- Email: [email protected]
Activators of ATP-sensitive K+ channels reduce anoxic depolarization in CA3 hippocampal neurons
0306-4522/90 $3.00 + 0.00 Pergamon Press plc 0 1990 IBRO
Neuroscimcc Vol. 37, No. I, pp. 55-60, 1990 Printedin Great Britain
ACTIVATORS OF ATP-SENSITIVE K+ CHANNELS REDUCE ANOXIC DEPOLARIZATION IN CA3 HIPPOCAMPAL NEURONS Y. BEN-ARI, K. KRNJEVI~* and V. CRBPEL INSERM
U. 29, HBpital de Port-Royal,
123 Bd de Port-Royal,
75014 Paris, France
AbstracG-In
CA3 hippocampal neurons of the rat, brief anoxic episodes produce a depolarization which is probably due to a synaptic release of glutamate. Diazoxide, an activator of ATP-sensitive K+ channels (K+ ATP), blocks the anoxic depolarization and has no effect in control oxygenated artificial cerebrospinal fluid. The hormone somatostatin which activates K+ ATP channels in the pancreas also reduces the anoxic depolarization in CA3 neurons. We suggest that drugs that open K+ ATP channels may constitute a novel approach to selectively reducing the deleterious effects of excessive release of glutamate during anoxia without producing a generalized blockade of glutamatergic synaptic transmission.
cut using a McIllwain tissue chopper and incubated at room temperature in ACSF for at least 1h before use. Individual @ices were transferred to a chanber in which they were kept submerged and superfused with ACSF a1 2.5-3.0 ml/min at 34°C. Mossy fibre excitatory postsynaptic potentials were produced by electrical stimulation of the granule cell region with a bipolar tungsten wire electrode. Intracellular recordings were made with 3 M KC1 electrodes (resistance of 40-80 MR). Current was passed through the recording electrode by means of an Axoclamp 2 amplifier. Bridge balance was checked repeatedly during the experiment. For voltage clamp experiments, a single electrode voltage clamp amplifier was used and the sampling frequency was 3-4 kHz, 30% duty cycle. This was set to operate at a sampling frequency of 3-5 kHz, gain 8-25 nA/mV, and output bandwidth O-l kHz. Capacitance neutralization was optimized to ensure a satisfactory performance of the single electrode voltage clamp, the voltage at the head stage amplifier being monitored throughout on a separate oscilloscope. Anoxia was produced by switching to ACSF equilibrated with 95% Nz and 5% CO,; because of the small dead space of the system, the exchange solution had an initial delay of only 15-20 s and it was 90% comnlete bv 40-60 s. The followine drugs were bath applied: te&xlotoiin (Sigma), kynurenate (gift of Dr Herrling, Sandoz), diazoxide (gift of Unicet), somatostatin (Sigma).
Even brief periods of anoxia have a dramatic effect on brain function.1° In addition to the reversible blockade of synaptic transmission, repeated or prolonged anoxic episodes are followed by delayed neuronal death, especially in the hippocampus.‘* There is now good evidence that this is caused by excessive release of glutamate, resulting in a cytotoxic entry of Ca*+ through the N-methyl-D-aspartate (NMDA) receptor coupled Ca2+ permeable channel.536 The present trend in the search for substances that protect neurons from anoxic death is to design drugs that antagonize glutamate at the NMDA receptor-channel complex. Clearly an alternative approach would be to prevent the cytotoxic release of glutamate produced by anoxia/ischemia. We now report that diazoxide and the hormone somatostatin, agents known to activate K+ channels inhibited by intracellular ATP’,’ (KiT,, channels1.‘6) reduce the anoxic depolarization of CA3 neurons due to the release of glutamate. These effects are not associated with a change in the postsynaptic response to glutamate agonists. Previously, the hyperglycemic hormone galanin (which activates K+ ATP channels in the pancreas*) was found to reduce the effects of anoxia on hippocampal neurons.2.4 These observations have been presented in brief elsewhere.’ EXPERIMENTAL
RESULTS In agreement with recent observations2,4 intracellular recordings from CA3 neurons of hippo-
campal slices‘ (n = 34) showed that 2-4 min anoxic episodes had a depolarizing effect on CA3 neurons (20 out of 25 cells and in 76 out of 96 anoxic tests). This is probably due to glutamate release, since it was blocked by tetrodotoxin (1 PM, n = 3), or by the broad spectrum glutamate antagonist kynurenate (1 mM, n = 3 also see Ref. 2). The depolarization was often preceded by a variable hyperpolarization associated with a fall in input resistance. The hyperpolarization was probably due to activation by anoxia of a calcium-dependent K+ conductance.“,” Following return to oxygenated solution, there was a characteristic late hyperpolarization due to the
PROCEDURES
Adult male Wistar rats were anaesthetized with ether and decapitated. The brain was removed and submerged in artificial cerebrospinal fluid (ACSF) of the following composition (in mM): NaCl 126, KC1 3.5, CaCl, 2, MgCl, 1.3, NaHCO, 25 and glucose IO (PH 7.3) and gassed with 95% 0, and 5% CO,. Approximately 600-pm-thick slices were
*On leave from Anaesthesia Research Department, McGill University, Montrtal, Canada. Abbreviorions: ACSF, artificial cerebrospinal fluid; NMDA, N-methyl-D-aspartate. 55
Fig. I. Djazoxide has no hy~ia~zin~ effect on CA3 pyramjda~ neuron, but it ran eliminate depolarizing effects of anoxia. thus revealing underline anoxic and posr-anoxic by~~~a~zations. Recordingwith 3 M KC1electrodeshows
(Nat
K+) AT&W
PnmP:.“.‘S In crxygenated ACSF, diazoxide ~~.87mM~ induced no tn = 6 neuror~s) or minima! change (3.4 & I .69mii d~~l~~j~~~on. mean f S.E. II = 3 neurons) in membrane potential. A small increase in input resistance (Fig. 1A) was observed in nine out af I1 tests (IO.4 f 2.53%%),in contrast. in the presence of oxygen deficient ACSF (Fig. I Bj. diazonidc blocked
the anaxic de~lar~~tion~ thus revealing the anoxic and p~~anoxic hy~r~~ari~tions. At the same time. the maximum reduction in input resistance caused by anoxia feB from 80 to 60% in the presence of diazoxide. Rec5very from the effects of diazaxide was slow, typically being incomplete even after 3Qmin (Eig, iC-D). Similar clear reductions in anoxic de~o~a~~a~~o~s or inward currents in single electrode voltage clamp experiments (see Fig. 2) were
57
Activators of ATP-sensitive K+ channels reduce anoxic depolarization CA3
VH -7OmV AV -2OmV
iN,+
DIAZOXIDE
(Li
17 mln
RECOVERY 2Rmln
Fig. 2. Diazoxide reduces inward current generated during anoxia. Membrane current of CA3 pyramidal neuron was recorded by single electrode voltage clamp, at holding potential (Vn) of - 70 mV. In A, lower trace shows membrane current and upper trace voltage commands (-20 mV, lasting 200 ms) as recorded at fast and slow chart paper speeds. During the 3-min period of anoxia (N2) note slow inward shift and conductance increase by 160%, as well as sharp inward transients, associated with variable degree of uncontrolled firing (no tetrodotoxin was applied to block Na currents). (B) Prolonged application of diazoxide in bath (0.87 mM) caused only very small changes-a 30 pA inward shift in holding current and 8.3% reduction in input conductance-but inward current transients evoked by anoxia were greatly reduced. Maximum conductance increase during anoxia fell to +69%. (C) Return of large anoxic inward transients and firing 29 min after end ofdiazoxide application; conductance now increased by 115% during anoxia. Note that in C the anoxic episode had a duration of 3 min (as in other records).
observed in eight tests out of nine; only in one cell where anoxia had a very large depolarizing effect (>30mV), diazoxide failed to reduce the depolarization. As shown in Table 1, the overall shifts in membrane potential obtained by graphical integration of depolarizing and hyperpolarizing shifts indicate a statistically significant reduction of the anoxic depolarization by diazoxide (P -c 0.02, paired r-test). Unlike diazoxide, somatostatin, applied via oxygenated ACSF, often had (n = 12 out of 21) a hyperpolarizing effect (Fig. 3) (3.8 &-0.57 mV) associated
with a fall in cell input resistance (20.7 + 4.79%). In the remaining neurons, somatostatin induced either no change in membrane potential (Fig. 4, n = 4) or a small depolarization (+5.1 + 2.1 mV, n = 5). In oxygen deficient ACSF, somatostatin reduced or eliminated the anoxic depolarization (Figs 3 and 4, in nine out of 13 tests in eight neurons), even when somatostatin per se had no effect in oxygenated ACSF (Fig. 4). Somatostatin or diazoxide did not block glutamate receptors. In fact, the depolarizations (Fig. 5)
Table I. Overall shifts in V m during and immediately after 2-4-min periods of anoxia, observed before, during and after superfusion with diazoxide (0.87 mM) Controls before diazoxide (A)
During diazoxide (B)
Diazoxide-induced changes (B-A)
Controls after diazoxide (C)
Post-diazoxide changes (C-B)
t-O.61
-1.01 (kO.54) n=9
- l.63* (kO.50)
-0.041 (2 0.302)
+ 1.64** (kO.34)
( f 0.437) n=9
n=9
n=7
n=7
Values given (in V s. means ? SE.) were obtained by graphical integration of depolarizing and hyperpolarizing shifts from the baseline, and then summing the resulting positive and negative totals. For purposes of comparison with current clamp data, outward and inward current shifts recorded by single electrode voltage clamp (n = 3) were converted to voltage changes by dividing them by the input conductance (in a first approximation assumed to remain constant). Asterisks indicate paired comparison; statistically significant difference (*P < 0.02, +*p c 0.005).
58
Y. BEN-ARIel a/.
A
20ms
N2
c
t
t
0.5s i,,
D ~ 6 7
soM
N2
I
2rain
i ............................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
t
N2
E
t
~
14/ -.69 . . . .
,
I
.
I I
1
'
Fig. 3. Somatostatin can also reduce or eliminate N2-evokecl depolarization of CA3 pyramidal neuron, while producing only small changes in Vm, R or excitatory postsynaptic potentials. (A, B) Excitatory postsynaptic potential and spike evoked by stimulating dentate gyrus, before and after 9 rain of superfusion with 1 taM somatostatin (SOM). (C) In same cell, before application of somat~tatin, anoxia caused early hyperpolarization, with superimposed depolarization and firing. (D) Somatostatin (1 taM) itself producecl a small hyperpolarization and a 10% fall in input resistance while the anoxic depolarization and firing disappeared; (E) 3 rain after the end of 14-min somatostatin application, a 3-rain anoxic test had a marked depolarizing effect. or inward currents (not shown) evoked by the potent glutamate agonist quisqualate (10#M) were not reduced by diazoxide (0.87 mM, n = 2). Furthermore, the mossy fibre excitatory postsynaptic potentials evoked by electrical stimulation of the granular layer were not depressed by somatostatin (Fig. 3A-B, n = 3) (or diazoxide, Fig. 5A, n --- 4). These observations further suggest that the effects of K + ATP activators occur at a presynaptic level. DISCUSSION
Thus, in CA3 neurons, diazoxide and somatostatin had effects similar to those of galanine2"4 in reducing anoxic depolarizations. It is relevant to mention that the effects of galanin are blocked by the antidiabetic sulfonylurea glibenclamide2"4 which selectively blocks K ~ channels in the pancreasJ 9 Taken together these observations suggest that ATP K + channels play a role in the anoxic responses of CA3 neurons. The present results obtained with diazoxide fully accord with this notion since this drug is known to activate K~,rp channels in pancreatic B cells and
other tissuesJ "2° Although somatostatin had more complex effects than diazoxide including a hyperpolarization (presumably due to the activation of an M-type K + channeP 4' mr), the reduced anoxic depolarization in the presence of somatostatin may be explained by a diazoxide like action on K + ATP channels. We suggest that the anoxic effects of K + ATP channel activation were generated presynaptically primarily via the mossy fibers. This would be consistent with the high density of high affinity glibenclamide binding sites in the mossy fibre region ~5 and with the observations previously made with galanin showing that in the presence of tetrodotoxin or kynurenate galanin had no effect on postsynaptic response of the pyramidal neurons to anoxia, as well as the absence of postsynaptic effects of tolbutamide or diazoxide in CA 1 pyramidal neurons. 2"ts~ This would also be consistent with the lack o f effects of diazoxide on the anoxic response of dentate granule cells from which the mossy fibres originateJ 3 In oxygenated ACSF, these channels are presumed to be closed. The sulphonylurea antagonists or K + ATP
A
N2
....s.o.m............................................................................ N2
B
-66mY
C
N2 -64mV Jlomv lmin
Fig. 4. Somatostatin eliminates N 2-evoked depolarization. In this CA3 neuron, a brief (2 min) anoxic episode induced a post-anoxic depolarization (A) which was fully blocked by somatostatin (B) and recovered after wash (C). A
B
auls
20 +
I
I i [~
'~ ~
DIAZO
---
+
I
lmin
~
'
II~HIIIt, t
.,~,,.,~,.~
Fig. 5. Diazoxide has little effect on evoked excitatory postsynaptic potentials or quisqualate (QUIS)induced depolarizations. Recording was from CA3 pyramidal layer neuron, with 3 M KC1 electrode. (A) Each trace shows from, left to right, excitatory postsynaptic potential and spike(s) evoked by dentate stimulus, and the voltage change produced by hyperpolarizing current pulses. Upper trace is initial control; middle, was obtained after 3 min of superfusion with diazoxide (0.87 mM); note slight increase in R N ( + 12%), and in evoked excitatory postsynaptic potential; lower trace was recorded 7 min after end of superfusion with diazoxide. Vm remained within 1 mV of - 6 2 mV throughout; hyperpolarizing current pulses were 0.7, 0.7 and 1.0 nA from top to bottom. (B) In another CA3 neuron, marked excitation (and depolarization) evoked by 1-min applications of QUIS (10/tM) were not reduced by diazoxide (0.87 mM). Upper and lower traces are pre- and post-control runs; resting membrane potential remained at - 73 mV throughout, and - 0 . 4 nA current pulses were applied at regular intervals.
Y. BEN-ARI el al.
60
[K];” would gradually depolarize the nerve terminals and thus enhance on-going transmitter release: but as the ATP levels fall, activation of K,+,, opposes terminal depolarization and the increase in release. Sulphonylurea drugs now become effective, because they eliminate the opposing action of KiTp and therefore facilitate depolarization and transmitter release. Whereas diazoxide and other activators channels further reduce the depolarizing of K;v tendency, and thus diminish release.
galanin and somatostatin or by the administration of diazoxide-like drugs, it may be possible to reduce significantly the anoxia triggered release of excitotoxic amino acids. This approach seems, a priori, preferable to the indiscriminate blocking of glutamate receptors which are widely involved in synaptic transmission throughout the central nervous system.
CONCLUSION
We suggest that by stimulating the production and/or secretion of endogenous peptides such as
Acknowledgements-We are grateful to Dr E. Cherubini and A. Nistri for their suggestions. Supported by INSERM.
REFERENCES
potassium channels. A. Rea. Neurosci. 11, 97-118. 2. Ben-Ari Y. (1990) Effects of galanin and glibenclamide on the anoxic response of rat hippocampal neurons in I. Ashcroft F. M. (1988) Adenosine 5’-triphosphate-sensitive
vitro.
Eur. J. Neurosci. 2, 62-68.
3. Ben-Ari Y. and KrnjeviC K. (1989) The ATP sensitive K+ channel opener diazoxide reduces anoxic depolarization in rat CA3 hippocampal neurones in isolated slices (P.) J. Physiol. 418, l92P. 4. Ben-Ari Y. and Lazdunski M. (1989) Galanin protects hippocampal neurons from the functional effects of anoxia. Eur. J. Pharmac. 165, 1331-1332.
5. Benveniste H., Drejer J., Schousboue A. and Diemer N. H. (1984) Elevation of the extracellular concentration of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdial. J. Neurochem.
43, 1369-1374.
6. Choi D. W. (1988) Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci. 11, 465469.
I. De Weille J., Schmid-Antomarchi H., Fosset M. and Lazdunski M. (1988) Regulation of ATP-sensitive K+ channels in insulinoma cells: activation by somatostatin and protein kinase C and the role of cCAMP. Proc. nam. Acad. Sci. U.S.A. 86, 2971-2976.
8. De Weille J., Schmid-Antomarchi H., Fosset M. and Lazdunski M. (1988) ATP-sensitive K+ channels that are blocked by hypoglycemia-inducing sulfonylureas in insulin-secreting cells are activated by galanin, a hyperglycemia-inducing hormone. Proc. nafn. Acnd. .Sci. U.S.A. 85, 1312-1316. 9. Fujiwara N.. Higashi H., Shimoji K. and Yoshimura M. (1987) Effects of hypoxia on rat hippocampal neurones by tlitro. J. Physiol., Land. 384, I3 I -I 5 I. 10. Hansen A. J. (1985) Effect of anoxia on ion distribution in the brain. Physiol. Rec. 65, 101-148. Il. Hansen A. J.. Hounsgaard J. and Jahnsen H. (1982) Anoxia increases potassium conductance in hippocampal nerve cells. Acra physiol. stand. 115, 301-310. 12. Krnjevie K. and Ben-Ari Y. (1989) Anoxic changes in dentate granule cells. Neurosci Lefr. 107, 89-93. 13. KrnjeviC K. and Leblond J. (1987) Anoxia reversibly supresses neuronal calcium currents in rat hippocampal slices. Can. J. Physiol. 65, 2 157-2 161. 13a. KrnjeviC K. and Leblond 1. (1988) Are there hippocampal ATP-sensitive K channels that are activated by anoxia’! Pjltigers Arch. ges. Physiol. 411, suppl., R145. 14. Moore S. D., Madamba S. G., Joels M. and Siggins G. R. (1988) Somatostatin augments the m-current in hippocampal neurons. Science 239, 279-280. 15. Mourre C.. Ben-Ari Y.. Bernardi H.. Fosset M. and Lazdunski M. (1989) Antidiabetic sulfonylureas: localization of binding sites in the brain and effects on the hyperpolarization induced by anoxia in hippocampal slices. Brain Res. 486, 159-164. 16. Noma A. (1983) ATP regulated K+ channels in cardiac muscle. Nature 305, 147-148.
17. Pittman Q. J. and Siggins G. R. (1981) Somatostatin hyperpolarizes hippocampal pyramidal cells
in Glro.
Brain Res.
221, 402-408. 18. Pulsinelli W. A. (1985) Selective neuronal vulnerability: morphological and molecular characteristics.
Prog. Brain Res.
63, 29-37.
Schmidt-Antomarchi H.. De Weille J.. Fosset M. and Lazdunski M. (1987) The receptor for antidiabetic sulfonylureas controls the activity of the ATP-modulated K+ channel in insulin-secreting cells. J. biol. Chem. 262, 15,840-15.844. 20. Trube G.. Rorsman 0. and Ohno-Shosaku T. (1986) Opposite effects of tolbutamide and diazoxide on the ATP-dependent KC channel in mouse pancreatic B cells. I’piigers Arch. ges. Physiol. 407, 493-499. 21. Trube G. and Hescheler J. (1984) Inward-rectifying channels in isolated patches of the heart cell and membrane: ATP-dependence and comparison with cell-attached patches. P/tigers Arch. ges. Physiol. 401, 178-184. 19.
(Accepted 24 January 1990)
Neuroscimcc Vol. 37, No. I, pp. 55-60, 1990 Printedin Great Britain
ACTIVATORS OF ATP-SENSITIVE K+ CHANNELS REDUCE ANOXIC DEPOLARIZATION IN CA3 HIPPOCAMPAL NEURONS Y. BEN-ARI, K. KRNJEVI~* and V. CRBPEL INSERM
U. 29, HBpital de Port-Royal,
123 Bd de Port-Royal,
75014 Paris, France
AbstracG-In
CA3 hippocampal neurons of the rat, brief anoxic episodes produce a depolarization which is probably due to a synaptic release of glutamate. Diazoxide, an activator of ATP-sensitive K+ channels (K+ ATP), blocks the anoxic depolarization and has no effect in control oxygenated artificial cerebrospinal fluid. The hormone somatostatin which activates K+ ATP channels in the pancreas also reduces the anoxic depolarization in CA3 neurons. We suggest that drugs that open K+ ATP channels may constitute a novel approach to selectively reducing the deleterious effects of excessive release of glutamate during anoxia without producing a generalized blockade of glutamatergic synaptic transmission.
cut using a McIllwain tissue chopper and incubated at room temperature in ACSF for at least 1h before use. Individual @ices were transferred to a chanber in which they were kept submerged and superfused with ACSF a1 2.5-3.0 ml/min at 34°C. Mossy fibre excitatory postsynaptic potentials were produced by electrical stimulation of the granule cell region with a bipolar tungsten wire electrode. Intracellular recordings were made with 3 M KC1 electrodes (resistance of 40-80 MR). Current was passed through the recording electrode by means of an Axoclamp 2 amplifier. Bridge balance was checked repeatedly during the experiment. For voltage clamp experiments, a single electrode voltage clamp amplifier was used and the sampling frequency was 3-4 kHz, 30% duty cycle. This was set to operate at a sampling frequency of 3-5 kHz, gain 8-25 nA/mV, and output bandwidth O-l kHz. Capacitance neutralization was optimized to ensure a satisfactory performance of the single electrode voltage clamp, the voltage at the head stage amplifier being monitored throughout on a separate oscilloscope. Anoxia was produced by switching to ACSF equilibrated with 95% Nz and 5% CO,; because of the small dead space of the system, the exchange solution had an initial delay of only 15-20 s and it was 90% comnlete bv 40-60 s. The followine drugs were bath applied: te&xlotoiin (Sigma), kynurenate (gift of Dr Herrling, Sandoz), diazoxide (gift of Unicet), somatostatin (Sigma).
Even brief periods of anoxia have a dramatic effect on brain function.1° In addition to the reversible blockade of synaptic transmission, repeated or prolonged anoxic episodes are followed by delayed neuronal death, especially in the hippocampus.‘* There is now good evidence that this is caused by excessive release of glutamate, resulting in a cytotoxic entry of Ca*+ through the N-methyl-D-aspartate (NMDA) receptor coupled Ca2+ permeable channel.536 The present trend in the search for substances that protect neurons from anoxic death is to design drugs that antagonize glutamate at the NMDA receptor-channel complex. Clearly an alternative approach would be to prevent the cytotoxic release of glutamate produced by anoxia/ischemia. We now report that diazoxide and the hormone somatostatin, agents known to activate K+ channels inhibited by intracellular ATP’,’ (KiT,, channels1.‘6) reduce the anoxic depolarization of CA3 neurons due to the release of glutamate. These effects are not associated with a change in the postsynaptic response to glutamate agonists. Previously, the hyperglycemic hormone galanin (which activates K+ ATP channels in the pancreas*) was found to reduce the effects of anoxia on hippocampal neurons.2.4 These observations have been presented in brief elsewhere.’ EXPERIMENTAL
RESULTS In agreement with recent observations2,4 intracellular recordings from CA3 neurons of hippo-
campal slices‘ (n = 34) showed that 2-4 min anoxic episodes had a depolarizing effect on CA3 neurons (20 out of 25 cells and in 76 out of 96 anoxic tests). This is probably due to glutamate release, since it was blocked by tetrodotoxin (1 PM, n = 3), or by the broad spectrum glutamate antagonist kynurenate (1 mM, n = 3 also see Ref. 2). The depolarization was often preceded by a variable hyperpolarization associated with a fall in input resistance. The hyperpolarization was probably due to activation by anoxia of a calcium-dependent K+ conductance.“,” Following return to oxygenated solution, there was a characteristic late hyperpolarization due to the
PROCEDURES
Adult male Wistar rats were anaesthetized with ether and decapitated. The brain was removed and submerged in artificial cerebrospinal fluid (ACSF) of the following composition (in mM): NaCl 126, KC1 3.5, CaCl, 2, MgCl, 1.3, NaHCO, 25 and glucose IO (PH 7.3) and gassed with 95% 0, and 5% CO,. Approximately 600-pm-thick slices were
*On leave from Anaesthesia Research Department, McGill University, Montrtal, Canada. Abbreviorions: ACSF, artificial cerebrospinal fluid; NMDA, N-methyl-D-aspartate. 55
Fig. I. Djazoxide has no hy~ia~zin~ effect on CA3 pyramjda~ neuron, but it ran eliminate depolarizing effects of anoxia. thus revealing underline anoxic and posr-anoxic by~~~a~zations. Recordingwith 3 M KC1electrodeshows
(Nat
K+) AT&W
PnmP:.“.‘S In crxygenated ACSF, diazoxide ~~.87mM~ induced no tn = 6 neuror~s) or minima! change (3.4 & I .69mii d~~l~~j~~~on. mean f S.E. II = 3 neurons) in membrane potential. A small increase in input resistance (Fig. 1A) was observed in nine out af I1 tests (IO.4 f 2.53%%),in contrast. in the presence of oxygen deficient ACSF (Fig. I Bj. diazonidc blocked
the anaxic de~lar~~tion~ thus revealing the anoxic and p~~anoxic hy~r~~ari~tions. At the same time. the maximum reduction in input resistance caused by anoxia feB from 80 to 60% in the presence of diazoxide. Rec5very from the effects of diazaxide was slow, typically being incomplete even after 3Qmin (Eig, iC-D). Similar clear reductions in anoxic de~o~a~~a~~o~s or inward currents in single electrode voltage clamp experiments (see Fig. 2) were
57
Activators of ATP-sensitive K+ channels reduce anoxic depolarization CA3
VH -7OmV AV -2OmV
iN,+
DIAZOXIDE
(Li
17 mln
RECOVERY 2Rmln
Fig. 2. Diazoxide reduces inward current generated during anoxia. Membrane current of CA3 pyramidal neuron was recorded by single electrode voltage clamp, at holding potential (Vn) of - 70 mV. In A, lower trace shows membrane current and upper trace voltage commands (-20 mV, lasting 200 ms) as recorded at fast and slow chart paper speeds. During the 3-min period of anoxia (N2) note slow inward shift and conductance increase by 160%, as well as sharp inward transients, associated with variable degree of uncontrolled firing (no tetrodotoxin was applied to block Na currents). (B) Prolonged application of diazoxide in bath (0.87 mM) caused only very small changes-a 30 pA inward shift in holding current and 8.3% reduction in input conductance-but inward current transients evoked by anoxia were greatly reduced. Maximum conductance increase during anoxia fell to +69%. (C) Return of large anoxic inward transients and firing 29 min after end ofdiazoxide application; conductance now increased by 115% during anoxia. Note that in C the anoxic episode had a duration of 3 min (as in other records).
observed in eight tests out of nine; only in one cell where anoxia had a very large depolarizing effect (>30mV), diazoxide failed to reduce the depolarization. As shown in Table 1, the overall shifts in membrane potential obtained by graphical integration of depolarizing and hyperpolarizing shifts indicate a statistically significant reduction of the anoxic depolarization by diazoxide (P -c 0.02, paired r-test). Unlike diazoxide, somatostatin, applied via oxygenated ACSF, often had (n = 12 out of 21) a hyperpolarizing effect (Fig. 3) (3.8 &-0.57 mV) associated
with a fall in cell input resistance (20.7 + 4.79%). In the remaining neurons, somatostatin induced either no change in membrane potential (Fig. 4, n = 4) or a small depolarization (+5.1 + 2.1 mV, n = 5). In oxygen deficient ACSF, somatostatin reduced or eliminated the anoxic depolarization (Figs 3 and 4, in nine out of 13 tests in eight neurons), even when somatostatin per se had no effect in oxygenated ACSF (Fig. 4). Somatostatin or diazoxide did not block glutamate receptors. In fact, the depolarizations (Fig. 5)
Table I. Overall shifts in V m during and immediately after 2-4-min periods of anoxia, observed before, during and after superfusion with diazoxide (0.87 mM) Controls before diazoxide (A)
During diazoxide (B)
Diazoxide-induced changes (B-A)
Controls after diazoxide (C)
Post-diazoxide changes (C-B)
t-O.61
-1.01 (kO.54) n=9
- l.63* (kO.50)
-0.041 (2 0.302)
+ 1.64** (kO.34)
( f 0.437) n=9
n=9
n=7
n=7
Values given (in V s. means ? SE.) were obtained by graphical integration of depolarizing and hyperpolarizing shifts from the baseline, and then summing the resulting positive and negative totals. For purposes of comparison with current clamp data, outward and inward current shifts recorded by single electrode voltage clamp (n = 3) were converted to voltage changes by dividing them by the input conductance (in a first approximation assumed to remain constant). Asterisks indicate paired comparison; statistically significant difference (*P < 0.02, +*p c 0.005).
58
Y. BEN-ARIel a/.
A
20ms
N2
c
t
t
0.5s i,,
D ~ 6 7
soM
N2
I
2rain
i ............................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
t
N2
E
t
~
14/ -.69 . . . .
,
I
.
I I
1
'
Fig. 3. Somatostatin can also reduce or eliminate N2-evokecl depolarization of CA3 pyramidal neuron, while producing only small changes in Vm, R or excitatory postsynaptic potentials. (A, B) Excitatory postsynaptic potential and spike evoked by stimulating dentate gyrus, before and after 9 rain of superfusion with 1 taM somatostatin (SOM). (C) In same cell, before application of somat~tatin, anoxia caused early hyperpolarization, with superimposed depolarization and firing. (D) Somatostatin (1 taM) itself producecl a small hyperpolarization and a 10% fall in input resistance while the anoxic depolarization and firing disappeared; (E) 3 rain after the end of 14-min somatostatin application, a 3-rain anoxic test had a marked depolarizing effect. or inward currents (not shown) evoked by the potent glutamate agonist quisqualate (10#M) were not reduced by diazoxide (0.87 mM, n = 2). Furthermore, the mossy fibre excitatory postsynaptic potentials evoked by electrical stimulation of the granular layer were not depressed by somatostatin (Fig. 3A-B, n = 3) (or diazoxide, Fig. 5A, n --- 4). These observations further suggest that the effects of K + ATP activators occur at a presynaptic level. DISCUSSION
Thus, in CA3 neurons, diazoxide and somatostatin had effects similar to those of galanine2"4 in reducing anoxic depolarizations. It is relevant to mention that the effects of galanin are blocked by the antidiabetic sulfonylurea glibenclamide2"4 which selectively blocks K ~ channels in the pancreasJ 9 Taken together these observations suggest that ATP K + channels play a role in the anoxic responses of CA3 neurons. The present results obtained with diazoxide fully accord with this notion since this drug is known to activate K~,rp channels in pancreatic B cells and
other tissuesJ "2° Although somatostatin had more complex effects than diazoxide including a hyperpolarization (presumably due to the activation of an M-type K + channeP 4' mr), the reduced anoxic depolarization in the presence of somatostatin may be explained by a diazoxide like action on K + ATP channels. We suggest that the anoxic effects of K + ATP channel activation were generated presynaptically primarily via the mossy fibers. This would be consistent with the high density of high affinity glibenclamide binding sites in the mossy fibre region ~5 and with the observations previously made with galanin showing that in the presence of tetrodotoxin or kynurenate galanin had no effect on postsynaptic response of the pyramidal neurons to anoxia, as well as the absence of postsynaptic effects of tolbutamide or diazoxide in CA 1 pyramidal neurons. 2"ts~ This would also be consistent with the lack o f effects of diazoxide on the anoxic response of dentate granule cells from which the mossy fibres originateJ 3 In oxygenated ACSF, these channels are presumed to be closed. The sulphonylurea antagonists or K + ATP
A
N2
....s.o.m............................................................................ N2
B
-66mY
C
N2 -64mV Jlomv lmin
Fig. 4. Somatostatin eliminates N 2-evoked depolarization. In this CA3 neuron, a brief (2 min) anoxic episode induced a post-anoxic depolarization (A) which was fully blocked by somatostatin (B) and recovered after wash (C). A
B
auls
20 +
I
I i [~
'~ ~
DIAZO
---
+
I
lmin
~
'
II~HIIIt, t
.,~,,.,~,.~
Fig. 5. Diazoxide has little effect on evoked excitatory postsynaptic potentials or quisqualate (QUIS)induced depolarizations. Recording was from CA3 pyramidal layer neuron, with 3 M KC1 electrode. (A) Each trace shows from, left to right, excitatory postsynaptic potential and spike(s) evoked by dentate stimulus, and the voltage change produced by hyperpolarizing current pulses. Upper trace is initial control; middle, was obtained after 3 min of superfusion with diazoxide (0.87 mM); note slight increase in R N ( + 12%), and in evoked excitatory postsynaptic potential; lower trace was recorded 7 min after end of superfusion with diazoxide. Vm remained within 1 mV of - 6 2 mV throughout; hyperpolarizing current pulses were 0.7, 0.7 and 1.0 nA from top to bottom. (B) In another CA3 neuron, marked excitation (and depolarization) evoked by 1-min applications of QUIS (10/tM) were not reduced by diazoxide (0.87 mM). Upper and lower traces are pre- and post-control runs; resting membrane potential remained at - 73 mV throughout, and - 0 . 4 nA current pulses were applied at regular intervals.
Y. BEN-ARI el al.
60
[K];” would gradually depolarize the nerve terminals and thus enhance on-going transmitter release: but as the ATP levels fall, activation of K,+,, opposes terminal depolarization and the increase in release. Sulphonylurea drugs now become effective, because they eliminate the opposing action of KiTp and therefore facilitate depolarization and transmitter release. Whereas diazoxide and other activators channels further reduce the depolarizing of K;v tendency, and thus diminish release.
galanin and somatostatin or by the administration of diazoxide-like drugs, it may be possible to reduce significantly the anoxia triggered release of excitotoxic amino acids. This approach seems, a priori, preferable to the indiscriminate blocking of glutamate receptors which are widely involved in synaptic transmission throughout the central nervous system.
CONCLUSION
We suggest that by stimulating the production and/or secretion of endogenous peptides such as
Acknowledgements-We are grateful to Dr E. Cherubini and A. Nistri for their suggestions. Supported by INSERM.
REFERENCES
potassium channels. A. Rea. Neurosci. 11, 97-118. 2. Ben-Ari Y. (1990) Effects of galanin and glibenclamide on the anoxic response of rat hippocampal neurons in I. Ashcroft F. M. (1988) Adenosine 5’-triphosphate-sensitive
vitro.
Eur. J. Neurosci. 2, 62-68.
3. Ben-Ari Y. and KrnjeviC K. (1989) The ATP sensitive K+ channel opener diazoxide reduces anoxic depolarization in rat CA3 hippocampal neurones in isolated slices (P.) J. Physiol. 418, l92P. 4. Ben-Ari Y. and Lazdunski M. (1989) Galanin protects hippocampal neurons from the functional effects of anoxia. Eur. J. Pharmac. 165, 1331-1332.
5. Benveniste H., Drejer J., Schousboue A. and Diemer N. H. (1984) Elevation of the extracellular concentration of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdial. J. Neurochem.
43, 1369-1374.
6. Choi D. W. (1988) Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci. 11, 465469.
I. De Weille J., Schmid-Antomarchi H., Fosset M. and Lazdunski M. (1988) Regulation of ATP-sensitive K+ channels in insulinoma cells: activation by somatostatin and protein kinase C and the role of cCAMP. Proc. nam. Acad. Sci. U.S.A. 86, 2971-2976.
8. De Weille J., Schmid-Antomarchi H., Fosset M. and Lazdunski M. (1988) ATP-sensitive K+ channels that are blocked by hypoglycemia-inducing sulfonylureas in insulin-secreting cells are activated by galanin, a hyperglycemia-inducing hormone. Proc. nafn. Acnd. .Sci. U.S.A. 85, 1312-1316. 9. Fujiwara N.. Higashi H., Shimoji K. and Yoshimura M. (1987) Effects of hypoxia on rat hippocampal neurones by tlitro. J. Physiol., Land. 384, I3 I -I 5 I. 10. Hansen A. J. (1985) Effect of anoxia on ion distribution in the brain. Physiol. Rec. 65, 101-148. Il. Hansen A. J.. Hounsgaard J. and Jahnsen H. (1982) Anoxia increases potassium conductance in hippocampal nerve cells. Acra physiol. stand. 115, 301-310. 12. Krnjevie K. and Ben-Ari Y. (1989) Anoxic changes in dentate granule cells. Neurosci Lefr. 107, 89-93. 13. KrnjeviC K. and Leblond J. (1987) Anoxia reversibly supresses neuronal calcium currents in rat hippocampal slices. Can. J. Physiol. 65, 2 157-2 161. 13a. KrnjeviC K. and Leblond 1. (1988) Are there hippocampal ATP-sensitive K channels that are activated by anoxia’! Pjltigers Arch. ges. Physiol. 411, suppl., R145. 14. Moore S. D., Madamba S. G., Joels M. and Siggins G. R. (1988) Somatostatin augments the m-current in hippocampal neurons. Science 239, 279-280. 15. Mourre C.. Ben-Ari Y.. Bernardi H.. Fosset M. and Lazdunski M. (1989) Antidiabetic sulfonylureas: localization of binding sites in the brain and effects on the hyperpolarization induced by anoxia in hippocampal slices. Brain Res. 486, 159-164. 16. Noma A. (1983) ATP regulated K+ channels in cardiac muscle. Nature 305, 147-148.
17. Pittman Q. J. and Siggins G. R. (1981) Somatostatin hyperpolarizes hippocampal pyramidal cells
in Glro.
Brain Res.
221, 402-408. 18. Pulsinelli W. A. (1985) Selective neuronal vulnerability: morphological and molecular characteristics.
Prog. Brain Res.
63, 29-37.
Schmidt-Antomarchi H.. De Weille J.. Fosset M. and Lazdunski M. (1987) The receptor for antidiabetic sulfonylureas controls the activity of the ATP-modulated K+ channel in insulin-secreting cells. J. biol. Chem. 262, 15,840-15.844. 20. Trube G.. Rorsman 0. and Ohno-Shosaku T. (1986) Opposite effects of tolbutamide and diazoxide on the ATP-dependent KC channel in mouse pancreatic B cells. I’piigers Arch. ges. Physiol. 407, 493-499. 21. Trube G. and Hescheler J. (1984) Inward-rectifying channels in isolated patches of the heart cell and membrane: ATP-dependence and comparison with cell-attached patches. P/tigers Arch. ges. Physiol. 401, 178-184. 19.
(Accepted 24 January 1990)