342

Brain Research, 524 (1990) 342-346

Elsevier BRES 24220

Dendritic action potentials activated by NMDA receptor-mediated EPSPs in CA1 hippocampal pyramidal cells Nicholas P. Poo~os’,~,~and Jeffery D. Kocsis’-3 Departments of ‘Neurology and ‘Neuroanatomy, Yale University School of Medicine, ‘PVAIEPVA Neuroscience Research Center, VA Medical Center, West Haven, CT 06516 (U.S.A.) and 4Neuroscience Program, Stanford University, Stanford, CA 94305 (U.S.A.)

(Accepted 24 April 1990) Key words: Intradendritic; CA1 hippocampus; n-Methyl-n-aspartate; Short-term potentiation; Burst firing; Dendrite

Intradendritic recordings were obtained in rat CA1 hippocampal pyramidal cells. Repetitive stimulation produced substantial short-term potentiation of the dendritic excitatory postsynaptic potential (EPSP) which was partly attributable to activation of n-methyl-D-aspartate receptors. Accompanying the potentiated synaptic response were Na+-mediated spikes which appeared to originate at multiple sites in the dendritic arbor. These discrete dendritic action potentials are rarely distinguishable in somatic recordings, but may contribute to the subthreshold response at the pyramidal cell body. In addition, dendritic spikes may interact with other voltage-dependent dendritic conductances.

The highly arborized dendrites of the CA1 hippocampal pyramidal cell are the primary sites of excitatory synaptic input to the cell, yet comparatively little is known about their electrophysiology since dendritic potentials are obscured by the cable properties of the neuron when recorded at the soma”. Dendritic recording sites, however, have proved advantageous for studying synaptic potentials in hippocampal cells due to diminished electrotonic attenuation of voltage transients generated within the dendritic arbor. Several studies utilizing techniques in cerebellar intradendritic recording Purkinje2,3 and hippocampal pyramidal cells’*5,14 have revealed active properties of the dendritic membrane, including both sodium- and calcium-dependent action potentials, and the tendency to produce bursts of action potentials under conditions of blockade of y-aminobutyric acid (GABA)-mediated inhibition. n-Methyl-oaspartate (NMDA) receptors are another important feature of hippocampal dendritic physiology, having a role in synaptic plasticity and epileptiform seizure (for a review see ref. 6). In the present study, intradendritic recordings of rat CA1 hippocampal pyramidal cells were obtained in the in vitro slice preparation to examine NMDA receptor-mediated postsynaptic potentials during repetitive stimulation, and their contribution to dendritic spike activity. Microelectrodes were fabricated from 1.0 mm borosilicate glass capillaries using a Brown-Flaming puller, filled with 4 M potassium acetate solution, and beveled Correspondence:

to DC resistances of 70-200 MS. Dendritic impalements were obtained in the distal half of the stratum radiatum of CAl, at distances >200 pm from the pyramidal cell layer, using 400 pm rat hippocampal slices prepared by methods previously described’. Impaled dendrites were accepted for study if they exhibited a stable resting potential >-55 mV and excitatory postsynaptic potentials (EPSPs) with peak amplitude >lO mV. The impalements were considered intradendritic if they met the following criteria: (1) impalements were obtained in the distal half of stratum radiatum, (2) the initial action potential amplitude arising from the EPSP was small (~15 mV) in spite of large amplitude EPSPs and deep resting potentials and (3) the response did not exhibit high-frequency burst activity to synaptic inputs as do hippocampal interneurons. Finally, a limited (n) number of intradendritic impalements were confirmed with injections of lucifer yellow. Acceptable intradendritic recordings (n = 13) were maintained for 30 min to over 3 h. Standard intrasomatic impalements of CA1 pyramidal cells were obtained for the purposes of comparison, Stably impaled dendrites and somata did not differ significantly in their resting potential, with an average of 61.4 f 8.9 mV (n = 13) for dendrites and 69.9 f 7.9 (n = 19) for somata (P > 0.05 by two-tailed t-test). Intradendritic impalements showed a relatively high input resistance of 89.4 f 40.6 MQ (n = 9) compared to a range of 25-35 MQ obtained for intrasomatic impalements in our hands and elsewhere7x14. The high mean

J.D. Kocsis, Neuroscience Research (127A), VA Medical Center, West Haven, CT 06516, U.S.A

343 input resistance for dendritic impalements likely results from the relatively distal range of positions within the dendritic layer used for recording. Notable differences between dendritic and somatic impalements emerged with orthodromic stimulation. Single stimuli delivered to CA1 afferents in the stratum radiatum produced EPSPs in intradendritic recordings that were larger than their somatic counterparts, as would be expected from the closer proximity of the dendritic recording site to the locus of excitatory synaptic conductances (Fig. lA,B). Single subthreshold stimuli evoked fast-rising EPSPswith peak amplitudes of 15-30 mV. The dendritic EPSPwas of longer duration than its somatic counterpart but was similarly followed by a hyperpolarization, possibly an inhibitory postsynaptic potential (IPSP)‘. With increasing stimulation intensity, small spikes of up to 15 mV in amplitude often appeared superimposed on the dendritic EPSP, usually on its falling phase. Comparison with intrasomatic recordings suggests that these spikes are attenuated action potentials generated in the soma or axon initial segment. This conclusion is supported by the observations that the spike recorded in the dendrites occurs after the peak of the dendritic EPSP and has a small amplitude; since the electrotonically conducted somatic EPSP peaks later than its dendritic source, it initiates a somatic (or axon initial segment) action potential which is delayed and attenuated when recorded in the dendrites. In contrast, a locally generated dendritic action potential would be expected to fire on the rising edge of the EPSP as it reached threshold. In some cases,the identity of the intradendritically recorded spike was further confirmed by comparison with responses to antidromic stimulation (not shown). With further increases in stimulation intensity, additional action potentials of varying shapesare recruited in the dendritic response. An example is shown in Fig. lC, but this occurred in all intradendritic impalements (n = 13). At low intensity, only an EPSPis elicited. Increasing stimulation intensity produces much larger EPSPs on which are superimposed first a single spike, then, at the highest intensity, several spikes of differing amplitudes and widths. The highest amplitude spike appears on the rising phase of the EPSP,thus is likely of dendritic origin. In addition to this action potential, two small inflections can be distinguished, followed by a second large action potential whose smaller amplitude and greater width suggestselectrotonic conduction from a greater distance than the first spike. The family of dendritic responses to orthodromic stimulation in Fig. 1C demonstrates at least 3 distinct shapes of action potential, suggesting multiple sites of electrogenesis located at varying electrotonic distances

from the recording electrode. The variety of dendritic spike shapes may also reflect changes in activation properties of relevant channels. Except in caseswhere antidromic stimulation of the cell body was used for comparison, it was not possible to differentiate unambiguously between dendritic recordings of action potentials generated at the soma and those generated at other distal dendritic sites. However, the multiplicity of spike amplitudes and widths makes it unlikely that dendritically

A

I

I ‘:“\.. 20

mV

10

-xr

C

5

I4

mV I

“IF&\

Fig. 1. Comparison of intradendritically and intrasomatically recorded responses to a single orthodromic stimulus. Dendritic and somatic recordings are from different cells. A: somatic response shows an action potential of 60 mV arising at the peak of a somatic EPSP of 9 mV. Resting potential is -65 mV. B: dendritic response shows an EPSP with a relatively high amplitude of 15 mV and fast rise time compared to its somatic counterpart in A. An attenuated somatic spike (3 mV in amplitude) is superimposed on the dendritic EPSP. Resting potential is -63 mV. C: superimposed dendritic responses to single orthodromic stimuli of graded intensity. At low intensity, only an EPSP is evoked. With higher intensity, action potentials of differing amplitudes and widths are evoked. D: superimposed dendritic responses to the 1st and 5th stimulus in a train at 10 Hz. The fifth EPSP (arrow) is substantially potentiated compared to the first and has evoked a burst of attenuated action potentials.

B 20

msec

oca

2+

( 6mM

C

Hg2+ STX

20

mV I 20 mase

i Fig. 2. NMDA receptor activation during repetitive stimulation in dendritic recordings. All responses are from the same cell. A: superimposed dendritic responsesto the first and 5th (arrow) stimuli of a train delivered at 10 Hz. The 5th response shows significant short-term potentiation and multiple spikes of varying amplitude. B: in the presence of the NMDA receptor antagonist APV (100 yM), the 5th response (arrow) shows considerably less potentiation and spiking. Note that the control response is not significantly altered by APV. C: returning to normal solution restores much of the increase in short-term EPSP potentiation and spiking.

recorded spikes originated at a single somatic site. In these experiments, increasing stimulation intensity tended to elicit an increased number of action potentials of greater amplitude, perhaps reflecting recruitment of dendritic spike generation sites. Also evident in Fig. 1C is the more rapid onset and greater amplitude of an afterhyperpolarizing potential with increasing stimulus intensity. This phenomenon may stem from recruitment of inhibitory interneurons in the stratum radiatum by increasing afferent stimulation or an increase in Ca’+activated potassium conduction. Short trains of orthodromic stimuli led to substantial short-term potentiation’ of the dendritic EPSP and dendritic bursting. As shown in Fig. lD, with delivery of 5 stimuli at 10 Hz, the fifth dendritic EPSP has increased considerably in amplitude and duration compared to the control response. In addition, multiple attenuated spikes are superimposed on the EPSP. These action potentials are broader and more attenuated than those elicited by single stimuli, and may represent the summation of

Fig. 3. Dendritic responses to depolarizing current passage. All responses are from the same cell while passing an 0.9 nA, 100 ms current step. A: dendritic response in normal solution. Note fast action potentials of multiple sizes superimposed on large broad depolarization. B: dendritic response in 0 Ca*+/6 Mg*+ solution. Fast spikes of multiple amplitudes persist but absence of large depolarizations as in A implies that the latter are Ca’+-mediated. C: addition of STX to 0 Ca*+/6 Mg2+ solution abolishes fast spikes as well, demonstrating their mediation by voltage-sensitive Na+

multiple dendritic action potentials at distant sites which are activated by the potentiated EPSP. Note that similar burst responses are not evidenced in recordings at the CA1 cell body under normal conditions (i.e., intact inhibitory neurotransmission)“. This may reflect stronger inhibitory influences at the soma, and provides further evidence of the dendritic origins of CA1 pyramidal cell burst firing. Activation of NMDA receptors in part underlied the short-term potentiation of the EPSP and dendritic spikes produced by repetitive stimulation. In Fig. 2A, a train of 5 stimuli at 10 Hz elicited an increase in EPSP amplitude and multiple attenuated action potentials. This shortterm potentiated response was significantly reduced by application of the NMDA receptor antagonist 2-amino5phosphonovaleric acid (APV; 100 ,uM). In Fig. 2B, the 5th response to 10 Hz stimulation shows a smaller increase in EPSP amplitude and much less multiple spiking than the same cell in Fig. 2A. The contribution

345 of NMDA receptor-mediated conductances is primarily evidenced as a prolongation of the potentiated EPSP, allowing activation of multiple action potentials; these effects are consistent with the known slow onset and decay kinetics of NMDA receptor-mediated [email protected] activation of NMDA receptors during repetitive stimulation is considerably more evident in the dendrites than has been observed at somatic recording sites’, perhaps due to a relative masking of the attenuated EPSPby local IPSPsat the soma. Note that there is little appreciable NMDA receptor activation by single stimuli in the dendritic recordings, as shown by comparison of the control traces in Fig. 2A,B. The nature of the dendritic spikes observed during repetitive stimulation was further examined by passageof depolarizing current through the intradendritic recording electrode. As shown in Fig. 3A, a depolarizing current step produced both a train of spikes of varying amplitudes which appeared similar to those elicited by orthodromic stimulation, and a large, slow depolarization. The fast spikes occurred with even small depolarizing current steps, while the slow depolarization required larger current steps. When Ca*+ was replaced in the bathing solution with 6 mM Mg*+, the depolarizing waveform disappeared, while the multiple spikes remained largely unaffected (Fig. 3B), establishing the former as a calcium-mediated potential. Addition of saxitoxin (STX; 1 ,uM), a blocker of voltage-sensitive sodium channels, abolished the remaining fast spikes (Fig. 3C), implying that these spikes are mediated by voltage-sensitive sodium channels. These results demonstrate distinct responses of the dendritic arbor to short trains of repetitive stimuli. A substantial NMDA receptor-mediated component of the EPSP,largely absent in the responseto a single stimulus, develops during repetitive stimulation and leads to substantial short-term potentiation of the EPSP and dendritic burst firing. The afterhyperpolarizing potential observed in the dendrites, possibly an IPSP, less effectively truncates the dendritic EPSP than its somatic counterpart, allowing for more prolonged depolarization than is observed at the cell body. This would promote activation of voltage-dependent NMDA receptor-mediated channels. At the cell body, local IPSPsand intrinsic conductancesabbreviate the EPSPand prevent repetitive firing4*‘*. The finding of dendritic Na+-mediated spikes during

repetitive stimulation supports previous reports of dendritic sodium spikes in hippocampal pyramidal cells1T’4. The present data suggestthe existenceof multiple sites of action potentials distributed throughout the dendritic arbor; these sites can be recruited into the synaptic response with either increasing stimulus intensity or repetitive stimulation. Although Ca*+-mediated depolarizations were not routinely elicited by orthodromic stimulation in these experiments, they were evoked by direct depolarizing current, and such ‘subthreshold’ Ca*+ currents may also contribute to the response to orthodromic stimulations. Retrograde passive conduction of the somatic action potential also appears to occur, and may be represented by one of the several spike morphologies observed in the dendrites. It is unclear what role dendritic action potentials play in determining the pyramidal cell response to afferent input. The well-established voltage dependence of NMDA receptor-mediated conductances6suggestsa possible regenerative interaction between intrinsic and synaptic conductances, with local action potentials contributing to membrane depolarization and conductance of NMDA receptor-linked channels. The electrotonic attenuation of dendritic spikes would spatially limit such effects to small regions within the dendritic arbor, although for the Purkinje cell it has been suggestedthat attenuateddendritic spikes summate and contribute to the somatic EPSPwaveform*. Electrotonic attenuation would explain the usual lack of identifiable dendritic spikes on intrasomatic recordings (with the exception of the ‘fast prepotential’ describedby Spencerand Kandel13). The present data demonstrate the role of NMDA receptor-mediated conductances in the induction of dendritic spike activity during synaptic activation. These discrete dendritic action potentials become attenuated and summated in their passive conduction to the soma, and produce a smoothed somatic potential which is a composite of synaptic and intrinsic currents originating in the dendrites. The notable differences in synaptic potentials, as observed from the dendrites versus the cell body, indicate the importance of intradendritic analysis in the study of synaptic transmission in the hippocampal pyramidal neuron.

1 Benardo, L.S., Masukawa, L.M. and Prince, D.A., Electrophysiology of isolated hippocampal pyramidal dendrites, J. Neurosci., 2 (1982) 1614-1622. 2 Llinas, R. and Nicholson, C., Electrophysiological properties of dendrites and somata in alligator Purkinje cells, J. Neurophysiol., 34 (1971) 534-551.

3 Llinas, R. and Sugimori, M., Electrophysiological properties of in vitro Purkinje cell dendrites in mammalian cerebellar slices, J. Physiol., 305 (1980) 197-213. 4 Madison, D.V. and Nicoll, R.A., Control of the repetitive discharge of rat CA1 pyramidal neurones in vitro, J. Physiol., 354 (1984) 319-331.

This work was supported in part by the NINDS (NS06208) and the Medical Research Service of the Department of Veterans Affairs. N.P. was supported in part by a Public Health Service predoctoral fellowship.

346 5 Masukawa, L.M. and Prince, D.A., Synaptic control of excitability in isolated dendrites of hippocampal neurons, J. Neurosci., 4 (1984) 217-227. 6 Mayer, M.L. and Westbrook, G.L., The physiology of excitatory amino acids in the vertebrate nervous system, Progr. Neurobiol., 28 (1987) 197-276. 7 Pitler, T.A. and Landfield, P.W., Postsynaptic membrane shifts during frequency potentiation of the hippocampal EPSP, J. Neurophysiol., 58 (1987) 866-882. 8 Poolos, N.P. and Kocsis, J.D., Elevated extracellular potassium concentration enhances synaptic activation of NMDA receptors in hippocampus, Bruin Research, in press. 9 Poolos, N.P., Mauk, M.D. and Kocsis, J.D., Activity-evoked increases in extracellular potassium modulate presynaptic excitability in the CA1 region of the hippocampus, J. Neurophysiol.,

58 (1987) 404-416. 10 Purpura, D.P., Comparative physiology of dendrites. In G.C. Quarton, T. Melchuk and F.O. Schmidt (Eds.), The Neurosciences, Rockefeller Press, New York, 1967, pp. 372-393. 11 Schwartzkroin, P.A., Further characteristics of CA1 cells in vitro, Brain Research, 128 (1977) 53-68. 12 Schwartzkroin, P.A. and Prince, D.A., Changes in excitatory and inhibitory synaptic potentials leading to epileptogenic activity, Brain Research, 183 (1980) 61-76. 13 Spencer, W.A. and Kandel, E.R., Electrophysiology of hippocampal neurons. IV. Fast prepotentials, J. Neurophysiol., 24 (1961) 272-285. 14 Wong, R.K.S., Prince, D.A. and Basbaum, A.I., Intradendritic recordings from hippocampal neurons, Proc. Nurl. Acad. Sci. U.S.A.,

76 (1979) 986-990.