Effect of bicuculline and low [Cl-], on DRPs in the absence of M$+. A depression of the early component of the DRP and an augmentation of a late component was produced by exposure of the cord to medium containing bicuculline methiodide (BMI) (50 PM, 10 min) (A) or low [Cl-], (90% of Clreplaced by isethionate, 30 min) (B). Insets in B at faster sweep speed. Amplitude bar = 5 mV (A and B) 10 mV (insets B). Time bar = 10 s (A and B) 0.2 s (insets B).
64
J. (‘.
and R. A. I~AWI~O~F
tl.wKw\x
Table 2. Effects of amino acid antagonists
and low [Cl
1, on dorsal root potential3
OmM Mg” n Picrotoxin 100 p M BMI 50fiM Low [Cl-], 20 mM APV IOpM CNQX IOfiM Kynurenate 2 mM APV + CNQX
Amphtude
(3) (12) (3) (3) (3) (8) (3)
15-c 5 68 + 3 59* II 84 & 2 55+4 28 + 3 21 t6
l.OmM Mg:Duration
Area
,*
Amplitude
Duration
A rca
1175f67 472 f 68 254 + 62 83 _+7 75*21 12z 1 9_c4
1820 + 365 704 + I62 930 * I23 31 *u 32 + 9 5 + 0.3 50+ I8
(3) (3)
84 + 2 722 I2
252 f 54 24.3 _t- ‘7
22x .: 61 405 t 39
(3) (3) (3)
91 + I 35 f I ‘4 + 8
55 I 9 62 * 8 38 c 6
66 2 I _’ 27; I 12: 3
Numbers (n) of preparations tested are in parentheses. Values (mean f S.E.M.) are expressed as percentages of control values of the same cords.
that does not depend upon activation of an excitatory amino acid receptor. The lute medium
component
of the dorsal
root
potential
in
deooid q/ Mg2+
That the late, convulsant-resistant component of the DRP is sensitive to excitatory amino acid antagonists is demonstrated in Fig. 4A-C. Thus, addition of APV, CNQX and kynurenate) to the Ringer solution reduced the late phase of the DRP produced during superfusion with the GABA, antagonists bicuculline and picrotoxin (Fig. 4A-C Table 3). APV was much more effective than CNQX in reducing the duration of the late, convulsant-resistant component than was CNQX. Eficts
of a “physiological”
concentration
in Fig. ID, the DRP was substantially lower in amplitude (almost 23%) and considerably shorter in duration (almost 40%) (Table I).“” To ascertain whether or not reduction of transmitter release from presynaptic terminals of primary afferent fibers and intemeurons is responsible in part for the depression of the DRP in 1.0 mM Mgr-, we exposed the cord to Ringer solution containing an elevated (4.0mM) concentration of Car’. One would expect that elevation of the Ca2+ concentration in the superfusate would antagonize the ability of M$+ to depress presynaptic transmitter release.‘s~‘6~‘2The result of increasing the Ca’-. concentration was an increase of the amplitude
of Mg2+
The DRP recorded from cords bathed in medium containing a “physiological” concentration of Mg2-’ (1 .OmM) differed significantly from that seen in medium devoid of Mg2+ (Table 1). As illustrated
BMI+KYN
L---J KYN
KYN + BMI
KYN’+M&+
Fig. 3. Excitatory amino acid antagonists reduce DRPs in M$+-free Ringer solution. (A) APV (IOpM, 3Omin) preferentially reduced the late component of the DRP; the early component was relatively ur&ected. (B) In contrast, both the early and late components of the DRP were sensitive to CNQX (I .OFM, 30 mm). (C) Kynurenate (KYN, 2.0 mM, 15 min) substantially redused both compomnts of the DRP, but subsequent addition of 20 mM Mg’+ reduced the DRP still mote. Amplitude bar = 2.5 mV (A), 5.0 mV (B and C). Time bar = I .Os (A and B), 0.5 s (C).
Fig. 4. Interactions of excitatory amino acids and bicuculSpinal cords bathed in Ringer solution devoid of Mg’+. (A-C) Excitatory amino acid antagonists decrease DRPs exposed to bicuculhne (BMI, 5OpM). APV (10 PM, 30rnin) (A), CNQX (IOpM. 3Omin) (B), and kynurcnate (2.0 mM, line.
30min) (C) depressed DRPs evoked after exposure to bicuculline. (D) Tire effect of bicuculline applied after kynurenate. The residual DRP remaining after application of kynurenate (2.0 mM. 30 min) is rcduczd by bicuculline
(BMI, 50 PM, 2Omin). Amplitude bar = 2.5 mV (A-C), 5.0 mV (D). Time bar = I .Os (A and B), 2.5 s (C). I .Os (D).
65
Frog dorsal root potentials Table 3. Effects of excitatory amino acid antagonists on dorsal root potentials in cords exposed to bicuculline mkthiodide in the absence of M$+ Kynurenate 2.0 mM (8) APV 10pM (3) CNQX 10bM (3)
Amplitude
Duration
Area
10*2 78 f 8 63 & 8
9*1 55f23 55f6
2 k 0.6 8kO.2 45f19
Numbers of preparations tested are in parentheses. Ringer solution contained 0 mM M$+. Values (mean &S.E.M.) are expressed as percentages of values obtained in cords exposed to BMI (50 PM) for between 10 and 30 min.
(118 &-6%) and duration (172 &-31%, n = 3) of the DRP (not illustrated).’ However, physiological concentrations of Mg+ also have postsynaptic effects on reflex transmission.‘~7~21 For example, one would expect NMDAmediated synaptic transmission to be substantially blocked in medium containing 1.OmM M$+; and indeed, when applied in Ringer solution containing 1.OmM Mg2+ the NMDA antagonist APV had only minimal effects on the amplitude, and modest effects on the duration, of the DRP (Fig. SA, Table 2). In contrast, CNQX, which blocks non-NMDAmediated synaptic transmission, and kynurenate, which has effects on NMDA and non-NMDA receptors, still caused substantial reductions of the DRP (Fig. 5B,C, Table 2). GABA, antagonists like bicuculline cause a large increase in the late component of the DRP (Fig. 2A). Bathing spinal cords in 1.0 mM M$+ lessened this increase. Typical effects on the DRP of bicuculline
M#++AW
B
klg2+
C
rvlg2+
Mg2+ +KYN
M#++CNCU
Fig. 5. The effect on the DBP of amino acid antagonists in Ringer solution containing 1.0 mM MgZ+. (A) APV (10 PM, 30min) produced only a slight change. (B) Kynurenate (2.0 mM, 30 min) produced a decrement in both early and late components. (C) CNQX (IOpM, 20min) shortened and depressed the DBP. (D) Application of bicuculline (BMI, 50 PM, 30 min) attenuated the early component and considerably prolonged the late component.
(BMI, 50 PM) in medium containing 1.OmM Mg2+ are shown in Fig. 5D. As in OmM M&+, the early part of the DRP was significantly reduced in amplitude while its duration was still increased (Table 2). However, the increase in duration was much less than when frog cords were superfused with medium devoid of M&+ ions and the late component never achieved the amplitude seen in 0 mM Mg2+. Similar effects were seen with picrotoxin (Table 2). Addition of GABA (1 .OmM, 10-s applications) to M$+-free medium evoked DR depolarizations whose amplitudes were reduced to 82 + 9% (n = 6) of control values in 1.0 mM Mg2+-Ringer solution (not illustrated).’ DR depolarizations produced by elevations of [K+], (lOmM, 10s) were similarly depressed by l.OmM M$+ (83 + 8% of control, n = 6). With regard to [K+],, repetitive efferent volleys cause an increment of [K+], in the intermediate gray matter of the frog spinal cord. In Ringer solution containing no added Mg2+, the elicited rises in [K+], produced by supramaximal stimuli (25 Hz, 10 s tetanus) exceeded the resting level by 2.3 + 0.5 mM (n = 10). In l.OmM Mg2+, the increment of [KC], produced by repetitive afferent stimulation was reduced to 63 f 7% (n = 10) of the increase in solutions without Mg2+ (not illustrated). DISCUSSION
Amino acid transmitters and the two components of the dorsal root potential
The present observations support the supposition that the DRP evoked in the frog cord by DR stimulation consists of at least two different components. (1) An early component lasting 15&200 ms putatively produced by the synaptic release of GABA and the subsequent activation of GABA, receptors; in this regard, PAD in the frog spinal cord has been shown to be accompanied by a large conductance increase in primary afferent fibers;@ (2) a very longduration component presumably largely accounted for by an increase of [K+],.4,34*44,48 The finding that a concentration of Mg2+ sufficient to block chemical synaptic transmission2’ reduced both the early and the late components of the DRP by about 95% indicates that both components depend almost completely upon intact synaptic transmission.38~48 GABA receptors of both the GABA, and GABA, type are reported to be present on afferent terminals.‘3s42Both receptors may be activated by GABA released as a result of afferent volleys.” Indeed, the potent GABA, receptor antagonist 2-hydroxysaclofen increased the duration of the DRP, an indication that the late component of the DRP may be influenced by GABA, receptors. Our results also show that the non-selective excitatory amino acid antagonist kynurenate30 substantially reduced the DRP. Similar results in the in vitro rat cord have recently been reported.24 These data suggest that L-glutamate (or a related amino acid) is
66
J. C. HACKMAWand R. A. DAVIWFF
utilized by circuits responsible for generation of the DRP. However, it is unclear from our data whether or not the putative excitatory amino acid transmitter is released by afferent fibers and/or interneurons. Excitatory synaptic transmission not mediated by amino acids may also make a small contribution to the DRP since 20 mM Mg2+ further decreased DRPs already attenuated by kynurenate. In addition, the finding that bicuculline could reduce the small component of the DRP remaining after block of excitatory amino acid-mediated transmission by kynurenate may indicate that a small DRP component depends upon GABA, but not upon activation of excitatory amino acid receptors. We assume that the small DRP remaining after block of both GABA and excitatory amino acid receptors depends upon non-synaptic mechanisms. Both NMDA and non-NMDA receptors are implicated in spinal neurotransmission.i4 The present observations with the selective antagonists APV and CNQX indicate that both types of receptor are present on neurons involved in the generation of DRPs. In Mg2+-free medium, APV and CNQX26*5’ reduced the DRP by almost 70%. However, the differential effects of APV and CNQX on the early and late components of the DRP suggest that the early component of the DRP is produced by a series of steps that principally involve non-NMDA receptors and that the late component is mediated by both NMDA and non-NMDA receptors.24v4’ Alternatively, the NMDA receptor-mediated component of the DRP may be relayed through non-NMDA receptors at the synapses made by primary afferent terminals. Slow, picrotoxin-resistant DRPs may be produced either by elevated [K+], or by the actions of L-glutamate. 4*38 Findings that picrotoxin and bicuculline enhance the evoked release of spinal K’38,48,50 support the role of elevated [K+],. However, there is an intimate relationship between excitatory amino acids and the release of K+ by afferent volleys. For example, we have shown that the synaptic release of excitatory amino acids and the subsequent activation of both NMDA and non-NMDA receptors on interneurons is responsible for about 85% of the K+ released as a result of afferent stimulation.” Furthermore, L-glutamate increases [K+], in the spinal cord. 22,35,47 Although L-glutamate and other excitatory amino acids can depolarize afferent termina1s,2~9*22the depolarization of low threshold afferent fibers produced by excitatory amino acids has been postulated to be generated indirectly as a consequence of the release of K+ ions from depolarized neurons close to afferent fibers within the dorsal horn.22 In contrast, unmyelinated afferent fibers are selectively sensitive to kainate.23 There is also recent electrophysiological data indicating that NMDA receptors are present on the somata of rat afferent neurons3’ However, there is still insufficient information to determine if the bicuculline-
resistant, kynurenate-sensitive component of the DRP results directly from the release of K1 and the subsequent depolarization of afferent terminals or if it is a result of a direct action of L-glutamate on afferent terminals. The present results demonstrate that manipulations (e.g. application of APV, CNQX, kynurenate, or 1.OmM Mg*+) which depress the late, convulsantresistant component of the DRP presumably do so by blocking excitatory amino acid receptors. We have also shown that 1.OmM Mg2+ suppresses the evoked release of K+. It would therefore appear that intact excitatory amino acid-mediated synaptic transmission is necessary for the enhanced and prolonged, late, convulsant-resistant phase of the DRP and for the evoked release of K+. The late phase of the DRP seen after application of picrotoxin and bicuculline or superfusion with low [Cl-], medium may be the result of facilitated interneuron activity which presumably results from depression of GABA-mediated inhibition. Comparison of amphibian and mammalian dorsal root potentials
DRPs recorded with the sucrose gap technique from frog spinal cords bathed in 0 mM Mg2+ medium are considerably larger and longer than those recorded in vivo from the cat cord in the conventional manner with wire electrodes placed on two points on a DR. For example, in the cat spinal cord, DRPs evoked by stimulation of cutaneous and muscle nerves are reported to be about 200-500 NV in size and to last about 15&300ms.‘*~i9 In contrast, in the present experiments DRPs evoked by DR volleys in the frog spinal cord are sizable (up to 18 mV in amplitude) and long-lasting (up to 19 s in duration). The present observations show that the discrepancies in amplitude and duration of DRPs recorded in the in vitro frog spinal cord and in the in vivo cat spinal cord are dependent, at least in part. upon differences in amplifier coupling, recording techniques, temperature, and the concentration of Mg 2+ ions in the extracellular space in the two preparations. a.c.-coupled recording with a l-s time constant has been conventionally used to record DRPs.‘~ Such recording conditions can artifactually and substantially shorten the record by filtering out the slower components of the potential; such was the case in our a.c. recording experiments. In addition, recordings performed with surface wire electrodes measure only a fraction of the true potential difference between active and resting membrane of the presynaptic terminal because of shunting by low resistance extracellular paths; such recordings are thus of low amplitude. With sucrose gap recording, the flow of sucrose washes away most of the ions responsible for extracellular shunting of the current generated by the intraspinal portion of primary afferent fibers. This permits recording of stable, high resolution potential
Frog dorsal root potentials
changes which more closely approximate the size and duration of the true intracellular events underlying the DRP. Temperature is another significant variable. The present experiments were carried out at 18°C. Most other in vitro experiments using the frog cord have been performed at similar temperatures (12-18”C).L~3B~41~4s Most in oiuo experiments on the cat are carried out at 36-39°C. Raising the temperature of the frog cord substantially shortened the DRP. The mechanism for this change was not investigated, but it is known that cooling lowers the threshold for reflex discharges along DRs, a phenomenon presumably reflecting increased PAD.6 The effects Mg?+ ions
of a “physiological”
concentration
of
Mg*+ ions are not ordinarily added to frog Ringer solution, but frog cerebrospinal fluid contains the cation in a concentration of 0.92 mM.” We assume that the interstitial fluid surrounding frog neurons in uiuo contains Mgs+ in approximately that concentration as well. DRPs recorded from spinal cords bathed in medium containing 1.0 mM M$+ are smaller and shorter than the DRP recorded in medium which does not contain M$+ ions. The reduction of DRPs by 1.0 mM M$+ may be caused by one or more of the following mechanisms. Block of the ion channels associated with N-methylD-aspartate receptors. As discussed above, our data
indicate that release of L-glutamate and subsequent activation of excitatory amino acid receptors is necessary for the generation of DRPs. Exposure to micromolar concentrations of Mg*+ ions is sufficient to subject the NMDA receptor ion channels to voltage-dependent blockade.39 In the presence of a millimolar concentration of Mgz+ ions, the NMDA receptor-channel complex should be largely inactive;‘.& our experiments have shown that when 1.0 mM Mg*+ is added to the superfusate, the selective NMDA antagonist APV had limited effectiveness in attenuating the size and duration of the DRP. Because in uiuo interstitial fluid contains Mg*+ ions, one could presume that the NMDA receptor-channel complex is minimally involved in the generation of the DRP. Hence, activation of non-NMDA receptors is probably responsible for much, if not all, of the excitatory amino acid-mediated component of the DRP in 1.0 mM M&+ medium.” However, we
67
cannot dismiss the possibility that NMDA receptors can participate in DRPs during high-frequency synaptic transmission such as that which occurs with repetitive 6ring of aIferent fibers. Under such conditions, NMDA receptors are active.2s Presynaptic reduction of evoked transmitter release. Because Mg*+ ions in the interstitial space can
block the evoked release of transmitter at synaptic junctions,“@ it would be anticipated that changing the extracellular concentration of Mg*+ from 0 to 1.0 mM would reduce synaptically transmitted responses in the frog spinal cord. It would also be anticipated that M$+ ions’ depressant effect on presynaptic transmitter release would be reduced by elevation of the extracellular Ca2+ concentration.‘5v’6 In our experiments the effects on the DRP of raising the concentration of the two cations were as expected: raising Mgz+ to 1.0 mM reduced the DRP; the reduction was partly reversed by elevating the Ca*+ concentration in the superfusate. Decrease of GABA-induced depolarization of primary afferent fiber terminals. Our results show that
depolarizations of primary afferent fiber terminals produced by exogenous GABA were depressed by the presence of 1.0 mM Mgr+ in the Ringer solution.’ Although investigation of what caused the depression will have to await further investigation, we did note that in the presence of 1.O mM Mg*+ there was also a depression of the ability of K+ ions to depolarize afferent fibers. This indicates that the effect on GABA responses was not mediated by a selective action of Mg2+ on the GABA receptor or its ion channel!’ Decreased release of K+ by afferent uolleys. We
assume that the reduction of K+ release seen in 1.OmM Mg2+ in the present experiments results both from reduction of excitatory amino acid release from presynaptic terminals and postsynaptic effects of the cation on NMDA receptor cation channels.” Reduction of neuron excitability. High concentrations of Mg 2+ have a depressant effect on central neurons.33 This is presumably caused by an action on voltage-sensitive membrane to elevate electrical threshold. However, 1.0 mM Mgz+ is reported to be insufficient to produce an increase in the threshold of frog spinal neurons.*’ work was supported by USPHS grant NS 17577 and Veteran’s Administration Medical Center Funds (MRIS 1769 and 3369). Acknowledgements-This
REFERENCES
1. Ault B., Evans R. H., Francis A. A., Oakes D. J. and Watkins J. C. (1980) Selective depression of excitatory amino acid induced depolarizations by magnesium ions in isolated spinal cord preparations. J. Physiol., Lond. 307,413428. 2. Barker J. L. and Nicoll R. A. (1972) Gamma-aminobutyric acid: role in primary afferent depolarization. Science 176, 1043-1045. 3. Barker J. L. and Nicoll R. A. (1973) The pharmacology and ionic dependency of amino acid responses in the frog spinal cord. J. Phyxiol., Land. 228, 259-277. 4. Barker J. L., Niwll R. A. and Padjen A. (1975) Studies on convulsants in the isolated frog spinal cord. II. Effects on root potentials. J. Physiol., Land. 245, 537-548. 5. Barron D. H. and Matthews B. H. C. (1938) The interpretation of potential changes in the spinal cord. J. Physiof., Lond. 92, 276321.
68 6.
J. C.
HACKMAN and
R. A.
DAVIDOFF
Brooks C. M., Koizumi K. and Malcolm J. L. (1955), Effects of changes in temnerature on reactions of somai cord.
J. Neurophysiol. 18, 2055216.
7. Curtis D. R. and Gynther B. D. (1987) Divalent cations reduce depolarization of primary afferent terminations by GABA. Brain Res. 422, 192-195. 8. Czeh G. and Somjen G. G. (1989) Changes in extracellular calcium and magnesium and synaptic transmission in isolated mouse spinal cord. Brain Res. 486, 274-285. 9. Davidoff R. A. (1972) Gamma-aminobutyric acid antagonism and presynaptic inhibition. Science 175, 331-333. 10. Davidoff R. A. and Hackman J. C. (1985) GABA: presynaptic actions. In Neurotransmitter Acfions in the Vertebrale Neruous System (ed. Rogawski M. A. and Barker J. L.), pp. 3-32. Plenum Press, New York, 11. Davidoff R. A., Hackman J. C., Holohean A. M., Vega J. L. and Zhang D. X. (1988) Primary afferent activity, putative excitatory transmitters and extracellular potassium levels in frog spinal cord. J. Physiol., Land. 397, 291-306. 12. Davidoff R. A., Hackman J. C. and Osorio I. (1980) Amino acid antagonists do not block the depolarizing effects of potassium ions on frog primary afferents. Neuroscience 5, 117-126. 13. Davidoff R. A. and Sears E. S. (1974) The effects of Lioresal on synaptic activity in the isolated spinal cord. Neurolog) 24, 957-963. 14. Davies J. D. and Watkins J. C. (1983) Role of excitatory amino acid receptors in mono- and polysynaptic excitation in the spinal cord. Expl Brain Res. 49, 280-290. 15. Del Castillo J. and Engbaek L. (1954) The nature of the neuromuscular block produced by magnesium. J. Phy,riol.. Lond. 124, 370-384. 16. Del Castillo J. and Katz B. (1954) The effect of magnesium on the activity of motor nerve endings. J. Physiol., Land. 124, 553-559. 17. Eccles J. C. (1964) Presynaptic inhibition in the spinal cord. Prog. Brain Res. 12, 65-89. 18. Eccles J. C. and Kmjevic K. (1959) Potential changes recorded inside primary afferent fibres within the spinal cord. J. Physiol., Lond. 149, 250-273. 19. Eccles J. C., Magni F. and Willis W. D. (1962) Depolarization muscle. J. Physiol., Lond. 160, 62-93.
of central terminals of Group I afferent fibres from
20. Eccles J. C., Schmidt R. F. and Willis W. D. (1963) Pharmacological studies on presynaptic inhibition. J. Physiol., Land. 168, 500-530. 21. Erulkar S. D., Dambach G. E. and Mender D. (1974) The effect of magnesium at motoneurons of the isolated spinal cord of the frog. Brain Res. 66, 413424. 22. Evans R. H. (1980) Evidence supporting the indirect depolarization of primary afferent terminals in the frog by excitatory amino acids. J. Physiol., Lond. 298, 25-35. 23. Evans R. H. (1985) Kainate sensitivity of C-fibres in rat dorsal roots. J. Physiol., Lond. 358, 42P. 24. Evans R. H. and Long S. K. (1989) Primary afferent depolarization in the rat spinal cord is mediated by pathways utilising NMDA and non-NMDA receptors. Neurosci. Left. 100, 231-236. 25. Fatt P. (1954) Biophysics of junctional transmission. Physiol. Rev. 34, 674-710. 26. Fletcher E. J., Martin D., Aram J. A., Lodge D. and Honor6 T. (1988) Quinoxalinediones selectively block quisqualate and kainate receptors and synaptic events in rat neocortex and hippocampus and frog spinal cord in vitro. Br. J. Pharmac. 95, 585-597. 27. Hackman J. C. and Davidoff R. A. (1988) Mg*+, excitatory amino acids, and dorsal root potentials in the frog spinal cord. Sot. Neurosci. Abstr. 14, 789. 28. Herron C. E., Lester R. A. J., Coan E. J. and Collingridge G. L. (1986) Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic m&ha&m. Nature 3&I, 265-268.
29. Honor& T.. Davies S. N.. Dreier J.. Fletcher E. J.. Jacobsen P., Lodge D. and Neilsen F. E. (1988) Quinoxalinediones: potent competitive non-NMDA glutamate receptor antagonists. S&nce 241, 701-703. . _ 30. Jahr C. E. and Yoshioka K. (1986) la aBerent excitation of motoneurones in the in vitro new-born spinal cord is selectively antagonized by kynurenate. J. Physiol., Lond. 370, 515-530. 31. JimCnez I., Rudomin P. and Solodkin M. (1987) Mechanisms involved in the depolarization of cutaneous afferents produced by segmental and descending inputs in the cat spinal cord. Expl Brainkes. 69, 195-207. 32. Katz B. and Miledi R. (1963) A study of miniature potentials in spinal motoneurones. J. Physiol., Land. 168,389-422. 33. Kelly J. S.. Kmjevic K. and Somien G. (1969) Divalent cations and electrical properties of cortical cells. J. Neurobiol. 2, 1977208. 34. Kmjevic K. and Morris M. E. (1975) Correlation between extracellular focal potentials and K+ potentials evoked by primary afferent activity. Can. J. Physiol. Pharmac. 53, 912-922. 35. Kudo Y. and Fukoda H. (1976) Alteration in extracellular K+-activity induced by ammo acids in frog spinal cord. Jap. J. Pharmac. 26, 385-387. 36. Levy R. A. (1977) The role of GABA in primary afferent depolarization. Prog. Neurobiol. 9, 21 l-267. 37. Lovinger D. M. and Weight F. F. (1988) Glutamate induces a depolarization of adult rat dorsal root ganglion neurons
that is mediated by NMDA receptors. Neurosci. Lert. 94, 314-320. 38. Nicoll R. A. (1979) Dorsal root potentials and changes in extracellular potassium in the spinal cord of the frog. J. Physiol., Lond. 290, 113-127. 39. Nowak L., Bregstovski P., Ascher P., Herbet A. and Prochiantz A. (1984) Magnesium gates glutamate-activated channels in mouse central neurons. Nature 307, 462-465. 40. Padjen A. L. and Hashiguchi T. (1983) Primary afferent depolarization in frog spinal cord is associated with an increase in membrane conductance. Can. J. Physiol. Pharmac. 61, 626-631. 41. Padjen A. L. and Smith P. A. (1980) Specific effects of a-D,I_-aminoadipic acid on synaptic transmission in frog spinal cord. Can. J. Physiol. Pharmac. 58, 692-698. 42. Price G., Wilkin G. P., Turnbull M. J. and Bowery N. G. (1984) Are baclofen sensitive GABA, receptors present on primary afferent terminals of the spinal cord? Nature 307, 71-74. 43. Schmidt R. F. (1963) Pharmacological studies on the primary afferent depolarization of the toad spinal cord. PjzZgers Arch. ges. Physiol. 277, 325-346. 44. Shefner S. A. and Levy R. A. (1981) The contribution of increases in extracellular potassium to primary afferent depolarization in the bullfrog spinal cord. Brain Res. 205, 321-335.
Frog dorsal root potentials
69
45. Shirasawa Y. and Koketsu K. (1975) The effect of picrotoxin on the dorsal root potential of bullfrog spinal cords. Kurume Med. .I. 22, 14. 46. Smith P. A. (1982) The use of low concentrations of divalent cations to demonstrate a role for N-methyl-o-aspartate receptors in synaptic transmission in amphibian spinal cord. Br. J. Pharmac. 77, 363-373. 47. Sonnhof U. and Buhrle Ch. Ph. (1980) On the postsynaptic action of glutamate in frog spinal motoneurones. Pftigers Arch. ges. Physiol. 338, 101-109. 48. Sykova E. and Vyklicky L. (1977) Changes of extracellular potassium activity in isolated spinal cord of frog under high Mgz+ concentration. Neurosci. LQU. 4, 161-165. 49. Sykova E. and Vyklicky L. (1978) Effects of picrotoxin on potassium accumulation and dorsal root potentials in the
frog spinal cord. Neuroscience 3, 1061-1067. 50. ten Bruggencate G., Lux H. D. and Liebl L. (1974) Possible relationships between extracellular potassium activity and presynaptic inhibition in the spinal cord of the cat. PJtigers Arch. gei. Physiol. 349, 301-317.51. Wheatley P. L. and Collins K. J. (1986) Quantitative studies of N-methyl-D-aspartate, 2-amino-S-phosphonovalerate and cis-2,3-piperidine dicarboxylate interactions on the neonatal rat spinal cord in vitro. Eur. J. Pharmac. 121,257-263. (Accepted 12 September
1990)
NeuroscienceVol. 41, No. 1, pp. 6149, 1991 Printed in Great Britain
DORSAL SPINAL
ROOT POTENTIALS IN THE ISOLATED FROG CORD: AMINO ACID NEUROTRANSMITTERS AND MAGNESIUM IONS J. C. HACKMAN and R. A. DAMDOFF*
Department of Neurology, University of Miami School of Medicine and the Neurophysiology Laboratory, Veteran’s Administration Medical Center, Miami, FL 33101, U.S.A. Abstract-Sucrose gap techniques recorded dorsal root potentials evoked by supramaximal dorsal root stimulation in in oitro, hemisected frog spinal cords. In 0 mM Mg 2+ large (mean 13.0 mV), long lasting (mean 8.1 s) dorsal root potentials were recorded which consisted of two components: (1) an early component sensitive to picrotoxin, bicuculline, and low [Cl-], and presumably produced by activation of GABA, receptors; and (2) a long-duration second component enhanced and lengthened by picrotoxin, bicuculline and low [Cl-], and thought to result from increased intemeuron discharges resulting from depression of GABA-mediated pre- and postsynaptic inhibition. Both the early and late components were reduced by over 90% in amplitude and duration by 20mM Mg+ or by kynurenate and bicuculline. The early component of the dorsal root potential may depend mainly upon activation of non-N-methylp-aspartate receptors, but the late component requires both N-methyl-D-aspartate and non-iv-methyl-Daspartate receptors. Thus, the N-methyl-D-aspartate antagonist D-( -)-2-amino-S-phosphonovalerate caused only a modest reduction in the amplitude of the early dorsal root potential component while the non N-methyl-D-aspartate antagonist 6-cyano-7-nitroquinoxaline-2,3-dione caused a much more substantial reduction. Exposure of the spinal cord to a “physiological” concentration of M$+ (1.0 mM) greatly reduced the duration and somewhat reduced the amplitude of the dorsal root potential. The reduction of dorsal root potentials by 1.OmM M2+ appears to be caused by both pre- and postsynaptic factors. These include: (1) reduction of evoked transmitter release; the reduction of the dorsal root potential by 1.0 mM M+$+ was partly reversed by doubling [Ca*+],; (2) block of the N-methyl-o-aspartate receptor ibn channelq(3)
decrease of GABA-depolarization of primary tierent fiber terminals: GABA-denolarizations of terminals were depressed by l.OmM Mgr+; and (4) decreased release of K+ by’afferent volleys; exposure to 1.0 mM Mg+ reduced the increment in [K+], produced by repetitive afferent stimulation. D-( -)-2-amino-5-phosphonovalerate had only minimal effects on dorsal root potentials in 1.0 mM M$+ presumably because the N-methyl-o-aspartate receptor-ion channel complex is largely inactivated by the Mg’+. In contrast, kynurenate and 6-cyano-7-nitroquinoxaline-2.3-dione substantially diminished dorsal root potential amplitude and duration.
Afferent volleys produce a prolonged depolarization of the intraspinal portion of primary afferent fibersa process termed primary afferent depolarization (PAD).i7 PAD is electrotonically conducted from the intraspinal presynaptic terminal region out along afferent fibers where electrodes placed along the cut end of a dorsal root (DR) can record the change of potential as the dorsal root potential (DRP).’ There are two major hypotheses about the genesis of PAD. (1) PAD is a synaptic process produced by the action of a specific transmitter.25 Much pharmacological and chemical evidence has implicated GABA as the responsible transmitter.‘0s36 (2) Ionic changes external to the central terminals of DR fibers, in particular, an increase of [K+],, cause PAD.5
Studies using K+-selective microelectrodes have shown substantial increments of [K+], following afferent activity. 34*MThe two hypotheses are not mutually exclusive, for there is evidence that PAD consists of two components: an “early” phase caused by the synaptic release of GABA and a “late” phase produced by elevated [K+],.‘2,34,38*50 However, other mechanisms may be involved. For example, the increment in [K+], evoked by afferent volleys is estimated to account for l&40% of the potential change underlying the DRP,4,3*,44*” but data now indicate that activation of GABAergic or K+dependent processes may depend upon the stimulation paradigm used to activate afferent fiberq3’ and although the increase in [K+], is thought to be produced by intemeuronal discharges,” it is unclear whether the increase is evoked direclty as a result of an afferent volley, or indirectly by the action of a second transmitter released as a result of afferent activity. With regard to the GABA hypothesis, much weight has been placed on findings that the GABA antagonists picrotoxin and bicuculline reduce DRPs.“,~~ However, d.c. recordings from DRs of the
*To whom correspondence should be addressed. APV, D( -)-2-amino-S-phosphonovalerate; BMI, bicuculline methiodide; CNQX, 6-cyano-7nitroquinoxaline-2,3-dione; DR, dorsal root; DRP, dorsal root potential; 2-hydroxy-saclofen, 3-amino-2-(4chlorophenyl)-2-hydroxypropyl sulphonic acid; NMDA, N-methyl-D-aspartate; PAD, primary aBerent depolarization.
Abbreviations:
61
62
J. C. HACKMANand R. A. DAVIDOFF
isolated frog cord show that these agents reduce only the early phase of the DRP produced by stimulating DRs; the latter portion of the potential is actually enlarged and prolonged.4,3E~45,49 However, picrotoxin increases the K+ transients produced by DR stimulation,49r50 and accumulated interstitial K+ may account for the large, slow component of the DRP evoked after application of picrotoxin and bicuculline.34*49Alternatively, a second neurotransmitter (e.g. L-glutamate)4.22 resistant to picrotoxin and bicuculline may be involved in PAD. Interpretation of experimental data is further complicated by findings that both GABA and L-glutamate can increase extracellular K+ .22.‘5,47 There is also conflicting information about the roles played by N-methyl-Daspartate (NMDA) and non-NMDA excitatory amino acid receptors putatively activated by synaptically released L-glutamate.24.41 In addition to these physiological controversies, it must be noted that much of the data about PAD were obtained from experiments in amphibian spinal cord preparations bathed in Mg2+-free medium. Such preparations leave the NMDA receptor unblocked and in a presumably unphysiological state.’ DRPs recorded from such amphibian preparations are much larger and of longer duration than DRPs recorded from mammalian spinal cords. The reasons for the differences have never been investigated. The present experiments were performed to provide additional data about the pharmacology of the DRP. We used the isolated superfused amphibian spinal cord and d.c. sucrose gap recording from DRs. An account of some of these findings has been reported.27
The ninth DR was placed across a 3-mm sucrose gap 10 record the membrane potential of afferent terminals electrotonically conducted along the population of axons contained in the root. Differential d.c. recordings were obtained between the spinal cord bath and the distal end of the DR with calomel electrodes connected via agar--Ringer bridges to each recording site. The preparation was left ungrounded. Data were stored in an IBM AT computer. The integral [the area under the DRP or under the DR depolarization produced by application of GABA or by elevation of K + m the medium (mV ‘s)], the peak amplitude (mV). and the duration (s) of the DRP or other responses were determined by reprogrammed Asyst software; the potentials were plotted on a laser printer. In some cases three to six samples were averaged. In addition. signals were amplified and recorded with an oscilloscope and with a rectilinear pen recorder. With those recording conditions, an upward pen deflection represents depolarization of the intraspinal portion of afferent fibers. The fibers contained in the 10th DR were stimulated with supramaximal rectangular pulses (1 .Oms) delivered via bipolar silver/silver chloride electrodes. [K+l, was measured with double-barrel liquid eonexchanger microelectrodes. The electrode tips were posttioned in the intermediate portion of rhe gray matter 6O&lOOO~m from the cord dorsal surface. The sensitivity of the microelectrodes was tested with Ringer solution containing known concentrations of K+ before and after each experiment.
EXPERIMENTAL PROCEDURES Experiments were performed on adult frogs (Rmapipiem, Kons Scientific, Co., 3&55 g) chilled on ice until they were
thoroughly anesthetized. After decapitation and laminectomy, the lumbar cord with attached roots was quickly removed, hemisected sagittally (to reduce diffusional barriers), and one-half cord continuously superfused with HCO;-buffered Ringer solution [(in mM) NaCl, 114;CaCl,,
1.9; KCI, 2.0; NaHCO,, 10; glucose, 5.51.M$+ was either omitted from the medium to facilitate study of responses mediated by NMDA receptors,’ or added at a “physiological” concentration of l.OmM or at a concentration of 20mM to block synaptic transmission (see Results). The Ringer solution was kept at pH 7.2 rt 0.2 by gassing with 95% 0,/5% CO,. A Peltier cooling unit maintained its temperature at 18 k 1°C. Test substances were dissolved in Ringer solution and the pH and osmolarity adjusted as necessary. Electronically controlled solenoid valves were used for rapid switching between normal Ringer solution and superfusate solutions containing known concentrations of test substances.
RESULTS Large (up to 18 mV), stable, and long-lasting (up to 19) DRPs were evoked by DR volleys in the isolated, hemisected frog spinal cord maintained in medium devoid of Md+ ion$.“*” (Table 1). A typical sucrose gap recording of such a DRP elicited in response to supramaximal stimulation of an adjacent DR is illustrated in Fig. IA. When, as was customary in many laboratories,‘9 an a.c.-coupled amplifier system with a long time constant (I .O s) was used to record DR potential changes, the DRP was reduced in amplitude, considerably shortened, and followed by a prominent artifactual positive component (Fig. IA). The slow components of d.c.-recorded DRPs were very sensitive to the rate of stimulation and were reduced when the DR was stimulated at 5.0-30-s intervals (Fig. IB). For this reason, stimuli were delivered at I-min intervals in all experiments. Raising the temperature of the spinal cord by IO-C resulted in a significant shortening of the duration of the DRP (Fig. IC); amplitude was less affected. Synaptic transmission in the in vitro frog spinal cord was blocked by adding Mg*+ (20 mM).” Under such conditions, DRPs were attenuated to 6.9% of their amplitude and 7.4% of their duration in 0 mM Mg” (Fig. 1D, Table 1).‘8~4R
Table 1. Effects of magnesium ions on dorsal root potentials MS? OmM (64) MS+ 1.0 rnG (i3) Mg2+ 20 mM (4)
Amulitude (mV) .
Duration (s)
Area (mV.s)
13.0 + 0.5 10.1 f 0.5 0.9 + 0.2
8.1 f 0.5 3.3 + 0.4 0.620.1
23.7 + 2.9 5.9 f 0.8 0.4 f 0.2
All values are means + S.E.M. Numbers of preparations tested are in parentheses.
Frog dorsal root potentials
A dc.
record
O.C.record
l.OmM Mg2+
20mM Mg2+
H
Fig. I. Dorsal root potentials produced by single supramaximal dorsal root volleys. All records are sucrose gap recordings from the ninth DR of cords. AC are from cords superfused with Ringer solution devoid of Mg*+. Negativity is indicated by an upward deflection and signifies a depolarization of the intramedullary portion of afferent fibers. (A) DRPs recorded with d.c. and a.c. (time constant 1.0s) amplification. (B) DRPs recorded with d.c. amplification produced by stimuli delivered at 1.0~min (0.016 Hz) and 5-s (0.2 Hz) intervals. (C) DRP recorded with d.c. amplification at two different temneratures (12 and 22°C). (D) DRPs evoked in Ringer solution containing 0 mM M$’ ,‘1.OmM M$+ (30 min), and 20 mM Mg2+ (25 min). Time bar = 1 s (AC), 1.25 s (D). The early component of the dorsal root potential medium devoid of M&+
in
As demonstrated in previous studies involving amphibians,4,3*,4S exposure of the cord to medium containing bicuculline methiodide (50 PM) or picrotoxin (100 p M) progressively reduced the peak amplitude of the DRP by 15-30%. However, as shown in Fig. 2A, this reduction reflected a decrease in the size of the initial 150-200 ms of the DRP. Similar results were produced by bathing the cord in Ringer solution in which isethionate ions were substituted for 85% of
63
the chloride ions (Fig. 2B).3 The results are summarized in Table 2. These findings presumably reflect changes in GABA, sites, but GABA released as a result of afferent volleys may activate GABA, receptors as well as classical bicuculline-sensitive GABAA sites. Thus, when the GABA, receptor antagonist 3-amino-2-(4-chlorophenyl)-2-hydroxypropyl sulphonic acid (Zhydroxy-saclofen; 100 PM) was added to the superfusate, the area of the DRP was increased by 32 + 3% (n = 3), although there was no effect on the amplitude of the potential (100.3 f 2%). To determine which excitatory transmitter receptors might be involved in the synaptic generation of the DRP, we examined the effects of selective excitatory amino acid antagonists. Two antagonists were tested by bath application: the competitive NMDA receptor antagonist D( -)-Zamino-Sphosphonovalerate (APV; 10 PM) and the competitive non6-cyano-7-nitroantagonist NMDA receptor quinoxaline-2,3-dione (CNQX; 10 nM).29*s’ Both reduced the DRP, although the effects of the two antagonists on the early component differed. APV caused only a small (16%) reduction in the amplitude of the early component of the DRP, while CNQX caused a much more substantial reduction (45%) (Table 2, Fig. 3A,B). Both antagonists shortened the DRP and reduced the area of the potential (Table 2). The combination of APV and CNQX or kynurenate (2.0 mM), a non-selective NMDA and non-NMDA antagonist in the spinal cord,3o effectively and reversibly reduced both the amplitude (72%) and the duration (88%) of the DRP (Table 2, Fig. 3C). However, when 20 mM Mgr+ was subsequently added to the medium, the size of the residual potential was further reduced (Fig. 3C). From this finding one can infer that non-excitatory amino acid-mediated synaptic transmission may make a small contribution to the DRP. In addition, because bicuculline (50 PM, n = 3) reduced the amplitude (by more than 60%) and the duration (by more than 15%) of the small residual DRP remaining after exposure of the cord to kynurenate (2.0 mM) (Fig. 4D) it appears that there is a small GABAergic portion of the DRP
9 control
LOWcrlo
Fig. 2. Effect of bicuculline and low [Cl-], on DRPs in the absence of M$+. A depression of the early component of the DRP and an augmentation of a late component was produced by exposure of the cord to medium containing bicuculline methiodide (BMI) (50 PM, 10 min) (A) or low [Cl-], (90% of Clreplaced by isethionate, 30 min) (B). Insets in B at faster sweep speed. Amplitude bar = 5 mV (A and B) 10 mV (insets B). Time bar = 10 s (A and B) 0.2 s (insets B).
64
J. (‘.
and R. A. I~AWI~O~F
tl.wKw\x
Table 2. Effects of amino acid antagonists
and low [Cl
1, on dorsal root potential3
OmM Mg” n Picrotoxin 100 p M BMI 50fiM Low [Cl-], 20 mM APV IOpM CNQX IOfiM Kynurenate 2 mM APV + CNQX
Amphtude
(3) (12) (3) (3) (3) (8) (3)
15-c 5 68 + 3 59* II 84 & 2 55+4 28 + 3 21 t6
l.OmM Mg:Duration
Area
,*
Amplitude
Duration
A rca
1175f67 472 f 68 254 + 62 83 _+7 75*21 12z 1 9_c4
1820 + 365 704 + I62 930 * I23 31 *u 32 + 9 5 + 0.3 50+ I8
(3) (3)
84 + 2 722 I2
252 f 54 24.3 _t- ‘7
22x .: 61 405 t 39
(3) (3) (3)
91 + I 35 f I ‘4 + 8
55 I 9 62 * 8 38 c 6
66 2 I _’ 27; I 12: 3
Numbers (n) of preparations tested are in parentheses. Values (mean f S.E.M.) are expressed as percentages of control values of the same cords.
that does not depend upon activation of an excitatory amino acid receptor. The lute medium
component
of the dorsal
root
potential
in
deooid q/ Mg2+
That the late, convulsant-resistant component of the DRP is sensitive to excitatory amino acid antagonists is demonstrated in Fig. 4A-C. Thus, addition of APV, CNQX and kynurenate) to the Ringer solution reduced the late phase of the DRP produced during superfusion with the GABA, antagonists bicuculline and picrotoxin (Fig. 4A-C Table 3). APV was much more effective than CNQX in reducing the duration of the late, convulsant-resistant component than was CNQX. Eficts
of a “physiological”
concentration
in Fig. ID, the DRP was substantially lower in amplitude (almost 23%) and considerably shorter in duration (almost 40%) (Table I).“” To ascertain whether or not reduction of transmitter release from presynaptic terminals of primary afferent fibers and intemeurons is responsible in part for the depression of the DRP in 1.0 mM Mgr-, we exposed the cord to Ringer solution containing an elevated (4.0mM) concentration of Car’. One would expect that elevation of the Ca2+ concentration in the superfusate would antagonize the ability of M$+ to depress presynaptic transmitter release.‘s~‘6~‘2The result of increasing the Ca’-. concentration was an increase of the amplitude
of Mg2+
The DRP recorded from cords bathed in medium containing a “physiological” concentration of Mg2-’ (1 .OmM) differed significantly from that seen in medium devoid of Mg2+ (Table 1). As illustrated
BMI+KYN
L---J KYN
KYN + BMI
KYN’+M&+
Fig. 3. Excitatory amino acid antagonists reduce DRPs in M$+-free Ringer solution. (A) APV (IOpM, 3Omin) preferentially reduced the late component of the DRP; the early component was relatively ur&ected. (B) In contrast, both the early and late components of the DRP were sensitive to CNQX (I .OFM, 30 mm). (C) Kynurenate (KYN, 2.0 mM, 15 min) substantially redused both compomnts of the DRP, but subsequent addition of 20 mM Mg’+ reduced the DRP still mote. Amplitude bar = 2.5 mV (A), 5.0 mV (B and C). Time bar = I .Os (A and B), 0.5 s (C).
Fig. 4. Interactions of excitatory amino acids and bicuculSpinal cords bathed in Ringer solution devoid of Mg’+. (A-C) Excitatory amino acid antagonists decrease DRPs exposed to bicuculhne (BMI, 5OpM). APV (10 PM, 30rnin) (A), CNQX (IOpM. 3Omin) (B), and kynurcnate (2.0 mM, line.
30min) (C) depressed DRPs evoked after exposure to bicuculline. (D) Tire effect of bicuculline applied after kynurenate. The residual DRP remaining after application of kynurenate (2.0 mM. 30 min) is rcduczd by bicuculline
(BMI, 50 PM, 2Omin). Amplitude bar = 2.5 mV (A-C), 5.0 mV (D). Time bar = I .Os (A and B), 2.5 s (C). I .Os (D).
65
Frog dorsal root potentials Table 3. Effects of excitatory amino acid antagonists on dorsal root potentials in cords exposed to bicuculline mkthiodide in the absence of M$+ Kynurenate 2.0 mM (8) APV 10pM (3) CNQX 10bM (3)
Amplitude
Duration
Area
10*2 78 f 8 63 & 8
9*1 55f23 55f6
2 k 0.6 8kO.2 45f19
Numbers of preparations tested are in parentheses. Ringer solution contained 0 mM M$+. Values (mean &S.E.M.) are expressed as percentages of values obtained in cords exposed to BMI (50 PM) for between 10 and 30 min.
(118 &-6%) and duration (172 &-31%, n = 3) of the DRP (not illustrated).’ However, physiological concentrations of Mg+ also have postsynaptic effects on reflex transmission.‘~7~21 For example, one would expect NMDAmediated synaptic transmission to be substantially blocked in medium containing 1.OmM M$+; and indeed, when applied in Ringer solution containing 1.OmM Mg2+ the NMDA antagonist APV had only minimal effects on the amplitude, and modest effects on the duration, of the DRP (Fig. SA, Table 2). In contrast, CNQX, which blocks non-NMDAmediated synaptic transmission, and kynurenate, which has effects on NMDA and non-NMDA receptors, still caused substantial reductions of the DRP (Fig. 5B,C, Table 2). GABA, antagonists like bicuculline cause a large increase in the late component of the DRP (Fig. 2A). Bathing spinal cords in 1.0 mM M$+ lessened this increase. Typical effects on the DRP of bicuculline
M#++AW
B
klg2+
C
rvlg2+
Mg2+ +KYN
M#++CNCU
Fig. 5. The effect on the DBP of amino acid antagonists in Ringer solution containing 1.0 mM MgZ+. (A) APV (10 PM, 30min) produced only a slight change. (B) Kynurenate (2.0 mM, 30 min) produced a decrement in both early and late components. (C) CNQX (IOpM, 20min) shortened and depressed the DBP. (D) Application of bicuculline (BMI, 50 PM, 30 min) attenuated the early component and considerably prolonged the late component.
(BMI, 50 PM) in medium containing 1.OmM Mg2+ are shown in Fig. 5D. As in OmM M&+, the early part of the DRP was significantly reduced in amplitude while its duration was still increased (Table 2). However, the increase in duration was much less than when frog cords were superfused with medium devoid of M&+ ions and the late component never achieved the amplitude seen in 0 mM Mg2+. Similar effects were seen with picrotoxin (Table 2). Addition of GABA (1 .OmM, 10-s applications) to M$+-free medium evoked DR depolarizations whose amplitudes were reduced to 82 + 9% (n = 6) of control values in 1.0 mM Mg2+-Ringer solution (not illustrated).’ DR depolarizations produced by elevations of [K+], (lOmM, 10s) were similarly depressed by l.OmM M$+ (83 + 8% of control, n = 6). With regard to [K+],, repetitive efferent volleys cause an increment of [K+], in the intermediate gray matter of the frog spinal cord. In Ringer solution containing no added Mg2+, the elicited rises in [K+], produced by supramaximal stimuli (25 Hz, 10 s tetanus) exceeded the resting level by 2.3 + 0.5 mM (n = 10). In l.OmM Mg2+, the increment of [KC], produced by repetitive afferent stimulation was reduced to 63 f 7% (n = 10) of the increase in solutions without Mg2+ (not illustrated). DISCUSSION
Amino acid transmitters and the two components of the dorsal root potential
The present observations support the supposition that the DRP evoked in the frog cord by DR stimulation consists of at least two different components. (1) An early component lasting 15&200 ms putatively produced by the synaptic release of GABA and the subsequent activation of GABA, receptors; in this regard, PAD in the frog spinal cord has been shown to be accompanied by a large conductance increase in primary afferent fibers;@ (2) a very longduration component presumably largely accounted for by an increase of [K+],.4,34*44,48 The finding that a concentration of Mg2+ sufficient to block chemical synaptic transmission2’ reduced both the early and the late components of the DRP by about 95% indicates that both components depend almost completely upon intact synaptic transmission.38~48 GABA receptors of both the GABA, and GABA, type are reported to be present on afferent terminals.‘3s42Both receptors may be activated by GABA released as a result of afferent volleys.” Indeed, the potent GABA, receptor antagonist 2-hydroxysaclofen increased the duration of the DRP, an indication that the late component of the DRP may be influenced by GABA, receptors. Our results also show that the non-selective excitatory amino acid antagonist kynurenate30 substantially reduced the DRP. Similar results in the in vitro rat cord have recently been reported.24 These data suggest that L-glutamate (or a related amino acid) is
66
J. C. HACKMAWand R. A. DAVIWFF
utilized by circuits responsible for generation of the DRP. However, it is unclear from our data whether or not the putative excitatory amino acid transmitter is released by afferent fibers and/or interneurons. Excitatory synaptic transmission not mediated by amino acids may also make a small contribution to the DRP since 20 mM Mg2+ further decreased DRPs already attenuated by kynurenate. In addition, the finding that bicuculline could reduce the small component of the DRP remaining after block of excitatory amino acid-mediated transmission by kynurenate may indicate that a small DRP component depends upon GABA, but not upon activation of excitatory amino acid receptors. We assume that the small DRP remaining after block of both GABA and excitatory amino acid receptors depends upon non-synaptic mechanisms. Both NMDA and non-NMDA receptors are implicated in spinal neurotransmission.i4 The present observations with the selective antagonists APV and CNQX indicate that both types of receptor are present on neurons involved in the generation of DRPs. In Mg2+-free medium, APV and CNQX26*5’ reduced the DRP by almost 70%. However, the differential effects of APV and CNQX on the early and late components of the DRP suggest that the early component of the DRP is produced by a series of steps that principally involve non-NMDA receptors and that the late component is mediated by both NMDA and non-NMDA receptors.24v4’ Alternatively, the NMDA receptor-mediated component of the DRP may be relayed through non-NMDA receptors at the synapses made by primary afferent terminals. Slow, picrotoxin-resistant DRPs may be produced either by elevated [K+], or by the actions of L-glutamate. 4*38 Findings that picrotoxin and bicuculline enhance the evoked release of spinal K’38,48,50 support the role of elevated [K+],. However, there is an intimate relationship between excitatory amino acids and the release of K+ by afferent volleys. For example, we have shown that the synaptic release of excitatory amino acids and the subsequent activation of both NMDA and non-NMDA receptors on interneurons is responsible for about 85% of the K+ released as a result of afferent stimulation.” Furthermore, L-glutamate increases [K+], in the spinal cord. 22,35,47 Although L-glutamate and other excitatory amino acids can depolarize afferent termina1s,2~9*22the depolarization of low threshold afferent fibers produced by excitatory amino acids has been postulated to be generated indirectly as a consequence of the release of K+ ions from depolarized neurons close to afferent fibers within the dorsal horn.22 In contrast, unmyelinated afferent fibers are selectively sensitive to kainate.23 There is also recent electrophysiological data indicating that NMDA receptors are present on the somata of rat afferent neurons3’ However, there is still insufficient information to determine if the bicuculline-
resistant, kynurenate-sensitive component of the DRP results directly from the release of K1 and the subsequent depolarization of afferent terminals or if it is a result of a direct action of L-glutamate on afferent terminals. The present results demonstrate that manipulations (e.g. application of APV, CNQX, kynurenate, or 1.OmM Mg*+) which depress the late, convulsantresistant component of the DRP presumably do so by blocking excitatory amino acid receptors. We have also shown that 1.OmM Mg2+ suppresses the evoked release of K+. It would therefore appear that intact excitatory amino acid-mediated synaptic transmission is necessary for the enhanced and prolonged, late, convulsant-resistant phase of the DRP and for the evoked release of K+. The late phase of the DRP seen after application of picrotoxin and bicuculline or superfusion with low [Cl-], medium may be the result of facilitated interneuron activity which presumably results from depression of GABA-mediated inhibition. Comparison of amphibian and mammalian dorsal root potentials
DRPs recorded with the sucrose gap technique from frog spinal cords bathed in 0 mM Mg2+ medium are considerably larger and longer than those recorded in vivo from the cat cord in the conventional manner with wire electrodes placed on two points on a DR. For example, in the cat spinal cord, DRPs evoked by stimulation of cutaneous and muscle nerves are reported to be about 200-500 NV in size and to last about 15&300ms.‘*~i9 In contrast, in the present experiments DRPs evoked by DR volleys in the frog spinal cord are sizable (up to 18 mV in amplitude) and long-lasting (up to 19 s in duration). The present observations show that the discrepancies in amplitude and duration of DRPs recorded in the in vitro frog spinal cord and in the in vivo cat spinal cord are dependent, at least in part. upon differences in amplifier coupling, recording techniques, temperature, and the concentration of Mg 2+ ions in the extracellular space in the two preparations. a.c.-coupled recording with a l-s time constant has been conventionally used to record DRPs.‘~ Such recording conditions can artifactually and substantially shorten the record by filtering out the slower components of the potential; such was the case in our a.c. recording experiments. In addition, recordings performed with surface wire electrodes measure only a fraction of the true potential difference between active and resting membrane of the presynaptic terminal because of shunting by low resistance extracellular paths; such recordings are thus of low amplitude. With sucrose gap recording, the flow of sucrose washes away most of the ions responsible for extracellular shunting of the current generated by the intraspinal portion of primary afferent fibers. This permits recording of stable, high resolution potential
Frog dorsal root potentials
changes which more closely approximate the size and duration of the true intracellular events underlying the DRP. Temperature is another significant variable. The present experiments were carried out at 18°C. Most other in vitro experiments using the frog cord have been performed at similar temperatures (12-18”C).L~3B~41~4s Most in oiuo experiments on the cat are carried out at 36-39°C. Raising the temperature of the frog cord substantially shortened the DRP. The mechanism for this change was not investigated, but it is known that cooling lowers the threshold for reflex discharges along DRs, a phenomenon presumably reflecting increased PAD.6 The effects Mg?+ ions
of a “physiological”
concentration
of
Mg*+ ions are not ordinarily added to frog Ringer solution, but frog cerebrospinal fluid contains the cation in a concentration of 0.92 mM.” We assume that the interstitial fluid surrounding frog neurons in uiuo contains Mgs+ in approximately that concentration as well. DRPs recorded from spinal cords bathed in medium containing 1.0 mM M$+ are smaller and shorter than the DRP recorded in medium which does not contain M$+ ions. The reduction of DRPs by 1.0 mM M$+ may be caused by one or more of the following mechanisms. Block of the ion channels associated with N-methylD-aspartate receptors. As discussed above, our data
indicate that release of L-glutamate and subsequent activation of excitatory amino acid receptors is necessary for the generation of DRPs. Exposure to micromolar concentrations of Mg*+ ions is sufficient to subject the NMDA receptor ion channels to voltage-dependent blockade.39 In the presence of a millimolar concentration of Mgz+ ions, the NMDA receptor-channel complex should be largely inactive;‘.& our experiments have shown that when 1.0 mM Mg*+ is added to the superfusate, the selective NMDA antagonist APV had limited effectiveness in attenuating the size and duration of the DRP. Because in uiuo interstitial fluid contains Mg*+ ions, one could presume that the NMDA receptor-channel complex is minimally involved in the generation of the DRP. Hence, activation of non-NMDA receptors is probably responsible for much, if not all, of the excitatory amino acid-mediated component of the DRP in 1.0 mM M&+ medium.” However, we
67
cannot dismiss the possibility that NMDA receptors can participate in DRPs during high-frequency synaptic transmission such as that which occurs with repetitive 6ring of aIferent fibers. Under such conditions, NMDA receptors are active.2s Presynaptic reduction of evoked transmitter release. Because Mg*+ ions in the interstitial space can
block the evoked release of transmitter at synaptic junctions,“@ it would be anticipated that changing the extracellular concentration of Mg*+ from 0 to 1.0 mM would reduce synaptically transmitted responses in the frog spinal cord. It would also be anticipated that M$+ ions’ depressant effect on presynaptic transmitter release would be reduced by elevation of the extracellular Ca2+ concentration.‘5v’6 In our experiments the effects on the DRP of raising the concentration of the two cations were as expected: raising Mgz+ to 1.0 mM reduced the DRP; the reduction was partly reversed by elevating the Ca*+ concentration in the superfusate. Decrease of GABA-induced depolarization of primary afferent fiber terminals. Our results show that
depolarizations of primary afferent fiber terminals produced by exogenous GABA were depressed by the presence of 1.0 mM Mgr+ in the Ringer solution.’ Although investigation of what caused the depression will have to await further investigation, we did note that in the presence of 1.O mM Mg*+ there was also a depression of the ability of K+ ions to depolarize afferent fibers. This indicates that the effect on GABA responses was not mediated by a selective action of Mg2+ on the GABA receptor or its ion channel!’ Decreased release of K+ by afferent uolleys. We
assume that the reduction of K+ release seen in 1.OmM Mg2+ in the present experiments results both from reduction of excitatory amino acid release from presynaptic terminals and postsynaptic effects of the cation on NMDA receptor cation channels.” Reduction of neuron excitability. High concentrations of Mg 2+ have a depressant effect on central neurons.33 This is presumably caused by an action on voltage-sensitive membrane to elevate electrical threshold. However, 1.0 mM Mgz+ is reported to be insufficient to produce an increase in the threshold of frog spinal neurons.*’ work was supported by USPHS grant NS 17577 and Veteran’s Administration Medical Center Funds (MRIS 1769 and 3369). Acknowledgements-This
REFERENCES
1. Ault B., Evans R. H., Francis A. A., Oakes D. J. and Watkins J. C. (1980) Selective depression of excitatory amino acid induced depolarizations by magnesium ions in isolated spinal cord preparations. J. Physiol., Lond. 307,413428. 2. Barker J. L. and Nicoll R. A. (1972) Gamma-aminobutyric acid: role in primary afferent depolarization. Science 176, 1043-1045. 3. Barker J. L. and Nicoll R. A. (1973) The pharmacology and ionic dependency of amino acid responses in the frog spinal cord. J. Phyxiol., Land. 228, 259-277. 4. Barker J. L., Niwll R. A. and Padjen A. (1975) Studies on convulsants in the isolated frog spinal cord. II. Effects on root potentials. J. Physiol., Land. 245, 537-548. 5. Barron D. H. and Matthews B. H. C. (1938) The interpretation of potential changes in the spinal cord. J. Physiof., Lond. 92, 276321.
68 6.
J. C.
HACKMAN and
R. A.
DAVIDOFF
Brooks C. M., Koizumi K. and Malcolm J. L. (1955), Effects of changes in temnerature on reactions of somai cord.
J. Neurophysiol. 18, 2055216.
7. Curtis D. R. and Gynther B. D. (1987) Divalent cations reduce depolarization of primary afferent terminations by GABA. Brain Res. 422, 192-195. 8. Czeh G. and Somjen G. G. (1989) Changes in extracellular calcium and magnesium and synaptic transmission in isolated mouse spinal cord. Brain Res. 486, 274-285. 9. Davidoff R. A. (1972) Gamma-aminobutyric acid antagonism and presynaptic inhibition. Science 175, 331-333. 10. Davidoff R. A. and Hackman J. C. (1985) GABA: presynaptic actions. In Neurotransmitter Acfions in the Vertebrale Neruous System (ed. Rogawski M. A. and Barker J. L.), pp. 3-32. Plenum Press, New York, 11. Davidoff R. A., Hackman J. C., Holohean A. M., Vega J. L. and Zhang D. X. (1988) Primary afferent activity, putative excitatory transmitters and extracellular potassium levels in frog spinal cord. J. Physiol., Land. 397, 291-306. 12. Davidoff R. A., Hackman J. C. and Osorio I. (1980) Amino acid antagonists do not block the depolarizing effects of potassium ions on frog primary afferents. Neuroscience 5, 117-126. 13. Davidoff R. A. and Sears E. S. (1974) The effects of Lioresal on synaptic activity in the isolated spinal cord. Neurolog) 24, 957-963. 14. Davies J. D. and Watkins J. C. (1983) Role of excitatory amino acid receptors in mono- and polysynaptic excitation in the spinal cord. Expl Brain Res. 49, 280-290. 15. Del Castillo J. and Engbaek L. (1954) The nature of the neuromuscular block produced by magnesium. J. Phy,riol.. Lond. 124, 370-384. 16. Del Castillo J. and Katz B. (1954) The effect of magnesium on the activity of motor nerve endings. J. Physiol., Land. 124, 553-559. 17. Eccles J. C. (1964) Presynaptic inhibition in the spinal cord. Prog. Brain Res. 12, 65-89. 18. Eccles J. C. and Kmjevic K. (1959) Potential changes recorded inside primary afferent fibres within the spinal cord. J. Physiol., Lond. 149, 250-273. 19. Eccles J. C., Magni F. and Willis W. D. (1962) Depolarization muscle. J. Physiol., Lond. 160, 62-93.
of central terminals of Group I afferent fibres from
20. Eccles J. C., Schmidt R. F. and Willis W. D. (1963) Pharmacological studies on presynaptic inhibition. J. Physiol., Land. 168, 500-530. 21. Erulkar S. D., Dambach G. E. and Mender D. (1974) The effect of magnesium at motoneurons of the isolated spinal cord of the frog. Brain Res. 66, 413424. 22. Evans R. H. (1980) Evidence supporting the indirect depolarization of primary afferent terminals in the frog by excitatory amino acids. J. Physiol., Lond. 298, 25-35. 23. Evans R. H. (1985) Kainate sensitivity of C-fibres in rat dorsal roots. J. Physiol., Lond. 358, 42P. 24. Evans R. H. and Long S. K. (1989) Primary afferent depolarization in the rat spinal cord is mediated by pathways utilising NMDA and non-NMDA receptors. Neurosci. Left. 100, 231-236. 25. Fatt P. (1954) Biophysics of junctional transmission. Physiol. Rev. 34, 674-710. 26. Fletcher E. J., Martin D., Aram J. A., Lodge D. and Honor6 T. (1988) Quinoxalinediones selectively block quisqualate and kainate receptors and synaptic events in rat neocortex and hippocampus and frog spinal cord in vitro. Br. J. Pharmac. 95, 585-597. 27. Hackman J. C. and Davidoff R. A. (1988) Mg*+, excitatory amino acids, and dorsal root potentials in the frog spinal cord. Sot. Neurosci. Abstr. 14, 789. 28. Herron C. E., Lester R. A. J., Coan E. J. and Collingridge G. L. (1986) Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic m&ha&m. Nature 3&I, 265-268.
29. Honor& T.. Davies S. N.. Dreier J.. Fletcher E. J.. Jacobsen P., Lodge D. and Neilsen F. E. (1988) Quinoxalinediones: potent competitive non-NMDA glutamate receptor antagonists. S&nce 241, 701-703. . _ 30. Jahr C. E. and Yoshioka K. (1986) la aBerent excitation of motoneurones in the in vitro new-born spinal cord is selectively antagonized by kynurenate. J. Physiol., Lond. 370, 515-530. 31. JimCnez I., Rudomin P. and Solodkin M. (1987) Mechanisms involved in the depolarization of cutaneous afferents produced by segmental and descending inputs in the cat spinal cord. Expl Brainkes. 69, 195-207. 32. Katz B. and Miledi R. (1963) A study of miniature potentials in spinal motoneurones. J. Physiol., Land. 168,389-422. 33. Kelly J. S.. Kmjevic K. and Somien G. (1969) Divalent cations and electrical properties of cortical cells. J. Neurobiol. 2, 1977208. 34. Kmjevic K. and Morris M. E. (1975) Correlation between extracellular focal potentials and K+ potentials evoked by primary afferent activity. Can. J. Physiol. Pharmac. 53, 912-922. 35. Kudo Y. and Fukoda H. (1976) Alteration in extracellular K+-activity induced by ammo acids in frog spinal cord. Jap. J. Pharmac. 26, 385-387. 36. Levy R. A. (1977) The role of GABA in primary afferent depolarization. Prog. Neurobiol. 9, 21 l-267. 37. Lovinger D. M. and Weight F. F. (1988) Glutamate induces a depolarization of adult rat dorsal root ganglion neurons
that is mediated by NMDA receptors. Neurosci. Lert. 94, 314-320. 38. Nicoll R. A. (1979) Dorsal root potentials and changes in extracellular potassium in the spinal cord of the frog. J. Physiol., Lond. 290, 113-127. 39. Nowak L., Bregstovski P., Ascher P., Herbet A. and Prochiantz A. (1984) Magnesium gates glutamate-activated channels in mouse central neurons. Nature 307, 462-465. 40. Padjen A. L. and Hashiguchi T. (1983) Primary afferent depolarization in frog spinal cord is associated with an increase in membrane conductance. Can. J. Physiol. Pharmac. 61, 626-631. 41. Padjen A. L. and Smith P. A. (1980) Specific effects of a-D,I_-aminoadipic acid on synaptic transmission in frog spinal cord. Can. J. Physiol. Pharmac. 58, 692-698. 42. Price G., Wilkin G. P., Turnbull M. J. and Bowery N. G. (1984) Are baclofen sensitive GABA, receptors present on primary afferent terminals of the spinal cord? Nature 307, 71-74. 43. Schmidt R. F. (1963) Pharmacological studies on the primary afferent depolarization of the toad spinal cord. PjzZgers Arch. ges. Physiol. 277, 325-346. 44. Shefner S. A. and Levy R. A. (1981) The contribution of increases in extracellular potassium to primary afferent depolarization in the bullfrog spinal cord. Brain Res. 205, 321-335.
Frog dorsal root potentials
69
45. Shirasawa Y. and Koketsu K. (1975) The effect of picrotoxin on the dorsal root potential of bullfrog spinal cords. Kurume Med. .I. 22, 14. 46. Smith P. A. (1982) The use of low concentrations of divalent cations to demonstrate a role for N-methyl-o-aspartate receptors in synaptic transmission in amphibian spinal cord. Br. J. Pharmac. 77, 363-373. 47. Sonnhof U. and Buhrle Ch. Ph. (1980) On the postsynaptic action of glutamate in frog spinal motoneurones. Pftigers Arch. ges. Physiol. 338, 101-109. 48. Sykova E. and Vyklicky L. (1977) Changes of extracellular potassium activity in isolated spinal cord of frog under high Mgz+ concentration. Neurosci. LQU. 4, 161-165. 49. Sykova E. and Vyklicky L. (1978) Effects of picrotoxin on potassium accumulation and dorsal root potentials in the
frog spinal cord. Neuroscience 3, 1061-1067. 50. ten Bruggencate G., Lux H. D. and Liebl L. (1974) Possible relationships between extracellular potassium activity and presynaptic inhibition in the spinal cord of the cat. PJtigers Arch. gei. Physiol. 349, 301-317.51. Wheatley P. L. and Collins K. J. (1986) Quantitative studies of N-methyl-D-aspartate, 2-amino-S-phosphonovalerate and cis-2,3-piperidine dicarboxylate interactions on the neonatal rat spinal cord in vitro. Eur. J. Pharmac. 121,257-263. (Accepted 12 September
1990)
