NMDA receptor-mediated excitability in dendritically deformed dentate granule cells in pilocarpine-treated rats

Neuroscience Letters, 129 (1991) 69-73 © 1991 Elsevier Scientific Publishers Ireland Ltd. 0304-3940/91/$ 03.50 ADONIS 030439409100396E

69

NSL 07916

NMDA receptor-mediated excitability in dendritically deformed dentate granule cells in pilocarpine-treated rats M a s a k o Isokawa 1 and Luiz Eugenio A . M . Mello 2 1Brain Research Institute, Centerfor the Health Sciences, University of California Los Angeles, CA 90024-1761 (U.S.A.) and 2Departamento de Fisiologia, Escola Paulista de Medicina, Sao Paulo (Brazil) (Received 11 February 1991; Accepted 15 April 1991) Key words: Hippocampal slice; Membrane property; Long-lasting EPSP; Spine loss; Beaded dendrite Membrane properties and synaptic responses were analyzed in dentate granule cells in hippocampal slices prepared from pilocarpine-treated, chronically epileptic rats. Perforant path stimulation evoked a long-lasting excitatory postsynaptic potential (EPSP) with multiple spikes in a stimulus intensity-dependentfashion. The response was strongly facilitated by paired-pulse stimulation. Application of N-methyl-D-aspartate (NMDA) receptor antagonist, D-2-amino-5-phosphonovalerate (APV), not only blocked the paired pulse facilitation but also reduced the amplitude of the EPSP, indicating the involvement of the NMDA-receptor in synaptic responses of pilocarpine-treated dentate granule cells. Dendrites of these neurons showed loss of spines and beaded branches. These findings suggest that a degenerating dendrite could be a morphological substrate of neuronal hyperexcitability mediated by NMDA receptors, implicating possible in vivo glutamate toxicity as an underlying mechanism of chronic epilepsy.

Although cell loss [2] and neuroanatomical distortion [18] are well-documented epileptic pathology in humans, the morphological substrate for epileptiform discharges in animal models is still unclear. Some models such as the alumina cream focus [13], kainate lesion [15], and pilocarpine-induced status epilepticus [8] produce similar brain damages to those in human epilepsy, but others generate quite active epileptiform discharges without any apparent structural changes [6, 16]. Recent slice studies of human epileptic hippocampus revealed that Nmethyl-D-aspartate (NMDA) receptor-mediated longlasting excitatory postsynaptic potentials (EPSPs), which are indicative of epileptic hyperexcitability, were generated in neurons that showed loss of spines and nodule formed dendrites [10, 11]. However, limited opportunities for obtaining human tissue and the special circumstances of preparing human brain slices make it difficult to systematically study NMDA-receptor mediated epileptiform activities in relation to morphological alterations. In this paper, we report that a pilocarpine model of chronic limbic seizures produces dendritic spine loss and nodule formation similar to what have been observed in human epileptic hippocampal neurons. In such neurons, hyperexcitability was identified by evoking multiple firing by ortho- and antidromic stimulation, and by producing paired pulse facilitation. The Correspondence: M. Isokawa, Brain Research Institute, CHS, UCLA, Los Angeles, CA 90024-1761, U.S.A.

involvement of the NMDA receptor in synaptic transmission was indicated by reduction of the evoked response by the application of an NMDA receptor antagonist. These physiological and anatomical characteristics in pilocarpine-treated rat hippocampus appear to be desirable for a model to study a potential structural substrate of NMDA receptor-mediated neuronal hyperexcitability in epilepsy. Male Sprague-Dawley rats (170-250 g) received intraperitonial injection of pilocarpine (320-350 mg/kg), after being treated by methyl-scopolamine (1 mg/kg, i.p.) to minimize pilocarpine's peripheral effects. After establishing a condition under which chronic limbic seizures spontaneously recurred (6-20 weeks after the injection; see ref. 20 for details of techniques and electrographical and behavioral manifestations of this model), hippocampal slices were prepared with a thickness of 500/tm, and transferred to an interface type chamber. Artificial cerebrospinal fluid consisted of (in mM): NaCI 124, KC1 3, CaC12 2.4, NaHCO3 26, MgSO4 1.3, NaH2PO4 1.24, glucose 10, pH 7.4. Intracellular recording was performed in dentate granule cells (DGC) using glass microelectrodes filled with 0.1 M lithium chloride containing 5 % Lucifer Yellow (electrode resistance: 180-250 Mr2) or 2 M potassium acetate with 2 % biocytin (electrode resistance: 80-100 MI2). Tw o bipolar stimulating electrodes, made of twisted stainless-steel microwire (40 pm in diameter), were placed in the perforant path for orthodr,omic activation and in the CA3 stratum radiatum close to

70

the hilus for antidromic activation of DGCs (0.01-0.5 mA, 200/ts, 0.2 Hz). o-2-Amino-5-phosphonovalerate (APV; 50/zM) was given as bath application. The intracellularly recorded neurons were stained at the end of recording by applying negative current pulses (0.5-1.0 nA, 300 ms) for several minutes. Physiological data were stored on a video cassette recorder through a digital data adaptor (PCM, Unitrade) for later analyses. Lucifer yellow-filled neurons were viewed and photographed using fluorescent microscopy. Biocytin-filled neurons were

processed for visualization [9], then examined by light microscopy. No differences were found between Lucifer yellow-filled neurons and biocytin-filled neurons in physiological or anatomical results, so that the data were pooled together for the analyses. Fifteen DGCs in 4 pilocarpine-injected rats were compared with 21 DGCs in 8 normal rats, physiologically and morphologically. In normal DGCs, the average resting membrane potential (V~mp) was -59.2 mV + 1.5 S.E.M., and the average membrane resistance (Rm) was

a

'i!

0

"

15 mV

10 mV

I

5

mV

lOOm

b C

$ -,V

~.

"

11~ msec

1

lOu~

SmV ................................

i

lOmmc

d

•

•

.:,

..

•

.~,, ,~,

.

.i

J

I

IJ mV

6 lOmV 15 mV

I

S

Fig. 1. Membrane properties and synaptic responses of dentate granule cells (DGC) in normal and pilocarpine-treatcd rats. The current-voltage relationship of a normal D G C (a) and a pilocarpine D G C (d). In normal DGCs, pcrforant path stimulation produced EPSPs (b), and an action potential was accompanied by subsequent IPSPs (c). In pilocarpine DGCs, the amplitude of the EPSPs was more than doubled (e), and the increased stimulus intensity evoked an action potential and a late EPSP component (f). Paired pulse facilitation was observed in pilocarpine DGCs with an interstimulus interval of 100 ms (g), and this facilitation was blocked by APV (h). Note the voltage calibration for h is one-third of that in g. A burst discharge was elicited on a long-lasting EPSP (i). Hyperpolarizing current injection isolated an underlying EPSP from the burst firing (j). Multiple spike generation could be elicited by CA3 stimulation (k). Action potentials in a, c, d, f, g, i and k are truncated.

71 41.0 MI2 + 6.3 S.E.M. The current-voltage relationship of a normal DGC is shown in Fig. la. Upon direct depolarization, a fast sodium spike was elicited. Pefforant path stimulation produced EPSPs in 19 normal DGCs with an average latency of 4.4 ms + 0.3 S.E.M., an average duration of 11.7 ms + 2.6 S.E.M., and an average amplitude of 2.2 mV + 0.4 S.E.M. (Fig. lb). Inhibitory postsynaptic potentials (IPSP) were present subsequent to EPSPs in 8 neurons with an average duration of 92.2 ms ___ 24.9 S.E.M. and an average amplitude of 2.2 mV + 0.5 S.E.M. (Fig. lc). Confirmation of this hyperpolarization as an IPSP but not a hyperpolarizing afterpotential was accomplished by comparing the morphology of action potentials evoked synaptically and evoked by a direct positive current injection. In pilocarpine-treated rat DGCs, the average Vrmp was -61.1 mV + 2.9 S.E.M. and the average Rm was 29.3 MI2 + 3.4 S.E.M. (n = 15). The Rm was significantly lower in pilocarpine DGCs than in control DGCs (P<0.001). The currentvoltage relationship of .',o of piloearpine DGCs is shown in Fig. ld at Vnnp = - 6 1 mV (Rm = 32 MI2). Depolarizing current injection produced multiple firing with long-lasting depolarization presumably due to voltage dependent calcium conductance. When neurons were orthodromically activated, a primary response was an EPSP (Fig. le). No IPSPs were observed at resting potential. The amplitude and the duration of these EPSPs were considerably increased compared to those of normal DGCs. Weak stimulation produced EPSPs with an average latency of 4.6 ms __+ 0.9 S.E.M., an average duration of 51.0 ms _ 12.7 S.E.M., and an average amplitude of 6.8 mV _+ 1.0 S.E.M. (n = 13). When stimulus intensity was increased, a short-latency action potential was generated with a hardly identifiable EPSP-rising phase but with a long-lasting high amplitude EPSP (Fig. If). This late EPSP (arrowhead in f) had an average latency of 5.4 ms ___ 0.2 S.E.M., an average duration of 66.8 ms + 11.0 S.E.M., and an average amplitude of 18.7 mV + 14.3 S.E.M. Paired-pulse stimulation with interstimulus interval of 100 ms produced facilitation of the second response by increasing its amplitude threefold (Fig. lg). This paired pulse facilitation was blocked by the application of NMDA receptor antagonist, APV (Fig. lh). APV also reduced the overall amplitude of the EPSPs to both stimuli (Note, the voltage scale in h is one-third of that in g). A burst discharge was elicited from the long-lasting EPSP (Fig. li). Although the latency of these multiple spikes was very short so that an underlying EPSP rising phase was hard to identity, a direct negative current injection isolated the underlying EPSP (Fig. lj). Hyperpolarizing current injection strongly attenuated the duration and amplitude of the EPSP, indicating that this long-lasting EPSP was partial-

ly voltage-dependent, which supports the involvement of NMDA-receptor activation in the generation of this response. In some neurons, multiple spikes could be evoked by antidromic stimulation of CA3 stratum radiaturn (Fig. lk). Representative dendritic morphology of a normal DGC and pilocarpine DGC is presented in Fig. 2. A camera lucida drawing of the normal DGC, for which physiology is shown in Figure la-c, revealed well-developed dendritic arbors (Fig. 2a) and regularly-distributed spines as shown by Lucifer yellow in proximal (Fig. 2c) and distal dendrites (Fig. 2d). Dendritic spines were clearly observed in neurons filled with biocytin as well (Fig. 2e). In contrast, in pilocarpine DGCs, the proximal dendrite became extremely large in diameter dividing into only a few branches (Fig. 2b). A photomicrograph in Fig. 2f shows a soma and proximal dendrites of the neuron shown in Fig. 2b, for which physiology is shown in Fig. ld-h. The number of dendritic branches were decreased, the dendritic arbor was severely restricted, and most of the dendrites lost spines forming dendritic beading in both proximal (Fig. 2g, i) and distal (Fig. 2h, j) portions. However, these morphological deformities were not regional and showed no obvious relationship to the vascular supply or general tissue hypoxia. Dendrites were degenerated to varying degrees. In several cases, some dendrites lost spines and developed beading while others retained well-distributed spines. Low spine density appeared to be an initial stage to form nodules leading to heavy beading. No consistent proximal-distal gradient was observed in nodule formation. Even if a particular dendritic branch became beaded in distal parts, other branches at the same proximal-distal location often retained spiny configurations. The present findings demonstrate marked changes in membrane properties, synaptic responses, and dendritic morphology of DGCs in pilocarpine-treated rat hippocampus. Pilocarpine DGCs showed (1) a smaller Rm than controls, (2) multiple spike generation upon direct depolarization, (3) long-lasting and high-amplitude EPSPs that were voltage-dependent, stimulus intensitydependent and APV-sensitive, indicating the involvement of NMDA receptors in their synaptic transmission, (4) paired pulse facilitation, and (5) loss of spines and nodule formation in dendrites. The presence of an NMDA component in the EPSPs of pilocarpine DGCs is in agreement with results from the kindling model [14], bicuculline/picrotoxin-treated epileptic hippocampus [5, 7] and human epilepsy [21]. However, unlike kindled DGCs that show increased Rm with paired-pulse inhibition [14], pilocarpine DGCs show decreased Rm with paired-pulse facilitation. The decreased Rm observed in pilocarpine DGCs may be correlated with enhanced

72

i

b

Fig. 2. Somal and dendritic morphology of DGCs in normal and pilocarpine-treated rat hippocampus. A camera lucida drawing of a normal DGC (a) and a high magnification photomicrograph of dendritic spines of this neuron (c: Lucifer yellow). Dendritic spines were also present in further distal branches (d: Lucifer yellow). Spines were equally clearly observed in neurons injected with biocytin (e). A camera lucida drawing of a pilocarpine DGC is shown in b. The soma and proximal dendrites of the neuron in b are presented in the photomicrographs f, g, h and j. Dendrites formed nodules at proximal parts (g), and the pathological severityincreased at the distal ends showing beaded configurations (h and j). Beaded dendrites could also be visualized by Lucifer yellow(i).

N M D A receptor activation. One possible explanation is that the number of tonically active N M D A receptors [17] might be increased at resting potential, and these would be less likely to be influenced by steady state passive membrane voltage changes upon hyperpolarization. Such an increase in active N M D A receptors at rest could be due to a change in N M D A receptor sensitivity [4]. Alternatively, the participation of N M D A receptors in synaptic transmission and decreased R m may be attributed to altered dendritic morphology, since the geographical patterns of excitatory and inhibitory inputs are altered in degenerating neurons and such an alteration undoubtedly changes the input-output relationship of the neurons. Although neuronal necrosis has been documented in epileptic brains, the limited observation of cell pathology by light microscopy has made its interpretation difficult [1]. However, evidence has been accumulating in recent years to show that such cell death and morphological deformities do represent a 'true in vivo event' in epileptic brains and that they occur independently from energy

failure (for review see ref. 19). The neurotoxic effect of glutamate through the N M D A receptor has been welldocumented in cortical neurons in culture [3], retinal ganglion cells in vivo [4] and in vitro [12], indicating that increased N M D A receptor activation can induce dendritic spine loss, nodule formation and altered dendritic arbors. In the present study, we have shown N M D A receptor-mediated hyperexcitability in dendritically-deformed hippocampal neurons. This evidence indicates that N M D A receptors appear to be involved in both structural and functional alterations through glutamatemediated neurotransmission. Although it is still premature to draw any conclusions regarding whether the distorted neuron morphology observed here is a cause or an effect of neuronal hyperexcitability, it is possible that morphologically deformed n e u r o n s can generate N M D A - r e c e p t o r mediated hyperexcitability. These morphological changes might provide a long term environment in which the N M D A receptor could be persistently activated to play a key role in generating functional hyperexcitability in in vivo chronic epilepsy.

73

Supported by NIH Grant NS 02808. L.M. is a recipient of a FAPESP (Brazil) fellowship. We are grateful to D.M. Finch for his assistance in conducting this research and C.L. Wilson for his comments on the manuscript. I Brierley, J.B., Cerebral hypoxia. InW. Blackwood and J.A.N. Corsellis (Eds.), Greenfield's Neuropathology, 3rd edn. Arnold, London, 1976, pp. 43-85. 2 Brown, W.J., Structural substrates of seizure foci in the human temporal lobe. In M.A.B. Brazier (Ed.), Epilepsy, Its phenomena in man, Academic Press, New York, 1973, pp. 339-374. 3 Choi, D.W., Ionic dependence of glutamate neurotoxicity, J. Neurosci., (1987) 413~42. 4 Cline, H.T. and Constantine-Paton, M., NMDA receptor agonist and antagonists alter retinal ganglion cell arbor structure in the developing frog retinotectal projection, J. Nenrosci., 10 (1990) 11971216. 5 Dingledine, R., Hynes, M.A. and King, G.L. Involvement of Nmethyl-o-aspartate receptors in epileptiform bursting in the rat hippocampal slices, J. Physiol., 380 (1986) 175-189. 6 Goddard, G.V., Mclntyre, D.C. and Leech, C.K., A permanent change in brain function resulting from daily electrical stimulation, Exp. Neurol., 25 (1969) 295-330. 7 Hablitz, J.J., Picrotoxin-induced epileptiform activity in hippocampus: role of endogenous versus synaptic factors, J. Nenrophysiol., 51 (1984) 1011-1027. 8 Honchar, M.P., Olney, J.W. and Sherman, W.R., Systemic cholinergic agents induce seizures and brain damage in lithium-treated rats, Science, 220 (1983) 323-325. 9 Horikawa, K. and Armstrong, W.E., A versatile means of intracellular labeling: injection of biocytin and its detection with avidin conjugates, J. Neurosci. Methods, 1 (1988) 323-325. 10 Isokawa, M., Babb, T.L. and Engel, J. Jr., Physiological properties of dentate granule cells with dendritic deformities in human epileptic hippocampal slices, Soc. Neurosci. Abstr., 15 (1990) 309.

11 Isokawa, M., Levesque, M., Babb, T.L. and Engel Jr., J., Excitability and synaptic responses in morphologically deformed.dentate granule cells in human epileptic hippoeampal slices, Epilepsia, 31 (1990) 624. 12 Karschin, A., Aizenman, E. and Lipton, S.A., The interaction of agonists and noncompetitive antagonists at the excitatory amino acid receptors in rat retinal ganglion cells in vitro,'L Neurosci., 8 (1988) 2895-2906. 13 Kopeloff, L.M., Barrera, S.E. and Kopeloff, N., Recurrent convulsive seizures in animals produced by immunologic and chemical means, Am. J. Phychiat., 98 (1942) 881. 14 Mody, I., Stanton, P.K., Heinemann, U., Activation of N-methylD-aspartate receptors parallels changes in cellular and synaptic properties of dentate gyrus granule cells after kindling, J. Neurophysiol., 59 (1988) 1033-1054. 15 Nadler, J.V., Kainic acid as a tool for the study of temporal lobe epilepsy, Life Sci., 29 (1981) 2031-2042. 16 Prince, D.A., The depolarization shift in 'epileptic' neurons, Exp. Neurol., 21 (1968) 467-485. 17 Sah, P., Hestrin, S. and Nicoll, R.A., Tonic activation of NMDA receptors by ambient glutamate enhances excitability of neurons, Science, 246 (1989) 815-818. 18 Scheibel, M.E., Crandall, P.H. and Scheibel, A.B., The hippocampal-dentate complex in temporal lobe epilepsy, a Golgi study, Epilepsia, 15 (1974) 55-80. 19 Siesjo, B.K. and Wieloch, T., Epileptic brain damage: pathophysiology and neurochemical pathology. In A.V. Delgado-Escueta et al. (Eds.), Advances in Neurology, Vol 44, Raven, New York, (1986) pp. 813-847. 20 Turski, L., Ikonomidou, C., Turski, W.A., Bortolotto, Z.A. and Cavalheiro, E.A., Review: cholinerglc mechanisms and epileptogenesis. The seizures induced by pilocarpine: a novel experimental model of intractable epilepsy, Synapse, 3 (1989) 154-171. 21 Urban, L., Aitken, P.G., Friedman, A. and Somjen, G.G., An NMDA-mediated component of excitatory synaptic input to dentate granule cells in 'epileptic' human hippocampus in vitro, Brain Res., 515 (1990) 319-322.