Currents in Guinea Pig CA3 Pyramidal Cells

The Journal of Neuroscience, September 1994. 14(g): 55806589

Kp Opioid Receptors Inhibit NMDA Receptor-mediated Synaptic Currents in Guinea Pig CA3 Pyramidal Cells

Robert M. Caudle,’ Charles Chavkin,* and Ronald Dubner’ ‘Neurobiology and Anesthesiology Branch, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892 and *Department of Pharmacology, University of Washington,‘Seattle, Washington 98195

The role of the endogenous opioid peptide dynorphin (l- 17) dynorphin’s concentration in the spinal cord increasesfollowing in regulating NMDA receptor-mediated synaptic currents was peripheral tissue and nerve injury (Iadarola et al., 1988; Nahin examined in guinea pig hippocampus. Schaffer collateral/ et al., 1989; Kajander et al., 1990), and spinal cord injury (Faden commissural fiber-evoked NMDA synaptic currents were re- et al., 1985), and that its concentration in the hippocampus corded using whole-cell patch-clamp techniques in CA3 py- increases(Lee et al., 1987) and decreases(McGinty et al., 1986) ramidal cells. Dynorphin was found to have dual effects on following seizures.These studies of endogenousdynorphin have NMDA synaptic currents, increasing currents at low concen- also provided little information concerning dynorphin’s role in trations and decreasing currents at high concentrations. Only the CNS. In our present course of investigation, we have nar- the inhibitory action of dynorphin was sensitive to naloxone, rowed our studiesto structures that have dynorphin-containing indicating that this effect was mediated by an opioid recep- afferents that were demonstrated to release dynorphin under tor. The inhibitory effect was mimicked by bremazocine, but physiological conditions. The logic behind this approach is that not by U69,593, U50,466, [D-Ala*, KMe-Phe4, Gly-oll-en- an effect of exogenousdynorphin should be demonstrable for kephalin, or [D-PerV]-enkephalin. Bremazocine’s effect was endogenousdynorphin. Thus, if exogenousand endogenousdy- blocked by naloxone, but not by nor-binaltorphimine, cypro- norphin produce similar effects it is likely that a physiological dime, or naltrindole. These findings suggest that bremazo- function for dynorphin was demonstrated. tine’s effect was mediated by the K2 subtype of opioid re- The hippocampus of rodents contains a substantial concen- ceptor. In addition, 1 AM naloxone and antisera to dynorphin tration of dynorphin. This dynorphin resides primarily in the (l-1 7) were found to increase NMDA-mediated synaptic cur- mossy fiber pathway that extends from the dentate granule cells rents. Nor-binaltorphimine, cyprodime, naltrindole, and an- to the CA3 pyramidal cells (Gall et al., 1981; McGinty et al., tisera to met-enkephalin did not increase the NMDA synaptic 1983; McLean et al., 1987). The mossy fibers serve as the major current. These findings suggest that endogenous dynorphin excitatory input to the CA3 region. Previous attempts to dem- was acting at K2 receptors to inhibit NMDA receptor-medi- onstrate effectsof dynorphin and other K-S&&T opioids in the ated synaptic currents. Overall, these findings indicate that CA3 region of the rodent hippocampuswere either unsuccessful dynorphin is an endogenous agonist for K* receptors in the (Caudleand Chavkin, 1990) or ambiguous(Moises and Walker, CA3 region of the guinea pig hippocampus and that these 1985). On the other hand, we demonstrated that K agonists, receptors regulate NMDA receptor function. including dynorphin, inhibit excitatory transmitter releasefrom [Key words: opioids, dynorphin, kappa receptors, N-me- the perforant path onto dentate granule cells in guinea pig hip- thy/-D-aspartate receptors (NMDA), whole cell patch clamp, pocampus(Wagner et al., 1992,1993). Other groupshave found hippocampus, guinea pig, CA31 that dynorphin, through K receptors, inhibits calcium currents and calcium action potentials in cultured dorsal root ganglion The opioid peptide dynorphin is thought to be the endogenous cells(Werz and MacDonald, 1985; MacDonald and Werz, 1986; ligand for K-opioid receptors(Chavkin et al., 1982). This peptide Gross and MacDonald, 1987). Therefore, we proposedthat the has been examined in numerous models in order to determine inhibition of calcium currents was likely to be the mechanism its functional role in the CNS. The effects produced by exoge- for K receptor-mediated inhibition of excitatory amino acid re- nously applied dynorphin vary from neurotoxicity (Caudle and leasefrom the perforant path (Wagner et al., 1992, 1993). Isaac, 1987) to inhibition of calcium potentials (Werz and Mac- Using the K,-SekCtiVe ligand 3H-U69,593 for autoradiogra- Donald, 1985; MacDonald and Werz, 1986; Gross and Mac- phy, we found that virtually all of the specific binding was lo- Donald, 1987). These differing actions of exogenousdynorphin cated in the molecular layer of the guinea pig dentate gyrus have not provided a clear answerto dynorphin’s function in the (Wagner et al., 1991, 1992). This pattern of binding by 3H- CNS. Adding to the confusion are reports demonstrating that U69,593 explained the lack of effect of this compound in the CA3 region of the guinea pig hippocampus in our physiological studies(Caudle and Chavkin, 1990). However, Weisskopf et al. Received Sept. 13, 1993; revised Mar. 17, 1994; accepted Mar. 24, 1994. (1993) reported that U69,593 had inhibitory effectsin the guinea We thank Drs. Gene Williams and Ke Ren for their comments and advice in pig CA3. Given the lack of binding sites for U69,593 in this the preparation ofthis manuscript. C.C. is supported by U.S. Public Health Service region of the guinea pig hippocampus,Weisskopf et al.‘s findings Grant DA04 123. Correspondence should be addressed to R. M. Caudle, NAB, NIDR, NIH, are difficult to interpret. Building 49/Roam IA-1 1, 9000 Rockville Pike, Bethesda, MD 20892. Recently, evidence was presentedindicating the existence of Copyright 0 1994 Society for Neuroscience 0270-6474/94/145580-10$05.00/O at least two subtypes of K-opioid receptors (Neck et al., 1988, The Journal of Neuroscience, September 1994, M(9) 5581

1990, 1993; Zukin et al., 1988; Rothman et al., 1992). U69,593 Stimulating Electrode selectively binds to one of these sites and this site has been labeled the K, receptor. Compounds like ethylketocyclazocine and bremazocine, in the presence of p- and &opioid receptor blockers, bind to the U69,593 binding site and to another site with approximately equal affinity. The U69,593-insensitive site has been labeled the K2 receptor (Zukin et al., 1988; Rothman et al., 1992) or the t receptor (Neck et al., 1990, 1993). Auto- radiographic studies of K-Opioid binding sites using either 3H- ethylketocyclazocine or 3H-bremazocine in guinea pig hippocampus resulted in binding to both the molecular layer of the dentate gyrus and the stratum lucidum of the CA3 (McLean et al., 1986; Zukin et al., 1988). The binding in the CA3 contrasts with ‘H-U69,593 binding (Wagner et al., 199 1, 1992), and suggests that these sites are K2 or t receptors. The close proximity of the K2 binding sites in the guinea pig / CA3 to the terminals of the dynorphin-containing mossy fibers, Kappas Receptors as compared to the K, binding sites in the dentate gyrus, suggests that dynorphin may be an endogenous ligand for this receptor. Figure 1. Diagrammaticrepresentation of a guineapig hippocampal If endogenous dynorphin acts at K2 receptors, the question then slice.The cellthat the patchelectrode is connected to isa CA3 pyramidal cell.The stimulatingelectrode is placedin the Schaffercollateral/com- becomes what physiological function do K2 receptors perform missuralpathway. The gray shadedarea representsthe location of NMDA and how can we measure that function? Several studies have receptorsin the CA3 regionof the guineapig hippocampus.The hatched suggested that dynorphin may play a role in the regulation of urea representsthe locationof U69,593-insensitivebremazocine bind- NMDA receptors (Caudle and Isaac, 1988a,b; Bakshi and Fa- ing sites(K2 receptors)in the guineapig hippocampus. den, 1990a,b; Faden et al., 1990; Isaac et al., 1990; Shukla et al., 1992; Skilling et al., 1992). In these studies, selective an- ceptor physiology and dynorphin’s effectson K2 receptors, it was tagonists to the NMDA receptor complex blocked the effects of necessaryto dissectK, , K2, and nonopioid effectsfrom eachother. exogenously applied dynorphin. The results of these experi- To avoid the influence of K, receptors, the CA3 region of the ments suggest that dynorphin may potentiate NMDA receptor guinea pig hippocampus was chosen for study. As previously function. However, several other studies have demonstrated mentioned, this region of the guinea pig hippocampuslacks 3H- that dynorphin has inhibitory actions in the CNS. In addition U69,593,binding sites(Wagner et al., 1992) and is innervated to the K, inhibitory effects, several investigators have found neu- by a dynorphinergic fiber bundle (Gall et al., 198 1; McGinty et roprotective (Baskin et al., 1984; Garant et al., 1985), inhibitory al., 1983; McLean et al., 1987). Using whole-cell voltage-clamp (Stevens et al., 1987; Stewart and Isaac, 199 l), and both excit- techniques, we demonstrate that K2 receptors inhibit NMDA- atory and inhibitory effects of dynorphin (Moises and Walker, mediated synaptic currents in the CA3 region of the guinea pig 1985; Knox and Dickenson, 1987; Hylden et al., 199 1) that may hippocampus, and evidence is presentedthat endogenousdy- or may not be mediated by K-opioid receptors. The neuropro- norphin acts at thesereceptors. tective effects of dynorphin (Baskin et al., 1984; Garant et al., 1985) are similar to those of NMDA receptor antagonists (Choi Materials and Methods et al., 1988), suggesting that dynorphin may act, in the neuro- Electrophysiology. Hippocampalslices were prepared from 150-300gm protection experiments, by inhibiting the NMDA receptor com- maleHartley guineapigs as previously described (Caudle and Chavkin, 1990).Briefly, the guineapigs were anesthetized with pentobarbital(50 plex. mg/kg,i.p.) andthen decapitated.The brainswere removed and chilled In addition to dynorphin’s actions at opioid receptors, it was in ice-coldKreb’s bicarbonate buffer ofthe followingcomposition (mM): found to have nonopioid actions as well (Faden and Jacobs, NaCl(124),KC1 (4.9), KH,PO, (1.2),MgSO, (2.4), CaCl, (2.5) NaHCO, 1984; Moises and Walker, 1985). Supporting this finding are (25.6),and glucose(10). The brainswere then slicedon a vibratome receptor binding studies by Massardier and Hunt (1989) and (500pm). The hippocampalslices were dissected from the restof the tissueand placedin a tissuechamber. The sliceswere then superfused Shukla et al. (1992) demonstrating that dynorphin inhibits bind- (1 ml/min) with Kreb’sbicarbonate buffer. The temperaturewas main- ing of 3H-glutamate and ‘H-MK-801 to NMDA binding sites tainedat 34°Cand the bufferwas bubbled with 95% 0,/5% CO,. The in membranes. This effect of dynorphin was not sensitive to sliceswere allowed to equilibratein the chamberat 34°Cfor a minimum opioid antagonists, implying that the dynorphin was not acting of 1 hr beforeexperiments were started to ensurethat residualpento- barbitalwas washed out of the tissue.For recordingNMDA receptor- at an opioid receptor. The inhibition of the binding of these mediatedcurrents, the superfusionbuffer was changed to Kreb’sbicar- agents at NMDA receptors by dynorphin, rather than the en- bonatebuffer with nominallyzero magnesium, 10 PM6-cyano-7-nitro- hancement of binding, is not entirely consistent with the excit- quinoxaline-2,3-dione(CNQX), and 20 PM bicuculline.A concentric atory effects of dynorphin, but is compatible with the neuro- bipolarstimulating electrode (SNE 100,Rhodes Medical Instruments, protective actions of dynorphin (Baskin et al., 1984; Garant et WoodlandHills, CA) wasplaced in the stratumlacunosum moleculare (Schaffercollaterals) of the hippocampalslice, approximately at the al., 1985). However, until physiological studies are performed CAl/CA3 border(Fia. 1). Stimuliconsisted of sinale0.3 msecsquare the interpretation of the binding studies remains difficult. waveswith currentsr&g&g from 200to 400aA. Patch-clampelectrodes If K, receptors mediate inhibition of the voltage-sensitive cal- werepulled to resistancesof between 2 and 10MO. The electrodeswere cium currents, then perhaps K2 opioid receptors and the “non- filledwith (mM) CsCl(120), tetraethylammonium chloride (20), CaCl, (l), MgCl, (2), EGTA (lo), HEPES(lo), ATP (4), and GTP (0.5), and opioid dynorphin receptors” mediate the excitatory and inhib- the pH wasadjusted to 7.4 with CsOH.All datawas collected using an itory actions of dynorphin that are thought to be mediated Axopatch200 (Axon Instruments,Foster City, CA). The data wasdig- through the NMDA receptor complex. In order to study K2 re- itized and recordedon a personalcomputer for future analysis. 5582 Caudle et al. * K* Opioid Receptors in CA3 A.

CNQX + BIC + APV

Figure 2. A, Representative Schaffer collateral/commissural fiber-evoked NMDA receptor-mediated current in a guinea pig CA3 pyramidal cell. The current was measured in the absence of magnesium and in the presence of the non-NMDA excitatory amino acid antagonist CNQX (10 PM) and the GABA antagonist bicuculline (20 PM). In addition, the electrode contained cesium and tetraethylammonium chloride to CNQX + BIC block potassium channels. The holding potential while recording synaptic currents was + 5 mV. Depolarizing the cells inactivated voltage-gated sodium and calcium currents and reduced the am- "1 plitude of the NMDA-mediated synaptic current. As illustrated by the trac- 250 es, the current was blocked by the selective NMDA antagonist APV (50 PM). When the antagonist was washed out of the bath the current returned. 200 Each trace represents an average of five sweeps. The total length of the sweep was 960 msec. B, Dynorphin (l-l 7) 150 concentration-response relationship on NMDA receptor-mediated currents. The area over the curve was calculated loo for traces similar to those in A to determine the amount of charge (current multiplied by time) that entered the cells during the recordings. The charge entering the cell in the presence of dynorphin was then compared to the charge prior to dynorphin (Control) to determine the percent of control value. Each point represents between four and nine cells. Error bars represent SEM. rJkorPhid x.M

Experimentalprocedure. Guinea pig CA3 pyramidal cells were voltage calcium-mediated injury to the cells. The Schaffer collateral/commis- clamped and held at -65 mV. When the cell had stabilized the Kreb’s sural pathway was then stimulated and the synaptic current was re- bicarbonate buffer superfusing the slices was changed to buffer that was corded. Recording began 60 msec prior to the stimulation and continued the same in all respects except that it contained nominally zero mag- for 900 msec after the stimulus. The stimulus intensity was adjusted to nesium, 10 PM CNQX, and 20 PM bicuculline. Bicucullinewas used to achieve the maximal synaptic current with the lowest possible stimulus, block GABA,-mediated currents and CNOX was used to block non- generally between 200 and 400 PA. The rate of stimulation never ex- NMDA excitatory amino acid-mediated currents. The electrode con- ceeded one every 30 set in order to prevent long-term potentiation. In tained tetraethylammonium chloride and cesium to block both GABA, addition, the synaptic current had to be stable for 15 min prior to the and voltage-sensitive potassium currents. The Schaffer collateral/com- addition of any drugs to the bathing solution. missural pathway was chosen for stimulation because the mossy fiber/ Data recorded from the NMDA-mediated synaptic currents are pre- CA3 pyramidal cell synapse lacks NMDA receptor-mediated currents sented as charge entering the cell (current multiplied by time). Charge (Griffith. 1990) (see Fie. 1). Prior to stimulatine the Schaffer collateral/ is used, rather than peak current, because the amount of charge entering commissural pathwaythe’cell was stepped to-+5 mV and held there the cell determines how far the cell will be depolarized by the opening for 6 sec. The cells were depolarized for two reasons. First, depolariza- of the channels; the more charge entering the cell, the more depolarized tion was necessary to inactivate voltage-gated sodium and calcium chan- the cell would become if the cell were not voltage clamped. Thus, the nels that, if left active, seriously contaminate the NMDA receptor- total charge entering the cell represents a more reliable measure of mediated synaptic current. The complete inactivation of these currents channel activity than peak current. was evidenced by the holding current becoming steady after a few sec- Statistical analysis of data consisted of unpaired t tests, paired t tests, onds. Second, the cells were depolarized to reduce the amplitude of the and ANOVAs when appropriate; p < 0.05 was used as the criteria for NMDA receptor-mediated current. The peak NMDA synaptic currents significance. when the cells were held at + 5 mV were between 100 pA and 2 nA. At Drugs. Dynorphin (1- 17) was purchased from Sigma (St. Louis, MO) -65 mV the peak currents were typically in excess of 2 nA. The smaller and Peninsula Laboratories (Belmont, CA). Bicuculline methiodide, currents reduce the influx of calcium and decrease the likelihood of ATP, GTP, and naloxone were purchased from Sigma (St. Louis, MO). The Journal of Neuroscience, September 1994, 14(9) 5583

Rabbit antisera to met-enkephalin, rabbit antisera to dynorphin (l-17), and [n-Ala*, N-Me-Phe4, Gly-ol]-enkephalin (DAMGO) were pur- CNQX + SIC + I ,A, Naloxone

chased from Peninsula Laboratories (Belmont, CA). U69,593, U50,488, SPA bremazocine, nor-binaltorphimine (NorBNI), 2-amino-S-phosphonovalerate (APV), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), [D- Pen2J]-enkephalin (DPDPE), cyprodime, and naltrindole were purchased from Research Biochemicals Inc. (Natick, MA). All drugs were CNQX+BIC+lpMNd +1oOnMD,m dissolved and applied in the superfusion buffer. All drugs were superfused over the hippocampal slices for at least 10 min before measurements of synaptic responses were made.

Results ChQX + BIC + 5 pMDywrphin(l-17) CNQX + BIC + 1 @4 Naloxone Guinea pig CA3 pyramidal cells were voltage clamped as de- SPA scribed in Materials and Methods, and NMDA receptor-me- I diated synaptic currents were recorded. The average charge (current multiplied by time) moving into the pyramidal cells as a result of a maximal stimulus to the Schaffer collateral/commissural pathway was (51.3 + 8) x 10-l’ C (mean f SE, N = 21). This current was blocked by the selective NMDA antagonist 2-amino-5-phosphonovalerate (APV) (Fig. 2A; N = 3) and was stable for several hours in most cells. After collecting baselinecurrents, various concentrations of dynorphin (1- 17) were superfusedin the buffer and the effect of dynorphin on the NMDA-mediated current was determined. As illustrated by Figure 2B, dynorphin at concentrations be- SpMDyn+lpMNal * tween 10 nM and 1 FM increasedthe NMDA-mediated current. However, at a concentration of 5 PM dynorphin, the highest concentration tested, the current was inhibited. The holding current for the step to +5 mV was not altered by dynorphin: 209 f 60 pA before the addition of dynorphin and 218 & 70 pA in the presenceof 5 PM dynorphin (mean f SE, N = 5). The similar holding currents indicate the changein synaptic current wasnot likely to be due to alterations in membraneconductance. If 1 PM naloxone wasincluded in the buffer prior to and during the addition of dynorphin, 5 PM dynorphin increasedthe current Figure 3. Antagonismof dynorphin(1- 17)‘s effects by naloxone.Data (Fig. 3), and againthe holding current at + 5 mV wasnot altered werecollected as described in Figure2 for 5 PM dynorphinand 100nM by the drugs (199 f 110 pA, N = 3). The combination of 1 PM dynorphinin the presenceand absenceof 1 FM naloxone.The traces at naloxone and 5 I.LM dynorphin not only increasedthe NMDA- the top are sweepsfrom representativecells for eachcondition. Each mediated current, but also reducedthe survival time of the cells trace representsthe averageof five sweeps.The control responsesfor the dynorphin+naloxone-treatedcells were determined in the presence to lessthan 30 min following the addition of thesedrugs to the of 1 PM naloxone.Paired t testswere usedon the raw chargedata to bath (N = 3). In our hands, most pyramidal cells that were not comparecontrol currentsto currentsin the presenceof dynorphin.An exposed to this combination of drugs survived until the elec- asterisk indicatesthat the dynorphincurrent was significantly different trode was manually pulled from them. Although this observa- from its control (p < 0.05).Bars represent the meansof threeto nine tion was not quantified, the toxicity of overstimulated NMDA cells.Error barsrepresent SEM. receptors is well documented (Choi et al., 1988) and any enhancement of their activity is expected to be associatedwith To determine what opioid receptor subtype mediated the in- increasedcell mortality. One likely consequenceof this toxicity hibitory effect of dynorphin on synaptically mediated NMDA was that the current measuredin the presenceof 1 PM naloxone currents, a seriesof opioid agonistswere tested. The selective and 5 FM dynorphin was underestimateddue to the cells being K, agonistsU69,593 (1 PM ) and U50,488 (500 nM) (Neck et al., damagedprior to recording the synaptic currents. Naloxone had 1988, 1990, 1993; Zukin et al., 1988; Rothman et al., 1992) no effect on the excitatory actions of lower concentrations of had no effect on the NMDA-mediated current (Fig. 4). This dynorphin (Fig. 3). finding is consistentwith the lack Of K, binding sitesin the CA3 The biphasic nature of the dose-responserelationship indi- region of the guineapig hippocampus(Wagner et al., 1992). The cated that dynorphin (1- 17) had both an excitatory and an in- selective F-opioid receptor agonist [D-Ala*, N-Me-Phe4, Gly- hibitory action on NMDA currents. These effectscould be sep- ol]-enkephalin (DAMGO, 1 KM) increasedthe NMDA current, arated from each other by the concentration of dynorphin and as did the selective &opioid receptor agonist [D-Per+]-enkeph- by sensitivity to naloxone. The inhibitory action of dynorphin alin (DPDPE, 1 KM) (Fig. 4). The viability of the cells did not was sensitive to naloxone, suggestingthat this may be an effect appear to be compromised by the excitatory actions of DAM- at K-Opioid receptors. The excitatory effect of dynorphin was GO. However, the cells were clearly injured by 1 I.~M DPDPE not sensitive to naloxone, indicating that this effect was nono- in a manner similar to when the 1 PM naloxone/5 PM dynorphin pioid in nature and may be the result of dynorphin binding combination was applied. Becauseof the injury to the cells, the directly to the NMDA receptor complex (seeMassardier and increasein the NMDA-mediated current as a result of DPDPE Hunt, 1989; Shukla et al., 1992). The nonopioid effect was not was likely to be underestimated. Findings similar to these for studied further. DAMGO were reported previously (Chen and Huang, 199 1). 5584 Caudle et al. * 4 Opioid Receptors in CA3

1 @VI U69,593

1 pM Bremamcine

1 @I DPDPE

1 pMDAMG0

1 @I Bran + 1 @I Naloxone *

1 W Bran + 1 pW NorBNl

11&l Bran + 1 piI¶ Cyprodime

1 fl Bran + 1 fl Naltrindole

Figure 4. Effectof variousopioid agonistson NMDA receptor-mediatedcurrents. Data werecollected as described in Figure2. Pairedt Figure 5. A, Bremazocineconcentration-response relationships. Data testswere performed on the rawcharge data to compareNMDA currents were collectedas previouslydescribed for various concentrationsof without the opioidagonist to the currentsin the presenceof the opioid. bremazocine.The experimentswere then performedin the presenceof Barsrepresent means from threeto six cells.An asterisk indicatesthat 100 nM naloxone.Each point representsbetween three and six cells. the currentin the presenceof the opioid wassignificantly different from Error barsrepresent SEM. The lineswere significantlydifferent (p < its control(p < 0.05).Error bars represent SEM. The traces at the bottom 0.001,ANOVA). B, Antagonismof bremazocine’sinhibition of NMDA- are from a representativecell and demonstratethe inhibition of the mediatedsynaptic currents with opioidantagonists. Data were collected NMDA-mediatedcurrent by bremazocine.Each trace representsthe aspreviously described. For the experimentswith the antagonists,the averageof five sweeps. antagonistwas included in the bufferduring the control measurements. Eachbar representsthe meanof threeto six cells.Asterisk indicatesp < 0.05 whencompared to 1 PM bremazocineby unpairedt test.Error The relatively nonselective K agonistsbremazocine (1 PM) and barsrepresent SEM. 5 FM dynorphin (1- 17) inhibited the NMDA-mediated synaptic current (Fig. 4). The holding current at +5 mV wasnot affected by bremazocine (102 f 27 pA before bremazocine, 82 rfr 16 pA in the presenceof bremazocine, mean f SE, N = 5). The 7 to 10.5 nM (Rothman et al., 1992; Neck et al., 1993; Webster inhibitory effects of bremazocine and dynorphin and the lack et al., 1993). of inhibitory effects of U69,593, U50,488, DAMGO, and The inhibitory effects of 1 I.~M bremazocine were completely DPDPE suggestedthat the receptor mediating the inhibition of blocked by 1 PM naloxone, but not by the K,-SekCtiVe antagonist the NMDA synaptic current was likely to be the Kq opioid re- nor-binaltorphimine (NorBNI) (Fig. 5B). This finding is con- ceptor (seeNeck et al., 1990, 1993). sistent with binding studies indicating that NorBNI has little To determine whether bremazocinewas acting as a selective affinity for K2 binding sites(Neck et al., 1988, 1990, 1993; Zukin K2 agonistin this assay,concentration-response relationships in et al., 1988; Rothman et al., 1992). In addition, the p-selective the presenceand absenceof 100 nM naloxone were determined. antagonist cyprodime (Schmidhammer et al., 1990) and the As illustrated by Figure 5A, 100 nM naloxone shifted brema- S-selectiveantagonist naltrindole (Contreras et al., 1993) also zocine’s concentration-responserelationship to the right. Using had no effect on the inhibitory actions of bremazocine(Fig. 5s). the equation Ki = [antagonist]/(EC,,‘/EC,, - l), where EC,,’ is The results of theseopioid antagonist studiesand the previous the EC,, of bremazocinein the presenceof naloxone (Kosterlitz opioid agonist studiesindicate that bremazocine acts as a se- and Watt, 1968), a K, of 7.5 nM was calculated for naloxone lective K2 agonist in this assay. from the data in Figure 5A. This finding is consistent with Ki One observation that was made early in theseinvestigations values reported for naloxone at K2 binding sitesthat range from was that 1 PM naloxone produced an increasein the NMDA- The Journal of Neuroscience, September 1994, 74(9) 5585 mediated synaptic current (Fig. 6A). The holding current at + 5 mV was not altered by naloxone: 186 + 26 pA for cells not exposed to naloxone (mean f SE, N = 21) and 176 + 74 pA for cells in the presence of 1 PM naloxone (N = 9). The opioid antagonists NorBNI, cyprodime, and naltrindole did not influence the NMDA current. The currents were 95 f 3%, 136 + 20%, and 93 f 10% of control, respectively. Because of the Ao‘I? I- excitatory effect of naloxone, all of the antagonist experiments discussed above compared the currents measured in the presence of antagonist to currents recorded in the presence of both the antagonist and agonist. The excitatory effect of naloxone suggested that an endogenous agonist may be acting on the K~ receptors. As described in Materials and Methods, all recordings of NMDA synaptic currents were made in nominally zero magnesium. This low-magnesium buffer results in substantial spontaneous activity in the hippocampal slice; thus, it was likely that endogenous peptides were released continuously during the ex- periment. Because the mossy fibers contain dynorphin and ter- minate near the K2 receptors, and because we found dynorphin to be an agonist for K2 receptors, we hypothesized that dynorphin was the endogenous agonist. To test this hypothesis, rabbit an- Control 1 @I Naloxone tisera to dynorphin (l- 17) or antisera to met-enkephalin were added to the super-fusion buffer at a dilution of 1: 1000. Figure B. 6B shows that the antisera to dynorphin increased the NMDA- mediated current, while the antisera to met-enkephalin did not. The increase in the NMDA synaptic current was similar to that produced by naloxone, suggesting that dynorphin, or a dynorphin-like peptide, was being released by the spontaneous activity in the slice and that it acted upon K2 receptors. Previous work demonstrated that K, receptors inhibited excitatory amino acid release presynaptically (Wagner et al., 1992, 1993). The data presented above are entirely consistent with a presynaptic site for the K2 receptors; therefore, it was important to determine if excitatory amino acid release was altered by K2 receptor activation. To examine this question, the NMDA receptor antagonist APV (50 PM) was substituted for the non- Anti-Met-Enkephalin Anti-Dynorphin NMDA excitatory amino acid receptor antagonist CNQX in the Figure 6. A, Effectof naloxoneon NMDA-mediated synaptic currents. superfusion buffer. This substitution allowed the measurement NMDA-mediated currents were collected in the absence of any opioid of non-NMDA excitatory amino acid synaptic currents using antagonist (Control) (IV = 21) and in the presence of 1 PM naloxone (N the same protocol as was used for NMDA synaptic currents. = 9). Representative traces are presented at the top. Each trace represents the average of five sweeps. The total charge entering the cells during the Presumably, NMDA and non-NMDA excitatory amino acid recording was then calculated and the means plotted. Asterisk indicates receptors at the Schaffer collateral/CA3 synapse receive excit- p < 0.05 when compared to control by unpaired t test. Error bars atory amino acids from the same axon terminals. Thus, if the represent SEM. B, Effect of antisera to opioid peptides on NMDA- synaptic currents exhibit different pharmacologies, the effect mediated synaptic current. Synaptic currents were collected in the absence and presence of either antisera to met-enkephalin (N = 4) or most probably occurs postsynaptically rather than through a antisera to dynorphin (1-17) (N = 5). Paired t tests were performed on change in the release of excitatory amino acids presynaptically. the raw charge data comparing the charge before and after the addition Figure 7A illustrates a typical non-NMDA excitatory amino of the antisera to the bathing solution. Asterisk indicatesp < 0.05. Error acid-mediated current and its sensitivity to CNQX. The average bars represent SEM. charge entering the cell at + 5 mV following stimulation of the Schaffer collateraVcommissura1 fibers was (11.2 f 4) x lo-” C (mean + SE, N = 9). As Figure 7B demonstrates, 5 PM dynorphin (1- 17) had no effect on the non-NMDA excitatory amino acid synaptic current, while 1 ILM naloxone inhibited the Discussion current. The lack of effect of dynorphin and the inhibitory effect Previous studiesdemonstrated that dynorphin had both opioid of naloxone on non-NMDA-mediated synaptic currents indi- and nonopioid effects (Faden and Jacobs, 1984; Moises and cated that alterations in the release of excitatory amino acids Walker, 1985) that were difficult to separate.These opioid and were not responsible for the effects of these agents on NMDA- nonopioid effects were again demonstratedin this study; how- mediated synaptic currents. Therefore, we concluded that the ever, we were able to separatethese effects by altering the con- K2 receptors were likely to be situated postsynaptic to the Schaffer centration of dynorphin and by the useof naloxone. Our major collateral terminals. The K2 receptors may reside on the CA3 finding was that a receptor that bears similar pharmacology to pyramidal cells; however, further work will be needed to confirm that reported for the K,/t-opioid binding site (Neck et al., 1988, this hypothesis. 1990, 1993; Zukin et al., 1988; Rothman et al., 1992) inhibits 5566 Caudle et al. l cc2Opioid Receptors in CA3 A.

AF’V+ BIC + CNQX

50 pA

A.PV+ BIC

AFT’ + BIC (Washout CNQX)

Figure 7. A, Representative Schaffer collateral/commissural fiber-evoked non-NMDA excitatory amino acid- mediated synaptic current. The same protocol was used as for NMDA-mediated currents described in Figure 2A except that the non-NMDA excitatory amino acid antagonist CNQX was re- placed by 50 PM APV. The traces demonstrate that 20 PM CNQX blocks the current and that upon removal of CNQX from the superfusion solution the current returns. Traces represent the average of five sweeps. The length of the trace was 960 msec. B, Effect of naloxone (N = 5) and 5 PM dynorphin (l- 17) (N = 4) on the non-NMDA excit- atdry amino acid-mediated current. Data were collected as previously described. Paired t tests were performed on the raw charge data comparing the responses in the absence and presence of drug. Asterisk indicates p < 0.05. Error bars represent SEM. 1 pM Naloxone 5~Dynorp~

NMDA receptor-mediated synaptic currents. The data sup- et al., 1988; Rothman et al., 1992). However, the nonopioid porting this claim are threefold. First, the only opioid agonists excitatory effect of dynorphin probably antagonized the inhib- tested that inhibit the NMDA receptor-mediated synaptic cur- itory opioid effect, resulting in a reduction in dynorphin’s po- rent are bremazocine and dynorphin. The Kr-selective agonists tency. Second, the inhibitory effect of bremazocine on the U69,593 and U50,488 did not influence the NMDA receptor- NMDA-mediated current was blocked only by the nonselective mediated current. The selective p agonist DAMGO and the opioid antagonist naloxone. The K,-selective antagonist Nor- selective 6 agonist DPDPE enhanced rather than inhibited the BNI, the p-selective antagonist cyprodime, and the b-selective current. Concentrations of these opioid agonists above 1 I.LM antagonist naltrindole had no effect on bremazocine’s inhibition were not tested in this study because in previous work we found of the NMDA-mediated current. This pattern of antagonism is that 1 I.LM of these agents usually produces a maximum effect consistent with the binding selectivities of these agents for the (Caudle and Chavkin, 1990; Wagner et al., 199 1, 1992, 1993). K2 binding site (Neck et al., 1988, 1990, 1993; Zukin et al., 1988; In addition, opioid receptor selectivity for these agonists begins Rothman et al., 1992). The concentrations of the antagonists to break down at concentrations above 1 /IM. This pattern of used in this study were well above those previously shown to agonist selectivity is consistent with binding studies describing be effective at blocking opioid agonists (Caudle and Chavkin, theK2receptor(Nocketal., 1988, 1990, 1993;Zukinetal., 1988; 1990; Wagner et al., 1991, 1992, 1993). Also, opioid receptor Rothman et al., 1992). In addition, the relative potencies of selectivity of the antagonists breaks down at concentrations above bremazocine and dynorphin at inhibiting the NMDA current 1 PM. In addition, the selective antagonists had no effect on the (bremazocine > dynorphin) are consistent with their binding NMDA-mediated current when superfused over the slices alone, affinities for theK2 receptor (Nocket al., 1988, 1990,1993; Zukin indicating that the inhibitory effect of bremazocine and dynor- The Journal of Neuroscience, September 1994, f4(9) 5557

phin was not due to antagonism at p- or d-opioid receptors. centrations dynorphin enhanced the current. Whether or not Finally, the approximately loo-fold shift to the right of the endogenous dynorphin acted at the nonopioid site is difficult to bremazocine concentration-response relationship by 100 nM determine from the present data. It is possible that dynorphin naloxone indicates that naloxone’s K, for this receptor is ap- was released from the mossy fibers directly onto the K2 receptors proximately 7.5 nM. Recent binding studies have reported na- and that the dynorphin did not reach the nonopioid sites, which loxone’s K, for the K~ binding site to be in the range of 7-10.5 are presumably on the NMDA receptor complex (see Massardier nM (Rothman et al., 1992; Neck et al., 1993; Webster et al., and Hunt, 1989; Shukla et al., 1992) at some distance from the 1993). Based on these findings, we concluded that the inhibitory mossy fibers (Monaghan and Cotman, 1985) (see Fig. 1). This effect of bremazocine and dynorphin was mediated through the scenario would suggest that the excitatory effect observed in the K* opioid receptor. In preliminary studies 10 nM bremazocine presence of dynorphin antisera and naloxone was simply the was found to have no effect on NMDA receptor-mediated syn- loss of tonic K2 receptor inhibition without stimulation at the aptic currents. This contrasts with dynorphin’s excitatory effects nonopioid site. Another possibility is that endogenous dynor- at low concentrations and suggests that bremazocine may be a phin may bind to both sites, with its concentration being high selective agonist for the K2 receptor in this assay. Dynorphin, on enough for the opioid effect to overwhelm the nonopioid effect. the other hand, acts on at least two receptor systems in this In essence, endogenous dynorphin’s concentration was on the assay: an inhibitory opioid (KJ receptor and an as yet unchar- inhibitory side of the concentration-response relationship (Fig. acterized nonopioid excitatory receptor. 2B). Thus, the excitatory nonopioid effect predominated when At present, we do not know the mechanism of the K~ opioid dynorphin’s concentration was lowered by the antisera or when receptor inhibition of the NMDA receptor-mediated currents. the opioid receptors were blocked by naloxone. This second The K~receptors may be directly coupled to the NMDA receptor hypothesis seems less likely than the first because the excitatory complex. However, the K~ binding sites are located close to the nonopioid effect was clearly observed at low concentrations of terminals of the dynorphin-containing mossy fibers in the stra- exogenously applied dynorphin. If endogenous dynorphin was tum lucidum of the CA3 (McLean et al., 1986; Zukin et al., present at both receptors in sufficient concentration for the opioid 1988) whereas the NMDA receptors are located on the distal effect to predominate, as occurred when 5 KM dynorphin was dendrites (Monaghan and Cotman, 1985) (see Fig. 1). The dis- applied exogenously (Fig. 2B), the excitatory nonopioid effect tance between the K2 and the NMDA binding sites suggests that could not be observed without first blocking the opioid recep- some sort of diffusible, or transportable, second messenger is tors. Therefore, the most parsimonious explanation for our data utilized rather than a direct connection between the receptors. was that endogenous dynorphin, or a dynorphin-like peptide, Alternatively, activation of K2 receptors may alter the cable prop- was released by the mossy fibers onto adjacent K2 opioid recep- erties of the cell, increasing the electrotonic length. Our data tors, resulting in tonic inhibition of NMDA receptors while little suggest that this was not the case because membrane conduc- endogenous opioid reached the nonopioid site. We did not stim- tance was not altered by dynorphin or bremazocine. Also, non- ulate the mossy fibers in an attempt to release greater amounts NMDA currents were not affected in the same manner by these of endogenous dynorphin to determine if endogenous dynorphin agents as NMDA currents. If the cable properties of the CA3 acted at the nonopioid site. Until we understand what the non- pyramidal cells were altered by K2 receptor activation, we would opioid site is, and have a competitive antagonist for that site, expect that similar effects would be observed on both NMDA attempts to detect endogenous dynorphin’s effects on this re- and non-NMDA excitatory amino acid synaptic currents in receptor will be difficult. This task may be especially difficult in sponse to K~ opioid receptor activation. However, because the the presence of the K2 opioid receptor’s inhibitory actions. NMDA receptors and their associated ion channels are on the The results of this study and of previous studies with dynor- distal dendrites of the CA3 pyramidal cells, we could not be phin indicate that the role of endogenous dynorphin in the CNS certain of our voltage clamp in that region. Because it is unlikely is extremely complex. The results presented here indicate that that the dendrites were actually at +5 mV, it is possible that endogenous dynorphin can bind to K2 receptors to inhibit the unobserved changes in membrane properties or voltage-acti- current that flows through the NMDA receptor complex. This vated ion channels occurred in the distal dendrites, in response action provides a mechanism for mossy fiber regulation of a to K~agonists, that could be detected only by examining the large synapse that the fibers do not contact directly, the Schaffer col- NMDA receptor-mediated currents. Future studies will attempt lateral/commissural fiber CA3 pyramidal cell synapse. By re- to address this issue more thoroughly. leasing dynorphin on to K2 receptors in the stratum lucidum, the Data were also presented here suggesting that endogenous mossy fibers inhibit NMDA receptors in the distal dendrites. dynorphin acts at the K2 opioid receptor. Naloxone alone or We can only speculate as to the importance of this action of antibodies to dynorphin (1 - 17) added to the superfusion buffer endogenous dynorphin. However, it presumably would effect increased the NMDA-mediated current, suggesting that an en- the oscillatory nature of the CA3 pyramidal cells by dampening dogenous agonist, presumably dynorphin, was being displaced the recurrent circuitry. In other studies, we found that endog- from the receptor. The selective antagonists NorBNI, cypro- enous dynorphin-like peptides also bind to K, opioid receptors dime, and naltrindole as well as antisera to met-enkephalin did in the guinea pig dentate gyrus (Wagner et al., 1991) and that not influence the NMDA-mediated synaptic current. These find- the K, receptors inhibit excitatory amino acid release from the ings support the contention that endogenous dynorphin, or a perforant path (Wagner et al., 1992, 1993). This finding suggests dynorphin-like peptide, is an endogenous agonist for the K2 re- that endogenous dynorphin not only regulates NMDA receptors ceptor. in the CA3, but that it also regulates excitatory input to the cells In this study we demonstrated that exogenously applied dy- that contain the dynorphin: the dentate granule cells. Thus, norphin had a biphasic concentration-response relationship (Fig. through the two K-opioid receptors dynorphin provides a mech- 2B). At high concentrations exogenous dynorphin inhibited the anism for dentate granule cells to reduce the conduction of NMDA receptor-mediated synaptic current, and at low con- information through the hippocampus. 5588 Caudle et al. * K~ Opioid Receptors in CA3

MacDonald RL, Werz MA (1986) Dynorphin A decreases voltage- References dependent calcium conductance of mouse dorsal root ganglion neu- Bakshi R, Faden AI (1990a) Blockade of the glycine modulatory site rones. J Physiol (Lond) 377:237-249. of NMDA receptors modifies dynorphin-induced behavioral effects. Massardier D, Hunt PF (1989) A direct non-opiate interaction of Neurosci Lett 110:113-l 17. dynorphin-( l-l 3) with the N-methyl-D-aspartate (NMDA) receptor. Bakshi R, Faden AI (199Ob) Competitive and non-competitive NMDA Eur J Pharmacol 170: 125-l 26. antagonists limit dynorphin A-induced rat hindlimb paralysis. Brain McGinty J, Henriksen S, Goldstein A, Terenius L, Bloom F (1983) Res 507: 1-5. Dynorphin is contained within hippocampal mossy fibers: immu- Baskin DS, Hosobuchi Y, Loh HH, Lee NM (1984) Dynorphin (l- nochemical alterations after kainic acid administration and colchicine 13) improves survival in cats with focal cerebral ischaemia. Nature induced neurotoxicity. Proc Nat1 Acad Sci USA 80:589-593. 312551. McGinty JF, Kanamatsu T, Obie J, Dyer RS, Mitchell CL, Hong J Caudle RM, Chavkin C (1990) Mu opioid receptor activation reduces (1986) Amygdaloid kindling increases enkephalin-like immunoreac- inhibitory postsynaptic potentials in hippocampal CA3 pyramidal tivity but decreases dynorphin-A-like immunoreactivity in rat hip- cells of rat and guinea pig. J Pharmacol Exp Ther 252: 136 l-l 369. pocampus. Neurosci Lett 7 1:31-36. Caudle RM. Isaac L (1987) Intrathecal dvnornhin(l-13) results in an McLean S, Rothman RB, Jacobson AE, Rice KC, Herkenham M (1987) irreversible loss of the tail-flick reflex in-rats: Brain Res 435:1-6. Distribution of opiate receptor subtypes and enkephalin and dynor- Caudle RM, Isaac L (1988a) A novel interaction between dynorphin immunoreactivity in the hippocampus of the squirrel, guinea phin( 1-13) and an N-methyl-D-aspartate site. Brain Res 443:329- pig, rat, and hamster. J Comp Neurol 255:497-5 10. 332. Moises HC, Walker JM (1985) Electrophysiological effects of dynor- Caudle RM, Isaac L (1988b) Influence of dynorphin(l-13) on spinal phin peptides on hippocampal pyramidal cells in rat. Em J Pharmacol reflexes in the rat. J‘Pharmacol Exp Ther 246:508-5 13. 108:85-98. Chavkin C. James I. Goldstein A (1982) Dvnornhin is a specific en- Monaghan DT, Cotman CW (1985) Distribution of N-methyl-D-as- dogenous ligand of the kappa opibid receptor. Science 2 15~4 13-4 15. partate-sensitive L+H]glutamate-binding sites in rat brain. J Neu- Chen L, Huang L-YM (199 1) Sustained potentiation of NMDA re- rosci 5:2909-2919. ceptor-mediated glutamate responses through activation of protein Nahin RL, Hylden JLK, Iadarola MJ, Dubner R (1989) Peripheral kinase C by a p opioid. Neuron 7:3 19-326. inflammation is associated with increased dynorphin immunoreac- Choi DW, Koh J, Peters S (1988) Pharmacology of glutamate neu- tivity in both projection and local circuit neurons in the superficial rotoxicity in cortical cell culture: attenuation by NMDA antagonists. dorsal horn of the rat lumbar spinal cord. Neurosci Lett 96:247-252. J Neurosci 8: 185-196. Neck B, Rajpara A, O’Connor LH, Cicero TJ (1988) Autoradiography Contreras PC, Tam L, Drower E, Rafferty MF (1993) [3H]naltrindole: of pH]U-69593 binding sites in rat brain: evidence for K opioid re- a potent and selective ligand for labeling D-opioid receptors. Brain ceptor subtypes. Eur J Pharmacol 154:27-34. Res 604:160-164. Neck B, Giordano AL, Cicero TJ, O’Connor LH (1990) Affinity of Faden AI, Jacobs TP (1984) Dynorphin-related peptides cause motor drugs and peptides for U-69,593-sensitive and -insensitive kappa dysfunction in the rat through a non-opiate action. Br J Pharmacol opiate binding sites: the U69,593-insensitive site appears to be the 81:271-276. beta endorphin-specific epsilon receptor. J Pharmacol Exp Ther 254: Faden AI, Molineaux CJ, Rosenberger JG, Jacobs TP, Cox BM (1985) 412419. Endogenous opioid immunoreactivity in rat spinal cord following Neck B, Giordano AL, Moore BW, Cicero TJ (1993) Properties of traumatic injury. Ann Nemo1 17:386-390. the putative epsilon opioid receptor: identification in rat, guinea pig, Faden AI, Ellison JA, Noble LJ (1990) Effects of competitive and non- cow, pig and chicken brain. J Pharmacol Exp Ther 264:349-3591 - competitive NMDA receptor antagonists in spinal cord injury. Eur J Rothman RB. Bvkov V. Xue BG. Xu H. DeCosta BR. Jacobson AE. Pharmacol 175:165-174. Rice KC, Kleinman JE, Brady LS (1992) Interaction of opioid pep: Gall C, Brecha N, Karten HJ, Chang K-J (198 1) Localization of en- tides and other drugs with multiple kappa receptors in rat and human kephalin-like immunoreactivity to identified axonal and neuronal brain. Evidence for species differences. Peptides 13:977-987. populations of the rat hippocampus. J Comp Neurol 198:335-350. Schmidhammer H, Smith CFC, Erlach D, Kock M, Krassnig R, Schwetz Garant DS, Gale K (1985) Infusion of opiates into substantia nigra W, Wechner C (1990) Synthesis and biological evaluation of 14- protects against maximal electroshock seizures in rats. J Pharmacol alkoxymorphans. 3. Extensive study on cyprodime-related com- Exp Ther 234:45-48. pounds. J Med Chem 33:1200-1206. Griffith WH (1990) Voltage-clamp analysis of posttetanic potentiation Shukla VK, Bansinath M, Dumont M, Lemaire S (1992) Selective of the mossy fiber to CA3 synapse in hippocampus. J Neurophysiol involvement of kappa opioid and phencyclidine receptors in the an- 63:491-501. algesic and motor effects of dynorphin-A-( l-l 3)-Tyr-Leu-Phe-Asn- Gross RA, MacDonald RL (1987) Dynorphin A selectively reduces a Gly-Pro. Brain Res 59 1: 176-180. large transient (N-type) calcium current of mouse dorsal root ganglion Skilling SR, Sun X, Kurtz HJ, Larson AA (1992) Selective potentiation neurons in cell culture. Proc Nat1 Acad Sci USA 84:5469-5473. of NMDA-induced activity and release of excitatory amino acids by Hylden JLK, Nahin RL, Traub RJ, Dubner R (199 1) Effects of spinal dynorphin: possible roles in paralysis and neurotoxicity. Brain Res kappa-opioid receptor agonists on the responsiveness of nociceptive 575~272-278. superficial dorsal horn neurons. Pain 44:187-193. Stevens CW, Weinger MB, Yaksh TL (1987) Intrathecal dynorphins Iadarola MJ. Bradv LS. Draisci G. Dubner R (1988) Enhancement of suppress hindlimb electromyographic activity in rats. Eur J Phar- dynorphin gene expression in spinal cord following experimental in- macol 138:299-302. flammation: stimulus specificity, behavioral parameters and opioid Stewart P, Isaac L (199 1) Dynorphin-induced depression of the dorsal receptor binding. Pain 35:3 13-326. root potential in rat spinal cord: a possible mechanism for potentiation Isaac L, Van Zandt O’Malley T, Ristic H, Stewart P (1990) MK-801 of the C-fiber reflex. J Pharmacol Exp Ther 259:608-6 13. blocks dynorphin A (l-l 3binduced loss of the tail-flick reflex in the Wagner JJ, Evans CJ, Chavkin C (199 1) Focal stimulation ofthe mossy rat. Brain Res 531:83-87. fibers releases endogenous dynorphins that bind K, -opioid receptors Kajander KC, Sahara Y, Iadarola MJ, Bennett GJ (1990) Dynorphin in guinea pig hippocampus. J Neurochem 57:333-343. increases in the dorsal spinal cord in rats with a painful peripheral Wagner JJ, Caudle RM, Chavkin C (1992) k-Opioids decrease excit- neuropathy. Peptides 11:7 19-728. atory transmission in the dentate gyrus of the guinea pig hippocampus. Knox RJ, Dickenson AH (1987) Effects of selective and non-selective J Neurosci 12:132-141. K-opioid receptor agonists on cutaneous C-fibre-evoked responses of Wagner JJ, Terman GW, Chavkin C (1993) Endogenous dynorphins rat-dorsal horn neurones. Brain Res 4 15:2 l-29. inhibit excitatory neurotransmission and block LTP induction in the Kosterlitz HW, Watt AJ (1968) Kinetic parameters ofnarcotic agonists hippocampus. Nature 363:45 l-454. and antagonists, with particular reference to N-llylnoroxymorphone Webster JL, Polgar WE, Brandt SR, Berzetei-Gurske IP, Toll L (1993) (naloxone). Br J Pharmacol Chemother 33:266-276. Comparison of K2-opioid receptors in guinea pig brain and guinea pig Lee RJ, Hong J-S, McGinty JF, Lomax P (1987) Increased enkephalin ileum membranes. Eur J Pharmacol23 1:25 l-258. and dynorphin immunoreactivity in the hippocampus of seizure sen- Weisskopf MG, Zalutsky RA, Nicoll RA (1993) The opioid peptide sitive Mongolian gerbils. Brain Res 401:353-358. dynorphin mediates heterosynaptic depression of hippocampal mossy The Journal of Neuroscience, September 1994. 74(9) 5589

fibre synapses and modulates long-term potentiation. Nature 362: Zukin RS, Eghbali M, Olive D, Unterwald EM, Tempel A (1988) 423-421. Characterization and visualization of rat and guinea pig brain K opioid Were MA, MacDonald RL (1985) Dynorphin and neoendorphin pep- receptors: evidence for K, and K~opioid receptors. Proc Nat1 Acad Sci tides decrease dorsal root ganglion neuron calcium-dependent action USA 85:4061-4065. potential duration. J Pharmacol Exp Ther 234:49-56.