Cellular and synaptic basis of kainic acid-induced hippocampal epileptiform activity

Brain Research, 273 (1983) 97-109 Elsevier

97

Cellular and Synaptic Basis of Kainic Acid-Induced Hippocampal Epileptiform Activity G. L. WESTBROOK1and E. W. LOTHMAN2," 1National Institutes of Health (NICHHD) and 2Department of Neurology, Washington University School of Medicine, 660 South Euclid, St. Louis, MO 63110 (U.S,A.) (Accepted January 4th, 1983) Key words: epilepsy - - hippocampus - - kainic acid

The effects of kainic acid (KA) were studied using extracellular and intraceUular recordings in the hippocampal slice preparation. In sufficient concentrations, KA led to a loss of all evoked responses. However, the amount of drug needed for this varied according to anatomic region. CA3 was more sensitive (1/~M) than CA1 or the dentate gyrus (10/~M). These results can be understood in terms of a profound and long-lasting depolarization of neurons. Lower concentrations of KA (0.05--0.1/~M) did not change the resting membrane potential or input resistance of hippocampal pyramidal cells but produced spontaneous epileptiform activity which originated in CA3 and propagated to CA1. Epileptiform discharges were not present in the dentate gyrus. Coincident with the induction of paroxysms, the following changes were observed: (1) an increase in the excitability of CA3 and CA1 pyramidal cells as measured by a left shift in the input-output curves of evoked responses and a lowered threshold stimulus intensity necessary for activation of action potentials in single neurons; (2) augmentation and synchronization of bursting in pyramidal ceils; and (3) prolonged EPSPs without an increase in their amplitude. These findings indicate that multiple changes, involving both the properties of single neurons and synaptic connections, are involved in the development of hippocampal paroxysms and that CA3 and CA1 have different roles in the generation of these discharges. INTRODUCTION Kainic acid (KA) is a rigid dicarboxylic acid that excites neurons and, for the past decade, has been used extensively in neurobiological research22,26. More recently it has been recognized that K A is a convulsant with unique properties. Limbic structures readily display epileptiform activity when small amounts of K A are directly injected into the hippocampus or amygdala3.23,52. Systemic administrations of KA preferentially cause seizures in limbic areas and the hippocampus has a particularly low threshold for this epileptogenic effect 16.17. Such KA-induced seizures are long-lasting and gradually intensify over several hours during which time progressively more limbic structures become involved. This provides a means for studying the functional anatomy of various stages of limbic seizures in the intact brain. In addition, K A produces neuropathology similar to mesial temporal sclerosis which has been found in human patients with temporal lobe epilepsy27. Moreover, * To whom correspondence should be addressed.

K A causes epileptiform activity in isolated hippocampal slices 17. These properties make K A useful in epilepsy research, especially as a model for temporal lobe epilepsy. In particular, correlative studies, using a single agent and capitalizing on the advantages of both in vitro and in vivo preparations, may be employed to study basic mechanisms of epilepsy. Because of this and the evidence that suggests that the molecular mechanism of K A differs from that of convulsants which interfere with G A B A - m e d i a t e d inhibition4, 7, we chose to extend our electrophysiological studies of the epileptogenic actions of K A in the hippocampus slice. In the following sections we will consider in turn: (1) further details about the morphology, topography and triggering of KA-induced epileptiform events; (2) a quantitative analysis of how K A changes evoked potentials in various regions of the hippocampal formation; and (3) cellular correlates of KA-induced epileptiform events. The results indicate that various areas of the hippocampus have different roles in the generation of paroxysmal activ-

~S ity and that several changes are associated with KAinduced epileptiform activity. MATERIALS AND METHODS Hippocampal slices were obtained from albino rats (220-320 g). Techniques of preparation, maintenance, stimulation, and extracellular recording have been previously described tT. Specifically, slices were perfused with bathing media (composition: 127 mM NaCl, 2 mM KCI, 1.5 mM MgSO~, 1.5 mM CaCI 2, 25.7 mM NaHCO 3, 1.125 mM KH2PO 4, and 10 mM D-glucose) bubbled with 95% 02/5% CO 2, and maintained at 37 °C. In the initial experiments, electrodes were positioned in the stratum granulosum of the dentate gyrus and stratum pyramidale of CA1 and CA3 to record spontaneous activity and the population spikes evoked by monosynaptic activation, viz. perforant pathway (PP) to dentate gyrus, mossy fibers (MF) to CA3 and Schaffer collateral-commissural afferents (SC-C) to CA1. Additional studies of population spikes, population EPSPs and population afferent volleys were performed in CA1 according to the methods of Andersen et al.~ by simultaneously recording with two extracellular electrodes, one placed in the stratum pyramidale and one in the stratum radiatum, and stimulating the SC-C afferents. For these experiments, the perfusate flowed continuously at 1.5 ml/min. At this rate, the time required to exchange the fluid in the recording chamber was 5-7 min, as judged by introducing a dye-containing solution, so that multiple concentrations of KA could be studied for single slices in one experiment. Thirty to forty minutes of equilibration were allowed after each solution change before data were collected, and serial monitoring of evoked potentials assured stable responses. For quantitative analyses a range of responses was elicited by varying the stimulus intensity (0.1 ms duration, 1-30 V, 0.2-1 Hz) in control media, in various concentrations of KA, and after rinse with control media. For each condition (stimulus strength and concentration of KA) several (3-5) responses were measured from photographs of oscilloscope traces and averaged to construct individual points on input-output curves. In later experiments, intracellular recordings were obtained from the stratum pyramidale of CA1 and CA3 using 3 M potassium acetate micropipettes

(80-150 M ~ ) led to a capacity-compensated amplitier with bridge circuit for passing current pulses. Orthodromic and/or antidromic stimulation was used to activate and identify pyramidal cells in both regions. Resting membrane potential (RMP) was continuously monitored either on a chart recorder or digital volt meter. Cells were accepted for study if they had a steady RMP >1 50 mV, action potentials with overshoots, and typical behavior to orthodromic stimulation and current pulses34,35,47.51. Simultaneous extracellular field recordings were obtained with 2 M NaCI micropipettes (5-20 Mr2) positioned over the stratum pyramidale. Although a few cells were held throughout a change in perfusate, most neurons were studied in either control or in KA-containing solutions. Based on the initial dose-response relationships obtained with extracellular recordings, intracellular studies were done at 0.05-41.1 I¢M in CA3 and 0.05-0.3 p M in CA1. Most results were obtained at 0.05pM KA. In order to minimize the possibility of dislodging the intracellular electrodes, a perfusion rate of 0.2 ml/min was employed for the intracellular recordings. The stimulus strength in voltage (0.1 ms pulses applied to MF or SC-C for CA3 and CA1, respectively) which produced short-latency action potentials was taken as the 'threshold stimulus intensity'. The EPSP amplitude at a just subthreshold stimulus intensity (50% of trials producing action potentials) was measured from the responses which did not elicit action potentials. Input resistance was measured with hyperpolarizing current pulses (0.5 nA x 50-100 ms) with the bridge balance technique. Bridge balance was checked before and after penetrations. Only those impalements in which the balance was unambiguous and did not change during data collection were utilized in the analyses. RESULTS

Topography, morphology, timing and modulation of kainic-acid-induced spontaneous and triggered interictal spikes. Others have identified the extracellular field potentials consisting of multiple, tightly grouped, highfrequency discharges (bursts or interictal spikes) that appear in hippocampal slices after exposure to convulsants that block GABA-mediated inhibition as epileptiform events6,37-39. Similar interictal spikes (IIS)

99 occurred in KA. They appeared spontaneously (SIIS, see Fig. 1A) or could be elicited with electrical stimulation of afferents (T-IIS, see Fig. 1B). S-IIS were most prominent in the stratum pyramidale of the CA3 region where two temporal phases could be identified (Fig. 1A). S-IIS began with 20--40 ms of irregular, low-voltage undulations which were often associated with single unit action potentials (Fig. 1A, 2). Later there was a 60-100 ms positive wave of 2--4 mV amplitude with 7-20 superimposed peaked negative transients ('population spikes') with amplitudes up to 5 mV. After this, neuronal activity ceased for up to several seconds. This later phase of IIS usually contained 7-10 population spikes although paroxysms with as many as 20 were encountered. This later phase of the IIS is comparable to the epileptiform events previously stressed by others and corresponds

to the occurrence of paroxysmal depolarization shifts (PDS) with superimposed action potentials synchronized in a large population of neurons (see below). S-IIS occurred in the CA1 field, but were not always present in CA1, even if CA3 was actively generating S-IIS. Conversely, they were never seen in CA1 without coincident S-IIS in CA3. The S-IIS in CA1 differed from those in CA3. They had a shorter duration and fewer population spikes which began about 20 ms after the large amplitude population spikes of CA3. S-IIS were not detected in the dentate gyrus. The frequency of S-IIS depended on the concentration of KA present in the perfusate and duration of exposure to the drug tT. Low concentrations (0.05-1.0 a M KA) produced S-IIS that began sporadically, reached a stable rate 15-20 min after be-

A 1

2

CA 1

DG

cA3

cA3

B 1 CA 1

2 _

3 .--L_

I Fig. 1. Kainic-acid induced interictal spikes. A: morphology and topography of spontaneous kainic-acid-induced interictal spikes (S-IIS). 1, simultaneous records obtained from the stratum pyramidale of CA1 and CA3; 2, records from the stratum granulosum of the dentate gyrus and stratum pyramidale of CA3. Calibration: vertical, 5 mV for CA1 and DG, 2 mV for CA3, positive up in these and subsequent figures; horizontal, 40 ms. B: triggered interictal spikes (T-IIS) in hippocampal formation. Responses recorded simultaneously in stratum pyramidale of CA1 and CA3 during mossy fiber stimulation of two intensities, 4 V (1, 2) and 12 V (3), delivered at 1 Hz. Note the appearance of a late T-IIS with multiple high amplitude population spikes only on even number stimulus (2) and absence on odd number stimulus (1). Vertical calibration as in A; horizontal calibration, 20 ms. Concentration of KA 0.05 aM throughout.

100 ginning drug perfusion, and recurred at rates of 10/min for hours. Of note, S-1IS rate in these concentrations was not always perfectly regular but rather showed cycling, with groups of several S-IIS alternating with quiet periods. S-IIS were also inhibited for up to several seconds after tetanic stimulation of afferents. High concentrations (1.0/~M or more) produces S-IIS rates of 40-50/rain which were sustained for only a few minutes. Further results in this section were obtained with stable epileptiform discharges produced by 0.05-0.1 ~M KA. A striking effect of KA on evoked potentials was the production of multiple population spikes to afferent stimulation (T-IIS). T-IIS were recorded from CA3 and CA1 (Figs. 1 and 2) but not from the dentate gyrus. T-IIS (with 3 or more population spikes) could be elicited at 0.01-0.1/~M K A in CA3 with MF stimulation, whereas 0.3-1.0 p M was necessary to evoke T-IIS in CA1 with SC-C stimulation. In contrast, T-IIS were seen in CA1 with MF stimulation with 0.054).1 j~M KA, but were delayed, indicating that they arose as a consequence of a preceding T-IIS in CA3 (Fig. 1B). When stimuli were applied to the

PP, T-IIS appeared in CA3 with 0.01-0.1 /,M KA while the dentate gyrus evoked response was morphologically unchanged from controls. Low-intensity MF stimulation produced T-IIS in CA3 that had an initital buildup phase similar to S-IIS (Figs. 1A, 1 and 1B, 2). Stimuli at 0.5-1 Hz were alternatively effective or ineffective in eliciting delayed T-IIS in CA3 followed by a long-latency T-IIS in CA1. This alternating phenomenon was also observed in CA3 with low intensity PP stimuli. Transections to separate the various regions of the slice produced results identical to those obtained with penicillin-induced IIS 38. Transections between CA3 and CA1 blocked S-IIS in CA1, but not CA3, although SC-C stimulation distal to the cut still evoked T-IIS in CA1. S-IIS persisted in CA3 even after a second transection between dentate gyrus and CA3 completely isolated this region, confirming that KAinduced IIS are generated in CA3 and propagate to CA1. Low Ca2+/high Mg2+ solutions have been shown to abolish synaptic potentials and IIS in the slice37, but also may affect intrinsic membrane conductances im-

CA 1

CA 3

CONTROL

KA

.05

iiilI~l

I I / l ~

-

.

.--

10

CO.TROL

__y---.__.__

Fig. 2. Changes of CA1 and CA3 evoked potentials with KA. Responses recorded from stratum pyramidale of CA1 with SC-C stimulation (left) and CA3 with MF stimulation (right) in control, in various concentrations of KA (pM), and after rinse with control perfusate as indicated. Calibration: vertical, 4 mV for CA3, 8 mV for CA1 ; horizontal, 10 ms.

101 portant to PDS generation11, 47. Perfusion with 0.5 mM Ca2+/8.0 mM Mg 2÷ bathing media blocked KAinduced S-IIS. S-IIS were blocked before evoked potentials completely disappeared, and, on wash with a 1.5 mM Ca2+/1.5 mM Mg 2+ perfusate, S-IIS reappeared after evoked potentials, indicating partial blockade of synaptic transmission at the time of S-IIS suppression.

Effects of kainic acid on evoked potentials in the hippocampal slice In order to further investigate the means by which KA exerts its epileptogenic action we analyzed its effect on evoked potentials. The responses of dentate gyrus granule cells to PP input, CA3 pyramidal cells to MF input and CA1 pyramidal cells to SC-C inputs of various intensities were examined in control and KA-containing solutions using population spikes recorded with extracellular microelectrodes placed

CA 1 ~5 v

~-4 0, , 3

• : Control o : O.031.sMKA : 0.1 isM KA

/ / ~.,~o-.,c~n~c:-~

[] = 1.0 IsMKA • : IOIsMKA

/

0

Z

0

Intracellular studies on the effect of KA on hippocampal pyramidal cells.

STIMULUS(V) >8 E

n. 4 © I--

,-r- O, ~ o .

4

6

8

over the cell bodies in the appropriate regions. If the concentration of KA was high enough all evoked responses were blocked (Fig. 2), but returned to control levels if the tissue was rinsed with KA-free perfusate. The concentration of KA needed for this was 1 /~M in CA3 and 10 p M in both CA1 and the dentate gyrus. Below these concentrations the evoked responses were altered in a dose-dependent manner with variations in regional sensitivities. The first observed effect was a shift to the left in the input-output curves for population spikes which occurred with concentrations as low as 0.01/~M KA in CA3 with MF stimulation and 0.03/~M in CA1 with SC-C stimulation. With higher concentrations there was a further leftward shift and the appearance of multiple population spikes (Fig. 3). These changes were maximal with 0.1 ~M KA in CA3 and 1/~M in CA1. The responses in the dentate gyrus showed a shift to the left in the threshold for population spikes in 1-10/~M KA but no more than two population spikes were ever elicited with PP stimuli. The electrophysiological effects of K A were further characterized by measuring the population spikes, population EPSPs and presynaptic afferent population volleys in CA1 according to the method of Andersen et al. 1. In contrast to the impressive changes in population spikes, neither the presynaptic population volley nor the EPSP were altered by KA (Figs. 4 and 5) until a concentration of 10 ~M was reached when the population EPSP disappeared.

10 12

15

Fig. 3. KA dose-response for CA1 population spike input--Output curves. Ordinate represents responses (above - - number of population spikes, below - - amplitude of first population spike) evoked by different voltages SC-C stimulation (abscissae) in media with different concentrations of KA (see symbols in inset). Note progressive augmentation in population spike size and number up to 1 juM, loss of response at 10/~M and return of responses to normal with rinse in control medium.

A total of 55 pyramidal cells were accepted for study (Table I). Although our control perfusate had a lower K + concentration than previous publications, neuronal behavior as measured by input resistance, resting membrane potential, spontaneous activity, and response to orthodromic stimulation were comparable 34,35,47,51. The spontaneous activity of CA1 neurons (in the C A l b subregion where our recordings were obtained) consisted only of single action potentials (Fig. 6 and see ref. 20). In contrast, CA3 neurons showed several types of spontaneous discharges. All had bursts of 5-7 action potentials riding on a 10-15 mV, 50-75 ms depolarizing envelope, as well as single, double, or triple action potentials with little or no associated depolarizing wave (Fig. 6).

102

CONTROL

KA

SP SR

SR\ Fig. 4. Effect of KA on evoked responses in dendritic and cell body layers of CA1. Extracellular records from stratum pyramidale (SP) and stratum radiatum (SR) with 8 V SC-C stimulus at two different sweep speeds. Note augmentation of population spike without alteration of population volley or population EPSP. Calibration: vertical, 7 mV for SP, 3.5 mV for SR; horizontal, 20 ms (upper panels), 2 ms (lower panels). KA: 0.05 pM.

CA| 2 E

o>l o

3

4

6 8 10 15

30

~8

.---

3

4

6 8 10 15

30

~8

~ &

tu

,-

3

4

6

812

15

30

3 4.-6:~-10 ~ 15

30

Fig. 5. Effect of KA on different components of CA 1 field potential input-output curves. Amplitudes of population spikes, population EPSPs and presynaptic volleys are plotted along the ordinates for different intensity SC-C stimuli (abscissae) in control (triangles) and 0.05 aM KA (circles). Same experiment as in Fig. 4.

103 TABLE I

Comparison of CA3 and CA1 pyramidal cells in control and KA-containing solutions Results expressed as mean + S.D.

CA1 Control KA (0.05-0.3pM)

CA3 Control KA (0.05--0.1MM)

Resting membrane potential (mV)

Input resistance (M I2)

Threshold stimulus intensity (V)

56.1 (n = n.s. 56.7 (n =

23.5 (n = n.s. 27.2 (n =

+ 5.8 4)

6.4 + 1.1 (n = 10) 0.02 4.5 + 1.0 (n = 4)

30.1 + 6.2 (n = 5) n.s. 34.0 + 7.4 (n = 4)

8.0 + 1.4 (n = 10) 0.001 4.2 + 1.3 (n = 4)

+ 5.0 10) + 7.1 15)

57.8 + 8.2 (n = 19) n.s. 58.3 + 8.4 (n = 11)

+ 8.9 4)

EPSPs Amplitude (mV)

Duration (ms)

3.5 + 1.4

12.3 + 2.8 (n = 6)

n.s. 4.6 + 2.4

0.05 23.7 + 11.7 (n = 4)

4.2 + 1.8

9.0 + 3.5 (n =

10)

n.s. 4.1 + 1.8

0.001 37.4 + 10.5 (n = 7)

P values for two sided t-test; n.s., not significant.

D u r i n g any of these s p o n t a n e o u s single cell discharges, there was n e v e r an associated field p o t e n t i a l in CA1 or C A 3 . O r t h o d r o m i c s t i m u l a t i o n of b o t h C A 1 a n d C A 3

CA 1

CA 3 3

p r o d u c e d an E P S P - I P S P s e q u e n c e which was followed by a single action p o t e n t i a l as stimulus i n t e n sity was increased. T h e a m p l i t u d e a n d d u r a t i o n of the EPSPs at a stimulus i n t e n s i t y n e a r t h r e s h o l d (50% of stimuli p r o d u c i n g action potentials) was similar for CA1 a n d C A 3 ( T a b l e I). A c t i o n potentials were followed by b o t h an a f t e r h y p e r p o l a r i z a t i o n ( A H P ) of 2-5 m V a m p l i t u d e a n d 25-100 ms d u r a t i o n a n d a n aft e r d e p o l a r i z a t i o n ( D A P ) i n t e r p o s e d b e t w e e n the action p o t e n t i a l a n d the A H P 12,49. R a r e l y , very stong stimuli led to l o n g e r D A P s with a second action potential, but only in C A 1 n e u r o n s . N o m o r e t h a n two action potentials were ever seen in c o n t r o l recordings which were confined to the C A l b subregion. Depolarizing c u r r e n t pulses in C A 3 n e u r o n s p r o d u c e d bursts similar to s p o n t a n e o u s bursts11.32,47. In contrast, o n l y repetitive action p o t e n t i a l s (Fig. 6) were seen in C A l b n e u r o n s with depolarizing current20,35. C A 3 a n d C A 1 p y r a m i d a l cells r e c o r d e d in the concentrations of K A that p r o d u c e d e p i l e p t i f o r m discharges did n o t differ from control n e u r o n s in their input resistance or resting m e m b r a n e p o t e n t i a l (Table I). This o b s e r v a t i o n was s u p p o r t e d b y recordings from two C A 1 a n d two C A 3 cells held t h r o u g h a

_ _

I

!

I

I

Fig. 6. Activity of pyramidal cells in control media. On the left are recordings obtained from CA1 cells. Spontaneous single action potentials are displayed at end of upper most panel preceded by single action potential-IPSP sequence elicited with a SC-C stimulus. Below this the response of a CA1 cell to 0.3 nA depolarizing current is illustrated. On the right are the various forms of spontaneous discharges of CA3 cells. Calibration: vertical, 40 mV in 1, 20 mV in 2 and 3, 15 mV in 4 and 5; horizontal, 20 ms in 2, 3, 5, 40 ms in 4, 150 ms in 1. Action potentials retouched in this and subsequent figures as needed for clarity.

104

change from control to KA-containing solution during which time epileptiform activity developed. However, KA produced notable changes in the spontaneous and evoked activity of these neurons. K A (0.05-0.1 ~M) produced synchronization of cellular bursts in CA3 neurons so that IIS were observed simultaneously in the extracellular field. Synchronized bursting occurred spontaneously and in response to orthodromic stimultation (Figs. 7 and 8). The stimulus intensity necessary to produce action potentials (or bursts in KA) was also reduced in KA in both CA1 and CA3 neurons (Table I). This corresponded to the 'left shift' of the input-output curves for field responses (Figs. 3 and 5). Orthodromic stimulation of MF near threshold at 0.5-1.0 Hz. led, on successive stimuli, to prolonged (up to 50 ms) EPSPs alter-

nating with delayed PDS bursts in CA3 (Fig. 7A). This is the cellular correlate of the alternating T-IIS described above. In KA the EPSPs in both CA1 and CA3 were similar in amplitude to controls, but were greatly prolonged (Fig. 7, Table I). Spontaneous EPSPs were more frequent than in control, and prolonged depolarizations ('complex EPSPs') were seen (Fig. 8). These complex depolarizations were not due to IlS being 'seen' through the membrane as they were observed both in the absence and presence of field paroxysms. Complex EPSPs occurred spontaneously, in synchrony with field bursts in weakly bursting slices, or could be evoked with orthodromic stimulation (Fig. 8). These events are most likely polysynaptic potentials evoked by increased excitation in the net-

KA-CA 3 A

IC - ~

2

EC I il

Jk I

L

B EC

I1,,,, .

t_

IC Fig. 7. Cellular correlates of KA-induced paroxysms. A: intracellular recording of CA3 pyramidal cell (left) and extraceUular recording of field potential tight obtained just after coming out of cell (right). Traces are sequential responses to 5 V MF stimulus applied at 0.5 Hz., showing failure (1) and appearance (2) of delayed paroxysm. Note prolonged EPSP with amplitude similar to controls in first intracellular trace. B: simultaneous intraceltular (lower trace) and extracellular (upper trace) in CA3 with stronger intensity (10 V) stimulus to MF. Note short latency of paroxysm which was present with each stimulus. Calibration: vertical, 20 mV in intracellular traces in A, 8 mV extracellular traces in A, 10 mV in B; horizontal, 10 ms intracellular trace in A, 20 ms extracellular trace in A, 25 ms in B. KA 0.05/~M.

105 studied in detail, changes in afterpotentials following single action potentials or bursts were not detected in KA. In KA, depolarizing current pulses in CA3 neurons produced bursts similar to control. Current pulses in C A l b neurons which produced only repetitive APs in control perfusate produced bursts in KA solution (Fig. 8D). Fast prepotentials and 'd-spikes' were often present in isolation or underlying APs in bursting neurons, especially those showing non-PDS bursts. Higher concentrations of KA led to cessation of S-IIS in CA3 after a period of activity, and impalements revealed depolarized neurons 32. In this case, no S-IIS or spontaneous cellular bursts were seen in CA1, but SC-C stimulation led to T-IIS in the CA1 field and typical bursts in CA1 neurons (Fig. 8A).

work. They did not resemble the large, smooth depolarizing events seen in depolarized or injured neurons. No giant EPSPs (greater than 10 mV) were seen in KA. As previously noted, 0.05/~M KA sometimes produced S-IIS in CA3 before or when no S-IIS appeared in CA1. However, the effects of KA on CA1 neurons was still apparent as an increase in spontaneous and complex EPSPs and occasional spontaneous bursts in some neurons without accompanying field bursts (Figs. 8 and 9). When S-IIS appeared in CA1 intracellular recordings showed pyramidal cell bursts. Both 'PDS bursts' and bursts with little underlying depolarization ('non-PDS bursts') were seen as described with penicillin-induced epileptiform paroxysms37. Bursts were followed by an afterhyperpolarization lasting up to 1-2 s. Although this has not been

KA-CA 1 A

C 1

I|

2 B 0

x._

i

1

2

Fig. 8. Bursting induced in CA1 pyramidal cells by KA. A: PDS evoked in CA1 cell with SC-stimulus in 0.3/~M KA. B: spontaneous burst of CA1 neuron in 0.1/~M KA. C: spontaneousnon-PDSburst of CA1 pyramidalcells in 0.05/~M K A is registered in intraceilular recording (2) without associated field paroxysmsin simultaneous extracellular recording (1). D: burst induced in CA1 pyramidal cell (1) with 0.3 nA current pulse as indicated by current trace (2). Calibrations: vertical, 10 mV in A and B, 1.5 mV in C1, 15 mV in C2, 25 mV in D1; horizontal, 15 ms in A, 40 ms in B, 20 ms in C and D.

106

KA

CA3

CA1

1 2

B

C

E

I Fig. 9. Prolonged, complex depolarizations in pyramidal cells in KA. On the left are shown intracellular recordings from CAI showing complex EPSPs leading up to (A) and preceded by a spontaneous single action potential (B). Below (C) is shown a single action potential following by a complex EPSP following a SC-C stimulus. D: recordings from CA3 show complex depolarization and single action potential preceding a spontaneous paroxysm, indicated by the burst in the intracellular recording (t) and field IIS (2). E: intracellular recording of an isolated, complex, prolonged depolarization in a CA3 pyramidal cell. Calibrations: vertical 20 mV in A, B, C, D I, 15 mV in E, 10 mV in D2; horizontal, 5 ms in A, 10 ms in B, 25 ms in C, D, E. KA: 0.05/~M. DISCUSSION These studies show that K A has several actions in the hippocampal slice which depend on the dose used and region studied. The two major effects can be separated into depolarization and epileptogenesis. Using the disappearance of evoked responses as an index of profound depolarization, we found that 10#M KA was needed to produce this effect on CA1 pyramidal cells and dentate gyrus granule cells while 1 ~M was effective in CA3 pyramidal cells. The concentration needed to produce epileptiform discharges in both CA3 and CA1 was 0.05-0.1/~M K A or 20-200 times lower than the depolarizing concentration, depending on the region considered. Our observations agree with the intraceilular recordings of Robinson and Deadwyler32 who found that bath application of 0.1 ktM K A did not depolarize CA3 pyr-

amidai cells but increased their bursting. These authors also reported that 1/~M KA in CA3 and 10/~M KA in CA1 led to a profound and sustained depolarization of these cells along with a cessation of action potentials. While intracellular recordings of the effects of K A on dentate gyrus neurons are not available, it is likely that they too are tonically depolarized by 10/~M KA. The major changes observed at the epileptogenic concentration of K A were: (1) augmented bursting in pyramidal cells; (2) increased excitability of all pyramidal cells as measured by a left shift of the inputoutput curves in the field responses and a lowered threshold stimulus intensity for single cell firing; and (3) synchronization of burst discharges of CA3 pyramidal cells. The paroxysmal depolarizing shift (or PDS) is now accepted as the neuronal substrate of epileptiform in-

107 terictal spikes2,5,21,29-31. The propensity of hippocampal pyramidal cells to such behavior, often called bursting, has made the hippocampal slice particularly useful for experimental investigations of basic mechanisms of epileptogenesis. Several observations reported here show that concentrations of KA which induce spontaneous epileptiform discharges augment bursting in hippocampal cells. Since CA3 pyramidal cells burst spontaneously or with current injection in control media, the presence of burst activity in these neurons in KA does not represent a qualitative change. However, in the presence of KA, bursts could easily be triggered by orthodromic stimulation which was not the case for control neurons. This represents a quantitative change in the ease in which this paroxysmal event could be elicited. In contrast, C A l b pyramidal cells which do not normally burst20,35, are converted to a burst mode in KA. This represents a qualitative change in these neurons. Based on experiments with other convulsants than KA, particularly penicillin, several mechanisms have been offered for the cellular mechanisms involved in pyramidal cell bursting. There is good evidence to support the role of intrinsic membrane properties in hippocampal pyramidal cell bursting which could arise from imbalances in slow inward currents carried by calcium ions and outward currents carried by potassium i0ns9,11,30,31,36,40. Alternatively, bursting has been attributed to decreased GABA-mediated inhibition6,39,48 or altered excitatory postsynaptic potentials 10. The results of our experiments are consistent with several of these hypotheses, but do not conclusively prove any single basic mechanism for KA-induced epileptiform discharges. The ability of somatic current pulses to produce bursts in C A l b pyramidal cells suggests that the biophysical properties of individual hippocampal cells may change in the presence of KA. The appearance of 'd-spikes' and fast prepotentials could be due either to intrinsic membrane changes, unmasking by disinhibition4S, or to electronic connectionsl9. Since KA is commonly believed to be an analogue of putative excitatory amino acid neurotransmitters4,7,22, it might be expected that KA would affect EPSPs. We did not find evidence of an increase in EPSPs amplitude in KA, but cannot totally exclude an interaction of KA with excitatory synapses, since we observed prolonged EPSPs. The prolongation could represent a diminu-

tion of IPSPs6, 48. On the basis of binding and pharmacological considerations4,7,22, 26, one would not expect KA to interfere with GABA-mediated inhibition. Nontheless, Sloviter and Damiano 41 have suggested that KA does decrease inhibition in the hippocampal formation. These workers used a paired pulse stimulus paradigm and found that the magnitude of the second (test) population spike was increased after KA. However, their records also show an increase in the size of the population spike produced by the first (conditioning) stimulus. This fact precludes a straightforward conclusion of decreased inhibition 14. Thus the question of whether or not KA affects inhibition remains unanswered at present and is being studied in our laboratory with intracellular recordings. On the other hand, the in vivo studies of Sloviter and Damiano agree with our in vitro results that KA increases the excitability of hippocampal neurons. Whatever the basic mechanism of bursting in single pyramidal cells, these experiments show that both the spread of paroxysms and synchronization of bursting are also important in the expression of KA-induced epileptiform activity. Furthermore, synaptic processes seem critical in these events. The morphology of S-IIS, the relative timing of S-IIS simultaneous recordings from different hippocampal regions, and transection experiments indicate that various regions have different roles in KA-induced paroxysms as in penicillin-induced paroxysms39,45,46. The dentate gyrus does not participate. This may relate to the fact that these neurons are known to be resistant to bursting25,31. On the other hand, neurons in Ammon's horn are involved in KA-induced paroxysms with CA3 operating as a 'pacemaker' wherein S-IIS originate and CA1 as a 'follower' when bursts are simultaneously triggered in pyramidal cells by discharges relayed over Schaffer collaterals. Bursts of action potentials in CA3 pyramidal cells during S-IIS conducted to the Schaffer collateral terminals could account for CA1 pyramidal cell bursts. In addition, the increased excitability of CA1 pyramidal cells and unmasking of intrinsic membrane properties that lead to bursting (see above) would serve to amplify the epileptic activity. In either case, it is important to note that IIS do not arise spontaneously in CA1, and synchrony of bursting is dependent on synaptic input into CA1. The fact that S-IIS do not al-

108 ways occur in CA1 in spite of a strong input from CA3 (where IIS are occurring) could be due to cutting slices 'off line' or due to a less than 1:1 transfer of impulse traffic in the Schaffer collaterals to the CA1 subregion so that the critical input necessary to activate bursting in CA1 is not provided. In contrast, there are means for synchronization of bursting intrinsic to CA3. Several observations presented here are consistent with the idea that synaptic interactions are important in synchronization - - the initial buildup phase of S-IIS, the prolonged depolarizations in CA3 pyramidal cells preceding either spontaneous PDS or IIS triggered by low intensity stimuli, and the ease with which S-IIS are blocked by divalent cation manipulations. Both anatomical and physiological data suggest the existence of excitatory synaptic connections between CA3 pyramidal cells that could operate in this process 13,15,18,43,50. The increased excitability and augmented bursting of CA3 pyramidal cells would add a positive feedback in such a circuit until a critical mass of neurons generate a burst which is then terminated in an afterhyperpolarization. Other possible pathophysiological substrates for the synchronization include electronic connections between CA3 pyramidal cells 19,33.44, ephaptic interactions, or changes in extracellular ionic concentrations, particularly potassiumS, 42. Further studies are needed to clarify which mechanisms are involved. The basic pathophysiology of epileptogenesis in the hippocampal slice, both for K A and pen-

icillin, seem to involve alterations not only in the behavior of single neurons but also in the synaptic connections within a population of neurons 24,38,45,4% This suggests that a variety of fundamental alterations can activate a stereotyped sequence of events in hippocampal epileptogenesis. In addition to its ability to produce epileptiform discharges, K A is a potent agent for killing neurons in the hippocampal formation. There is a graded sensitivity to this effect, the order being CA3 > CA1 > D G 26. Two hypotheses have been offered for the neurotoxicity of KA, either that the drug kills neurons by tonic depolarization or by inducing epileptic discharges16, 23.27,2s. As discussed above, there is evidence that KA does strongly depolarize neurons. However, if this were a sufficient condition for the neurotoxicology of KA, then the potential for destruction of CA1 pyramidal cells should be the same as that for dentate gyrus granule cells which is not the case. On the other hand, the observations that K A produces more intense epileptiform activity in CA3 than in CA1, but none in D G , supports the epileptickill theory.

REFERENCES

6 Dingledine, R. and Gjerstad, L., Reduced inhibition during epileptiform activity in the in vitro hippocampal slice, J. Physiol. (Lond.), 305 (1980) 297-313. 7 Evans, R. H. and Watkins, J. C., Pharmacological antagonists of excitant amino acid action, Life Sci., 28 (1981) 1303--1308. 8 Fisher, R. S., Pedley, T. A., Moody, W. J. Jr. and Prince, D. A., The role of extracellular potassium in hippocampal epilepsy, Arch. Neurol., 33 (1976) 76-83. 9 Hotson, J. D. and Prince, D. A., Penicillin and barium-induced epileptiform bursting in hippocampal neurons: actions on Ca ++ and K ÷ potentials, Ann. NeuroL, 10 (1981) 11-17. 10 Johnston, D. and Brown, T. H., Giant synaptic hypothesis for epileptiform activity, Science, 211 (1981) 274--277. 11 Johnston, D., Hablitz, J. J. and Wilson, W. A., Voltage clamp discloses slow inward current in hippocampal burstfiring neurons, Nature (Lond.), 286 (1980) 39t-393. 12 Kandel, E. R. and Spencer, W. A., Electrophysiology of hippocampal neurons. II. Afterpotentials and repetitive firing, J. Neurophysiol., 24 (1961) 243-259.

1 Andersen, P. et al., Possible mechanisms for long-term potentiation of synaptic transmission in hippocampal slices from guinea pigs, J. Physiol. (Lond.), 302 (1980) 463--482. 2 Ayala, G. F. et al., Genesis of epileptic interictal spikes: new knowledge of cortical feedback systems suggests a neurophysiological explanation of brief paroxysms, Brain Research, 52 (1973) 1-17. 3 Ben-Ari, Y., Tremblay, E. and Otterson, O. P., Injections of kainic acid into the amygdaloid complex of the rat: an electrographic, clinical and histological study in relation to the pathology of epilepsy, Neuroscience, 5 (1980) 15-52. 4 Cotman, C. W., Foster, H. and Lanthorn, T., An overview of glutamate as a neurotransmitter. In G. Di Chiara and G. L. Gessa (Eds.), Glutamate as a Neurotransmitter, Raven Press, New York, 1981, pp. 1-27. 5 Dichter, M. and Spencer, W. D., Pencillin-induced interictal discharges from the cat hippocampus. I. Characteristics and topographical features, J. Neurophysiol., 32 (1969) 649-662.

ACKNOWLEDGEMENTS Supported in part by USPHS Grant NS-t4834 and a grant from the Institute of Medical Education and Research of the City of St. Louis. We thank Nancy Green and Kelley Kamer for secretarial assistance.

109

13 Lebovitz, R. M., Dichter, M. and Spencer, W. A., Recurrent excitation in the CA3 region of cat hippocampus, Int. J. Neurosci., 2 (1971) 99-108. I4 Lee, K., Dunwiddie, T. V. and Holler, B. J., Interaction of diazepam with synaptic transmission in the in vitro rat hippocampal preparation, Arch. Pharmacol., 309 (1979) 131-136. 15 Lorente de No, R., Studies on the structures of the cerebral cortex. II. Continuation of the study of the ammonic system, J. Physiol. Neurol., 46 (1934) 113-177. 16 Lothman, E. W. and Collins, R. C., Kainic acid induced limbic seizures: metabolic, behavioral, electroencephalographic and neuropathological correlates, Brain Research, 218 (1981) 299-318. 17 Lothman, E. W., Collins, R. C. and Ferrendelli, J. A., Kainic acid induced limbic seizures: electrophysiologic studies, Neurology, 31 (1981) 806-812. 18 MacVicar, B. A. and Dudek, F. E., Local synaptic circuits in rat hippocampus: interactions between pyramidal cells, Brain Research, 184 (1980) 220-223. 19 MacVicar, B. A. and Dudek, F. E., Electrotonic coupling between pyramidal cells: a direct demonstration in rat hippocampal slices, Science, 213 (1981) 782-785. 20 Masukawa, L. M., Benardo, L. S. and Prince, D. A., Variations in electrophysiological properties of hippocampal neurons in different subfields, Brain Research, 242 (1982) 341-344. 21 Matsumoto, H. and Ajmone Marsan, C., Cortical cellular phenomena in experimental epilepsy: interictal manifestations, Exp. Neurol., 9 (1964) 286-304. 22 McGeer, E. G., Olney, J. W. and McGeer, P. L. (Eds.), Kainic Acid as a Tool in Neurobiology, Raven, New York, 1978, p. 271. 23 Menini, C. et al., Sustained limbic seizures induced by intraamygdaloid kainic acid in the baboon: symptomatology and neuropathological consequences, Ann. Neurol., 8 (1980) 501-509. 24 Mesher, R. A. and Schwartzkroin, P. A., Can CA3 epileptiform discharge induce bursting in normal CA1 hippocampal neurons? Brain Research, 183 (1980) 472-476. 25 Misgeld, U., Klee, M. R. and Zeise, M. L., Differences in burst characteristics and drug sensitivity between CA3 neurons and granule cells. In M. R. Klee et al. (Eds,), Physiology in Pharmacology of Epileptogenic Phenomena, Raven Press, New York, 1982, pp. 131-140. 26 Nadler, J. V., Kainic acid: neurophysiological and neurotoxic actions, Life Sci., 24 (1979) 289-300. 27 Nadler, J. V., Kainic acid as a tool for the study of temporal lobe epilepsy, Life Sci., 79 (1981) 2031-2042. 28 Olney, J. W. et al., Kainic acid: a powerful neurotoxic analogue of glutamate, Brain Research, 77 (1974) 507-512. 29 Prince, D. A., The depolarization shift in epileptogenic neurons, Exp. Neurol., 21 (1968) 467-485. 30 Prince, D. A., Neurophysiology of epilepsy, Ann. Rev. Neurosci., 1 (1978) 395-415. 31 Prince, D. A., Connors, B. W. and Bernardo, L. S., Transactions from interictal to ictal epileptiform discharges. In A. V. Escueta et al. (Eds.), Status Epilepticus: Basic Mechanisms of Brain Damage and Treatment, Raven Press, New York, in press. 32 Robinson, J. H. and Deadwyler, S. A., Kainic acid produces depolarization of CA3 pyramidal cells in the in vitro hippocampal slice, Brain Research, 221 (1981) 117-127. 33 Schmalbruch, H. and Johnson, H., Gap function on CA3

pyramidal cells of guinea pig hippocampus shown by freeze fracture, Brain Research, 217 (1981) 175--178. 34 Schwartzkroin, P. A., Characteristics of CA1 neurons recorded intracellularly in the hippocampal in vitro slice prepararation, Brain Research, 45 (1975) 423-436. 35 Schwartzkroin, P. A., Further characteristics of hippocampal CA1 cells in vitro, Brain Research, 128 (1977) 53-68. 36 Schwartzkroin, P. A., Ionic and synaptic determinants of burst generation. In J. S. Lockard and A. A. Ward Jr. (Eds.), Epilepsy: A Window to Brain Mechanisms, Raven Press, New York, 1980, pp. 83-95. 37 Schwartzkroin, P. A. and Prince, D. A., Penicillin induced epileptiform activity in the hippocampal in vitro preparation, Ann. Neurol., 1 (1977) 463-469. 38 Schwartzkroin, P. A. and Prince, D. A., Cellular and field potentials properties of epileptogenic hippocampal slices, Brain Research, 147 (1978) 117-130. 39 Schwartzkroin, P. A. and Prince, D. A., Changes in excitatory and inhibitory synaptic potentials leading to epileptogenic activity, Brain Research, 183 (1980) 61-76. 40 Schwartzkroin, P. A. and Prince, D. A., Effects of TEA on hippocampal neurons, Brain Research, 185 (1980) 169-181. 41 Sloviter, R. S. and Damiano, B. P., On the relationship between kainic acid-induced epileptic activity and hippocampal neuronal damage, Neuropharmacology, 20 (1981) 1003-1011. 42 Somjen, G. G., Influence of potassium and neuroglia in the generation of seizures and their treatment. In G. H. Glaser, J. K. Penry, and D. J. Woodbury (Eds.), Antiepileptic Drugs -- Mechanisms of Action, Raven Press, New York, 1981, pp. 155-167. 43 Swanson, L. W. and Cowan, W. M., An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat, J. comp. Neurol., 172 (1977) 49-84. 44 Taylor, C. P. and Dudek, F. E., Synchronous neural afterdischarges in rat hippocampal slices without active chemical synapses, Science, 218 (1982) 810-812. 45 Traub, R. D. and Wong, R. K. S., Penicillin-induced epileptiform activity in the hippocampal slice: a model of synchronization of CA3 pyramidal cell bursting, Neuroscience, 6 (1981) 223-230. 46 Traub, P. and Wong, P. K. S., Cellular mechanisms of neuronal synchronization in epilepsy, Science, 216 (1982) 745-747. 47 Wong, R. K. S. and Prince, D. A., Participation of calcium spikes during intrinsic bursting in hippocampal neurons, Brain Research, 159 (1978) 385-390. 48 Wong, R. K. S. and Prince, D. A., Dendritic mechanisms underlying penicillin induced epileptiform activity, Science, 204 (1979) 1228-1231. 49 Wong, R. K. S. and Prince, D. A., Afterpotential generation in hippocampal pyramidal cells, J. Neurophysiol., 45 (1981) 86-97. 50 Wyss, J. M., Swanson, L. W. and Cowan, W. M., A study of the subcortical afferents to the hippocampal formation in the rat, Neuroscience, 4 (1979) 463-476. 51 Yamamoto, C., Activation of hippocampal neurons by mossy fiber stimulation on the brain sections in vitro, Exp. Brain Res., 14 (1972) 423-435. 52 Zaczek, R. and Coyle, J. T., Excitatory amino acid analogues: neurotoxicity and seizures, Neuropharmacology, 21 (1912) 15-26.