We compared the contribution of metabotropic glutamate receptors (mGluRs) to the generation and modulation of synaptic responses elicited in intracellularly recorded L5 motoneurons from neonatal rats by segmental and descending fibers. Dorsal root (DR) stimulation at high intensity (C-fiber strength) evoked long latency (2-5-s) depolarization in addition to early monosynaptic and polysynaptic responses. Stimulation of the descending ventrolateral funiculus (VLF) failed to evoke a late response in the same motoneuron. The mGluR antagonist (+)-α-methyl-4-carboxyphenylglycine (MCPG; 0.4 mM) selectively blocked the long latency DR response. This mGluR-mediated response persisted in α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate or N-methyl-d-aspartate (NMDA) antagonists, but not both, suggesting that glutamate transmission (either AMPA/kainate or NMDA) is required for mGluR-mediated inputs from small diameter sensory afferents to affect the motoneuron. Although MCPG inhibited the long latency DR response, it induced moderate facilitation of monosynaptic DR and VLF responses. The mGluR agonist 1s3r-ACPD induced motoneuron depolarization and depressed the monosynaptic DR and VLF responses. MCPG also facilitated the neurotrophin-3 and brain-derived neurotrophic factor induced strengthening of the monosynaptic DR responses (but only before P6, since neurotrophins are ineffective later at DR synapses and never at VLF synapses after birth). Our results suggest that mGluRs are involved in synaptic pathways to motoneurons made by DR but not VLF fibers. MCPG-induced facilitation of monosynaptic AMPA/kainate DR and VLF responses suggests the possibility of tonic mGluR-mediated inhibition of DR and VLF responses. We speculate that MCPG facilitates neurotrophin-induced strengthening of monosynaptic DR responses by reducing this tonic inhibition.
Lumbar motoneurons in neonatal rat spinal cord receive excitatory synaptic inputs from two major pathways: a segmental pathway that can be activated by electrical stimulation of the corresponding dorsal root (DR) and a descending pathway activated by stimulation of ventrolateral funiculus (VLF) axons (Pinco and Lev-Tov, 1994). Both these pathways elicit monosynaptic EPSPs via AMPA/kainate receptors. They also activate NMDA receptors, but this action is more complex since the properties of NMDA receptors on motoneurons change markedly during early postnatal development (Arvanian and Mendell, 2001a; Arvanian et al., 2004a) becoming more subject to Mg2+ block at resting membrane potential. Of considerable interest is the fact that the NMDA receptors associated with VLF inputs mature earlier than those associated with DR inputs on the same motoneuron (Arvanian et al., 2004a).
A third group of receptors activated by glutamate is the metabotropic class of glutamate receptors (for review, see Schoepp and Conn, 2002). Previous investigators have demonstrated a metabotropic component of DR-evoked responses in motoneurons (Jane et al., 1994, 1995; Thompson et al., 1995; King and Liu, 1996; Cao et al., 1997; Marchetti et al., 2003). However, there is no evidence concerning the metabotropic components produced by VLF stimulation. In this report, we demonstrate that VLF differs from DR in not eliciting a late component in motoneurons subject to mGluR antagonists. However, both the VLF and DR AMPA/kainate EPSPs are modulated by activating mGluR receptor components. Furthermore, we show that mGluR responsiveness of motoneurons requires function in either NMDA or AMPA/kainate receptors, but not both.
A final question considered here concerns the role of mGluR receptors in determining the ability of neurotrophins to acutely modulate synaptic responsiveness of motoneurons. In previous work, we demonstrated that the AMPA/kainate response was acutely enhanced by NT-3 (Arvanov and Mendell, 2000) and BDNF (with later inhibitory action by BDNF; Arvanian and Mendell, 2001b) and that this effect required the function of NMDA receptors in the motoneuron (Arvanian and Mendell, 2001a,b). Here, we demonstrate that mGluR receptors can also modulate the immediate effect of neurotrophins on the synaptic response of motoneurons.
Materials and Methods
Preparation of Hemisected Spinal Cord. Experiments were performed in vitro on hemisected lumbar spinal cord preparations from neonatal (P1-5, referred to as 1-week and P8-12, referred to as 2-week) Sprague-Dawley rats in accordance with protocols approved by the Institutional Animal Care and Use Committee at SUNY-Stony Brook. The spinal cord preparation and intracellular recording procedure have been described in detail previously (Arvanov et al., 2000; Arvanian et al., 2003). The youngest neonatal rats, postnatal days 1 to 3, were anesthetized via hypothermia, whereas older neonates were anesthetized in halothane, and all were decapitated. The spinal cord was removed and hemisected, and a section of the left hemicord (spanning segments from approximately T1 to S3) was transferred into the recording chamber and superfused (10 ml/min) with artificial cerebrospinal fluid (ACSF) containing 117 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 2.0 mM MgSO4, 25 mM NaHCO3, 1.2 mM NaH2PO4, and 11 mM dextrose, aerated with 95% O2, 5% CO2 (pH 7.4; 30°C). In experiments with reduced concentrations of Mg2+ in saline, corresponding equiosmolar changes in Na+ concentration were made. The L5 DR and ventral root as well as the cut VLF dissected from the spinal cord at T2 (Pinco and Lev-Tov, 1994) were placed in three different suction electrodes with silver-silver chloride internal wires connected to a Pulsemaster A300 (WPI, Sarasota, FL) via isolation units for stimulation.
Electrophysiology. Intracellular recordings were made from L5 lumbar spinal motoneurons with 70- to 110-MΩ microelectrodes, filled with 3 M potassium acetate, using an Axoprobe 1A (Axon Instruments Inc., Union City, CA) amplifier. Motoneurons were identified by their antidromic response to ventral root stimulation, and motoneuron input resistance was estimated by passing current pulses (100 ms) through the intracellular recording electrode as described previously (Fulton and Walton 1986; Arvanov et al., 2000). All cells displayed a resting membrane potential of -64 to -70 mV.
In this study, both DR and VLF responses were evoked at two intensities, monosynaptic A-fiber strength and high intensity C-fiber strength (Thompson et al., 1993; Arvanian et al., 2003, 2004a). The stimulus evoking the maximal monosynaptic potential was delivered at a rate of 0.1 Hz, 50-μs duration, and intensity of 60 to 120 μA (DR) and 80 to 150 μA (VLF). Ten responses of the motoneuron to these stimuli were averaged (pClamp 9; Axon Instruments Inc.). The high-intensity stimulus was 500 μs in duration and at 500-μA intensity for both DR and VLF; a stimulation rate of 0.01 Hz was chosen to prevent changes in either DR or VLF responses elicited by successive stimuli.
Data Analysis. Results are presented as mean values ± S.E.M. The significance of the difference between mean values was evaluated by t tests. The paired t test was used to compare mean values of EPSP amplitudes measured before and after drug administration in the same cell. In all cases, Bonferroni's correction was used for multiple comparisons.
Drugs. Most of the reagents used were obtained from Sigma-Aldrich (St. Louis, MO) and included the following: (+)-α-methyl-4-carboxyphenylglycine (MCPG), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), d-2-amino-5-phosphonovaleric acid (d-APV), strychnine, and (-)-bicuculline methchloride. 1S,3R-1-Aminocyclopentane-1,3-dicarboxylic acid (1S,3R-ACPD) and CGP46381 were purchased from Tocris Cookson Inc. (Ballwin, MO). Neurotrophins NT-3 and BDNF were received courtesy of Regeneron Pharmaceuticals, Inc. (Tarrytown, NY).
Contribution of mGluRs in Generation of DR- and VLF-Evoked Responses. To study the possible contribution of mGluRs to the synaptic response, the effects of the broad mGluR antagonist MCPG (Eaton et al., 1993) on the responses evoked by the low- and high-intensity stimulation of the DR and VLF in the same motoneuron were studied. In motoneurons recorded from animals younger than 1 week, stimulation of DR with increasing stimulus intensity (50 μs, 60-120 μA, maximal for monosynaptic component) induced the appearance of a short latency monosynaptic response (peak at 5-10 ms; Fig. 1A, DR, trace 1). The monosynaptic component is attributed to volleys in A-fibers (Thompson et al., 1993). Increasing the stimulus intensity up to about 200 μA (50-μs stimulus duration) did not affect the peak amplitude of the monosynaptic DR response. However, with stimuli of higher intensity and longer duration (about 250μA/200 μs) action potentials were noted at the peak of the monosynaptic component, and a long-lasting depolarization was observed in all cells studied (Fig. 1A, DR, compare traces 1 and 2; n = 27). This late response reached a maximum amplitude averaging 6.3 ± 2.1 mV (n = 27) at a stimulus intensity/duration of 300 μA/500 μs and is attributed to activation of C-fibers (Thompson et al., 1993). In this study, we used a stimulus of 500 μA/500 μs to examine the property of this late DR response (Fig. 1A, trace 2).
Stimulation of VLF with increasing stimulus intensity induced responses that peaked at 10 to 20 ms and required a stimulus intensity/duration of 80 to 150 μA/50 μs for the maximal VLF response (Fig. 1A, VLF, trace 1). As with the response evoked by high-intensity stimulation of DR, high-intensity (500 μA/500 μs) stimulation of VLF induced the appearance of action potentials on top of the VLF responses (Fig. 1A, VLF, trace 2). However, high-intensity stimulation of VLF never evoked the late response characteristic of high-intensity DR stimulation in the same motoneuron (n = 17; Fig. 2). VLF responses were lacking in the late component, even at an extremely high stimulus intensity of 900 μA/500 μs (n = 5).
MCPG (400 μM) completely and reversibly blocked the late component of the DR response (n = 6; Fig. 1A, DR, traces 3 and 4) with no effect on the initial response, including the action potentials associated with the early component. This suggests that mGluRs selectively mediate the late DR response. MCPG had no effect on VLF responses evoked by high-intensity stimulation (Figs. 1A, VLF, traces 2-4; and 2) or on motoneuron resting membrane potential (65.9 ± 1.3 mV in control and 65.4 ± 0.9 mV in MCPG; n = 22; p > 0.1) or input resistance (48.5 ± 3.6 MΩ in control and 49.1 ± 3.1 MΩ in MCPG; n = 5; p > 0.1).
As reported by others (Thompson et al., 1993), we noted an increase in synaptic noise associated with the late response to DR but not VLF stimulation; this synaptic noise was substantially diminished by MCPG when studied in 2 mM Mg2+ (Fig. 1, A and C; also see Fig. 3). In 0.5 mM Mg2+, where transmitter release was enhanced, we also saw spontaneous synaptic events (Fig. 1B), but these were not blocked by MCPG (not illustrated). These were not studied in detail.
A possible explanation for the lack of MCPG-sensitive late VLF EPSPs is that these responses might only occur in conjunction with NMDA receptor-mediated transmission. NMDA receptors in motoneurons from neonatal rats younger than 1 week are activated by DR but not VLF stimulation in 2 mM Mg2+ (Arvanov et al., 2000, 2004a). To evaluate this hypothesis, we performed experiments in motoneurons from rats younger than 1 week in reduced [Mg2+]o solution (0.5 mM), where NMDA components are evident at both DR and VLF connections at the resting membrane potential due to partial abolition of the Mg blockade of NMDA receptors (Arvanian and Mendell, 2001a; Arvanian et al., 2004a). In addition, we performed parallel experiments in motoneurons from 2-week-old rats in 2 mM Mg2+ ACSF, when NMDA components are absent at both DR and VLF connections due to Mg2+ block (Arvanian and Mendell, 2001a; Arvanian et al., 2004a). As shown in Fig. 1B, the high-intensity stimulus evoked the MCPG-sensitive late component at DR (5.1 ± 1.9 mV; n = 5), but not VLF connections (n = 5), in motoneurons from rats younger than 1 week measured in 0.5 mM Mg2+ in ACSF (where Mg2+ block is reduced and NMDA receptors are active at both DR and VLF connections). In motoneurons from rats older than 1 week studied in 2 mM Mg2+ ACSF (where NMDA receptors are blocked by Mg2+ at both DR and VLF connections), the high-intensity stimulus still evoked the late DR response (5.3 ± 1.7 mV; n = 6) and no late VLF responses (n = 6). Finally, we found that the MCPG-sensitive late component (4.3 ± 1.2 mV; n = 5) was clearly present in solutions containing the NMDA receptor antagonist d-APV (Fig. 3A1, trace 2; and 3A2, trace 4), which invariably depressed the earlier NMDAR-mediated components. Thus, NMDA receptor function is not required for the late MCPG component to be evoked.
To examine whether activation of mGluRs alone might mediate generation of the late DR component, we studied responses evoked by high-intensity stimulation of DR under conditions where all known excitatory and inhibitory receptors except mGluRs were blocked pharmacologically. In ACSF containing the AMPA/kainate antagonist CNQX, the NMDA antagonist d-APV, the GABAA antagonist bicuculline, the GABAB antagonist CGP46381, and the glycine receptor antagonist strychnine, high-intensity stimulation of DR failed to evoke any response in motoneurons (Fig. 3B, trace 1; n = 5). However, when the NMDA antagonist was removed from the cocktail, the high-intensity stimulation of DR evoked a response that consisted of early (latency 5-15 ms) and late (peak at 3-4 s) components (Fig. 3B, trace 2). Administration of MCPG under these conditions selectively blocked the late component with no effect on the early NMDA response in the presence of the non-NMDA antagonists cocktail (n = 5; Fig. 3B, trace 3). In the same motoneuron, high-intensity stimulation of VLF in the presence of non-NMDA antagonists cocktail failed to evoke a response (n = 5; not shown), which is consistent with previous results (Arvanian and Mendell, 2001a; Arvanian et al., 2004a).
mGluR-Induced Modulation of Monosynaptic DR and VLF Responses. We first examined the effect of the broad mGluR antagonist MCPG on the monosynaptic DR and VLF responses. In these experiments, we used a just maximal (A-fiber strength; see Materials and Methods) stimulation intensity to restrict the responses to subthreshold levels, i.e., no action potentials (see above). Surprisingly, administration of 400 μM MCPG alone induced moderate but consistent facilitation of the monosynaptic responses to both DR (119.7 ± 3.6%; n = 11; p < 0.05) and VLF (115.6 ± 3.3%; n = 11; p < 0.05) stimulation (Figs. 4A and 6). The amplitude of DR and VLF responses declined to pre-MCPG values 30 min after MCPG wash (Figs. 4A and 6A; n = 5).
Since lumbar motoneurons receive inhibitory inputs (Pinco and Lev-Tov, 1993, 1994; Peshori et al., 1998; Vinay and Clarac, 1999), and some of these inputs might interact with mGluRs (Marchetti et al., 2003), we examined whether blockade of inhibitory inputs affected the MCPG-induced facilitation of DR and VLF responses. As shown in Fig. 4B, administration of a cocktail of GABAA, GABAB, and glycine receptor antagonists induced a marked facilitation of the amplitude of both DR (195.1 ± 20%; n = 5; p < 0.05) and VLF (157.2 ± 13.1%; n = 5; p < 0.05) responses. The addition of MCPG to this cocktail further facilitated both DR (128.6 ± 9.9%; n = 5; p < 0.05) and VLF (118.9 ± 6.3%; n = 5; p < 0.05) responses (Fig. 4B). These results indicate that the MCPG-induced facilitation of DR and VLF responses does not depend on activity of the inhibitory inputs.
We then compared the effects of the mGluR agonist 13,3r-ACPD on short latency DR and VLF responses. ACPD (25 μM) induced motoneuron depolarization (12.3 ± 3.7 mV; n = 9) and the firing of action potentials (Fig. 5A). Consistent with previous reports (King and Liu, 1996), ACPD-induced depolarization was associated with enhanced frequency of spontaneous synaptic potentials and an increase in motoneuron input resistance (56.9 ± 4.3 MΩ in control and 89.4 ± 5.1 MΩ in ACPD; n = 5; p < 0.05). As shown previously with the mGluR1 agonist DHPG (Marchetti et al., 2003), addition of 0.5 μM tetrodotoxin, 40 μM APV, and 10 μM CNQX to restrict the agonist action to the motoneuron reduced the effect of ACPD and blocked the synaptic noise, but some depolarization was still observed (3.4 ± 1.1 mV; n = 4; not illustrated). This suggests that mGluR receptors are located both on motoneurons as well as on cells presynaptic to the motoneuron. ACPD-induced depolarization was transient with both membrane potential (n = 9) and input resistance (n = 5) recovering to pre-ACPD levels within 8 to 13 min in the presence of ACPD.
ACPD also induced depression of both DR and VLF monosynaptic responses. DR and VLF responses were measured in the presence of ACPD after recovery of the membrane potential to pre-ACPD levels, to avoid an effect of membrane depolarization on synaptic responses (Fig. 5, A, D, and E). Under these conditions, ACPD decreased the peak amplitude of monosynaptic DR responses to 26.9 ± 4% and VLF responses to 41.3 ± 6% (n = 9) of control values. These effects of ACPD were reversible, and the magnitude of both DR and VLF responses recovered after about 30 min of ACPD wash (Figs. 5B and 6). A second administration ACPD after a period of 30- to 45-min wash of the first application induced a depolarization and depression of DR and VLF responses comparable with the first application (n = 3; not shown).
We examined the ability of MCPG to block the action of ACPD in six cells. In four cells, MCPG was added 30 min after wash of the first administration of ACPD (Fig. 5, C-E), and in two cells MCPG was added before the first ACPD application (not shown). In all cells studied, MCPG (400 μM) alone induced facilitation of both DR and VLF EPSPs (Fig. 5; as described above) and completely blocked the ACPD-induced motoneuron depolarization and depression of DR and VLF responses (Figs. 5 and 6).
MCPG Enhances Facilitatory Action of Neurotrophins NT-3 and BDNF on Monosynaptic DR-Evoked Responses in Motoneuron from Neonatal 1-Week-Old Rats. We examined whether block of mGluR receptors by MCPG affects neurotrophin-induced long-lasting modulation of synaptic transmission. Thus, we compared the effects of NT-3 and BDNF on DR- and VLF-evoked responses in the presence and absence of MCPG. As described previously, addition of 200 ng/ml NT-3 for 10 min induced facilitation of the monosynaptic DR responses, and the peak amplitude remained elevated for at least 1 h after NT-3 wash (Arvanov et al., 2000). Although MCPG alone induced facilitation of about 19% in the DR responses (see above), the NT-3-induced facilitation of monosynaptic DR responses in MCPG-containing solution was significantly (p < 0.05) greater than that induced by NT-3 in the absence of MCPG (Fig. 7A). Interestingly, DR responses remained facilitated for at least 1 h at the MCPG + NT-3 level, even after wash of both NT-3 and MCPG (Fig. 7A). VLF responses were not facilitated by NT-3, and the addition of MCPG was without effect on the action of neurotrophins (Fig. 7B). In the presence of MCPG, the facilitatory action of BDNF was enhanced but the later inhibitory action was eliminated (Fig. 7C). VLF responses were not affected by BDNF, with or without MCPG (Fig. 7D). In 2-week-old animals (P8-P12), DR responses were not affected by NT-3 (n = 3) or BDNF (n = 3; not shown).
We have confirmed that the very long latency response in motoneurons evoked by C-fiber stimulation of DR is mediated by mGluR receptors. Volleys of similar intensity delivered to VLF did not evoke a late mGluR-mediated response. We cannot be sure of the location of all the mGluR receptors responsible for these long-latency MCPG-inhibited responses from DR since various subtypes of mGluR receptors are located throughout the dorsal horn as well as on motoneurons (Anneser et al., 1999; Berthele et al., 1999; Alvarez et al., 2000; Carlton et al., 2001; Valerio et al., 2002), and activity in dorsal horn interneurons affects motoneurons (Thompson et al., 1993). Although VLF did not elicit an mGluR-mediated response, at least some mGluR receptors associated with the DR response may be associated with the motoneuron itself (as indicated by ACPD-induced depolarization of the motoneuron in the presence of tetrodotoxin, APV, and CNQX) since as we have seen previously for NMDA receptors (Arvanian and Mendell, 2001a; Mendell et al., 2001; Arvanian et al., 2004a), synaptic receptors activated by DR are not necessarily activated by VLF fibers terminating on the same motoneuron.
The finding that no mGluR response was observed when both AMPA/kainate and NMDA receptors were blocked suggests that the monosynaptic input to motoneurons does not activate mGluRs. This indicates that the mGluR-mediated response is multisynaptic. Presumably, either AMPA/kainate or NMDA receptors alone are required to ensure transmission through this multisynaptic pathway from DR to motoneurons, with the late component being driven additively by mGluR receptors. The enhanced frequency of synaptic noise observed in association with the late MCPG-sensitive response elicited by DR, as well as the abolition of this noise by MCPG, suggests (but does not prove) that some mGluRs responsible for the long-latency input were located in cells presynaptic to the motoneuron.
Our results indicate an additional role for mGluR receptors in controlling synaptic transmission to motoneurons. We found that inhibiting mGluR receptors with MCPG increased the amplitude of the monosynaptic EPSP in motoneurons, whereas ACPD reduced its amplitude. Of considerable interest was the finding that the monosynaptic inputs to motoneurons from VLF as well as DR was modified by modulating mGluR transmission (Figs. 5 and 6), unlike the long-latency mGluR-mediated depolarization that was associated only with DR (Fig. 1). This suggests a different network of mGluR neurons than that responsible for the long-latency DR-activated response of motoneurons. Since these MCPG-induced changes took place in the absence of changes in motoneuron properties, we speculate that they were the result of presynaptic changes on fibers mediating the monosynaptic response from both VLF and DR to motoneurons. We cannot specify the location of the mGluR receptors, only that they are involved in the network responsible for this action. The fact that MCPG acted to enhance the response is taken as evidence that this presynaptic inhibitory system is tonically active and that administration of MCPG inhibits it leading to disinhibition of the monosynaptic input.
Our interpretation of the enhanced potentiation of the monosynaptic EPSP by NT-3 in the presence of MCPG is that it was subject to less tonic inhibition and therefore could be facilitated to a greater extent by the NT-3. However, NT-3 was unable to potentiate VLF responses after treatment with MCPG, presumably because VLF still did not activate NMDA receptors known to be necessary for NT-3 to have its effect (Arvanian and Mendell, 2001a). The same explanation applies to the inability for MCPG to facilitate NT-3 potentiation of DR-EPSPs in neonates older than 1 week. The effect of MCPG on the actions elicited by BDNF in rats younger than P6 is more difficult to interpret. The enhancement of facilitation probably has the same interpretation as that advanced to explain its effects on the response to NT-3. However, the decrease in inhibition is not readily explained. Probably the simplest explanation is that the enhanced facilitation overcomes the later inhibitory effect. However, we cannot rule out the possibility that there is a direct effect of MCPG on the interneurons responsible for the inhibitory effect of BDNF. These have been suggested to act via axo-axonal synapses leading to presynaptic inhibition (Arvanian and Mendell, 2001b). The mechanism responsible for the enhancement of neurotrophin-induced facilitation of synaptic transmission by MCPG requires further studies.
Metabotropic receptors are thought to differ from ionotropic receptors in being directly coupled to second messengers (Schoepp and Conn, 2002) and therefore being able to change the properties of affected cells over long periods. The effects of ACPD lasted for at least several minutes. Interestingly, its effect on motoneuron depolarization seemed to be self-limiting, whereas its inhibitory effect on transmission from DR and VLF in the same experiment persisted considerably longer. This implies that different second messenger systems are activated or that activation of a given system has different outcomes in different cells. Such considerations will require further experimentation, probably involving selective agonists for mGluR1-8 subtypes and intracellular recordings from identified populations of interneurons.
We thank Alyssa Tuthill for technical support. We also thank the Statistical Consulting Unit at SUNY-Stony Brook for assistance and Regeneron Pharmaceuticals Inc. for a gift of NT-3 and BDNF.
- Received July 29, 2004.
- Accepted September 21, 2004.
This study was supported by National Institutes of Health Grants 2RO1 NS16996 and 5PO1 NS39420, the Christopher Reeve Paralysis Foundation (to L.M.M.), and the New York State Spinal Cord Injury Foundation (to V.L.A.).
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: DR, dorsal root; VLF, ventrolateral funiculus; EPSP, excitatory postsynaptic potential; AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; NMDA, N-methyl-d-aspartate; mGluR, metabotropic glutamate receptor; ACSF, artificial cerebrospinal fluid; MCPG, (+)-α-methyl-4-carboxyphenylglycine; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; d-APV, d-2-amino-5-phosphonovaleric acid; 1s3r-ACPD, 1S, 3R-1-aminocyclopentane-1,3-dicarboxylic acid; NT-3, neurotrophin-3; BDNF, brain-derived neurotrophic factor; CGP46381, (3-aminopropyl) (cyclohexylmethyl)phosphinic acid.
- The American Society for Pharmacology and Experimental Therapeutics