Suppression of presynaptic calcium currents by hypoxia in hippocampal tissue slices

Brain Research, 573 (1992) 70-70 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/92/$05.00

70 BRES 17440

Suppression of presynaptic calcium currents by hypoxia in hippocampal tissue slices J.N. Young I and G.G. Somjen 2 1Division of Neurosurgery and 2Department of Cell Biology, Duke University Medical Center, Durham, NC 27710 (U.S.A.) (Accepted 1 October 1991)

Key words: Stroke; Hypoxia; Synaptic blockade; Calcium channel; N-Methyl-D receptor; Glutamate receptor; Spreading depression

We tested the hypothesis that suppression of inward calcium current in presynaptic terminals is the cause of failure of synaptic transmission early during cerebral hypoxia. Postsynaptic responses in CA1 zone of hippocampal tissue slices were blocked either by the combined administration of 6,7-dinitroquinoxaline-2,3-dione (DNQX) and 3-((+)-2-carboxypiperazine-4-yl)-propyM-phosphonicacid (CPP) or by lowering extracellular calcium concentration ([Ca2+]o). Repetitive orthodromic activation of central neurons caused transient decrease of [Ca2+]o (measured by ion selective microelectrodes) in neuropil, attributable to influx of Ca 2+ in presynaptic terminals. Presynaptic [Ca2÷]o responses were rapidly and reversibly suppressed when oxygen was withdrawn from hippocampal tissue slices. The 'resting' baseline level of [Ca2÷]o declined at first gradually, then precipitously as in spreading depression (SD). Presynaptic volleys during high frequency train stimulation were also depressed somewhat before SD began. We conclude that (1) presynaptic Ca z+ currents fail during hypoxia, perhaps because 'resting' intracellular free Ca2÷ activity is increased and, in part, also because of partial failure of presynaptic impulse conduction; (2) the influx of Ca2+ into brain cells in hypoxic spreading depression is not mediated by glutamate/aspartate dependent channels. INTRODUCTION W h e n blood stops flowing to the brain, subjects faint within 10 s and transmission at all synapses in the forebrain becomes blocked in minutes 32. Correlated consistently in time with hypoxic transmission failure, inhibitory and excitatory synaptic potentials (IPSPs and EPSPs) are depressed (reviewed in ref. 31 and 32). The depression of synaptic potentials could be due to impairment either of postsynaptic receptors or of the release of transmitter substance from presynaptic terminals. In the early stages of oxidative energy shortage, (postsynaptic) neurons respond normally to exogenous transmitters 1'22'26. Therefore, impaired presynaptic transmitter release is the most likely cause of hypoxic synaptic block. A short time after synaptic failure severe depolarization, resembling spreading depression (SD), sets in 14'33. A d a m s et al. 1 and we 35 have suggested that hypoxic failure of transmitter release is due to blockade of voltage gated calcium channels in presynaptic terminals. Voltage d e p e n d e n t calcium currents have indeed been found to be blocked in neuron cell bodies deprived of oxygen or glucose, or exposed to cyanide 9'24'36. Moreover, raising [Ca2+]o delayed the onset of synaptic blockade in hypoxic spinal cords 6. These observations are consistent with failure of Ca 2+ currents during hypoxia but none demonstrates the actual b l o c k a d e of voltage gated

calcium channels in presynaptic terminals during oxygen deprivation in m a m m a l i a n cerebral tissue. The experiments r e p o r t e d here provide this missing proof. Repetitive activation of neuron populations causes a decrease of [Ca2÷]o attributed to the m o v e m e n t of Ca z+ ions from interstitial fluid into the cytosol of neurons 3°. H e i n e m a n n and associates 1s'19 have shown that pre- as well as postsynaptic Ca 2÷ currents contribute to the [Ca2+]o response, and that the presynaptic c o m p o n e n t can still be r e c o r d e d after elimination of the postsynaptic response. We used this approach to observe the effect of oxygen shortage on presynaptic Ca 2÷ currents. MATERIALS AND METHODS Hippocampal tissue slices, 400/~m thick, were prepared from the brains of anesthetized rats3'7. The slices were maintained in an interface chamber in flowing artificial cerebrospinal fluid (ACSF) of the following composition: (in mmol/l): NaCI 130, KC1 3.5, NaH2PO 4 1.25, NaHCO 3 24, CaC12 1.2, MgSO 4 1.2, glucose 10, saturated with 95% 02-5% CO2, pH 7.4 at 34.5°C. Voltage and [Ca2+]o were recorded with double-barreled ion selective microelectrodes from stratum (st.) radiatum of CA1 region and in some experiments also from st. pyramidale. The Fluka calcium ionophore 'cocktail A' filled one tip of the double pipette. Electrodes were calibrated before and after use in the presence of physiological sodium and potassium concentrations at 3 different calcium concentrations, at constant ionic strength. The properties of such electrodes have repeatedly been described (e.g. refs. 8, 19 and 30). Orthodromic stimuli were applied to the Schaffer collateral-commissural bundle, alternated in some experiments with antidromic

Correspondence: G.G. Somjen, Box 3709 Duke University Medical Center, Durham NC 27710, U.S.A. Fax: (1) (919) 684-5481.

71 stimulation of the alveus. Stimulus trains of 20 Hz frequency and 5 s train duration, 140 /~A, 0.05-0.1 ms pulses (evoking maximal population spikes) were used to evoke transient decreases of [Ca2+]o. To block glutamate receptors, 6,7-dinitroquinoxaline-2,3dione (DNQX, 10 #M) and 3-((-)-2-carboxypiperazine-4-yl)-propyl-l-phosphonic acid (CPP, 20/~M) were administered in the bathing fluid for 60-90 min until postsynaptic orthodromically evoked responses were suppressed. Low-calcium solution contained 0.12 mmol/l Ca2+ and 2.4 mmoF1 Mg2+. Hypoxia was induced by replacing 95% 02-5% CO2 by 95% N2-5% CO2 in the gas phase of the chamber 3. Voltage and ion selective signals were recorded on chart paper and on VCR tape. The taped data were processed by the Axotape program. The potential of the ion selective electrode was transformed to read millimoles concentration on a linear scale with the aid of the AXUM computer program.

RESULTS Fig. 1 illustrates the abolition by D N Q X plus CPP of orthodromic postsynaptic evoked potentials and the orthodromicaUy evoked [Ca2+]o responses in st. pyramidale. The presynaptic volley (Fig. 2A,B), the orthodromic [CaZ÷]o response in st. radiatum as well as the antidromic responses in st. pyramidale were still evoked after blockade of postsynaptic responses. Figs. 2C and 3 show tracings of [Ca2+]o in st. radiatum (layer of apical dendrites) after such blockade of

postsynaptic responses, during hypoxia. Soon after the withdrawal of oxygen the A[Ca2÷]o responses evoked by orthodromic train stimuli began to decrease and they were invariably severely depressed before the onset of SD (Figs. 2C and 3a-f). The orthodromically evoked A[Ca2+]o responses were 10% depressed after an average of 106 s (range: 35-208 s), while SD began on the average after 279 s (range: 216-520 s; n = 20) following oxygen withdrawal. The last orthodromic stimulus delivered before SD always failed to elicit any A[Ca2+]o response (see last point of Fig. 4, representing stimuli delivered approximately 10 s before SD onset); the stimulus before the last also frequently failed to evoke a A[Ca2+]o response (Figs. 2 and 3). Antidromically evoked A[Ca2+]o responses recorded in st. pyramidale were depressed more slowly than were presynaptic A [Ca2+]o responses. The presynaptic volley (afferent compound action potential) evoked by low frequency stimulation as well as the first volley in a train remained of normal amplitude and wave form almost until the onset of SD, and then disappeared (Fig. 2A). The relative resistance of the presynaptic volley to hypoxia has been noted before 27'29. The presynaptic volleys evoked by the later pulses during high frequency train stimulation were, however, de-

Orthodromic Evoked Potential

[Ca2+]

Antidromic Evoked Potential

[Ca 2+]

Control St. p. Blocked

Control St. R.

>

Blocked 1 5 ms

20 sec

Tstim. Fig. 1. Administration of 6,7-dinitroquinoxaline-2,3-dione (DNQX) (10/~M) plus 3-((+)-2-carboxypiperazine-4-yl)-propyl-l.phosphonic acid (CPP) (20/~M) blocks postsynaptic responses, but preserves presynaptic and antidromic responses. Control records were obtained in normal ACSF, blocked recordings in the presence of the glutamate antagonist drugs. St.P.: recordings in stratum pyramidale; St.R.: recordings in stratum radiatum. Arrows mark moment of stimulation. Orthodromic volleys evoked by stimulation of Schaffer collateral-commissural bundle in st. radiatum of CA1; antidromic volleys by stimulation of alveus. A[Ca2+]o responses evoked by trains of 20 Hz orthodromic stimuli for 5 s indicated by horizontal bars below traces. Left side calibration marks refer to evoked potentials; right side calibrations refer to linearly scaled recordings of interstitial calcium, [Ca2++]o. All recordings are extracellular. Downward deflection in voltage traces indicates negativity. Note that trace of [Ca2+]o during stimulus train is distorted by stimulus artefacts that have been partially filtered by computer processing.

72

J ¢, I ¸,

q }Ss

'1

~ gss

P

c-

r

155s i

,. " , _

-. . . .

-_ _ ~ . _ _

~f 'A .......... I i

4

/-

.--

_

....

--

.~-

,i

'./

,/

,i

:!34s

r: m

'1

Fig. 2. Presynaptic impulse volleys are more slowly depressed during hypoxia than are presynaptic d[Ca2+]o responses. All recordings obtained after blockade of postsynaptic responses. A: presynaptic volleys evoked in st. radiatum by first pulse in a 20 Hz, 5 s orthodromic train stimulus. B: the last presynaptic volleys in the same trains. C: polygraph pen recordings of presynaptic A[Ca2+]o responses evoked in st. radiatum by the same orthodromic trains; timing of train indicated by horizontal bar under lowest trace. Control: before oxygen deprivation. Times following oxygen withdrawal indicated by numbers at end of traces in column A. Letters (a-f) correspond to those used in Fig. 3.

pressed during hypoxia (Fig. 2B). As is clear from comparing the recordings in Fig. 2B and C, the presynaptic A[Ca2+]o responses were depressed faster than were presynaptic impulses. The discrepancy is greatest in row 'e' of Fig. 2, containing recordings made 275 s after oxygen withdrawal, when presynaptic A[Ca2÷]o failed entirely, while the presynaptic volley was still evoked albeit smaller than normal. The mean readings from 20 experiments are plotted in Fig. 4. Because the rate at which hypoxia took effect varied greatly from slice to slice, for the purpose of calculating mean values the readings of [CaZ+]o were lined up with the onset of SD taken to be zero time. Standard errors calculated in this manner were much smaller than when readings were averaged relative to the m o m e n t of oxygen withdrawal. Since the time of stimulation and hence the times of measurement varied relative to 'zero' time, error bars for both timing (abscissa) and magnitude (ordinate) are shown in Fig. 4; moreover, the number of measurements averaged for each point varies. The 'resting' or baseline level of [Ca2+]o began to decline gradually at about the same time as the orthodromically evoked A[Ca2+]o responses (Figs. 3 and 4). In some experiments the slow decrease of [Ca2+]o was preceded by a small increase (not illustrated, but contribut-

ing to the small rise of the m e a n [Ca2+]o baseline seen during -360 to -240 s in Fig. 4). The baseline [Ca2+]o decrease gradually accelerated thereafter, until its precipitous drop and the coincident large negative sustained potential shift signalled the onset of SD-like hypoxic depolarization (Fig. 3 and refs. 14 and 33). Following timely reoxygenation the [Ca2+]o baseline rose rapidly to between 0.9 and 1.1 mM, then decreased again somewhat and finally slowly and hesitantly returned to the control level of 1.2 m M over several minutes. In this entire period of moderately reduced [Ca2+]o the responses evoked by orthodromic train stimuli remained depressed, but they recovered to pre-hypoxic control amplitude when the baseline [Ca2+]o returned to normal (Fig. 3). Hypoxic changes evolved faster in normal A C S F than in the presence of glutamate receptor blocking drugs. The difference was consistently noted when a dual-well chamber was used, with one slice in normal A C S F and the other in A C S F containing D N Q X and CPP, as well as in experiments when the measurements were made in the same slice first in control A C S F and then in the presence of the glutamate antagonist compounds. The SD-like depolarization began after an average of 189 (S.E.M. + 16, n = 10) s hypoxia in control A C S F and 289 s ( + 31; P < 0.001; n = 10, NB: only paired com-

73

C i

I

d .............

25

1 0590

230

250

1~265

12a5

~g. . . . . . . .

1305

1.~25

1.4

b .............

1.25

210

13 . . . . .

1.0

i0570 125

11o

~o

~o

08

/

c~ (3

(3

.........

0,6

)

0.4

I.t5~

N2 0.2

, °%

!

1 02

,;

.

0.0 -200

,

0

200

,

400

600 800 SECONDS

,

1000

1200

r

1400

Fig. 3. Presynaptic A[fa2+]o responses are suppressed and baseline [Ca2+]o declines during hypoxia. The central graph traces [Ca2+]o in st. radiatum after blockade of postsynaptic responses (as in Fig. 1) during hypoxia. The solid arrows indicate trains of orthodromic stimuli. Inserts marked a-g are expanded from corresponding segments of main graph, illustrating individual d[Ca2+]o responses evoked by trains of orthodromic stimuli (a-d are computer-processed, linearly scaled versions of the similarly marked responses in Fig. 2C). Oxygen was withdrawn at zero time, and supplied again at the open arrow. 1220

20

3

8

12

1 200

'~

L]:

-Z~

~

11 ~

19

1140 1120

c~

1100

14

11 1 4 0

T j~

11 60

-~

160

12

~

1180 ;

~

16

i

C

1 20

~

l

loo~=

/ T

so

~"

"-

:

}

6o ~'

1 060 1 O4O 1 020 1 000

0 -360

-300

-240

-1 80

-1 20

-60

0

Time before SD [sec]

Fig. 4. Mean [Ca2+]o baseline (upper graph) and presynaptic d[CaZ+]o responses (bars) recorded at corresponding moments in 20 experiments similar to the one of Figs. 2 and 3. Abscissa shows time before SD-like depolarization. C: control response before oxygen withdrawal, all other responses during hypoxia. Numbers above graph show numbers of measurements. Error bars show standard errors of the means for both [Ca2+]o and time. Time of measurements and numbers of observations vary because of variations in the timing of the recordings relative to SD onset.

parison trials included) in the presence of the drugs. [Ca2+]o d r o p p e d during SD to 0.199 in control solution and only to 0.298 m M when glutamate receptors were blocked (n = 8), but the difference was not consistent and statistically not significant. Because D N Q X and CPP are selective antagonists of glutamate receptors, a minor contribution of post-synaptic Ca 2÷ currents due to other transmitters could not be ruled out in slices treated with these agents. To reliably block all transmission, in 4 experiments the slice was bathed in low Ca2+-medium with Mg 2÷ concentration doubled. Tracings of [Ca2+]o generally showed m o r e noise in r e d u c e d Ca 2+ m e d i u m than in control ACSF. W h e n oxygen was withdrawn, the orthodromically e v o k e d presynaptic [Ca2÷]o responses were depressed and the baseline [Ca2+]o declined, similarly to the slices treated with D N Q X plus CPP.

74 DISCUSSION The suppression of the orthodromically evoked decreases of [Ca2+]o during hypoxia confirms that presynaptic voltage dependent Ca 2+ currents are suppressed early during hypoxia. Several mechanisms could account for this failure. Impulses could have failed to arrive at presynaptic terminals due to conduction block at branch points, as has been suggested for neuromuscular junction ~7"21 and demonstrated recently in hypoxic dorsal root ganglia 36. In the case of the hippocampal Schaffer collateral-pyramidal cell junction this seems less likely, because synapses here are said to be formed by boutons en passage and not by branched terminals 37. The depression of presynaptic volleys evoked by high frequency train stimulation (Fig. 2B) nevertheless indicates blockade of impulse conduction in some presynaptic fibers, or else that the mean unit impulse amplitude was depressed. That presynaptic conduction failure could not have been the only factor, is evident from the slow suppression of presynaptic volleys compared to the more rapid loss of orthodromic A[Ca2+]o responses (Fig. 2). It should also be remembered that synaptic transmission of single volleys also fails during hypoxia at a time when single, low frequency presynaptic volleys are not yet depressed at all. Unfortunately, the presynaptic A[Ca2+]o responses evoked by single volleys are too small to be resolved without averaging, and averaging is not practical during the rapid changes induced by hypoxia. We can conclude that both impulse conduction block and failure of membrane Ca 2+ channels play a part in synaptic failure, but the relative weight of the two mechanisms remains to be determined. Another factor may have been depletion of adenosine triphosphate (ATP), for voltage gated Ca 2+ channels are believed to depend on ATP levels2°. A third possible mechanism could be the activation of potassium conductance, known to occur during hypoxia in some but not all neuron membranes 5'a5'23'24. Dorsal root ganglion (DRG) cells are believed to express many of the properties of presynaptic terminals. In freshly dissociated DRG cells exposure to low glucose or to cyanide caused enhancement of potassium currents t°, but in similar cells within the (undissociated) ganglion no hyperpolarization was seen during hypoxia36. Perhaps the most important factor was a rise in the 'resting' or baseline level of free intracellular calcium, [Ca2+]i, as voltage gated Ca 2+ channels are also known to be inactivated by a rise of [Ca2+]i 16. As shown in Figs. 3 and 4, the baseline or 'resting' [Ca2+]o was declining slowly, while the stimulus-induced responses were being suppressed. This early, gradual decline of the baseline level of [Ca2+]o suggests influx of C a 2+ into cells, among

them into presynaptic terminals. Duchen et al.tM have recorded a reversible increase of [Ca2*]i in isolated cells during energy failure due to glucose deprivation and cyanide poisoning. These observations are as expected, if a rise of [Ca2+]~ is the cause of the inactivation of presynaptic Ca 2+ channels. It may seem paradoxical that [Ca2+]i should increase, while the entry of Ca 2+ ions through membrane channels is blocked. Yet the 'resting' or 'background' level of [Ca2+]o was declining at the same time as the stimulus-induced [Ca2+]o responses were suppressed (Figs. 3 and 4). It seems that Ca 2+ ions were leaking into cells through transport other than the voltage dependent Ca 2+ channels, which were becoming blocked. The influx may reflect failure of the ATP dependent and/or Na+-coupled transport which normally moves Ca 2+ outward through the plasma membrane. Besides leaking inward, [Ca2+]i may also rise by being released from intracellular storage sites. The increase of the "resting' level of [Ca~+]i is expected to stimulate 'spontaneous' or background release of transmitter substance, but at the same time to inactivate the voltage dependent surges responsible for transmitter release. The reported increased overflow of glutamate and other transmitters during hypoxia in the central nervous system (CNS) 4'28 m a y , in part, be caused by the increased 'resting' [Ca2+]i. If so, then this excess release would be independent from impulse activity. Hypoxia interferes with membrane function in more than one way. G16tzner 13 described hyperpolarization of neurons in ischemic cerebral cortex. A similar hyperpolarization of hypoxic CA1 pyramidal cells has been attributed to increased membrane potassium conductance 12'15'25, possibly by Ca 2+ dependent K + channels. Not all central neurons show, however, hyperpolarization 5'~2"23'34. In the cases where potassium conductance does increase, it undoubtedly hastens the failure of excitatory synaptic transmission, but because of its inconsistency it cannot fully explain it. In hippocampal CA1 pyramidal cells increase of potassium conductance is frequently observed 12"15'24, whereas in dentate granule cells it is not 23'34. This difference may account for the faster failure of synaptic transmission in CA1 compared to dentate gyrus3. The questions remain, whether failure of EPSPs could account for the early, reversible blockade of synaptic transmission, and whether the suppression of the presynaptic A[Ca2+]o response could account for the failure of EPSPs. The orthodromic presynaptic A[Ca2+]o responses disappeared after 3-5 min of hypoxia, whereas the transmission of orthodromic population spikes is blocked in 1-2 min in the same preparation 3. The discrepancy may in part be due to the presence of the glutamate antago-

75 nist drugs, which slowed not only the onset of SD but also the prodromal slow decrease of [Ca2+]o. Besides, congruence of the depression of orthodromic population spikes and A[Ca2+]o is not expected, because of the nonlinearities in the system. The input-output function linking fEPSP to orthodromic population spike amplitude 2 suggests that the spike should be blocked completely when the fEPSP still retains 25-30% of its control magnitude. The relationship of presynaptic Ca 2÷ flux to EPSP is also not linear. While the exact function linking Ca 2÷ flux to EPSP at this synapse is not known, data show that reducing [Ca2+]o by 50% (from the normal 1.2 mM to 0.6 mM), depresses fEPSP to about 20% of normal 8. Assuming that [Ca2+]i remains constant, [Ca2+]o represents the driving force for Ca 2÷ current. In view of the double non-linearity, a relatively small impairment of presynaptic Ca 2÷ flux is expected to disproportionately jeopardize synaptic function. In a report based on intracellular recordings from CA1 neurons, Fujiwara et al. 12 concluded that EPSPs are relatively resistant to hypoxia. As we just emphasized, EPSPs need not disappear altogether in order to block orthodromic impulse transmission. Fujiwara et al.'s 12 study was not designed to detect quantitative changes of EPSPs and it was based on a limited sample of cells. The suppression of synaptic potentials was observed in all other studies expressly designed to follow their course quantitatively in a neuron pool (reviewed in refs. 31 and

REFERENCES 1 Adams, D.J., Takeda, K. and Umbach, J.A., Inhibitors of calcium buffering depress evoked transmitter release at the squid giant synapse, J. Physiol., 369 (1985) 145-159. 2 Balestrino, M., Aitken, EG. and Somjen, G.G., The effects of moderate changes of extracellular K + and Ca 2+ on synaptic and neural function in the CA1 region of the hippocampal slice, Brain Research, 377 (1986) 229-239. 3 Balestrino, M., Aitken, P.G. and Somjen, G.G., Spreading depression-like hypoxic depolarization in CA1 and fascia dentata of hippocampal slices: relationship to selective vulnerability, Brain Research, 497 (1989) 102-107. 4 Benveniste, H., Drejer, J., Schousboe, A. and Diemer, N.H., Elevations of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis, J. Neurochem., 43 (1984) 1369-1374. 5 Caspers, H., Speckmann, E.J. and Lehmenkiihler, A., DC potentials of the cerebral cortex, Rev. Physiol. Biochem. Pharmacol., 106 (1987) 127-178. 6 Cz6h, G. and Somjen, G.G., Hypoxic failure of synaptic transmission in the isolated spinal cord, and the effects of divalent cations, Brain Research, 527 (1990) 224-233. 7 Dingledine, R., Brain Slices, Plenum, New York, 1984. 8 Dingledine, R. and Somjen, G.G., Calcium dependence of synaptic transmission in the hippocampal slice, Brain Research, 207 (1981) 218-222. 9 Duchen, M. and Somjen, G., Effects of cyanide and low glucose on the membrane currents of dissociated mouse primary sensory neurones, J. Physiol., 401 (1988) 61P-61E

32). Depression of EPSPs therefore seems to be a major factor and possibly the most important one in interrupting transmission at central synapses. Synaptic inhibition is reportedly suppressed during hypoxia faster than is synaptic excitation 12'22'25. The difference may be due in part or in whole to the presence of an interneuron in inhibitory pathways. The depression of excitation of the interneuron is added to the direct depression by hypoxia on the inhibitory presynaptic terminal. A specific vulnerability of inhibitory synaptic terminals cannot, however, be ruled out. In tracings such as Fig. 3 the early gradual decrease of [Ca2+]o appears continuous with the sudden, precipitous drop associated with the onset of the SD-like depolarization. The two may, but need not, be different phases of the same non-physiological influx of Ca 2÷ ions into cells. Blocking concentrations of DNQX plus CPP did not prevent the SD-like precipitous decrease of [Ca2+]o, although they did postpone its onset and perhaps attenuate it slightly. It seems therefore that Ca 2+ enters cells during the hypoxic SD-like response mainly through paths other than quisqualate or NMDA receptor controlled channels. Acknowledgments. We thank Dr. Uwe Heinemann for suggesting this technique for recording presynaptic calcium currents. Ms. Manuela Fernandez manufactured ion selective electrodes. The work was supported by Grants NS 18670 and NS 06233 from the NIH, NINDS.

10 Duchen, M.R., Effects of metabolic inhibition on the membrane properties of isolated mouse primary sensory neurones, J. Physiol., 424 (1990) 387-409. 11 Duchen, M.R., Valdeolmillos, M., O'Neill, S.C. and Eisner, D.A., Effects of metabolic blockade on the regulation of intracellular calcium in dissociated mouse sensory neurones, J. Physiol., 424 (1990) 411-426. 12 Fujiwara, N., Higashi, H., Shimoji, K. and Yoshimura, M., Effects of hypoxia on rat hippocampal neurones in vitro, J. Physiol., 384 (1987) 131-151. 13 G16tzner, F., Intracellul~re Potentiale, EEG und kortikale Gleichspannung an der sensorimotorischen Rinde der Katze bei akuter Hypoxie, Arch. Psychiatr. Nervenkr., 210 (1967) 274296. 14 Hansen, A.J., Effects of anoxia on ion distribution in the brain, Physiol. Rev., 65 (1985) 101-148. 15 Hansen, A.J., Hounsgaard, J. and Jahnsen, H., Anoxia increases potassium conductance in nerve cells, Acta Physiol. Scand., 115 (1982) 301-310. 16 Hille, B., Ionic Channels of Excitable Membranes, Sinauer Associates, Sunderland, Massachusetts, 1984. 17 Hubbard, J.F. and LCyning, Y., The effects of hypoxia on neuromuscular transmission in a mammalian preparation, J. Physiol., 185 (1966) 205-223. 18 Jones, R.S.G. and Heinemann, U., Pre- and postsynaptic K ÷ and Ca 2+ fluxes in area CA1 of the rat hippocampus in vitro: effects of Ni 2+, TEA and 4-AP, Exp. Brain Res., 68 (1987) 205-209. 19 Konnerth, A. and Heinemann, U., Effects of GABA on presumed presynaptic calcium entry in hippocampal slices, Brain Research, 270 (1983) 185-189.

76 20 Kostyuk, P.G., Metabolic control of ionic channels in the neuronal membrane, Neuroscience, 13 (1984) 983-989. 21 Krnjevi6, K. and Miledi, R., Presynaptic failure in neuromuscular propagation in rats, J. Physiol., 149 (1959) 1-22. 22 Krnjevi6, K., Xu, Y.Z. and Zhang, L., Anoxic block of GABAergic IPSPs, Neurochem. Res., 16 (1991) 279-284. 23 Krnjevi6, K. and Ben-Ari, Y., Anoxic changes in dentate granule cells, Neurosci. Lett., 107 (1989) 89-93. 24 Krnjevi6, K. and Leblond, J., Changes in membrane currents of hippocampal neurons evoked by brief anoxia, J. Neurophysiol., 62 (1989) 15-30. 25 Leblond, J. and Krnjevi6, K, Hypoxic changes in hippocampal neurons, J. Neurophysiol., 62 (1989) 1-14. 26 Rader, R.K., Lanthorn, T.H. and Lipton, P., Effects of hypoxia on responses to acidic amino acids in the in vitro hippocampal slice, Soc. Neurosci. Abstr., 13 (1987) 1495. 27 Schiff, S.J. and Somjen, G.G., Hyperexcitability following moderate hypoxia in hippocampal tissue slices, Brain Research, 337 (1985) 337-340. 28 Schmidt-Kastner, R. and Freund, T.E, Selective vulnerability of the hippocampus in brain ischemia, Neuroscience, 40 (1991) 599-636. 29 Sick, T.J., Solow, E.L. and Roberts, E.L., Extraeellular potassium ion activity and electrophysiology in the hippocampal slice: paradoxical recovery of synaptic transmission during anoxia, Brain Research, 418 (1987) 227-234. 30 Somjen, G.G., Stimulus-evoked and seizure-related responses of extracellular calcium activity in spinal cord compared to those in cerebral cortex, J. Neurophysiol., 44 (1980) 617-632.

31 Somjen, G.G., Basic mechanisms in cerebral hypoxia and stroke: background, review and conclusions. In G.G. Somjen (Ed.), Mechanisms of cerebral hypoxia and stroke. Plenum, New York, 1988, pp. 447-466. 32 Somjen, G.G., Mechanism of the reversible arrest of function during transient cerebral hypoxia and ischemia. In A. Schurr and B.M. Rigor (Eds.), Cerebral lschemia and Resuscitation, CRC Press, Boca Raton, 1990, pp. 301-317. 33 Somjen, G.G., Aitken, P.G., Balestrino, M.. Herreras, O. and Kawasaki, K., Spreading depression-like depolarization and selective vulnerability of neurons. A brief review, Stroke, 21 (1990) III179-III183. 34 Somjen, G.G., Aitken, P.G., Balestrino, M. and Schiff, S.J., Uses and abuses of in vitro systems in the study of the pathophysiology of the central nervous system. In A. Schurr, T. Teyler and M. Rigor (Eds.), Brain Slices: Fundamentals, Applications and Implications, Karger, Basel, 1986, pp. 89-104. 35 Somjen, G.G., Schiff, S.J., Aitken, P.G. and Balestrino, M., Forms of suppression of neuronal function. In N. Chalazonitis and M. Gola (Eds.), Inactivation of Hypersensitive Neurons, Alan R.Liss, New York 1987, pp. 137-145. 36 Urb~in, L. and Somjen, G.G., Reversible effects of hypoxia on neurons in mouse dorsal root ganglia in vitro, Brain Research, 520 (1990) 36-42. 37 Westrum, L.E. and Blackstad, T.W., An electron microscopic study of stratum radiatum of the rat hippocampus (regio superior, CA1) with particular emphasis on synaptology, J. Comp. Neurol., 119 (1962) 281-309.