Cellular and synaptic physiology and epileptogenesis of developing rat neocortical neurons in vitro

Developmental Brain Research, 34 (1987) 161-171

161

Elsevier BRD 50565

Research Reports

Cellular and synaptic physiology and epileptogenesis of developing rat neocortical neurons in vitro A.R. Kriegstein, T. Suppes and D.A. Prince Department of Neurology, Stanford University School of Medicine, Stanford, CA 94305 (U.S.A.) (Accepted 16 December 1986)

Key words: Epileptogenesis; Neocortex; Developing Neuron; Cortical inhibition; Neurogenesis; Pyramidal Neuron

The cellular and synaptic physiology of developing rat neocortical neurons was studied using the in vitro slice method. Rats aged 1-28 days were used for analysis. During the first two postnatal weeks several sequential changes occur in membrane properties and evoked synaptic potentials. Immature neurons had higher input resistances, more linear 1-V characteristics, longer membrane time constants, and slower rising and falling phases of action potentials. The developmental increase in rate of rise of the action potential suggests an increasing density of voltage-dependent Na+-channels are inserted in neuronal membranes during postnatal development. The higher input resistance of young cells might be due to their small size and differences in membrane properties. The long time constant indicates a higher specific membrane resistivity of immature neurons. Postsynaptic potentials (PSPs) recorded in young neurons were longer in latency, longer in duration, and more fragile during repetitive activation than their mature counterparts. In addition, PSPs evoked in neurons of animals less than 1 week old did not contain inhibitory postsynaptic components. These physiological features of immature neocortical neurons help explain the pattern of epileptogenesis in young animals. When neonatal cortical slices were exposed to the y-aminobutyric acid (GABA) antagonists penicillin or bicuculline, the frequency of occurrence of discharges resembling epileptiform depolarization shifts approached that found in mature slices only during the second postnatal week. Depolarization shifts at younger ages were less stereotyped and more sensitive to stimulus parameters than those in mature neurons.

INTRODUCTION A l t h o u g h the anatomic d e v e l o p m e n t of m a m m a lian neocortex has been e x a m i n e d in detail in a number of species 10'18"19"22"24"26"30'31, there is very little c o m p a r a b l e information about the physiological properties of cortical neurons during ontogenesis. Information about the sequence of d e v e l o p m e n t of intrinsic m e m b r a n e p r o p e r t i e s and synaptic responses in neocortical neurons will contribute to an understanding of n o r m a l and d i s o r d e r e d function in the immature nervous system. Previous physiological studies of immature cortical neurons have generaly employed in vivo recording m e t h o d s where it has been difficult to obtain stable intracellular recordings 28.29. W e have used the in vitro slice p r e p a r a t i o n to study intracellular activities of neurons from rat neocortex

in animals aged 1-28 days. Since rat pups are relatively immature at birth, this a p p r o a c h has allowed us to observe some of the sequential changes in m e m b r a n e properties and e v o k e d synaptic potentials which take place during cellular differentiation. A preliminary description of these results has been presented in abstract form 16. MATERIALS AND METHODS

Electrophysiological techniques Experiments were p e r f o r m e d on 300-400 /~mthick slices of rat s e n s o r i m o t o r cortex p r e p a r e d and maintained in vitro in an interface type c h a m b e r as described elsewhere 5.34.45. The composition of the normal saline used as perfusion m e d i u m was (in mM): NaCI 124,KC1 5, NaH2PO4 1.25, MgSO4 2,

Correspondence: A.R. Kriegstein, Department of Neurology, Stanford University School of Medicine, Stanford, CA 94305, U.S.A. 0165-3806/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

162 CaCl 2,NaHCO 3 26, and dextrose 10. The pH was between 7.35 and 7.40. Recordings were usually obtained at 35 + 0.5 °C for animals less than one week of age; otherwise the temperature was maintained at 37 + 0.5 °C. These differences in temperature did not significantly affect the sequence of electrophysiological changes described below, as judged by preliminary experiments at 37 °C in the young age group, and other data from our laboratories 41 (Thompson and Prince, unpublised observations). Orthodromic stimulation was applied through bipolar sharpened tungsten electrodes placed in the subcortical white matter underlying the recording site. Intracellular recordings from non-labelled cells were made with glass microelectrodes filled with 3 M potassium acetate. Electrodes had resistances between 50 and 100 MQ. Current was injected through the recording electrode via an active bridge circuit and balance was continually adjusted by cancelling the rapid voltage jump at the onset and offset of the current pulse. I - V data were obtained by passing current pulses of different amplitudes through the microelectrode and recording the neuronal membrane response. The signals were amplified and recorded on magnetic tape (DC to 5 kHz). Some of the data were digitized and analyzed off-line on a MINC-23 computer (Digital Equipments). All cells included for analysis met the following criteria: input resistances (RN) higher than 20 M ~ ; resting membrane potential (VM) at least -60 mV; and overshooting action potentials (APs). These criteria were establised to help insure that healthy neurons were included for study, and were not too restrictive as we were able to obtain cells from the youngest postnatal animals that met or exceeded these criteria. Neurons were grouped according to the postnatal age of the animal from which slices were obtained; a P3 neuron or a neuron ~3 days old' would be one recorded 3 days after the date of birth.

Intracellular labelling For those experiments in which cells were labelled with horseradish peroxidase (HRP), microelectrodes were made from fiber-filled glass capillary tubes (WP Instruments, Inc., New Haven, CT), filled with 3% HRP (Sigma Type VI) and bevelled to resistances of 50 to 150 M ~ . H R P was ejected by applying depolarizing current pulses of 2 to 5 nA and 250 to 500 ms

duration at a frequency of 1 to 2 Hz for at least 5 min. Tissue labelled with H R P was immersion fixed in 2% glutaraldehyde-2% paraformaldehyde, rinsed in phosphate buffer, and processed with a cobalt intensified diaminobenzidine method as described elsewhere 15. When cells were labelled with Lucifer Yellow CH (LY) 38, microelectrodes were filled with 3% LY and 2 M lithium chloride and bevelled to resistances of 50 to 150 Mff2. LY was iontophorectically injected by 3 to 4 nA of steady hyperpolarizing current for 3 to 10 min. Tissue containing LY-filled cells was fixed in 4% buffered formalin overnight and then dehydrated through graded alcohols, cleared, and mounted in methylsalicylate. H R P and LY-filled cells were viewed and photographed under a Leitz Dialux 20 microscope equipped with epifluorescent filter (Leitz E3) and a camera lucida drawing tube. RESULTS

Morphological characterization Since mammalian neocortex is composed of a heterogeneous population of morphological and physiological cell types, it was important to determine whether the cells characterized here represented the same cell type at different developmental stages, or a sampling of different cell types. We therefore recorded from 12 cells with dye-filled microelectrodes containing either Lucifer yellow (LY) or horseradish peroxidase HRP). After analyzing a cell physiologically, dye was injected and the neuron subsequently recovered and characterized morphologically. Representative camera lucida drawings of cells recovered from such experiments are shown in Fig. 1. One HRP-filled cell (Aa) and 2 cells filled with LY (Ab and c) from 1-week-old rat cortex, are illustrated as well as 2 HRP-filled cells from 2-week-old cortex (B). All the neurons recovered from animals less than 2 weeks of age could be assigned to the pyramidal cell category based on pyramidal shaped perikarya, prominent apical dendrites, and spiny apical and basilar dendritic processes (spines were observed but not drawn in the LY-stained neurons in Fig. 1). A variety of non-pyramidal cell types including multipolar and bipolar stellate cells were recovered from experiments with tissue from animals 2 weeks or older (Fig. 1Bb). We tentatively conclude that most of our physiological observations from neu-

163

B

A

\ a

b -

b

100 IJm Fig. 1. Camera lucida drawings of 3 neurons at 1 week (A) and two neurons at 2 weeks (B) of age. Cell a at 1 week is drawn from an HRP-filled ceil, while cells b and c are reconstructed from Lucifer yellow-injected cells. Spines which were visible in the LY-fiiled cells were not transcribed in the drawings. The two neurons at 2 weeks of age are drawn from HRP-filled cells. All 3 cells at 1 week of age and cell a at 2 weeks of age can be assigned to the pyramidal cell class based on pyramidal shaped perikarya, prominent apical and basilar dendrites, and spiny dendritic processes. Cell b at 2 weeks of age is an aspiny or sparsely spiny stellate cell.

rons of animals less than two weeks old apply to pyramidal type n e u r o n s in various stages of differentiation.

where they are compared with those from adult guinea pig neurons obtained in a previous study using similar techniques 6. Data on n e u r o n s from adult rat cortex do not differ significantly from those obtained

M e m b r a n e properties We arbitrarily divided n e u r o n s impaled in the first few weeks of postnatal cortical development into two groups in order to compare changes in physiological properties. These results are summarized in Table I

in guinea pig (B.W. Connors, personal communication). In comparison to adult neurons, young n e u r o n s (less than 1 week postnatal) had higher input resistances (RN = 64.1 + 18.4 Mr2 vs. 24.3 ___ 13.7 Mf~ in adults), slower rising and falling phases of action po-

TABLE I Membrane properties of immature neocortical neurons

I'M, membrane potential; RN, input resistance. Age

No.

v (ms)

V M (mV)

Rlv (M~2)

R M (~2.crr2)

Young (< 1 week) Juvenile (2-3 weeks) Adulta

9 6 20

22.1 + 8.4* 13.5 + 3.9** 8.2 + 4.7

-68.7 + 9.8 -73.6 + 5.0 -75.3 + 7.8

64.1 + 18.4 35.2 + 12.0 24.3 + 13.7

~22,100 ~13,500 ~ 8,200

a Adult values from ref. 6; *P < 0.001. **P < 0.005.

164 A

t

!

B

-20

Vm (mY)

• i

t

i

i

-10

.f,

><

~

~-50 msec

oo JE=

i

[-

i

0

~

13d

=

10

I (nA)

••

==

• L -100

Fig. 2, Membrane properties of a P3 cortical neuron. A: mem brane responses (top trace) to a depolarizing and hyperpolarizing applied current pulse (lower trace). The action potential in young neurons is characteristically slow (5.5 ms measured at the base in this example) and the input resistance relatively high (60 Mfl here). B: The current-voltage relationship for the cell shown in A: the 1-V relationship of young neurons tends to be relatively steep and linear, however this neuron showed rectification in the depolarizing quadrant.

tentiais (APs) (Fig. 5), and longer membrane time constants (r = 22.1 + 8.4 ms vs 8.2 + 4.7 ms in adults). Neurons from juvenile animals (2-3 weeks postnatal) had measurements of r and R N that were intermediate between young and adult values. Resting membrane potentials of young neurons (-68.7 + 9.8 mV) were significantly different from adult values (-75.3 + 7.8 mV), possibly because of the fragility of the immature cells and damage by the microelectrode. Fig. 2 shows several of the characteristic features of the membrane properties of a 3-day-old (P3) neuron. Although the resting potential of -65 mV and the AP amplitude of 70 mV would not be unusual for an adult cell, the AP duration of 5.5 ms (measured at the AP base) and RN of 60 MQ are both significantly larger than expected for a mature cortical neuron. Fig. 2 illustrates the data from which values of r were calculated for a P3 and P13 neuron. Young neurons have long time constants (r=24 ms in the P3 neuron of Fig. 3), averaging almost 3 times the value for adult cells. Another indication of postnatal changes in membrane properties is provided by comparing current-voltage ( l - V ) relationships of cells at different ages. Fig. 4 shows that neurons from younger animals (e.g. P4) tended to have a steep relationship between applied current and membrane voltage, which reflects a high R N a s shown in Table I, and an apparent relative linearity of the I - V relationship. In older animals, membrane depolarization often activates instrinsic voltage-dependent currents that appear as non-linearities in the I - V curves (Fig. 4: P6, P14) 5. This is illustrated by comparison of the rectification

0 Z

oN o'~O

-2o

~) O -J

o~

_

. ~ 1 3d \

~ . . ~ d

T = 14 msec 30

n

L 10

I __

1__ 20

L

T=24

J~_.__~_ 30

= 40

msec ~_____J

50

Time (msec) Fig. 3. The membrane time constants of neurons aged 3 and 13 days are compared. The inset to the right shows that applied square-wave hyperpolarizing currents (top traces) evoked characteristic membrane voltage responses (lower traces) which form the basis for the data points plotted in the graph. Membrane time constants for young neurons were significantly longer than those for more mature cells.

ratio 14 obtained by dividing the slope resistance at -55 mV with the slope resistance at - 7 0 mV in immature vs mature neurons. The mean rectification ratio in 9 neurons from animals 3-5 days old was 0.95 + 0.11 with a range of 0.75-1.13 compared to a mean r

P4

P4

?

:

L

i-I J

~

i

.

I

±

•

r •• 1"

• • .'e° i

1

L P6

P14

r

vm (my) -20

i • II 0

I (nA)

f

= • f

t

i - 1.0

i

~

,°

io • "

1.0

-100

Fig. 4. Current-voltage (I-V plots obtained from four neurons (P4, P4, P6, P14). The origin of each plot is at resting potential. I-V relationships at 4 days are relatively linear. Note the steep slope in the upper left plot. Some delayed rectification is present in the P6 cell and depolarizing inward rectification in the 14 day cell. Bars in P14 apply to all plots.

165 of 1.47 + 0.51 with a range of 1.06-1.83 in neurons 12-19 days old. The rectification ratios suggest that there is an apparent change in Rr~ at depolarized Vms which is due to activation of subthreshold currents in more mature neurons. The duration of APs gradually decreased with increasing age. In Fig. 5A, APs of 3 neurons ages P4, P7 and P19 are shown for comparison, and in Fig. 5B the differentiated traces of these APs are illustrated. The rate of rise (dV/dt) of the AP increased in the older neurons. The dV/dt of the rising phase of the AP was measured in 24 neurons ranging in age from P4 to P19. The resulting graph (Fig. 5C) indicates an age-related increase in dV/dt. The rate of repolarization of the AP also appears to increase with age, howB

A

ever; this feature was more variable. As has been previously reported 28, the frequency of spike discharges tends to be significantly lower in neurons from immature vs mature animals. In response to large intracellular depolarizing current pulses the minimum interspike interval in neurons less than 7 days old averaged 9.3 ms (+2.7, n = 5) compared to 4.5 ms (+0.5, n = 4) in cells 7-12 days old. None of the young (< 1 week) neurons from Table I (n = 9), or from a large sample of cells of the same ages which were less 'healthy' as judged by Vm, RN or spike height, generated burst discharge during depolarizing current pulses. This finding has been confirmed recently in a larger sample of presumed pyramidal cells from all layers (McCormick and Prince, in preparation).

-1.

Synaptic physiology P19

/

P4 /

J

II

i

/'

/

i

P7 ~

'/

-

10 mV

/

,

P19

25 V / s e c

25 msec C

(n=8) 200

o

160

>

120

G)

(n=a)~ (n=4)

(n-3)~....../~

80 40 0

.

0

2

.

.

4

.

6

8

10

12

i

i

i

I

14

16

18

20

Age (days) Fig. 5. A: characteristic action potentials elicited by depolarizing current pulses in neurons at 3 different ages are shown. The progressive decrease in duration of the action potentials is apparent. B: differentiating the voltage traces shown in A revealed that the rate of rise of the action potential (dV/dt) increases progressivly with neuron age. dV/dt is measured between the baseline and the peak of the downward deflection. C: a plot of the measured dV/dt for 24 neurons aged 4-19 days. The mean data points + S.E.M. are shown. Time bar in A apllies to A and B.

Synaptic events in young cortical neurons were also examined. Synaptic potentials could be evoked by stimulating underlying white matter in animals at all postnatal ages. Several features distinguished the synaptic responses of immature neurons from their adult counterparts. The latencies of the responses were characteristically quite long, exceeding 200 ms in some cases, and responses were very long lasting. The duration of PSPs elicited by stimulation of subcortical white matter averaged 231 + 101 ms (n = 9) in cells 3-7 days old, compared to only 69.4 + 37 ms (n = 8) in cells 7-14 days old. PSPs were invariably depolarizing at resting potential in cells less than 10 days old, and with sufficiently intense stimulation, PSP amplitude often increased above threshold and APs were generated. Figs. 6 and 7 illustrate some of these features. In Fig. 6, the PSPs evoked by progressively increasing intensities of subcortical stimulation in neurons from P6 and P28 animals are compared. PSPs from the P6 neuron (Fig. 6A) had several components, including an initial fast depolarization that may have been generated by depolarization of adjacent electrically5 or ephaptically 4° coupled neurons, and a long duration very slow depolarization. At higher stimulus intensities a multiphasic depolarizing event was evoked that ultimately triggered an AP. The overall PSP lasted about 200 ms. By contrast, the PSP in the P28 neuron (Fig. 6B) lasted less than 50 ms and had fewer depolarizing components. These differences in PSP

166

A P6

ir

'

W

~

~,

.~r

A

B P28

~ - -

.~

~_

7--

"

"

"

%

7

"

i

0

O4

• 50 msec Fig. 6. Superimposed sweeps of responses to increasing intensity subcortical white matter stimulation in a P6 (A) and P 28 (B) neuron. The PSP is longer and more complex in the younger neuron.

duration were too large to be explained by the differences in membrane time constants of immature vs mature neurons (Table I). PSPs in cells less than 7 days old were also very fragile during repetitive activation. Successive PSP amplitudes invariably decreased at stimulation fre-

A

P5

•

B

5 0 msec P6

0.1 Hz - - ' ~ m u m ~ - - - " -n"~--' --%?? " -._. ---. %.,o..

•

o t~l

100 msec

Fig. 7. Orthodromic responses to white matter stimulation in neurons from a P5 (A) and a P6 (B) animal. Four successive stimuli delivered at 0.5 Hz evoked progressively smaller PSPs in A. In B, similar but less marked decreases in response occur following stimulation at 0.2 Hz (1) and 0.1 Hz (2). Voltage bar for A and B is shown in B2.

quencies greater than 0.1 Hz. Fig. 7 shows the effect of stimulus frequency on PSPs evoked in young neurons. In the P5 cell of Figure 7A, stimulation every 2 s produced a significant decrease in the amplitude of fast and slow components of the PSP. In the P6 cell of Fig. 7B, the amplitude of the PSP decreased significantly at stimulus frequencies greater than 0. I Hz (cf. B1 and B2). PSPs evoked in neurons from animals less than one week old (n = 12) did not contain the excitatory postsynaptic potential-inhibitory postsynaptic potential (EPSP-IPSP) sequence typical of evoked responses in mature neocortical neurons. Instead, PSPs with only depolarizing components were observed. Depolarizing the cells to levels just below spike threshold did not uncover significant hyperpolarizing PSP components that might have been masked by concurrent excitation or have represented IPSPs which were depolarizing at rest (e.g. ref. 6, Fig. 12).

Epileptogenesis in immature cortical slices When GABAergic inhibition is blocked by bathing adult neocortical slices in solutions containing penicillin (3.5 mM) or bicuculline (10/~M), epileptiform depolarization shifts (DSs), consisting of large-amplitude long-duration depolarizations associated with bursts of spikes, can be orthodromically evoked in large populations of pyramidal cells. We found that when neocortical slices from animals 3-5 days old were exposed to penicillin only approximately 60% of those cells from which good quality recordings were obtained developed DS-like discharges (n = 9) while other cells from the same slices did not (n = 5). In contrast, all 'healthy' neurons aged 7-11 days (n = 9) developed DS-like discharges in penicillin solution. DSs in neurons less than 2 weeks old tended to be longer in duration (60-200 ms), less stereotyped and more sensitive to stimulus parameters than those in mature cells, and were not associated with the high-frequency spike discharges that characterize neuronal firing in acute epileptiform loci of adult animals (Figs. 8C, 9; see refs. 20, 28). While orthodromic stimulation at 1 Hz can regularly evoke stereotyped DSs in mature neocortical slices it, very long interstimulus intervals (up to 20 s) are necessary to evoke successive DSs of similar configuration in slices from animals in the first week of life. For example, in the P3 neuron of Fig. 8A, which is from a slice bathed in a

167

A

B

!

E
Pll

msec

30 msec

C

D Pll

o 75 msec

o

•

75 msec

Fig. 8. A: the first of a pair of orthodromic stimuli (triangle) 12 s apart evokes a DS in a cell from a bicuculline-treated slice of a P7 animal. The second stimulus (superimposed trace) evokes only a small PSP. Superimposed responses to hyperpolarizing current pulses (top traces) show a large conductance increase at the peak of the DS. B-D: responses to orthodromic stimulation of neurons at different ages from slices bathed in bicuculline. Note the graded responses with increasing stimulus intensity in B and D. Small spikes in C and in Fig. 9B are intracellular reflections of an extracellular unit. solution containing bicuculline (5 ~ M ) , the first stimulus e v o k e d a DS; however, a second stimulus 12 s later only elicited a PSP. The amplitude of the evoked DSs in cells less than 7 days old (n = 8) de-

A

B B,oo .L,NE

t

!!

sEc ,NTE V..

75

msec

Fig. 9. Recordings from a P3 neuron in a slice bathed in bicuculline. A: successive stimuli at increasing intensities evoke a graded depolarization. B: the same neuron as in A, showing a decrease in evoked depolarization amplitude with successive stimuli of the same intensity delivered at intervals of 15 s.

creased progressively even when stimuli were delivered 15 s apart, as illustrated in Fig. 9B for a P3 cell in a bicuculline-treated slice. Conductance pulses superimposed on these responses show that during the peak of the DS, there was a very large increase in m e m b r a n e conductance (Fig. 8A), as is the case in adult animals 11. A n o t h e r important feature of epileptogenesis in immature slices is shown in Fig. 8B, D and 9A. In mature cortex in vivo 2°,27 and in vitro 1~ the DS m a y occur at full a m p l i t u d e and long latency as the low-frequency o r t h o d r o m i c stimulus reaches threshold; with increases in stimulus intensity the DS latency shortens, but its configuration m a y remain relatively stereotyped. By contrast, stepwise increases in orthodromic stimulus intensity usually evoke graded depolarizations with a relatively constant onset latency in neurons of convulsant-treated i m m a t u r e slices (Fig. 8B, D; Fig. 9A). The g r a d e d nature of these fixed latency depolarizations closely r e s e m b l e d that seen in

168 the same neurons when constant intensity orthodromic stimuli were delivered at low frequencies (e.g. compare Fig. 9A and B). The occurrence of similar behaviors in response to an orthodromic stimulus among neurons impaled close to one another in the same slice, and the occasional fortuitous recording of an extracellular bursting unit roughly synchronous with the depolarization and spikes of the impaled neuron (Figs. 8C and 9B, small biphasic spikes) suggested that the abnormal evoked neural activity in convulsant-treated slices was occurring in populations of cells. DISCUSSION Our cellular physiological studies of neonatal rat cortex indicate that several intrinsic membrane properties and synaptic responses change during postnatal development. During the first 2 postnatal weeks the duration of APs decreased while the dV/dt of the rising phase increased, R N became smaller, and PSPs became shorter in duration and exhibited less decrement in amplitude during low frequency orthodromic stimulation. Since the dV/dt of the spike rising phase is largely dependent on Na+-conductance, the developmental increase in dV/dt, together with recent evidence from studies of [3H]saxitoxin binding 2, and 22Na+-flux assays 7 suggest that there is an increasing density of voltage-dependent Na+-channels in the membrane during postnatal development. Similar increases in Na+-channel density occur in acutely dissociated cortical neurons during the first weeks of life 13. The developmental increase in action potential rise rate reported here closely resembles a similar change observed in dissociated mouse spinal cord neurons during the first three weeks in culture 43 which parallels an increase in estimated Na+-channel density 17. Other data also indicate that changes in the electrical properties of the membrane are occurring during pyramidal cell differentiation. As shown in Fig. 3, and Table I, membrane time constants in young neurons were substantially longer than those in mature cells. Immature neurons would be expected to have a high input resistance because of their small size, however, the long time constant of these cells is not a direct result of a large Rrq, since r is a product of specific resistivity and specific capacitance. It is reasonable to assume that the specific ca-

pacitance of membranes of immature and mature neurons is the same. Therefore, the specific resistivity of immature neurons must be higher than that of adult cells. This suggests that fundamental changes in neuronal membrane structure are occurring postnatally, although their precise nature cannot be determined from our data. From the data of Fig. 4 and the rectification ratios of immature neurons, it appears that there are changes in voltage-dependent subthreshold currents during development as well. The relative linearity of I - V relationships at younger ages may indicate that activation of voltage-dependent ion channels is not significant within the range of membrane potentials examined, or that such channels are activated but the resulting inward and outward currents are relatively balanced. Recent results in acutely dissociated immature neurons 12 (Hamill et al., in preparation) do indicate that there is a sequential development of some subthreshold voltage-dependent K +- and Na+-currents during ontogenesis. PSPs evoked in immature neurons of neocortex in vitro by stimulation of subcortical white matter or pia are similar to those found in immature neocortex in vivoa8'29 and hippocampus in vitro 36'37. The outstanding characteristics of these PSPs are their long duration, their tendency to become progressively smaller in amplitude, even at low rates of stimulation, and their lack of hyperpolarizing components resembling inhibitory postsynaptic events. The mechanisms for this marked sensitivity to stimulus frequency are not known; however, neuroanatomical studies do show that cortical presynaptic terminals are structurally immature and contain many fewer synaptic vesicles at young ages 3, suggesting that the failure of transmission may be due to presynaptic factors in the axonal terminal arborization. Other possible factors that might contribute to the frequency-dependent fall-off in PSPs, such as peculiarities of calcium currents in presynaptic terminals, unusual receptor properties, or developmental alterations in transmitter release and reuptake mechanisms remain to be explored. We found a lack of inhibitory components of PSPs in young cortical neurons. Pyramidal tract neurons in neonatal kitten cortex in vivo also generate mainly EPSPs following cerebellar afferent stimulation; EPSP/1PSP sequences develop only after 3 days 25. Hyperpolarizing IPSPs also appear relatively late in the course of functional synaptogenesis in the devel-

169 oping rabbit hippocampus 36. Neocortical inhibition is presumably mediated by y-aminobutyric acid (GABA) through activation of GABAergic interneurons. Most biochemical markers for GABAergic activity, including high-affinity GABA uptake, glutamic acid decarboxylase (GAD) activity, and Ca 2+dependent GABA release, are present at birth in rat cortex but increase markedly in the second and third postnatal weeks 8"33. Also, while inhibitory interneurons have recently been detected by GAD-like immunoreactivity in prenatal rat neocortex 44, the characteristic 'baskets' of GAD-positive axons surrounding pyramidal cells were detectable only during the second postnatal week. Ultrastructural studies of rat cortex also show that axosomatic synapses with symmetrical membrane specializations, commonly thought to be inhibitory, develop following birth 1'3~21. Therefore biochemical, anatomical, and histochemical studies support the physiological data that functional inhibition develops relatively late during postnatal stages of rat corticogenesis. However, as reported here, G A B A antagonists such as penicillin and bicuculline can increase the excitability of some neocortical neurons even at very young ages when IPSPs are not apparent. The mechanisms for this effect are not known. Possible explanations include a tonic level of junctional or nonjunctional G A B A release in neonatal cortex, very slow uptake mechanisms, a small 'hidden' GABAergic hyperpolarizing component of the PSP, or even G A B A actions to depress Ca 2+ currents 9. Epileptogenesis in immature neocortex In immature cortex we noted that some neurons did not generate DS-like discharges when exposed to GABA antagonists, and others generated graded non-stereotyped depolarizations and DSs as the stimulus intensity was increased (Figs. 8B and 9A). Stereotyped DSs with long and variable iatencies of the sort easily evoked in mature neocortical slices li are unusual in slices from immature animals < 2 weeks of age. This may be attributed to the paucity of intrinsic cortical circuitry necessary to amplify intracortical synaptic activities that contribute to DS generation (see refs. 11 and 42). It also seems possible that the delayed development of voltage-dependent subthreshold currents in immature neurons might have an important influence on epileptogenesis. For ex-

ample, the burst generating neurons found in mature slices and postulated to be pacemakers for epileptogenesis 4'11 presumably depend upon slow inward currents to initiate bursts. Such neurons have not been identified in recordings from immature slices to date. Absence of bursting pacemakers could have an important effect on initiation and synchronization of epileptiform discharge (e.g. ref. 35). In contrast to results in the hippocampus 37,39, sustained ictal discharges were not seen in immature neocorticai slices treated with bicuculline (10 ~tM), and the mechanism for generation of DSs was less robust. These data suggest that the 'normal' features of neuronal activities in immature cortex, including slow rate of spiking, long duration of PSPs, the safety factor for eliciting PSPs with repetitive stimuli, and perhaps the paucity of burst-generating neurons, influence the characteristics of interictal epileptogenesis. Some of the functional consequences of the immature neuronal properties discussed above are seen in epileptogenesis following application of convulsant drugs to the cortical surface in vivo or to slices in vitro. The epileptiform events recorded from the pial surface following application of penicillin to the neocortex of immature rabbits have a variety of morphologies23, suggesting that the neuronal aggregate generating these interictal discharges changes from second to second or that different sequences of activation of cortical elements occur. The interictal 'spikes' differ from the more or less stereotyped potentials seen in comparable preparations in adult animals and are generally smaller in amplitude. Other evidence for relative asynchrony in the epileptic neuronal aggregate of immature cortex includes variations in the amplitude and duration of DSs in neurons and the variable relationship between neuronal depolarizations and particular E E G epileptiform events 28.

ACKNOWLEDGEMENTS We are grateful to our colleagues Barry Connors and Michael Gutnick who performed initial pilot experiments in immature neocortical slices, and to John Huguenard, Owen HamiU, and David McCormick for helpful comments on the manuscript. We thank Jay Kadis for technical assistance and Cheryl Joo and Marie Holt for manuscript preparation. This study

170 was s u p p o r t e d by N I H G r a n t s NS 21223, NS 00887 ( A R K ) and NS 06477 ( D A P ) f r o m the N I N C D S ; by

REFERENCES 1 Bahr, S. and Wolff, J.R., Postnatal development of axosomatic synapses in the rat visual cortex: morphogenesis and quantitative evaluation, J. Comp. Neurol., 233 (1985) 405-420. 2 Baumgold, J.I., Zimmerman, and Bambrick, L., Appearance of 3H-saxitoxin binding sites in developing rat brain, Dev. Brain Res., 9 (1983( 405-407. 3 Blue, M.E. and Parnavelas, J.G., The formation and maturation of synapses in the visual cortex of the rat. I. Qualitative analyses, J. Neurocytol., 12 (1983) 599-616. 4 Connors, B.W., Initiation of synchronized neuronal bursting in neocortex, Nature (London), 310 (1984) 685-687. 5 Connors, B.W., Benardo, L.S. and Prince, D.A.., Coupling between neurons of the developing rat neocortex, J. Neurosci., 3 (1983) 773-782. 6 Connors, B.W., Gutnick, M.J. and Prince, D.A., Electrophysiological properties of neocortical neurons in vitro, J. Neurophysiol., 48 (1982) 1302-1320. 7 Couraud, F., Martin-Moutot, N., Koulakoff, A. and Berwald-Netter, Y., Neurotoxin-sensitive sodium channels in neurons developing in vivo and in vitro, J. Neurosci., 6 (1986) 192-198. 8 Coyle, J.T. and Enna, S.J., Neurochemical aspects of the ontogenesis of GABAergic neurons in the rat brain, Brain Res., 111 (1976) 119-133. 9 Deisz, R.A. and Lux, H.D., Gamma-aminobutyric acid-induced depression of calcium currents of chick sensory neurons, Neurosci. Lett., 56 (1985) 205-210. 10 Eayers, J.T. and Goodhead, B., Postnatal development of the cerebral cortex of the rat, J. Anat., 93 (1959) 385-402. 11 Gutnick, M.J., Connors, B.W. and Prince, D.A., Mechanisms of neocortical epileptogenesis in vitro., J. Neurophysiol., 48 (1982) 1321-1335. 12 Hamill, O.P., Huguenard, J.R., Enayati, E.F. and Prince, D.A., Single channel currents underlying slow threshold Na + conductances in rat neocortical neurons, Neurosci. Abstr. 12 (1986) 950 (No. 261.5). 13 Huguenard, J.R., Hamill, O.P., Enayati, E.F. and Prince, D.A., Developement of Na + conductance in neocortical neurons of the rat, Soc. Neurosci. Abstr., 12 (1986) 950 (No. 261.6). 14 Kandel, E.R. and Tauc, L., Anomalous rectification in the metacerebral giant cells and its consequence for synaptic transmission, J. Physiol. (London), 183 (1966) 287-304. 15 Kitai, S.T. and Bishop, G.A., Horseradish peroxidase. Intracellular staining of neurons. In L. Heimer and M.J. Robards (Eds.), Neuroanatomic Tract-Tracing Methods, Plenum, New York, 1981, pp. 263-277. 16 Kriegstein, A.R., Suppes, T. and Prince, D.A., Cellular physiology of the developing rat neocortex in vitro, Soc. Neurosci. Abstr., 9 (1983) 602 (No. 178.7). 17 MacDermott, A.B. and Westbrook, G.L. Early development of voltage-dependent sodium currents in cultured mouse spinal cord neurons, Dev. Biol., 113 (1986) 317-326. 18 Marin-PadiUa, M., Prenatal and early postnatal ontogenesis of the human motor cortex: A Golgi study. I. The se-

T h e K l i n g e n s t e i n F u n d ( A R K ) , and the M o r r i s R e search Fund.

quential development of the cortical layers, Brain Res., 23 (1970) 167-183. 19 Marin-Padilla, M., Prenatal and early postnatal ontogenesis of the human motor cortex: A Golgi study. II. The basket-pyramidal system, Brain Res., 23 (1970) 185-191. 20 Matsumoto, H. and Ajmone-Marsan, C., Cortical cellular phenomena in experimental epilepsy: interictal manifestations. Exp. Neurol., 9 (1964) 286-304. 21 Miller, M. and Peters, A., Maturation of rat visual cortex. II. A combined Golgi-electon microscope study of pyramidal neurons, J. Comp. Neurol., 203 (1981) 555-573. 22 Morest, D.K., A study of neurogenesis in the forebrain of opossum pouch young, Z. Anat. Entwicklungsgesch.. 130 (1970) 265-305. 23 Mutani, R., Futamachi, K.J. and Prince, D.A., Potassium activity in immature cortex, Brain Res., 75 (1974) 27-39. 24 Noback, C.R. and Purpura, D.P., Postnatal ontogenesis of neurons in cat neocortex, J. Comp. Neurol., 117 (1961) 291-307. 25 Oka, H., Samejima, A., and Yamamoto, T., Post-natal development of pyramidal tract neurons in kittens, J. Physiol. (London), 363 (1985) 481-499. 26 Peters, A. and Feldman, M., The cortical plate and molecular layer of the late rat fetus, Z. Anat. Entwick!ungsgesch., 141 (1973) 3-37. 27 Prince, D.A., Modification of focal cortical epileptogenic discharge by afferent influences, Epilepsia, 7 (1966) 181-201. 28 Prince, D.A. and Gutnick, M.J., Neuronal activities in epileptogenic foci of immature cortex, Brain Res., 45 (1972) 455-468. 29 Purpura, D.P., Schofer, R.J. and Scarff, T., Properties of synaptic activities and spike potentials of neurons in immature neocortex, J. Neurophysiol., 28 (1965) 925-942. 30 Rakic, P. and Goldman-Rakic, P.S., Development and modifiability of cerebral cortex, Neurosci. Res. Prog. Bull., 20 (1982) 429-611. 31 Ramon y Cajal, S., Histologie du Systeme Nerveux de l'Homme et des Vertebres, 2 Vols. (L. Azoulay, trans.). (Reprinted by Instituto Ramon Y Cajal del C.S.I.C., Madrid, 1952-1955), (1905-1911). 32 Ransom, B,R., Yamate C.L. and Connors, B.W., Activitydependent shrinkage of extracellular space in rat optic nerve: A developmental study, J. Neurosci., 5 (1985) 532-535. 33 Redburn, D.A., Broome, D., Ferkany, J. and Enna, S.J., Development of rat brain uptake and calcium-dependent release of GABA, Brain Res., 152 (1978) 511-519. 34 Schwartzkroin, P.A., Characteristics of CA1 neurons recorded intracellularly in the hippocampal in vitro slice preparation, Brain Res., 83 (1975) 423-436. 35 Sehwartzkroin, P.A. and Prince, D.A., Cellular and field potential properties of epileptogenic hippocampal slices, Brain Res., 147 (1978) 117-130. 36 Schwartzkroin, P.A., Development of rabbit hippocampus: physiology, Dev. Brain Res., 2 (1982) 469-486. 37 Schwartzkroin, P.A. and Kunkel, D.D., Electrophysiology and morphology of the developing hippocampus of fetal rabbits, J. Neurosci., 2 (1982) 448-462.

171 38 Stewart, W.W., Functionalconnections between cells as revealed by dye-coupling with a highly fluorescent naphthalimide tracer, Cell, 14 (1978) 741-759. 39 Swann, J.S. and Brady, R.J., Penicillin-induced epileptogenesis in immature rat CA3 hippoeampal pyramidal cells, Dev. Brain Res., 12 (1984) 243-254. 40 Taylor, C.P. and Dudek, F.E., Synchronization without active chemical synapses during hippocampal afterdischarges, J. Neurophysiol,, 52 (1984) 143-155. 41 Thompson, S.M., Masukawa, L.M. and Prince, D.A., Temperature dependence of intrinsic membrane properties and synaptic potentials in hippocampal CA 1 neurons in vitro, J. Neurosci., 5 (1985) 817-824. 42 Traub, R.D. and Wong, R.K.S., Synchronized burst discharge in disinhibited hippocampal slice. II. Model of cel-

lular mechanism, J. Neurophysiol., 49 (1983) 459-471. 43 Westbrook, G.L. and Brenneman, P.E., The development of spontaneous electrical activity in spinal cord cultures. In F. Caciagli, E. Giacobini and R. Paoletti (Eds.), Devel-

opmental Neuroscience: Physiological, Pharmacological and Clinical Aspects, Elsevier, Amsterdam, 1984, pp. 11-17. 44 Wolff, J.R., Bottcher, H., Zetzsche, T., Oertel, W.H. and Chronwall, B.M., Development of GABAergic neurons in rat visual cortex as identified by glutamate decarboxylaselike immunoreactivity, Neurosci. Lett., 47(1984) 207-212. 45 Yamamoto, C., Intracellular study of seizure-like afterdischarges elicited in thin hippocampal sections in vitro, Exp. Neurol,, 35 (1972) 152-164.