Peptides.Vol. 12, pp. 47-51. ©Pergamon Press plc, 1991. Printed in the U.S.A.
0196-9781/91 $3.00 + .00
Peptidergic Transmission in the Brain. III. Hippocampal Inhibition by the Amygdala T. S M O C K , D. A L B E C K A N D P. M c M E C H E N
Center for Neurosciences, Department of Psychology, Campus Box 345 University of Colorado, Boulder, CO 80309 R e c e i v e d 6 M a r c h 1990
SMOCK, T., D. ALBECK AND P. McMECHEN. Peptidergic transmission in the brain. IlL Hippocampal inhibMon by the amygdala. PEPTIDES 12(1) 47-51, 1991.--The mechanism of arginine vasopressin (AVP) action in the rat hippocampus has been determined. The peptide activates inhibitory intemeurons and constricts cerebral microvessels. In the whole animal, each of these direct actions has secondary consequences for the excitability of pyramidal cells. Recent studies have shown that a peptide similar to AVP mediates endogenous neurotransmission in the hippocampus. Here we report experiments showing that the endogenous peptide activates the same mechanisms as exogenously applied AVP. The endogenous AVP-like peptide has no effect on the presynaptic fiber volley, or on pure somatic and dendritic postsynaptic potentials. These results are taken to exclude presynaptic mechanisms as explanations for the peptide's action. The endogenously released peptide inhibits individual pyramidal cells in single unit recording and excites presumed interneurons, just as AVP itself is known to act. The endogenous peptide is released only by stimuli applied to a nucleus that contains immunoreactive AVP and projects to the hippocampus. Arginine vasopressin (AVP)
Hippocampus
Amygdala
Peptide transmitter
A peptide similar to arginine vasopressin (AVP) acts as a transmitter in the rat hippocampus (2,14). The peptide is released from fibers that originate in the medial amygdaloid nucleus (AME) and project to the ventral hippocampus and lateral septum (3). Stimulation of the AME in the anesthetized rat yields a depression in the amplitude of the evoked population spike in area CA1, an effect that is identical to that produced by ventricular application of AVP (2). The action of the endogenous peptide transmitter is blocked by an antioxytocic receptor antagonist but not an antipressor receptor antagonist. We have interpreted these data to mean that the native transmitter is similar to vasopressin or oxytocin in structure, but is probably not AVP itself (14). The action of this peptidergic system can be modeled by study of AVP effects on the rat hippocampal slice (1, 8, 9, 12). In the isolated preparation AVP had the same action on the evoked pyrarnidal cell population spike as it has in the whole animal (1). Comparison with intracetlular records showed that the depression of the population spike reflected fewer individual action potenrials per stimulus rather than current shunting or refractoriness. The possibility that the depression of evoked activity could be produced by presynaptic mechanisms (i.e., by reduction in the excitability of the afferents in stratum radiatum) was eliminated by demonstration of AVP depression of antidromic field potential spikes (1). Single unit recording from the slice showed that AVP reduced the activity of pyramidal cells and increased the activity of inhibitory interne~cms (8,9). The action on interneurons was deemed to be direct, since it persisted in conditions of synaptic blockade (8,9). Taken together, the data indicate that AVP inkib-
its pyramidal cell somata by increasing the discharge rate of tonically active inhibitory interneurons, and by no other mechanism. The purpose of the present study is to extend the same type of analysis to the mechanism activated by the endogenously released AVP-like peptide. We will show that AVP stimulation depresses the population spike without having effect on presynaptic excitability or on pure somatic or dendritic postsynaptic potentials. With single unit techniques, we will show that AME stimulation inhibits pyramidal ceils and excites intemeurons, just as does AVP itself when applied to the slice. Finally, we will provide evidence that the AME influence comes about because of activation of the peptidergic fibers projecting to the hippocampus and not by release of AVP into the circulation. METHOD Experiments were conducted on 41 male rats (300-350 g). For field potential recording, methods duplicated those described in previous reports (2,14). Animals were acutely anesthetized (0.75 rot/100 g, 20% urethane). A test pulse was delivered to stratttm radiatum with a concentric bipolar stimulating electrode and the hybrid postsynaptic potential and population spike were recorded in dorsal C A 1. The coordinates for the stimulating electrode were - 3.9 mm posterior, 3.5 mm lateral and 4.0 mm down. The coordinates for the recording electrode were - 4 . 4 mm posterior, 2.9 nun lateral, 2.0 nun down. Stimulus intensity was always adjusted to provide a response of less than the largest that could be obtained and in some cases was reduced to a level subthresh-
47
48
SMOCK, ALBECK AND McMECHEN
,a,
_1
B
FIG. 1. Presynaptic fiber volley is unchanged with AME stimulation. Traces are (L to R) immediately before, 15 s after, and 20 min after AME stimuation (13 V, 100 ms train, 0.5 ms pulses at 100 Hz). Bars = 10 ms, 2 mV.
hold for production of a population spike. A digital average of eight test pulses delivered at 0.2 Hz was obtained every 5 minutes. A second 500 ixm bipolar concentric stimulating electrode was placed in the ipsilateral AME and delivered a conditioning pulse train (0.5 ms, 100 Hz, 100 ms) at a submaximal intensity so as to permit repetitive activation and internal controls (14). Single unit extracellular recording employed insulated steel electrodes (Frederick Haer, 5-7 M ~ ) and capillary glass pipettes filled with 4 M NaC1 (tip resistance 10-30 M~). "Complex spike" and "theta" cells were recognized by their spontaneous activity (6,11). Studies have shown that most complex spike cells are pyramidal cells and that there is a very large overlap between the class of theta cells and the class of interneurons (6,7). The cells we accepted as complex spike cells fired phasically in bursts of 3-5 spikes that diminished in size during the burst. For the purposes of this study, we assumed that these were pyramidal cells. The cells we accepted as theta cells fired tonically at high frequency and did not display bursting behavior. For the purposes of this study, we assumed that these were interneurons. Single unit activity was fed into a window discriminator which was used to drive an audio monitor and chart recorder. To examine the specificity of AME activation, electrolytic lesions (15 milliamps for 15 seconds) were made to mark electrode position. The positions of the lesions were determined by histology of cresyl violet-stained sections and comparison with the atlas of Paxinos and Watson (10). RESULTS In 11 animals the evoked population spike and synaptic potential were accompanied by a small deflection shortly after the artifact. This we took to be the fiber volley set up by action potentials in the excitatory afferents in stratum radiatum. With only two exceptions, this presynaptic fiber volley never changed during AME inhibition of the population spike. Figure 1 shows an example of the AME effect on the population spike for one placement in which the fiber volley is relatively well isolated from the stimulus artifact and synaptic potential onset. Here, as for the other field potential records, the traces represent the evoked responses (L to R) before, during, and after recovery from AME stimulation. Though the AME effect is quite large, no alteration in the fiber volley is detectable. Use of prevolley amplitude as an indicator of input efficacy is difficult since the synaptic potential is rarely completely distinct from the volley temporally. Therefore, change in synaptic potential amplitude may mask alteration in the prevolley. Furthermore, little of the prevolley likely reflects presynaptic calcium influx and this is the parameter of presynaptic physiology most intimately connected to transmitter release. Therefore, we used two other approaches to test for possible presynaptic mechanisms for
1 FIG. 2. Somatic postsynaptic potential is unchanged with AME stimulation. (A) AME stimulation (15 V, 100 ms train, 0.5 ms pulses, 100 Hz) inhibits the population spike of CA t pyramidal neurons. Traces are (L to R) immediately before, 15 s after, and 15 min after AME stimulation. (B) In the same preparation, stimulation intensity for the test pulse was decreased to yield a synaptic potential with no population spike. Traces are (L to R) immediately before, 15 s after, and 5 min after AME stimulation (15 V, 100 ms train, 0.5 ms pulses at 100 Hz). Bars = 10 ms, 4 mV. the AME effect. Pure postsynaptic potentials, uncontaminated by active processes in the soma and axon, may be taken to reflect intensity of transmitter release. We measured these in two ways. Since the AME effect is robust and reproducible within animals (14) it is possible to demonstrate action on population spike amplitude with a superthreshold test pulse, then make the test pulse subthreshold to examine the AME effect on the pure somatic synaptic potential. These data are shown in Fig. 2, which represents results obtained in 13 animals. Figure 2A shows the consequence of AME stimulation on a superthreshold test pulse and Fig. 2B shows that identical stimulation has no effect on a test pulse of amplitude subthreshold for eliciting a population spike. Another approach is illustrated in Fig. 3, representing results obtained in 3 animals. Here an AME response was obtained (Fig. 3A) and then the recording electrode was lowered 200 t~m into the dendritic region of CA l . Stimulation conditions for the test pulse in stratum radiatum and for the conditioning pulse in the AME remained constant. The synaptic potential appears inverted because the electrode is close to the current sink generated by the action of the synaptic transmitter. AME stimulation has no effect on the pure dendritic synaptic potential (Fig. 3B). In Fig. 3A and B the recovery trace is somewhat larger than the baseline trace. This may reflect a latent and long-lasting consequence of vascular mechanisms activated by the AVP-like peptide (14). In the examination of AVP action on field potentials in the slice, it was shown that the peptide had no action on very fast antidromic population spikes, i.e., those that arise less than 2 ms after the stimulus (1). The present results with orthodromic stimulation provide an interesting parallel with these data. Of 100 experiments in the present and previous studies, 30 had a population spike of less than 5-ms latency. These fast spikes were significantly less susceptible to AME inhibition than spikes of normal latency (p
0196-9781/91 $3.00 + .00
Peptidergic Transmission in the Brain. III. Hippocampal Inhibition by the Amygdala T. S M O C K , D. A L B E C K A N D P. M c M E C H E N
Center for Neurosciences, Department of Psychology, Campus Box 345 University of Colorado, Boulder, CO 80309 R e c e i v e d 6 M a r c h 1990
SMOCK, T., D. ALBECK AND P. McMECHEN. Peptidergic transmission in the brain. IlL Hippocampal inhibMon by the amygdala. PEPTIDES 12(1) 47-51, 1991.--The mechanism of arginine vasopressin (AVP) action in the rat hippocampus has been determined. The peptide activates inhibitory intemeurons and constricts cerebral microvessels. In the whole animal, each of these direct actions has secondary consequences for the excitability of pyramidal cells. Recent studies have shown that a peptide similar to AVP mediates endogenous neurotransmission in the hippocampus. Here we report experiments showing that the endogenous peptide activates the same mechanisms as exogenously applied AVP. The endogenous AVP-like peptide has no effect on the presynaptic fiber volley, or on pure somatic and dendritic postsynaptic potentials. These results are taken to exclude presynaptic mechanisms as explanations for the peptide's action. The endogenously released peptide inhibits individual pyramidal cells in single unit recording and excites presumed interneurons, just as AVP itself is known to act. The endogenous peptide is released only by stimuli applied to a nucleus that contains immunoreactive AVP and projects to the hippocampus. Arginine vasopressin (AVP)
Hippocampus
Amygdala
Peptide transmitter
A peptide similar to arginine vasopressin (AVP) acts as a transmitter in the rat hippocampus (2,14). The peptide is released from fibers that originate in the medial amygdaloid nucleus (AME) and project to the ventral hippocampus and lateral septum (3). Stimulation of the AME in the anesthetized rat yields a depression in the amplitude of the evoked population spike in area CA1, an effect that is identical to that produced by ventricular application of AVP (2). The action of the endogenous peptide transmitter is blocked by an antioxytocic receptor antagonist but not an antipressor receptor antagonist. We have interpreted these data to mean that the native transmitter is similar to vasopressin or oxytocin in structure, but is probably not AVP itself (14). The action of this peptidergic system can be modeled by study of AVP effects on the rat hippocampal slice (1, 8, 9, 12). In the isolated preparation AVP had the same action on the evoked pyrarnidal cell population spike as it has in the whole animal (1). Comparison with intracetlular records showed that the depression of the population spike reflected fewer individual action potenrials per stimulus rather than current shunting or refractoriness. The possibility that the depression of evoked activity could be produced by presynaptic mechanisms (i.e., by reduction in the excitability of the afferents in stratum radiatum) was eliminated by demonstration of AVP depression of antidromic field potential spikes (1). Single unit recording from the slice showed that AVP reduced the activity of pyramidal cells and increased the activity of inhibitory interne~cms (8,9). The action on interneurons was deemed to be direct, since it persisted in conditions of synaptic blockade (8,9). Taken together, the data indicate that AVP inkib-
its pyramidal cell somata by increasing the discharge rate of tonically active inhibitory interneurons, and by no other mechanism. The purpose of the present study is to extend the same type of analysis to the mechanism activated by the endogenously released AVP-like peptide. We will show that AVP stimulation depresses the population spike without having effect on presynaptic excitability or on pure somatic or dendritic postsynaptic potentials. With single unit techniques, we will show that AME stimulation inhibits pyramidal ceils and excites intemeurons, just as does AVP itself when applied to the slice. Finally, we will provide evidence that the AME influence comes about because of activation of the peptidergic fibers projecting to the hippocampus and not by release of AVP into the circulation. METHOD Experiments were conducted on 41 male rats (300-350 g). For field potential recording, methods duplicated those described in previous reports (2,14). Animals were acutely anesthetized (0.75 rot/100 g, 20% urethane). A test pulse was delivered to stratttm radiatum with a concentric bipolar stimulating electrode and the hybrid postsynaptic potential and population spike were recorded in dorsal C A 1. The coordinates for the stimulating electrode were - 3.9 mm posterior, 3.5 mm lateral and 4.0 mm down. The coordinates for the recording electrode were - 4 . 4 mm posterior, 2.9 nun lateral, 2.0 nun down. Stimulus intensity was always adjusted to provide a response of less than the largest that could be obtained and in some cases was reduced to a level subthresh-
47
48
SMOCK, ALBECK AND McMECHEN
,a,
_1
B
FIG. 1. Presynaptic fiber volley is unchanged with AME stimulation. Traces are (L to R) immediately before, 15 s after, and 20 min after AME stimuation (13 V, 100 ms train, 0.5 ms pulses at 100 Hz). Bars = 10 ms, 2 mV.
hold for production of a population spike. A digital average of eight test pulses delivered at 0.2 Hz was obtained every 5 minutes. A second 500 ixm bipolar concentric stimulating electrode was placed in the ipsilateral AME and delivered a conditioning pulse train (0.5 ms, 100 Hz, 100 ms) at a submaximal intensity so as to permit repetitive activation and internal controls (14). Single unit extracellular recording employed insulated steel electrodes (Frederick Haer, 5-7 M ~ ) and capillary glass pipettes filled with 4 M NaC1 (tip resistance 10-30 M~). "Complex spike" and "theta" cells were recognized by their spontaneous activity (6,11). Studies have shown that most complex spike cells are pyramidal cells and that there is a very large overlap between the class of theta cells and the class of interneurons (6,7). The cells we accepted as complex spike cells fired phasically in bursts of 3-5 spikes that diminished in size during the burst. For the purposes of this study, we assumed that these were pyramidal cells. The cells we accepted as theta cells fired tonically at high frequency and did not display bursting behavior. For the purposes of this study, we assumed that these were interneurons. Single unit activity was fed into a window discriminator which was used to drive an audio monitor and chart recorder. To examine the specificity of AME activation, electrolytic lesions (15 milliamps for 15 seconds) were made to mark electrode position. The positions of the lesions were determined by histology of cresyl violet-stained sections and comparison with the atlas of Paxinos and Watson (10). RESULTS In 11 animals the evoked population spike and synaptic potential were accompanied by a small deflection shortly after the artifact. This we took to be the fiber volley set up by action potentials in the excitatory afferents in stratum radiatum. With only two exceptions, this presynaptic fiber volley never changed during AME inhibition of the population spike. Figure 1 shows an example of the AME effect on the population spike for one placement in which the fiber volley is relatively well isolated from the stimulus artifact and synaptic potential onset. Here, as for the other field potential records, the traces represent the evoked responses (L to R) before, during, and after recovery from AME stimulation. Though the AME effect is quite large, no alteration in the fiber volley is detectable. Use of prevolley amplitude as an indicator of input efficacy is difficult since the synaptic potential is rarely completely distinct from the volley temporally. Therefore, change in synaptic potential amplitude may mask alteration in the prevolley. Furthermore, little of the prevolley likely reflects presynaptic calcium influx and this is the parameter of presynaptic physiology most intimately connected to transmitter release. Therefore, we used two other approaches to test for possible presynaptic mechanisms for
1 FIG. 2. Somatic postsynaptic potential is unchanged with AME stimulation. (A) AME stimulation (15 V, 100 ms train, 0.5 ms pulses, 100 Hz) inhibits the population spike of CA t pyramidal neurons. Traces are (L to R) immediately before, 15 s after, and 15 min after AME stimulation. (B) In the same preparation, stimulation intensity for the test pulse was decreased to yield a synaptic potential with no population spike. Traces are (L to R) immediately before, 15 s after, and 5 min after AME stimulation (15 V, 100 ms train, 0.5 ms pulses at 100 Hz). Bars = 10 ms, 4 mV. the AME effect. Pure postsynaptic potentials, uncontaminated by active processes in the soma and axon, may be taken to reflect intensity of transmitter release. We measured these in two ways. Since the AME effect is robust and reproducible within animals (14) it is possible to demonstrate action on population spike amplitude with a superthreshold test pulse, then make the test pulse subthreshold to examine the AME effect on the pure somatic synaptic potential. These data are shown in Fig. 2, which represents results obtained in 13 animals. Figure 2A shows the consequence of AME stimulation on a superthreshold test pulse and Fig. 2B shows that identical stimulation has no effect on a test pulse of amplitude subthreshold for eliciting a population spike. Another approach is illustrated in Fig. 3, representing results obtained in 3 animals. Here an AME response was obtained (Fig. 3A) and then the recording electrode was lowered 200 t~m into the dendritic region of CA l . Stimulation conditions for the test pulse in stratum radiatum and for the conditioning pulse in the AME remained constant. The synaptic potential appears inverted because the electrode is close to the current sink generated by the action of the synaptic transmitter. AME stimulation has no effect on the pure dendritic synaptic potential (Fig. 3B). In Fig. 3A and B the recovery trace is somewhat larger than the baseline trace. This may reflect a latent and long-lasting consequence of vascular mechanisms activated by the AVP-like peptide (14). In the examination of AVP action on field potentials in the slice, it was shown that the peptide had no action on very fast antidromic population spikes, i.e., those that arise less than 2 ms after the stimulus (1). The present results with orthodromic stimulation provide an interesting parallel with these data. Of 100 experiments in the present and previous studies, 30 had a population spike of less than 5-ms latency. These fast spikes were significantly less susceptible to AME inhibition than spikes of normal latency (p
