Acetylcholine Causes Rapid Nicotinic Excitation in the Medial Habenular Nucleus of Guinea Pig, in Vitro

The Journal of Neuroscience, March 1987, 7(3): 742-752

Acetylcholine Causes Rapid Nicotinic Excitation in the Medial Habenular Nucleus of Guinea Pig, in vitro

David A. McCormick and David A. Prince Department of Neurology, Stanford University School of Medicine, Stanford, California 94305

The actions of ACh in the medial habenular nucleus (MHb) The medial habenular nucleus (MHb) is a collection of closely were investigated using extra- and intracellular recording spacedcells on the dorsal-medial surfaceof the thalamus.This techniques in guinea pig thalamic slice maintained in vitro. nucleus receives its major input via the stria medularis from Applications of ACh to MHb neurons resulted in rapid ex- the postcommissuralseptum (Gottesfeld and Jacobowitz, 1979; citation followed by inhibition. Neither of these responses Herkenham and Nauta, 1979; Contestabileand Fonnum, 1983) was abolished by blockade of synaptic transmission, indi- and gives riseto a large axonal projection through the fasciculus cating that they are consequences of ACh action directly retroflexus to the interpeduncular nucleus at the baseof the on MHb cells. Local applications of the nicotinic agonists brain (Herkenham and Nauta, 1979). Both the afferents and nicotine and cytisine caused long-lasting excitation, while efferents of the MHb have been proposedto be at least in part applications of another nicotinic agonist, 1,l -dimethyl4- cholinergic (Kataoka et al., 1973, 1977; Kuhar et al., 1975; phenylpiperazinium caused both the excitatory and inhibi- Gottesfeld and Jacobowitz, 1979; Vincent et al., 1980; Con- tory responses. Applications of the muscarinic agonists DL- testabile and Fonnum, 1983; Houser et al., 1983; Villani et al., muscarine and acetyl-B-methylcholine did not consistently 1983; Keller et al., 1984; Woolf and Butcher, 1986). Recent cause either the excitatory or inhibitory response. Adding immunohistochemical studies for choline acetyltransferase the nicotinic antagonist hexamethonium to the bathing me- (ChAt), the synthesizing enzyme for ACh, showthat the ventral dium blocked both the excitatory and inhibitory ACh re- portions of the MHb as well as many fibers in the fasciculus sponses, while addition of the muscarinic antagonists atro- retroflexus are heavily reactive (Houser et al., 1983). Lesionsof pine or scopolamine had no effect. These results indicate the septalregion, stria medularis,MHb, or fasciculusretroflexus that the effects of ACh on MHb neurons are mediated by reduce ChAt activity in both the medial habenular and inter- nicotinic receptors. peduncular nuclei (Kataoka et al., 1973, 1977; Gottesfeld and Intracellular recordings revealed that ACh or nicotine Jacobowitz, 1978, 1979; Vincent et al., 1980; Contestabileand cause an increase in membrane conductance associated Fonnum, 1983; Contestabile and Villani, 1983). These results with depolarizations that had an average reversal potential imply that the MHb may receive a cholinergic projection from of - 16 to - 11 mV. These results indicate that the ACh- the diagonal band/posterior septal region, while the interpe- induced excitation is due to an increase in membrane cation duncular nucleus may receive a choline& input from this re- conductance. gion as well as from the medial habenularnuclei (however, see The inhibitory response that follows ACh-induced depo- Woolf and Butcher, 1985). larization and repetitive firing was associated with a hyper- Autoradiographic localization of putative cholinergic recep- polarization and an increase in membrane conductance. tors showsthat the MHb contain a high density of nicotinic, Similar postexcitatory inhibition could also be elicited by but not muscarinic, receptors(Rotter et al., 1979; Martin and direct depolarization or by applications of glutamate, indi- Aceto, 1981; Clarke et al., 1984, 1985; Rainbow et al., 1984; cating that the hyperpolarizing response to ACh may be an London et al., 1985). Furthermore, recent autoradiographic la- endogenous postexcitatory potential that is not directly cou- beling of the presumeda-subunit of neural nicotinic receptors pled to activation of nicotinic receptors. These results sug- showsthat the MHb is heavily reactive (Boulter et al., 1986; gest that cholinergic transmission in the MHb may be largely Goldman et al., 1986). These results indicate that cholinergic of the nicotinic type. This nucleus may be of one of the transmissionin the MHb may be largely nicotinic. major regions of the nervous system through which nicotine Intracellular studiesof the postsynaptic actions of ACh in the mediates its central effects. CNS have been almost exclusively limited to those involving muscarinic receptors (see referencesin KrnjeviC, 1975, and McCormick and Prince, 1986a, b). Traditional nicotinic recep- Received Apr. 25, 1986; revised Aug. 25, 1986; accepted Aug. 26, 1986. tor-mediated excitation, similar to that seenin the Renshaw We thank Dr. B. W. Connors for comments on the manuscrid. J. Radis for technical assistance, and Walter Carlini for help with the potassium-sensitive cells of the spinal cord (Curtis and Ryall, 1966), has been ob- microelectrodes. This study was supported by NIH Grants NS 12151 and NS served in only a limited number of nuclei, including the inter- 06477 (D.A.P.) and NS 07331 (D.A.M.) from the National Institute of Neuro- peduncular nucleus(Brown et al., 1983; Takagi, 1984)and per- lo&al Communicative Diseases and Stroke, by the Morris Research Fund, and b; fellowships from the Huntington’s Disease Foundation of America and the haps some nuclei of the brain stem (Bradley and Dray, 1972), Giannini Foundation (D.A.M.). hypothalamus(Cobbett et al., 1986),and thalamus(Phil& 1971). Correspondence should be addressed to David A. McCormick, Ph.D., Depart- In this study, we describea rapid nicotinic excitatory response ment of Neurology, Rm. C338, Stanford University Medical Center, Stanford, CA 94305. to ACh in the MHb that appearsto be due to an increasein Copyright 0 1987 Society for Neuroscience 0270-6474/87/030742-l 1$02.00/O membrane cation conductance. The Journal of Neuroscience, March 1997, 7(3) 743

Figure I. Extracellular recordings of responses to ACh. A, Application of ACh causes a rapid excitation followed by inhibition. A second application of ACh during the inhibitory period results in an additional exci- A tatory-inhibitory sequence, indicat- ACh ACh Glu 5 ing that the inhibition is not a result of depolarization block. Application of glutamate to this neuron also causes excitation followed by inhibition. B Normal , B, Anolication_- of ACh to this MHb neuron in normal solution causes the typical excitation-inhibition sequence. Stimulation in the region of L./-b the stria medularis fiber pathway (filled circle) consistently activates an action r, potential (B: right, arrow). C, After approximately 30 min in Mn*+ (5 mM) . and 0 Ca*+ containing solution, the neuron still responds to ACh with an excitation and inhibition, although the spontaneous firing rate in this neuron 1 has decreased. Stimulation of synap- low Ca++, Mn++, , tic inputs (C, right) does not activate action potentials, indicating that synaptic transmission is blocked. The ACh-induced excitatory response is reduced in Mn*+, low CA2+, perhaps due to hyperpolarization of the neuron, as indicated by the reduced spontaneous firing rate, or some change in spike threshold during 5 mM Mn2+ 0 , 4 perfusion. Calibrations under first and A second segments of C are also for ACh 5 set lO?Zec comparable segments of B.

from Sigma. The MHb could be readily identified in the thalamic slice Materials and Methods under the dissecting microscope. It lay on the most dorsomedial aspect Procedures for preparing thalamic slices and obtaining intracellular re- of the thalamus. ”iust dorsal and medial to the end of the fasciculus cordings from them have been described previously (Jahnsen and Llinls, retroflexus. 1984; McCormick and Prince, 1986b). Male or female adult guinea pigs Intracellular recordings were obtained with microelectrodes fabri- were anesthetized with sodium pentobarbital (30 mg/kg) and decapi- cated on a Brown and Flaming P-go/PC Duller and beveled to a final tated. A block of thalamus and cortex containing the MHb was rapidly resistance of 150-l 75 MQ. Bridge balance was continuously monitored. removed, placed in cold (5°C) slice solution, and cut into 400-500 PM Only neurons with steady resting membrane potentials of at least -50 slices on a vibratome (Lancer Corporation). Slices were maintained in mV and overshooting action potentials were included in the intracellular an interface type recording chamber at a temperature of 36 + 1°C. The study. Extracellular recordings were obtained with broken intracellular control bathing medium contained the following (in mM): NaCl, 124; microelectrodes filled with the bathing medium. These microelectrodes KCl, 2.5 or 5;MgSO,, 2; NaH,PO,, 1.25; NaHCO,, 26; CaCl,, 2;‘and had a resistance of 10-25 MQ. All data were recorded on tape (O-5 KHz dextrose. 10. When divalent cations (e.g.. Mn2+) were added, PO, and for intracellular recordings; 300-5 KHz for extracellular units) and a SO, were omitted to prevent precipita& and the solution contained strip-chart recorder. In some experiments, extracellular [K] was mea- the following: NaCl, 132; KCl, 5; MgCl,, 2; NaHCO,, 22; and dextrose, sured using valinomycin K+-sensitive microelectrodes (Oehme and Si- 10. The slice solution was saturated with 95% 0,/5% CO, and buffered mon, 1976) connected to an axoprobe- 1 amplifier (Axon Instruments). to pH 7.4. These K+ -sensitive microelectrodes registered no response when placed All agonists were applied using the pressure pulse technique. Brief in the ACh-containing application solution, indicating that our [K], oulses of oressure lannroximatelv 40 nsi (280 KPa). 10 msecl were used measurements were not contaminated by applications of this agent. io extrude approximately O. l-5-pl of agonist-containing slice solution from a broken micropipette. Repeated applications of ACh at different depths in the slice were used to find a region from which robust responses were obtained. In experiments in which acetyl-@-methylcholine (MCh) Results was to be applied, the site of best response to ACh was first determined and a second microelectrode containing MCh was lowered to approx- Extracellular single-unit recordings indicated that the neurons imately the same point in the slice. To further insure an adequate ap- of the MHb are spontaneously active in the slice, firing in a plication of MCh, the volume of application was increased to 10-l 5 pl regular, repetitive manner at approximately 2-6 Hz (Fig. 1). at a concentration of 5 mM. Nicotine, cytisine, and 1,l -dimethyl-4- phenylpiperazinium (DMPP) were applied only to the surface of the Application of ACh (OS-l.0 mM) to MHb neurons resulted in slice since this was found to be sufficient to cause large responses. An- a short-latency excitation followed by a period of inhibition of tagonists were applied in the bathing medium. All drugs were obtained spike discharge (Fig. 1A). Similar responses were obtained from 744 McCormick and Prince * ACh in Medial Habenula

4 ACh

4 DMPP

4 Nicotine

Figure 2. Pharmacological character- ization of receptors mediating responses to ACh in the MHb. A, Appli- cation of ACh to this neuron causes the 4 typical excitation-inhibition sequence. Cytisine B, Application of the nicotinic agonist DMPP also causes excitation followed E by inhibition. C and D, In contrast, application of the nicotinic agonists nicotine and cytisine causes prolonged excitation with little or no inhibition. Application of the muscarinic agonist acetyl-fl-methylcholine (MCh) causes only a weak excitation (E), while DL- muscarine causes no response Q. All data in A-F are from the same MHb 4 5 set neuron. MCh DL-Muscarine neurons at all locations within the nucleus (e.g., ventral versus However, the action potentials usually beganto recover before dorsal).The inhibitory responsealways occurredafter the initial the period of inhibition, indicating that the late period of spike excitation and was never seenin isolation. The duration and inhibition was not merely the result of depolarization block intensity of both the excitation and inhibition becamegreater (Brown, 1980). This conclusion was supported by the finding with increasingdoses of ACh. Applications of the excitatory that additional applications of ACh or of Glu during the late amino acid glutamate (Glu, 1 mM) resultedin a similar excitato- period of inhibition resulted in an additional excitatory-inhib- ry-inhibitory response(Fig. 1A). The excitatory responsesto itory sequence(Fig. lA, second and third ACh applications). ACh and Glu were often associatedwith a transient decreasein Interestingly, in many cells, the first few spikesafter the period action potential amplitude and an increasein action potential of inhibition had amplitudes that were generally larger than width. The action potential amplitude could even be transiently those of spikesbefore ACh application (e.g., Fig. 1, A, B). This eliminated, indicating that depolarization block may have oc- may indicate that the membrane potential is hyperpolarized curred (for example, see Fig. 34 normal and wash-arrow). during the period of inhibition. The Journal of Neuroscience, March 1997, 7(3) 745

Hexamethonium Scopolamine

5 set

ACh Figure 3. Effectof nicotinicand muscarinicantagonists on the responseof MHb neuronsto ACh. A, Applicationof ACh to this MHb neuronin normalsolution results in the typical excitation-inhibitionsequence. Washing in of the nicotinic antagonisthexamethonium (250 MM) greatly diminishesthe excitatoryand inhibitory responses.Washing out of the hexamethoniumreinstates astrong excitation-inhibition response. The lack of actionpotentials during the midportionof the excitatoryresponse (arrow) isprobably due to depolarizationblock. The lossof a second,smaller- amplitudeunit during the recordingcontributed to the changein appearanceof the responseto ACh in washversus normal segments. B, Bath applicationof the muscarinicantagonist scopolamine (10 PM) for morethan 45 min did not affect the excitatoryor inhibitory responseto ACh.

To test whether the excitatory-inhibitory responsesto ACh Krnjevic, 1975; Brown, 1980; Volle, 1980; Takagi, 1984; were direct on the cells recorded or were mediated through the McCormick and Prince, 1985), while inhibitory responsesare releaseof another neurotransmitter, sagittal thalamic sliceswere usually associatedwith receptors of the muscarinic type (e.g., bathed in solutions containing Mn2+ (3-5 mM) and low Ca2+ seeKrnjevic, 1975; McCormick and Prince, 1986a). We inves- (0.0-0.5 mM) to block synaptic transmission. Before block of tigated the pharmacologicalprofile of the receptorsmediating synaptic transmission,both stimulation of the stria medularis the responsesof MHb neuronsto ACh, using typical nicotinic fiber pathway (Fig. lB, right) and application of ACh (Fig. lB, and muscarinic agonistsand antagonists.Local application of left) causedshort-latency excitatory responsesof MHb neurons. the nicotinic agonistsnicotine (1 mM; n = 10) or cytisine (1 mM; The ACh-induced excitation was followed by the typical period n = 4) resulted in prolonged excitation, which was gradual in of inhibition (Fig. le). After approximately 30 min of exposure both its rise to peak firing frequency and decay back to baseline to the Mn2+, low Ca*+ solution, the synaptically evoked excit- firing rates (Fig. 2, C, D). Inhibition was absent or only very atory responsewas completely blocked, but the ACh-induced weak. In contrast, application of the nicotinic agonist DMPP excitation and inhibition were not (n = 6; Fig. lc). In fact, (Chen et al., 1951; seeTable 2 in Gyermek, 1980) (1 mM; n = bathing the slicesfor more than 1 hr in 0.0 mM Ca*+, 5 mM 7) causedexcitation followed by prolongedinhibition (Fig. 2B). Mn2+ did not block either spontaneousactivity or the excitatory In general,the excitatory responsesto DMPP were strongerand and inhibitory ACh responses.These results indicate that both more rapid in onset and offset than responsesto a similar ap- the excitatory and the inhibitory responsesof MHb neuronsto plication of nicotine or cytisine but not as rapid as responsesto ACh are a consequenceof direct action on the neuron studied ACh. The ability of DMPP, but not nicotine, to generate the and not mediatedthrough the releaseof other neurotransmitters. inhibitory responsemay be related to the prolonged action of Furthermore, the persistenceof spontaneousactivity in the nu- nicotine due to its uptake and slow releaseby brain tissue(Brown cleus following synaptic blockade indicates that this activity is et al., 1971). intrinsic to MHb neurons and is not due to external synaptic In contrast to applications of nicotinic agonists,application drive. of the muscarinic agonistMCh (5 mM; n = 9) (Curtis and Ryall, Rapid excitatory responsesto ACh have been associatedwith 1966) most often did not causea response,though weak exci- nicotinic aswell asmuscarinic receptors (Curtis and Ryall, 1966; tation or weak inhibition was occasionallyseen with large doses 746 McCormick and Prince l ACh in Medial Habenula

Figure 4. Characteristics of intracel- lularly recorded, spontaneously firing MHb neurons. A, Spontaneously oc- curring action potentials are followed by at least 2 phases of hyperpolarization. B, Segment of A expanded for de- tail. Dashed line drawn in at -55 mV. C C, Reversal potential of the afterhy- perpolarization (AHP). Activation of 2 action potentials by a short (30 msec) depolarizing pulse is followed by a sub- -60 - stantial AHP that reverses polarity at about -95 mV (5 mM [K],). The am- -60 plitude of the injected current pulse was > adjusted so as to cause the generation E of 2 action potentials at each membrane -100 L v) potential. Action potential amplitudes ol are not fully shown. 0.25 set

Vm Vm

A A& ACh ACh Nick B

Figure 5. ACh causes short-latency nicotinic depolarization of MHb neurons. A, Application of a brief pulse of ACh to a MHb neuron causes a rapid depolarization (A, left). Lowering the pressure and prolonging the duration of application results in a more square-shaped response (A, middle). Adjusting the injected DC so as to compensate for the ACh-induced depolarization shows that ACh causes an increase in membrane conductance during the response associated with a peak inward current of approximately 170 pA (A, right). Upper truces are the current monitor, with hyperpolarizing current pulses being delivered at a rate of 1 Hz. B, Application of ACh to another MHb neuron causes typical depolarization, while application of the muscarinic agonist acetyl-@-methylcholine (MCh) evokes only a very small response. C, Application of a small amount of nicotine to the surface of the slice causes a large and prolonged depolarization. D, Manual voltage clamp so as to maintain a relatively constant membrane potential following nicotine application. Nicotine causes a large increase in membrane conductance, as indicated by a decrease m membrane electrotonic responses to hyperpolarizing current pulses. Current monitor shows that nicotine induces an inward current with a peak amplitude of about 660 PA. Calibration in B also for A and in D also for C and current trace of A. The Journal of Neuroscience, March 1987, 7(3) 747

(Fig. 2E). Similarly, applications of DL-muscarine (10 mM; n = 4) resulted in no response (Fig. 2fl. These results indicate that the ACh-induced excitation and inhibition is the result of activation of nicotinic receptors. To confirm this hypothesis we - -50 added the specific nicotinic antagonist, hexamethonium (Paton and Zaimis, 195 1; Brown, 1980), 100-500 PM, to the bathing - -60 F medium (Fig. 3A). Hexamethonium reversibly antagonized both 3 the excitatory and inhibitory responses to ACh (n = 5). Bath - -70 5 applications of the muscarinic antagonists scopolamine or atro- pine (10 MM), on the other hand, did not affect the ACh-induced excitation or inhibition (Fig. 3B; n = 4). These findings support . -60 the conclusion that ACh responses of MHb neurons are the result of the activation of nicotinic receptors. - -90

Intracellular recordings Intracellular recordings from MHb neurons in the ventral half of the nucleus indicated that these cells are remarkable for the presence of rather large input resistances of 50-500 MB (x = 144 t- 110 MB) and large afterhyperpolarizations (AHPs) (14- 45 mV, 0.2 to > 1 set at firing threshold in 2.5 mM K), which followed even a single action potential (Fig. 4, A, B). Sponta- neous activity was characterized by action potentials followed by AHPs that appeared to have both fast and slow components (Fig. 4, A, B). The large AHPs were associated with a substantial increase in membrane conductance and had a reversal potential of -88.1 mV (k3.3 mV, n = 6) in 5.0 mM extracellular K (Fig. 4C) and - 106.7 mV (k3.9 mV, n = 4) in 2.5 mM K. This shift in equilibrium potential (62 mV per lo-fold change in [K],) -100 -60 -60 -40 -20 +10 indicates that the AHP is probably due to activation of a K Vm (mV) current, as in other regions of the CNS (Hotson and Prince, Figure 6. Estimationof the reversalpotential of the rapiddepolarizing 1980; Madison and Nicoll, 1984). MHb neurons completely responseto ACh. A, Applicationsof ACh to 2 neuronsat different lacked the low-threshold, voltage-sensitive, slow Ca2+ spike membranepotentials (scale at right) indicatethat the cholinergicde- characteristic of other thalamic neurons (Jahnsen and Llinas, polarizationbecomes smaller at morepositive V,,,‘s.B, Plot of the data in A showsthat the extrapolatedreversal potentials for Cells I and 2 1984; Wilcox et al., 1985). are - 25 and + 5 mV, respectively.Reversal potentials for these2 neu- To prevent the confounding factors associated with sponta- ronsrepresent 2 of the moreextreme values for all neuronstested. neous neuronal discharge during applications of ACh, all cells were hyperpolarized with intracellular injection of DC before applications of agonists. After a few minutes of hyperpolarizing > 7 nS (Fig. 5A, third application). Much larger increasesin G,., current injection, most MHb neurons spontaneously hyperpo- occurred with larger applications of ACh (e.g., seeFig. 7A). larized to such an extent that resting potential remained below The nicotinic nature of this responsewas confirmed with firing threshold after removal of the holding current, even though intracellular recordings by applications of nicotine (Fig. 5, C, the same cells were originally spontaneously active upon initial D) and MCh (Fig. 5B). Small applications of nicotine to the stabilization of the membrane potential (e.g., Fig. 4). Under surface of the slice resulted in large depolarizations (Fig. 5C). these circumstances, the average resting membrane potential When the injected current was adjusted during theseresponses was -63 mV (& 6.5 mV) and action potential amplitude was so as to maintain a relatively constant I’,,,, it was apparent that 76 mV (k 10.6 mV). nicotine causeda large increasein G, of 5 to > 10 nS (Fig. SD). Brief (ca. 10 msec)pressure pulse applications of ACh to MHb In contrast, applicationsof the muscarinic agonistMCh resulted neuronsalways resultedin a rapid depolarization with an onset in either no responseor only very smalldepolarizations, whereas latency that could be as brief as the delay imposedby the ap- ACh causedthe typical rapid excitatory responsewhen applied plication system (i.e., approximately 10 msec) and a duration to the sameneurons (Fig. 5B). of approximately 0.5-5 seconds(Fig. 5, A, B). In most cells, these applications of ACh were kept small so as to avoid the Reversalpotential of ACh-induced depolarization generation of action potentials. Larger ACh applications, which The rapid excitatory responseto ACh varied in amplitude when generated action potentials, were associatedwith large mem- tested at various V,‘s (Fig. 6). The responsewas largestat hy- brane depolarizations followed by a hyperpolarization (seebe- perpolarized V,,,‘s, smallest at depolarized Vm’s,and had ex- low). When long-duration (3-6 set), low-pressureACh pulses trapolated reversal potentials of -40 to +5 mV (2 = - 16.7 + were applied, square-shapeddepolarizations were evoked (Fig. 18.1 mV, n = 5) (Fig. 6, A, B). In most MHb neurons,the input 5A, second response).When intracellular injection of hyper- resistanceappeared to be relatively constant over the range of polarizing current was usedto compensatefor the ACh-induced V,‘s tested (e.g., cell 2, Fig. 6A). However, in some neurons, depolarization, it was clear that ACh causedan increasein ap- R, decreasedsubstantially with depolarization (e.g., Cell 1, Fig. parent membrane conductance (G,), which ranged from 3 to 6A). The possibleeffects of thesenonlinearities in R, upon the 748 McCormick and Prince * ACh in Medial Habenula

F&we 7. Characteristicsof ACh-inducedhyperpolarization. A, Large ap- nlicationofACh to a MHb neuronheld hearfiring threshold(-65 mV) causes a largedepolarization and the generation of actionpotentials followed by a hyperpolarization.Both responsesare associatedwith substantialincreases in GN.A secondapplication of ACh at -90 mV resultsin a relatively steadymem- branepotential during the periodof the ACh I hyperpolarizingresponse of the - 65 mV 1 set applicationas G, graduallyreturns to normal. Application of ACh with the cellhyperpolarized to - 105mV results in a slowdepolarization during the inhibitory period,indicating that the hy- perpolarizationhas been reversed. B, Plot of percentagepeak hyperpolarization versus percentage peak change in G,. Datawere obtained from measure- mentsat the peak of the hyperpolarization (100%)and as the membrane potentialreturned to normalfor 8 cells that had first been depolarized to near firing threshold with intracellular injection of DC. Change in conductance (AG,) was calculated as the input conductance during the hyperpolarization minus the input conductance at a comparable membrane potential after the response was over. The data are ex- pressed as a percentage of the AG, at 4=j$y~, . , , . . . , the peak of the hyperpolarization. Note that the 2 measures are highly corre- lated (r = 0.92), and as the change in conductance declines to zero, so does 0 20 40 60 80 100 the hyperpolarization.Different sym- bols represent different MHb neurons. % peak AG, projected reversal potential were corrected by dividing the am- excitatory hyperpolarization (X = 7.4 f 4.0 set; n = 14) was plitude of the ACh responseby R, at each V, tested. The pro- similar to that of the inhibition seenin the extracellular re- jected reversal potential for ACh-induced depolarizations in cordings (e.g., seeFigs. l-3). The similarity in duration and the MHb neuronsafter correction for nonlinearitiesin R, was - 14.7 ability of the hyperpolarization to inhibit ongoing spike activity mV (f 18.5).Furthermore, given that the amplitude of the ACh indicated that this responseunderlies the postexcitatory inhib- responsevaried as a relatively linear function of V,,,, it was also itory phase seen with ACh applications during extracellular possibleto calculate the extrapolated reversal potential of the recordings (see Fig. 1). Therefore, we refer to this period of ACh induced depolarization utilizing the manual voltage clamp hyperpolarization as “ACh-induced inhibition or hyperpolar- technique and the following equation: V,,,, = Vclamp + I,,,,/ ization” while making no assumptionsas to whether this effect AG, (Ginsborg, 1967). Using this approach, the calculated re- is direct (e.g., direct coupling of ACh receptorsto K+ channels) versal potential of the ACh-induced rapid excitation was -28 or indirect (e.g., caused by the previous generation of action to+8mV(8=-11.3? 13SmV,n=4),inagreementwith potentials). the above current-clamp results. A comparison of the amplitude of V, changesevoked by current test pulsesduring the ACh-induced hyperpolarization, Postexcitatory hyperpolarization versusthose at similar V,‘s after the responsewas over, revealed Applications of ACh that generated large depolarizations and that the agonist-inducedhyperpolarization was associatedwith action potentials caused a postexcitatory hyperpolarization an increasein membrane conductance of 3-45 nS (z = 16.8 + (n = 14; Fig. 7A; -65 mV application). When MHb neurons 12.4 nS) at its peak. When ACh was applied with the neuron were allowed to be spontaneouslyactive, application of ACh held at various V,,,‘s, the evoked hyperpolarization became resultedin a rapid depolarization associatedwith the generation smallerat more negative values and could be reversedto a slow of action potentials,followed by a hyperpolarization during which depolarization (Fig. 7A). The reversal potential of the ACh- the spontaneousspike activity ceased.The duration of this post- induced hyperpolarization varied dependingupon whether the The Journal of Neuroscience, March 1987, 7(3) 749

7.5 [Kl 0 5.0

2.5

4 , 4 ACh 1 set Figure 8. Simultaneousrecording of [K], measuredwith a valinomycinK+-sensitive microelectrode and responseof MHb neuronsto application of ACh. ACh causesthe typical excitation-inhibitionsequence. During the excitation,[K], is seento increaseto approximately8 mM. Thisincrease in [K]. graduallyreturns to normal,[K], beingin the rangeof 4-7.5 mM during the periodof inhibition of spikeactivity. The [K], of the bathing mediumwas 2.5 mM.

early or late componentsof the responsewere measured.The driving voltage between V, and the reversal potential during early portions (Fig. 7A, first arrow) reversed at a much more different parts of the response,decay of a possibleresidual in- positive membrane potential than the later components (Fig. fluence of the previous depolarization, or [K], changes.The 7A, secondarrow). These resultsindicate that the ACh-induced Pearson’scorrelation coefficient for thesedata rangedfrom 0.90 depolarization and hyperpolarization interact substantially dur- to 0.99 for individual cells and 0.92 (p < 0.001, G!!= 49) for ing at leastthe early phasesof the inhibitory response.However, all cells normalized to the same peak response(Fig. 7B) even it was usually possibleto find a membrane potential at which in the presenceof this nonlinear relationship. This resultis most the later componentsof the hyperpolarization were at a stable readily explained if the hyperpolarization is causedby an in- equilibrium potential as GNincreased back to normal (Fig. 7A, creasein membrane conductance to K or Cl, as opposed to -90 mV application). This equilibrium potential varied from activation of an electrogenic ionic pump. -75 to -90 mV in 2.5 mM K-containing bathing medium. The range at which the ACh-induced hyperpolarization re- The presenceof an increasein membrane conductanceand a versed (- 75 to - 90 mV) is significantly positive to expected demonstrablereversal potential would seemto indicate that the EK (-106.7) in 2.5 mM [K], (as measuredfrom the reversal hyperpolarizing responseis due to an increase in membrane potential ofthe spike AHP). However, given the closeapposition conductance to one or more ions, as opposed to activation of of MHb neuronsto one another, often without intervening glial an electrogenicionic pump (Brown et al., 1972; Nicoll and Alger, processes(Tokunaga and Otani, 1978), and the apparent in- 1981; Thompson and Prince, 1986). However, similar results creasein cation conductance during the ACh-induced depolar- could be obtained if there were substantialoverlap between an ization, a substantial increasein extracellular [K] might be ex- ACh-induced excitation and a strong electrogenic ionic pump. pected, thereby altering E,. We tested this possibility by In this situation the apparent increase in membrane conduc- performing simultaneousextracellular [K] measurementswith tance during the hyperpolarization could be causedby the re- valinomycin K+-sensitive microelectrodes and single-unit re- sidualincrease in cation conductanceto ACh, while the apparent cordings with a secondindependent microelectrode during ap- reversal potential would represent the V, at which the depo- plications of ACh (Fig. 8). ACh causedan increasein [K], up larizing inlhtence of the cation conductance is at equilibrium to peak levels of 15 mM from a baselinelevel of 2.5 mM during with the hyperpolarizinginfluence of the electrogenicionic pump. the period of excitation. This increasein [K], gradually returned To obtain someindication ofwhether the increasein membrane to normal over the courseof many seconds(Fig. 8), with [K], conductance was related to the generation of the hyperpolar- often being in the range of 5-8 mM during the period of inhi- ization or merely reflected the residual influence of ACh upon bition (Fig. 8). Given an [K], of 5-8 mM, a reversal potential cation channels,we plotted the relative amplitude of the change of - 75 to - 90 mV is exactly as expectedfor increasesin mem- in neuronal input conductance (calculated as the difference in brane K conductance (seeabove and McCormick and Prince, G, during the ACh responseand that at a similar V, after the 1986a), indicating that the ACh-induced inhibition could be responsewas over) versusthe relative amplitude of the hyper- entirely due to an increasein gK. polarizing phase (Fig. 7B). Such plots revealed that these 2 measuresof the responseparallel eachother quite well, although Discussion in a somewhatcurvilinear manner (Fig 7B). This nonlinearity The results of the present study indicate that MHb neurons may be due to a number of factors, including changesin the possessnicotinic receptorsthat, when activated, causea rapid 750 McCormick and Prince l ACh in Medial Habenula excitation through an increase in membrane conductance. The Prince, 1986a, b). Therefore, a direct mediation of the inhibitory average extrapolated reversal potential of this response (- 11 to response by nicotinic receptors in the MHb would be highly - 16 mV) agrees well with that for ACh nicotinic responses in unusual. Rather, our data suggest that the inhibitory response the PNS and neuromuscular junction, where ACh is known to is an indirect consequence of the preceding ACh-induced de- cause an increase in membrane conductance to cations (Takeu- polarization. Indeed, applications of another excitatory agent, chi and Takeuchi, 1960; Dennis et al., 1971; Gallagher et al., Glu, or direct depolarization alone was also associated with 1982). These results suggest that the ACh-induced rapid exci- postexcitatory periods of inhibition (Figs. lA, 4C). In other tation in the MHb is also mediated by an increase in membrane regions of the nervous system, trains of action potentials are conductance to cations. The presence of strong nicotinic re- often associated with an AHP (Hotson and Prince, 1980; Morita sponses and the apparent lack of muscarinic components fol- et al., 1982; Madison and Nicoll, 1984; McCormick et al., 1985) lowing ACh applications to the MHb is consistent with auto- that appears to be due in large part to activation of specialized radiographic findings that show that this nucleus has a high K conductances by increases in intracellular [Ca*+] during the density of nicotinic, but not muscarinic, receptors (Rotter et al., preceding depolarization. Glutamate-induced depolarizations 1979; Martin and Aceto, 1981; Clarke et al., 1984, 1985; Rain- of hippocampal pyramidal cells may also cause activation of a bow et al., 1984; London et al., 1985). The block of ACh re- Ca-sensitive K current (Nicoll and Alger, 198 1). MHb neurons sponses by hexamethonium indicates that the nicotinic receptor respond to direct depolarization with an unusually large AHP in the MHb may be of the ganglionic (versus neuromuscular) (see Fig. 4). The reversal potential of this AHP is consistent type (Brown, 1980; Goldman et al., 1986). with the presumed equilibrium potential for K-mediated slow A cholinergic synaptic input to the MHb has not yet been hyperpolarizations in other thalamic neurons (McCormick and proven. It has been proposed that the MHb receives a cholin- Prince, 1986a), although its Ca*+ dependency is not yet known. ergic input from the postcommissural septal region (Kataoka et The presence of this large AHP following even a single action al., 1977; Gottesfeld and Jacobowitz, 1979; Contestabile and potential (Fig. 4, A, B) raises the possibility that the inhibition Fonnum, 1983). Recent double-labeling experiments (retro- which follows ACh-induced excitation may merely be a pro- grade fluorescent dyes and ChAt immunoreactivity or AChE longed response of the same type. staining) indicate that the habenular nuclei may also recieve a The most parsimonious explanation for our data is that the cholinergic input from the horizontal limb of the diagonal band, ACh-induced hyperpolarization is due to activation of some magnocellular preoptic area, and the pedunculopontine and dor- hyperpolarizing current. One possibility, therefore, is that the solateral tegmental nuclei of the brain stem (Woolf and Butcher, inhibitory response may in large part be due to a K conductance 1986). However, the latter study did not determine which of activated by Ca entry through the nicotinic channels (Dwyer et the habenular nuclei (medial versus lateral) were the recipients al., 1980) opened by ACh. If one takes into account the increase of this presumed cholinergic innervation. Other anatomical in [K], during the ACh-induced inhibitory responses (5-8 mM; studies confirm the presence of a projection to the MHb from Fig. 8), then a reversal potential in the range of -75 to -90 the horizontal limb of the diagonal band (Domesick, 1976; Her- mV is exactly what would be expected if the hyperpolarization kenham and Nauta, 1977). Much of the evidence for a cholin- were due predominately to activation of a K conductance. How- ergic input to the MHb is based upon the loss of ChAt activity ever, we were unable to block the inhibitory phase of ACh in the MHb after lesions of the septal region or stria medularis responses by bathing the slices in 0.0 Ca*+ and 5 mM Mn*+, a fiber pathway. Since some MHb neurons themselves appear to procedure that should greatly reduce Ca entry into depolarized possess intrinsic ChAt activity (Houser et al., 1983; Keller et neurons and therefore reduce Ca-activated K currents (e.g., Hot- al., 1984), it is possible that these lesion-induced reductions in son and Prince, 1980; Madison and Nicoll, 1984). Although not ChAt levels are transynaptic and not due to the loss of cholin- conclusive, these data suggest that the ACh-induced inhibition ergic terminals per se. Given the possible cholinergic nature of is mediated by an increase in gK other than that associated with MHb neurons, an additional possible source of cholinergic input entry of Ca*+ into the MHb cells. is recurrent collaterals within the nucleus itself (Iwahori, 1977). Extensive exposure (over minutes) of neurons in the periph- Confirmation of a cholinergic pathway to the MHb must await eral ganglia to cholinergic agonists causes a postexcitation in- combined ChAt immunocytochemistry and electron micros- hibition. This inhibition is due to activation ofelectrogenic ionic copy. Our data do indicate that activation of such an input would pumps (Brown et al., 1972; Volle, 1980). Similarly, large ap- almost certainly cause rapid nicotinic excitation. If this ex- plications of Glu to hippocampal pyramidal cells also result in citation were of sufficient amplitude, a postexcitatory period of activation of an electrogenic ion pump (Thompson and Prince, inhibition would be expected, as reported here. 1986). These postexcitatory hyperpolarizations, in contrast to A number of different mechanisms may underlie the ACh- those reported here, are associated with little or no increase in induced inhibition including: (1) direct coupling of nicotinic input conductance. Our limited data would argue against the receptors to inhibitory mechanisms; (2) an indirect response major involvement of such ionic pumps in the ACh-induced due to an increase in intracellular [Ca*+] that activates a Ca- inhibition of MHb neurons. However, it is entirely possible that sensitive K conductance (gK&); (3) activation of an electrogenic all of the mechanisms mentioned above are involved to varying ionic pump due to changes in distribution of ions across the cell degrees (e.g., see Baylor and Nicolls, 1969), with the relative membrane during the previous depolarization; and (4) activa- contribution of each perhaps depending upon the length of ex- tion of voltage-dependent inhibitory mechanisms by the pre- posure to ACh. The neural mechanisms underlying this hyper- vious depolarization alone. polarizing response will require further study. Where they have been extensively investigated, direct inhib- Although nicotine is one of the drugs most widely used by itory responses to ACh have been associated largely with mus- modem man, the mechanisms by which it produces effects on carinic receptors (e.g., see Krnjevic, 1975; McCormick and the CNS are largely unknown. To date, only a few subcortical The Journal of Neuroscience, March 1987, 7(3) 751 regions of the CNS are known to contain neurons that have sympathetic neurons in the heart of the hog. Proc. R. Sot. London rapid nicotinic excitatory responses. These include the MHb [Biol.] 177: 509-539. Dome&k, V. B. (1976) Projections of the nucleus of the diagonal (this study), interpeduncular nucleus (Brown et al., 1983; Taka- band of Broca in the rat. Anat. Rec. 184: 391-392. gi, 1984) spinal cord (Curtis and Ryall, 1966), and perhaps Dwyer, T. M., D. J. Adams, and B. Hille (1980) The permeability of some regions of the thalamus (Phillis, 197 l), brain stem (Bradley the endplate channel to organic cations in frog muscle. J. Gen. Physiol. and Dray, 1972), and hypothalamus (Cobbett et al., 1986). Nico- 75: 469-492. Gallagher, J. P., W. H. Griffith, and P. Shinnick-Gallagher (1982) Cho- tinic receptors or receptors having mixed nicotinic-muscarinic linergic transmission in cat parasympathetic ganglia. J. Physiol. (Lond.) properties have also been implicated in the regulation of syn- 332: 473-486. aptic transmission in the hippocampus and interpeduncular nu- Ginsborg, B. L. (1967) Ion movements in junctional transmission. cleus (Rovira et al., 1983; Brown et al., 1984). Pharmacol. Rev. 19: 289-3 16. Given the prevalence of nicotinic receptors in the MHb and Goldman, D., D. Simmons, L. W. Swanson, J. Patrick, and S. Heine- mann (1986) Mapping of brain areas expressing RNA homologous interpeduncular nuclei and their known connections with the to two different acetylcholine receptor a-subunit cDNAs. Proc. Natl. forebrain limbic system, as well as their possible involvement Acad. Sci. USA 83: 4076-4080. in many different complex behaviors, it is reasonable to suggest Gottesfeld, Z., and D. M. Jacobowitz (1978) Cholinetgic projection that some of the central actions of nicotine may be mediated of the diagonal band to the interpeduncular nucleus of the rat brain. through increased activation of these nuclei. Brain Res. 156: 329-332. Gottesfeld, Z., and D. M. Jacobowitz (1979) Cholinergic projections from the septal-diagonal band area to the habenular nuclei. Brain Res. 176: 291-394. References Gyermek, L. (1980) Methods for the examination ofganglion-blocking Baylor, D. A., and J. G. Nicolls (1969) After-effectsof nerveimpulses activity. In Pharmacolonv of Ganglionic Transmission, D. A. Khar- on signalling in the central nervous system of the leech. J. Physiol. kevich; ed., Handbook ofEiperi&ental Pharmacology 53: 63-l 2 1. (Lond.) 203: 571-589. Herkenham. M.. and W. J. H. Nauta (1977) Afferent connections of Boulter. .I.. K. Evans. D. Goldman. G. Martin. D. Treco. S. Heinemann. the habenular’nuclei in the rat. A horseradish peroxidase study, with and J: Patrick (1986) Isolation of a cDNA clone coding for a possible a note on the fiber-of-passage problem. J. Comp. Neurol. 173: 123- neural nicotinic acetylcholine receptor o-subunit. Nature 319: 368- 146. 374. Herkenham, M., and W. J. H. Nauta (1979) Efferent connections of Bradley, P. B., and A. Dray (1972) Short latency excitation of brain the habenular nuclei in the rat. J. Comp. Neurol. 187: 19-48. stem neurones in the rat by acetylcholine. Br. J. Pharmacol. 45: 372- Hotson, J. R., and D. A. Prince (1980) A calcium-activated hyper- 374. polarization follows repetitive firing in hippocampal neurons. J. Neu- Brown, D. A. (1980) Locus and mechanism of action of ganglion- rophysiol. 43: 409-419. - - - - blocking agents. In Pharmacology of Ganglionic Transmission, D. A. Houser. C. R.. G. D. Crawford. R. P. Barber. P. M. Salvaterra. and .I. Kharkevich, ed., Handbook of Experimental Pharmacology 53: 185- E. Vaughn ‘( 1983) Organization and morphological characteristics 236. of cholinergic neurons: An immunocytochemical study with a mono- Brown, D. A., J. V. Halliwell, and C. N. Scholfield (1971) Uptake of clonal antibody to choline acetyltransferase. Brain Res. 266: 97-l 19. nicotine and extracellular space markers by isolated rat ganglia in Iwahori, N. (1977) A golgi study on the habenular nucleus of the cat. relation to receutor activation. Br. J. Pharmacol. 42: 100-l 13. J. Comp. Neurol. 171: 319-344. Brown, D. A., M: J. Brownstein, and C. N. Scholfield (1972) Origin Jahnsen, H., and R. Llinls (1984) Electrophysiological properties of of the after-hyperpolarization that follows removal of depolarizing guinea pia thalamic neurones: An in vitro studv. J. Phvsiol._ (Land.). I agents from the isolated cervical ganglion of rat. Br. J. Pharmacol. j49: 2651226. 44: 651-671. Kataoka, K., Y. Nakamura, and R. Hassler (1973) Habenulo-inter- Brown, D. A., R. J. Docherty, and J. V. Halliwell (1983) Chemical nedunctdar tract: A possible cholineraic neuron in rat brain. Brain transmission in the rat interpeduncular nucleus in vitro. J. Physiol. kes. 62: 264-267. - (Lond.) 341: 655-670. Kataoka, K., M. Sorimachi, S. Okuno, and N. Mizuno (1977) Cho- Brown, D. A., R. J. Docherty, and J. V. Halliwell (1984) The action liner& and GABAeraic fibers in the stria medularis of the rabbit. of cholinomimetic substances on impulse conduction in the haben- Brain Res. Bull. 2: 461-464. ulointerpeduncular pathway of the rat in vitro. J. Physiol. (Lond.) Keller, F., K. Rimvall, and P. G. Waser (1984) Slice cultures confirm 353: 101-109. the presence of cholinergic neurons in the rat habenula. Neurosci. Chen, G., R. Portman, and A. Wickel (195 1) Pharmacology of 1,l- L&t. 52: 299-304. dimethyl-4-phenylpiperazinium iodide, a ganglion stimulating agent. Kmjevic, K. (1975) Chemical nature of synaptic transmission in ver- J. Pharmacol. Exp. Ther. 103: 330-336. tebrates. Physiol. Rev. 54: 4 18-540. Cobbett, P., W. T. Mason, and D. A. Poulain (1986) Intracellular Kuhar, M. J., R. N. DeHaven, H. I. Yamamura, H. Rommel-Spacher, analysis of control of supraoptic neurone (SON) activity in vitro by and J. R. Simon (1975) Further evidence for cholinergic habenulo- acetylcholine (ACh). J. Physiol. (Lond.) 3jI: 2 16P. interpeduncular neurons: Pharmacologic and functional characteris- Clarke. P. B. S.. C. B. Pert. and A. Pert. (1984) Autoradioaranhic tics. Brain Res. 97: 265-275. distribution of nicotine receptors in rat brain. Brain Res. 323: 390- London, E. D., S. B. Waller, and J. K. Wamsley (1985) Autoradio- 395. graphic localization of [“HI nicotine binding sites in the rat brain. Clarke, P. B. S., R. D. Schwartz, S. M. Paul, C. B. Pert, and A. Pert Neurosci. Lett. 53: 179-184. (1985) Nicotine binding in rat brain: Autoradiographic comparison Madison, D. V., and R. A. Nicoll (1984) Control ofrepetitive discharge of [)H] acetylcholine, [3H] nicotine, and [1*51]-a-bungarotoxin. J. Neu- of rat CA 1 pyramidal neurones in vitro. J. Physiol. (Land.) 354: 3 19- rosci. 5: 1307-1315. 331. Contestabile. A.. and F. Fonnum ( 1983) Cholinergic and GABAergic Martin, B. R., and M. D. Aceto (1981) Nicotine binding sites and forebrain projections to the habenula and nucleus interpeduncularis: their localization in the central nervous system. Neurosci. Biobehav. Suraical and kainic acid lesions. Brain Res. 275: 287-297. Rev. 5: 473-478. Conteitabile, A., and L. Villani (1983) The use ofkainic acid for tracing McCormick, D. A., and D. A. Prince (1985) Two types of muscarinic neuroanatomical connections in the septohabenulointerpeduncular response to acetylcholine in mammalian cortical neurons. Proc. Natl. system of the rat. J. Comp. Neurol. 214: 459-469. Acad. Sci. USA 82: 6344-6348. Curtis, D. R., and R. W. Ryall (1966) The excitation of Renshaw cells McCormick, D. A., and D. A. Prince (1986a) Acetylcholine induces by cholinomimetics. Exp. Brain Res. 2: 49-65. burst firing in thalamic reticular neurones by activating a potassium Dennis, M. J., A. J. Harris, and S. W. Kuffler (197 1) Synaptic trans- conductance. Nature 319: 402-405. mission and its duplication by focally applied acetylcholine in para- McCormick, D. A., and D. A. Prince (1986b) Mechanisms of action 752 McCormick and Prince l ACh in Medial Habenula

of acetylcholine in the guinea pig cerebral cortex, in vitro. J. Physiol. Takagi, M. (1984) Actions of choline@ drugs on cells in the inter- (Lond.) 375: 169-194. peduncular nucleus. Exp. Neurol. 84: 358-363. McCormick, D. A., B. W. Connors, J. W. Lighthall, and D. A. Prince Takeuchi, A., and N. Takeuchi (1960) On the permeability of end- (1985) Comparative electrophysiology of pyramidal and sparsely plate membrane during the action of transmitter. J. Physiol. (Lond.) spiny stellate neurons of the neocortex. J. Neurophysiol. 54: 782-806. 154: 52-67. Morita, K., North, R. A., and T. Tokimasa (1982) The calcium-ac- Thompson, S. M., and D. A. Prince (1986) Activation of the electro- tivated potassium conductance in guinea-pig myenteric neurones. J. genie sodium pump in hippocampal CA1 neurons following gluta- Physiol. (Lond.) 329: 341-354. mate-induced depolarization. J. Neurophysiol. 56: 507-522. Nicoll, R. A., and B. E. Alger (198 1) Synaptic excitation may activate Tokunaga, A., and K. Otani (1978) Fine structure of the medial ha- a calcium dependent potassium conductance in hippocampal pyrami- benular nucleus in the rat. Brain Res. 150: 600-606. dal cells. Science 212: 957-959. Villani, L., Constestabile, A., and F. Fonnum (1983) Autoradiographic Oehme, M., and W. Simon (1976) Microelectrode for potassium ions labeling of the cholinergic habenulo-interpeduncular projection. Neu- based on a neutral carrier and comparison of its characteristics with rosci. Lett. 42: 261-266. a cation exchange sensor. Anal. Chim. Acta 86: 21-25. Vincent, S. R., W. A. Staines, E. G. McGeer, and H. C. Fibiger (1980) Paton, W. D. M., and E. J. Zaimis (1951) Paralysis of autonomic Transmitters contained in the efferents of the habenula. Brain Res. ganglia by methonium salts. Br. J. Pharmacol. 6: 155-168. 195: 479-484. Phillis, J. W. (1971) The pharmacology of thalamic and geniculate Volle, R. L. (1980) Nicotinic ganglion-stimulating agents. In Phar- neurons. Int. Rev. Neurobiol. 14: l-48. macology of Ganglionic Transmission, D. A. Kharkevich, ed., Hand- Rainbow, T. C., R. D. Schwartz, B. Parsons, and K. J. Kellar (1984) book of Experimental Pharmacology, 53: 28 l-3 12. Quantitative autoradiography of nicotinic [3H].acetylcholine binding Wilcox, K. S., M. J. Gutnick, and G. R. Christoph (1985) Burst gen- sites in rat brain. Neurosci. Lett. 50: 193-196. eration in lateral habenular neurons recorded in vitro. Sot. Neurosci. Rotter, A., N. J. M. Birdsall, A. S. V. Burgen, P. M. Field, E. C. Hulme, Abstr. 11: 509. and G. Aisman (1979) Muscarinic receptors in the central nervous Woolf, N. J., and L. L. Butcher (1985) Cholinergic systems in the rat system of the rat. I. Technique for autoradiographic localization of brain: II. Projections to the interpeduncular nucleus. Brain Res. Bull. the binding of [3H] propylbenzilylcholine mustard and its distribution 14: 63-83. in forebrain. Brain Res. Rev. 1: 141-165. Woolf, N. J., and L. L. Butcher (1986) Choline& systems in the rat Rovira, C., Y. Ben-Ati, E. Cherubuni, K. Krnjevic, and N. Ropert brain: III. Projections from the pontomesencephalic tegmentum to (1983) Pharmacology of the dendritic action of acetylcholine and the thalamus, tectum, basal ganglia, and basal forebrain. Brain Res. further observations on the somatic disinhibition in the rat hippo- Bull. 16: 603-637. campus in situ. Neuroscience 8: 97-106.