- Email: [email protected]
Chlorpromazine protects brain tissue in hypoxia by delaying spreading depression-mediated calcium influx
Brain Research, 385 (1986) 219-226 Elsevier
219
BRE 12074
Chlorpromazine Protects Brain Tissue in Hypoxia by Delaying Spreading Depression-Mediated Calcium Influx MAURIZIO BALESTRINO and GEORGE G. SOMJEN
Departmentof Physiology, Duke University Medical Center, Durham, NC27710 (U.S.A.) (Accepted 18 March 1986)
Key words: Hypoxia - - Cerebral hypoxia - - Cerebrovascular disorder - - Chlorpromazine - - Spreading depression - - Calcium
We have investigated the possible protective effect of chlorpromazine in hypoxia of brain tissue, using rat hippocampal slices maintained at 35-36 °C. The recovery of synaptic transmission along the Schaffer collaterals to the CA1 pathway after 9 min hypoxia was compared in chlorpromazine-treated and in control slices. Recovery upon reoxygenation was the exception in control slices, while it was observed in approximately 50 and 100% of slices treated with 7 and 70/~M chlorpromazine, respectively. Chlorpromazine also significantly delayed the occurrence of the hypoxia-induced spreading depression (SD). Recovery took place when SD occurred late during hypoxia, not when it occurred early. In those slices in which 7~tM chlorpromazine afforded no protection, SD occurred as early as it did in control slices. In further experiments, we deliberately induced SD during hypoxia in 70/~M-treated slices by topically applying a drop of high-K+ artificial cerebrospinal fluid (ACSF). Recovery was not observed when SD was induced early, but it was observed when it was induced near the end of the hypoxic period. Slices exposed to the same period of hypoxia in Ca2+-free ACSF recovered synaptic transmission (even without chlorpromazine treatment) despite early induction of SD. We conclude that: (1) chlorpromazine protects brain tissue from hypoxia-induced irreversible loss of synaptic transmission; (2) it does so by delaying the occurrence of SD, and hence shortening the time spent in the SD-induced depolarized state; and (3) the harm done by SD in hypoxia is related to the influx of Ca 2+ into neurons. INTRODUCTION Knowledge of drugs that can p r e v e n t the d a m a g e induced by hypoxia in the central nervous system could lead to new t h e r a p e u t i c a l a p p r o a c h e s , and also to a better understanding of the mechanisms involved in such damage. A m o n g the theories to explain neuronal d a m a g e due to hypoxia, one of the most widely consid e r e d 23'31'35 is the intracellular accumulation of Ca 2+. Despite a large a m o u n t of indirect evidence, the 'Ca 2+ hypothesis' is, so far, u n p r o v e n 19'36. A l s o , the possible role of the hypoxia-induced spreading depression ('anoxic d e p o l a r i z a t i o n ' ) in m e d i a t i n g both intracellular Ca 2÷ accumulation and hypoxic d a m a g e remains hypothetical 19. W e decided to study possible protection by chlorpromazine because it has been r e p o r t e d that this drug mitigates ischemic d a m a g e in liver 16 and heart 18,4°.
W e tested its possible protective effect against brain tissue hypoxia using in vitro h i p p o c a m p a l slices. In this p r e p a r a t i o n ' p u r e ' hypoxia can be induced, without the systemic, vascular and o t h e r cerebral effects that occur in in vivo models. W e found that chlorpromazine protects brain slices from hypoxia-induced irreversible loss of synaptic transmisson; an analysis of this protection suggests that spreading depression is a m a j o r determinant of hypoxic damage. W e also found additional evidence that spreading depression-induced damage is linked to the massive influx of Ca 2÷ that it determines. Portions of this w o r k have been r e p o r t e d in abstract form 9,38. MATERIALS AND METHODS W e used 180-250 g female S p r a g u e - D a w l e y rats
Correspondence: M. Balestrino, Department of Physiology, Box 3709-N, Duke University Medical Center, Durham, NC 27710, U.S.A. 0006-8993/86/$03.50 ~ 1986 Elsevier Science Publishers B.V. (Biomedical Division)
220 from Charles Rivers. Slices were prepared and maintained as described earlier 11'17. Briefly, 400#m transverse slices were cut from one hippocampus and maintained at 35-36 °C in an interface chamber. Artificial cerebrospinal fluid (ACSF) used for perfusion contained (in mM): NaC1, 130; KCI, 3.5; NaH2PO 4, 1.25; N a H C O 3, 24; CaC12, 1.2; MgSO4, 1.2; glucose 10. This medium was bubbled with 95% 02-5% CO2 to reach a pH of 7.4. The same mixture was used for the gas phase. Potentials evoked by stimulation of the Schaffer collaterals with a tungsten microcathode were recorded from both stratum (st.) radiatum and st. pyramidale of the CA1 region with glass micropipettes, 1-6 Mf2 resistance, filled with 150 mM NaCI solution. Population spike amplitude (measured as described by Aitken I and expressed as a percentage of prehypoxic value) was used to evaluate recovery of synaptic transmission. Ground-referenced extracellular voltage was also continuously recorded from the electrode in st. pyramidale, and sometimes also from the electrode in st. radiatum, using a DCcoupled amplifier. Hypoxia always consisted of switching the gas phase to 95% N2-5% CO2 for 9 min. To investigate a possible protective effect of chlorpromazine, 3 groups of slices were studied. The first group served as a control, the other two were perfused with ACSF containing 7 or 70/~M chlorpromazine (chlorpromazine hydrochloride, Sigma). 'Treated' slices were taken from rats that had received an i.p. injection of 30 mg/kg chlorpromazine 30 min before sacrifice 18'4°. Rats for the control group received an equivalent volume of i.p. saline. The following protocol was used. For every experiment, 3-8 slices were cut from the brain of each rat and placed in the same recording chamber. The number of slices cut from every rat, as well as rats' weights, were not statistically different in the 3 groups. After 90 min of incubation, viability of all slices was checked. Occasional slices that did not generate a population spike of at least 1 mV were discarded. One slice from the batch was then selected for continuous testing of evoked and sustained potentials. Evoked responses were tested with a stimulus strength adjusted to evoke, at the beginning of the experiment, a population spike of about 75% maximal amplitude. Hypoxia was induced after a 30 min control period, that is, 120 min following slice preparation. The period of hypoxia lasted 9 min. Re-
sponses were sampled 5 and 2 min before the beginning of hypoxia, every minute during hypoxia, and 5, 10, 15 and 20 min after the beginning of reoxygenation. Follow-up after hypoxia was 1 h (if postsynaptic potentials recovered within this period) or 2 h (if no recovery was observed). After this time, the other slices in the batch (that had been exposed in the same chamber to the same degree of hypoxia) were checked again for the presence of synaptic transmission. Absence of postsynaptic responses after hypoxia was always confirmed by increasing stimulation intensity. To deliberately induce spreading depression at a desired time, we used a broken glass micropipette 4 to deliver a drop of high-K + ACSF to st. pyramidale of CA1. Preliminary experiments with fully oxygenated slices showed that this method reliably induces a typical, fully reversible episode of spreading depression. High-K + ACSF was identical in composition to the ACSF used for slices perfusion (including the appropriate addition of chlorpromazine or the appropriate changes in [Ca 2÷] and [Mg2+]), except for complete replacement of NaC1 with KCI (see ref. 28). When Ca2+-free perfusion was desired, we allowed 30 min after changing perfusion medium for equilibration before inducing hypoxia. In Ca2+-free ACSF Mg 2+ concentration was raised to 6 mM (see ref. 2). We switched back to perfusion with normal ACSF when we started reoxygenation. RESULTS
Electric responses during hypoxia and reoxygenation The responses evoked by stimulation of Schaffer collaterals consisted of a presynaptic volley6, a focal postsynaptic potential (fEPSP) 3° and a compound action potential (or 'population spike') 5. They are shown in Figs. 1A and 4A. Postsynaptic responses were abolished in treated as well as control slices within 3 min after oxygen withdrawal. The fEPSP disappeared later than the population spike, and the time of onset of synaptic block was not different in control and chlorpromazine treated slices. The presynaptic volley proved more resistant to hypoxia, disappearing only 3-7 min after oxygen withdrawal, often coincident with the occurrence of spreading depression (see below). One to 3 min after oxygen withdrawal there was a slow, negative shift of potential in
221
both st. p y r a m i d a l e and st. radiatum. This negative shift could reach 2 - 5 m V amplitude, and was followed by a sudden large negative shift of potential (Fig. 2), again occurring at the same time in both strata. This is the h a l l m a r k of the hypoxia-induced spreading depression 15,26'27.
C Chlorpromozine;SD by high [K*]
,~ Control
I - -
N2
I
9' Hypolio
Protective effect o f chlorpromazine Since stimulation during h y p o x i a m a y m a k e a difference in the establishment of hypoxic d a m a g e , for statistical analysis we s e p a r a t e d continuously monBEFORE HYPOXIA
DURING HYPOXIA
AFTER HYPOXIA
A I0 msec SA~ Control
Pv~ EPSP
EPSP ImVl~
SA
lOmse¢ SA PS
SA
I N2
1
N2
02
D Coz~- tree ACSF with Glulomole;SD by high [K ÷]
E] Chlorpromozine
U p o n reoxygenation, the presynaptic volley always recovered within 1 min. Postsynaptic responses either r e a p p e a r e d within 10 min after reoxygenation, or they did not r e a p p e a r at all (Figs. 1 and 4). The focal E P S P recovered earlier than the p o p u l a t i o n spike. Extracellular potential r e t u r n e d rapidly toward its control level in all slices, regardless of recovery of synaptic function, sometimes slightly undershooting, sometimes slightly overshooting it (Fig. 2). These observations are consistent with previous reports 19,29.
I
02
I 02
I N2
I O?
Fig. 2. DC-coupled tracings showing extracellular ground-referenced voltage from st. pyramidale. Negative deflection is downward in all tracings. Each time bar represents 9 rain. N 2, time of induction of hypoxia; 0 2, beginning of reoxygenation. A: control slice. B: slice treated with chlorpromazine as in Fig. 1. C: same as in B, but spreading depression provoked 4 min after beginning of hypoxia by a small drop of high-K+ medium. D: slice perfused with chlorpromazine-free, Ca2+-free, 6 mM [Mg2÷] medium, 16#M glutamate added to the medium, 0.1 Hz antidromic stimulation during hypoxia; spreading depression induced as in C 3 min after beginning of hypoxia.
itored slices (in which responses were s a m p l e d during hypoxia, see Methods) from those in which the responses were checked only before and after hypoxia. A few control slices in the latter group recovered, while no control slice did so in the former group. H o w e v e r , results from these two groups largely agreed. They are r e p o r t e d in Table I. A s it can be seen, recovery, virtually absent in the control groups, was seen in many of the t r e a t e d slices and the difference was statistically significant. Also, the n u m b e r of recovered slices was d o s e - d e p e n d e n t . T h e TABLE I
Effects of chlorpromazine on recovery
CPZ
lOmse¢
i V
-U
Slices were classified as 'recovered' or 'not recovered' according to whether or not focal EPSP and population spike reappeared after hypoxia. Hypoxia time was 9 min in all experiments. Monitoring of slices and chlorpromazine treatment are detailed in the text. Probability shown are for X2-tests. CPZ, chlorpromazine.
Slices (recovered~totalnumber tested) Control Fig. 1. Potentials evoked in CA1 by stimulation of Schaffer collaterals before, during and after hypoxia. Negative deflection downward in all tracings. A: control slice. B: slice dissected from rat injected with chlorpromazine, superfused by ACSF containing 70 MM of the drug. In each condition, upper tracings are from stratum (st.) radiatum, lower tracings are from st. pyramidale. SA, stimulus artifact; PV, presynaptic volley; EPSP, focal EPSP; PS, population spike.
Monitored continuously 0/8 Monitored only before and after hypoxia 7/36
CPZ (7ltM)
CPZ (70ktM)
5/8
5/5
P < 0.01
17/33
25/27
P < 0.01
222 0
0
540-
0
o
-1-
-100
"~ 440-
o c~ cf')
OlD E
~ 340-
000
-200 ~'
o•
g
o
D
~ ec
•
D D
E2~
D
~" 240-
t00 co
DD
E o
D °
~
C3
E
D 140
I ~ l
t__J
Slices
Slices o:C0ntr01s
100
N01 Recovered
Recovered •
-
CPZ 7~M
©= CPZ 70 #M
Fig. 3. Relationship between latency of spreading depression from beginning of hypoxia and recovery of synaptic transmission. Left-side scale represents time (in seconds) between beginning of hypoxia and spreading depression (SD). Right side scale represents time between onset of SD and reoxygenation. Left column of data points represents slices that recovered synaptic transmission upon reoxygenation. Right column represents slices that did not. It should be noted that one slice in the '7 ~M' group is not represented in this figure, since no DC-coupied recording was made in that experiment.
potential that signals the onset of the hypoxia-induced spreading depression-like response 15,2°'26"2v. Spreading depression could be studied only in the continuously monitored slices. Latency between induction of hypoxia and the negative shift of potential was (in seconds, mean _+ S.D.) 237 _ 43 in the control group, 319 +_ 42 in the 7 ~M chlorpromazine group, and 418 _ 99 in the 70/*M chlorpromazine group (P < 0.01, one-way analysis of variance). Examples of DC-coupled tracings from these experiments are shown in Fig. 2, A and B. Chlorpromazine protection was clearly associated with relatively longer latencies of spreading depression. As shown in Fig. 3, early occurrence of spreading depression resulted in no recovery even with chlorpromazine treatment. Also, no slice showed irreversible loss of synaptic transmission if spreading depression occurred after 325 s from the beginning of hypoxia. Since the duration of hypoxia was constant, the time from the beginning of hypoxia to spreading depression was inversely correlated to the time from onset of spreading depression to reoxygenation (Fig. 3, right side scale).
The role of spreading depression We next tried to determine whether the improved BEFORE HYPOXIA
amplitude of the recovered population spike could be compared to the control amplitude in the continuously monitored slices only, since in those that were checked before and after hypoxia the recording site and the distance between the electrodes did not remain constant. In the continuously monitored slices the mean _+ S.D. of the recovered spike was 114 + 39% of the prehypoxic value with 7/~M chlorpromazine (not recovered slices excluded from computation); and 115 _+ 103% in slices treated with 70/tM. Examples of waveforms are shown in Fig. 1. However, there was variability among slices and recovery was incomplete in a number of instances. Protection may have been partial in these, chlorpromazine not preventing the loss of some neurons. Spontaneous changes or 'drift' in population spike amplitude may also have played a role in this variability. Another effect of chlorpromazine was a delay of the sudden, large negative shift of the extracellular
DURING HYPOXIA
AFTER HYPOXIA
ImVL SA-pv / E
CPZ SD ot 4'
SA
SA PV
ImvLEPSP a SA SA PS
B
(2o2-* flee ACSF SO at 4 '
Fig. 4. A: slice treated with chlorpromazine as in Fig. IB, but spreading depression (SD) provoked 4 min after beginning of hypoxia by a small drop of high-K* medium. B: slice subjected to hypoxia in Ca2+-free, 6 mM [Mg2+], 16#M glutamate medium; no chlorpromazine used, 0.1 Hz antidromic stimulation during hypoxia; SD provoked 3 rain afIer beginning of hypoxia as in A. Recordings labeled as in Fig. 1.
223 survival rate of chlorpromazine-treated slices actually depends on the delay of the onset of spreading depression, or whether protection against hypoxia and against spreading depression depends on two different actions of the drug. We deliberately provoked spreading depression at various times during hypoxia in slices treated, as above, with 70/~M chlorpromazine (see Methods for the technique used). A sample record is shown in Fig. 2C. In 5 out of 5 slices in which spreading depression was provoked 4 min after oxygen withdrawal (i.e. at the same time when, on the average, it would occur in untreated slices, see above), recovery was not observed in spite of the presence of chlorpromazine (example in Fig. 4A). However, 3 out of 3 slices in which spreading depression was induced at 7 min after withdrawal of oxygen (2 min before reoxygenation) did recover. This difference is statistically significant (P < 0.05, Ze-test)i In the latter, 'recovered' group, posthypoxic (mean _+ S.D.) population spike was 68 + 34% of prehypoxic value.
synaptic transmission upon reoxygenation (4 slices from 4 rats). In 4 additional slices, similarly treated, we provoked spreading depression at 3 min hypoxia using high-K + ACSF (example in Fig. 2D) and all 4 slices recovered (example in Fig. 4B). In two of the slices incubated in Ca2+-free ACSF we observed spontaneous paroxysmal firing during the early phases of hypoxia 1°. In summary, a total of 6 out of 6 slices in which spreading depression was provoked either after 3 or 4 min hypoxia in Ca2+-free medium did show recovery (mean + S.D. of population spike amplitude being 85+67% of the prehypoxic value). This outcome is significantly different (P < 0.01, Z~-test) from the 5 out of 5 slices (see preceding section) in which recovery failed when spreading depression was provoked in the same way and after an equal duration of hypoxia in Ca2+-containing medium. In addition, 4 other slices recovered in Ca2+-free medium despite the spontaneous occurrence of spreading depression within 4 min of the onset of hypoxia.
Mechanism of damage by spreadingdepression
DISCUSSION
As suggested in the literature ~9 increase in intracellular free Ca 2÷ may be the reason why spreading depression is harmful in hypoxia. We tested this hypothesis by inducing hypoxia in Ca2+-free ACSF. [Mg2+]o was raised to 6 mM to maintain normal electric excitability of neurons 2. Four minutes after induction of hypoxia (see above) spreading depression was provoked by the local application of high-K + ACSF. We observed recovery in both of two such experiments. Still, it was possible that other effects of Ca2+-free ACSF, especially its effect on quantal release of neurotransmitters 24, including glutamate, were at work and somehow accounted for at least part of this protection. In fact, synaptic transmission 32 and hypoxia-induced glutamate release 13,33 have been suggested to play a role in hypoxic damage. In 4 additional experiments, we therefore added 16 ~M glutamate (maximal average concentration in ischemic brain, as calculated from ref. 13) to the Ca2÷-free, high-Mg 2+ ACSF. We also stimulated antidromically CA1 pyramidal cells at 0.1 Hz during hypoxia, to cause periodic cell firing in spite of blocked synaptic transmission. Slices so treated underwent spreading depression without provocation after approximately 4 rain hypoxia, yet recovered
Our data indicate that chlorpromazine can protect brain tissue against hypoxic damage. Whether or not the mechanism of protection is similar to that operating in other organs ~6'~s'4°, remains to be seen. In brain tissue, synaptic recovery was clearly correlated with the delay in the onset of the negative shift in sustained potential during hypoxia. This is often referred to as 'anoxic depolarization' (refs. 15, 19). However, there is little doubt that its mechanisms are similar to those of Leao's 'spreading depression' (refs. 15, 19, 25-27). That chlorpromazine can delay the hypoxia-induced spreading depression has already been noted in 1957 by Benesova et al. 12, but these authors did not investigate the effect of the drug on functional recovery. Indirect evidence linking spreading depression to irreversible damage is, however, already available in the literature. In 1957 Bures and Buresova TM demonstrated that cold delayed the occurrence of the spreading depression-like 'terminal depolarization'. Since Hirsch had earlier shown 21 that cooling also protects brain tissue from ischemic damage, Bures and Buresova speculated that any therapeutic intervention able to delay terminal depolarization would
224 improve the chances of recovery of brain function after ischemia. Moreover, it has been found in this laboratory 34 that brain slices submerged in ACSF, incubated at 29 °C and made hypoxic for as long as 45 min do not show spreading depression and do recover synaptic transmission after reoxygenation (these slices show a state of posthypoxic hyperexcitability). Others have reported that in ischemic brain necrotic phenomena take place only in those areas where ion derangements typical of spreading depression are observed during deprivation of blood, and not in those areas (notably the 'ischemic penumbra') where these derangements are not observed 7,39. While all these data agree with the notion of a causal relationship between spreading depression and irreversible damage, they still do not prove it. By deliberately inducing spreading depression at various times during hypoxia we could show a true causal relationship: when we induced it early enough after withdrawal of oxygen, chlorpromazine no longer afforded any protection. Protection in our slices was due to the delay in spreading depression. It is of course well known that in normally oxygenated brain tissue spreading depression can be induced, both in vivo and in vitro, without causing irreversible damage 15'37. It therefore seems that spreading depression interacts with oxygen deprivation in bringing about irreversible damage. Clues to the mechanisms involved in this interaction may be found by noting that this damage can be modulated by both the timing of spreading depression during hypoxia and the presence of extracellular Ca 2+. As to the former, spreading depression was well tolerated when it occurred late during a fixed period of hypoxia, not when it occurred early. In principle, the critical variable could be either the time between onset of hypoxia and of spreading depression, or the time between the onset of spreading depression and of reoxygenation. The latter is more likely: the critical factor for irreversible damage may be the length of time spent in the spreading depression-induced depolarized state 19. Two observations seem to support this view. Symon 39 reported that brain damage in vivo correlated with the time during which the tissue had been exposed to flow rate lower than that at which 'significant increases in potassium activity are first seen during progressive brain ischemia'. These 'significant increases in potassium activity' are proba-
bly those that are recorded by others as spreading depression 2°. Also, Lipton and Kass have found (personal communication) that superfusion of brain slices with 60 mM [K +] medium for 5 min results in irreversible loss of synaptic transmisson, while after superfusion for 3 min some slices recover, some do not. Superfusion with 60 mM [K +] is likely to have caused spreading depression. If we assume so, the time spent in the spreading depression-induced depolarized state by the slices that did not recover was remarkably similar in Lipton and Kass's and our experiments. The question may be raised, whether the small drops of high-K + ACSF used to provoke spreading depression could have, in themselves, been the cause of irreversible damage to neurons. This was almost certainly not the case, since synaptic transmission always recovered when such droplets were applied to fully oxygenated slices, or to hypoxic slices in Ca 2+free medium, or to hypoxic slices in chlorpromazinecontaining ACSF 2 min before reoxygenation. As to the second of the above-mentioned modulating factors, hypoxic spreading depression requires the presence of extracellular C a 2+ to exert its harmful effect. Neither decreased cell firing in Ca2+-free ACSF nor a lack or decrease of Ca2+-dependent glutamate release in the course of hypoxia could explain this protection: we observed recovery in Ca2+-free medium even from slices that were treated with high doses of extracellular glutamate and were forced to fire during hypoxia by antidromic stimulation, as well as from those that showed hypoxia-induced paroxysmal firing. Our data therefore support the view that spreading depression is harmful in hypoxia because it allows Ca 2+ to cross the cell membrane and to accumulate in the intracellular space 19. Kass and Lipton 23 also explained the protection they observed in Ca 2÷free ACSF by a lack of intracellular accumulation of this ion. The recently reported 22 protection by high[Mg2+] ACSF has also been explained in terms of Ca 2+ antagonism by this ion. These notions may lead to a clinical application in brain protection during hypoxic-ischemic states. From this point of view, it must be noted that it is not yet certain how loss of synaptic transmission in vitro relates to neuron damage in vivo; nor whether the recovery of synaptic function reliably predicts longterm survival in the post-hypoxic period. However,
225 our data seem strong enough to encourage further exploration of the clinical potential of chlorprom-
spreading depression) may prove useful in favoring postischemic recovery in intact brains.
azine and related drugs. Concentrations of chlorpromazine of 10/.tM and more (a certain a m o u n t being b o u n d by proteins 3) have b e e n reported in the ce-
ACKNOWLEDGEMENTS
rebrospinal fluid of some patients treated with the drug for psychiatric conditions8'41. P r e t r e a t m e n t with chlorpromazine (or other drugs capable of delaying
REFERENCES 1 Aitken, P.G., Kainic acid and penicillin: differential effects on excitatory and inhibitory interactions in the CA1 region of the hippocampal slice, Brain Research, 325 (1985) 261-269. 2 Aitken, P.G. and Somjen, G.G., The sources of extracellular potassium accumulation in the CA1 region of hippocampal slices, Brain Research, 369 (1986) 163-167. 3 Alfredsson, G. and Sedvall, G., Protein binding of chlorpromazine in cerebrospinal fluid and serum, Int. Pharrnacopsychiat., 15 (1980) 261-269. 4 Alger, B.E., Dhanjal, S.S., Dingledine, R., Garthwaite, J., Henderson, G. et al., Brain slice methods. In R. Dingledine (Ed.), Brain slices, Plenum Press, New York, 1984, pp. 381-437. 5 Andersen, P., Bliss, T.V.P. and Skrede, K.K., Unit analysis of hippocampal population spikes, Exp. Brain Res., 13 (1971) 208-221. 6 Andersen, P., Silfvenius, H., Sundberg, S.H., Sveen, O. and Wigstrom, H., Functional characteristics of unmyelinated fibers in the hippocampal cortex, Brain Research, 144 (1978) 11-18. 7 Astrup, J., Siesjo, B.K. and Symon, L., Thresholds in cereberal ischemia; the ischemic penumbra, Stroke, 12 (1981) 723-725. 8 Axelsson, S., Jonsson, S. and Nordgren, L., Cerebrospinal fluid levels of chlorpromazine and its metabolites in schizophrenia, Arch. Psychiat. Nervenkr., 221 (1975) 167-170. 9 Balestrino, M., Mechanism of protection by chlorpromazine: the role of spreading depression in hypoxic brain damage, Abstr. 30th Congress of IUPS, Vancouver, Canada, in press. 10 Balestrino, M. and Aitken, P.G., Paroxysmal firing in hippocampal slices during the early phase of hypoxia, Soc. Neurosci. Abstr., 11 (1985) 433. 11 Balestrino, M., Aitken, P.G. and Somjen, G.G., The effects of moderate changes of extracellular K+ and Ca2+ on synaptic and neural function in the CA1 region of the hippocampal slice, Brain Research, in press. 12 Benesova, O., Buresova, O. and Bures, J., Die Wirkung des Chlorpromazins und der Glykamie auf das elektrophysiologisch kontrollierte Uberleben der Hirnrinde bei verschiedenen Korpertemperaturen, Arch. Exp. Pathol. Pharmakol., 231 (1957) 550-561. 13 Benveniste, H., Drejer, J., Schousboe, A. and Diemer, N.H., Elevation 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.
This work was supported by U S P H S G r a n t NS 18670.
14 Bures, J. and Buresova, O., Die anoxische Terminaldepolarisation als Indicator der Vulnerabilitat der Grosshirnrinde bei Anoxie und Ischamie, Pfliigers Arch., 264 (1957) 325-334. 15 Bures, J., Buresova, O. and Krivanek, J., The Mechanisms and the Applications of Leao's Spreading Depression of Electroencephalographic Activity, Academia, Prague and Academic Press, New York, 1974. 16 Chien, K.R., Abrams, J., Pfau, R.G. and Farber, J.L., Prevention by chlorpromazine of ischemic liver cell death, Am. J. Pathol., 88 (1977) 539-558. 17 Dingledine, R. and Somjen, G.G., Calcium dependance of synaptic transmission in the hippocampal slice, Brain Research, 207 (1981) 218-222. 18 Edoute, Y., Van Der Merwe, E.L., Sanan, D., Kotze, J.C.N., Van Niekerk, I. and Lochner, A., Normothermic ischaemic cardiac arrest and reperfusion of the isolated working heart: effect of chlorpromazine on functional, metabolic and morphological recovery, J. Mol. Cell. Cardiol., 15 (1983) 603-620. 19 Hansen, A.J., Effect of anoxia on ion distribution in the brain, Physiol. Rev., 65 (1985) 101-148. 20 Hansen, A.J. and Zeuthen, T., Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex, Acta Physiol. Scand., 113 (1981) 437-445. 21 Hirsch, H., Erholung und Wiederbelebung des Warmblutergehirns bei Normo- und Hypothermie, Proc. 20th Int. Physiol. Congr., Brussels, 1956, p. 426. 22 Kass, I.S. and Cottrell, J.E., Effects of blocking calcium entry with nimodipine and magnesium in anoxic hippocampal slices, Soc. Neurosci. Abstr., 11 (1985) 375. 23 Kass, I. and Lipton, P., Mechanisms involved in irreversible anoxic damage to the in vitro rat hippocampal slice, J. Physiol. (London), 332 (1982) 459-472. 24 Katz, B. and Miledi, R., The timing of calcium action during neuromuscular transmission, J. Physiol. (London), 189 (1967) 535-544. 25 Leao, A.A.P., Spreading depression of activity in the cerebral cortex, J. Neurophysiol., 7 (1944) 359-390. 26 Leao, A.A.P., Further observations on the spreading depression of activity in the cerebral cortex, J. Neurophysiol., 10 (1947) 409-414. 27 Leao, A.A.P., The slow voltage variation of cortical spreading depression of activity, Electroencephalogr. Clin. Neurophysiol., 3 (1951) 315-321. 28 Leao, A.A.P. and Morison, R.S., Propagation of spreading cortical depression, J. Neurophysiol., 8 (1945) 33-45. 29 Lipton, P. and Whittingham, T.S., The effect of hypoxia on evoked potentials in the in vitro hippocampus, J. Physiol. (London), 287 (1979) 427-438.
226 30 Lomo, T., Patterns of activation in a monosynaptic cortical pathway: the perforant path input to the dentate area of the hippocampal formation, Exp. Brain Res., 12 (1971) 18-45. 31 Meldrum, B.S., Simon, R.P., Griffiths, T. and Evans, M.C., Hyperexcitability and brain protection, Clin. Neuropharmacol., Suppl. 1,7 (1984) 500-501. 32 Rothman, S.M., Synaptic activity mediates death of hypoxic neurons, Science, 220 (1983) 536-537. 33 Rothman, S., Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death, J. Neurosci., 4 (1984) 1884-1891. 34 Schiff, S.J. and Somjen, G.G., Hyperexcitability following moderate hypoxia in hippocampal tissue slices, Brain Research, 337 (1985) 337-340. 35 Siesjo, B.K., Cell damage in the brain: a speculative synthesis, J. Cereb. Blood Flow Metabol., 1 (1981) 155-185. 36 Siesjo, B.K. and Wieloch, T., Molecular mechanisms of ischemic brain damage: Ca2+-related events. In F. Plum and W. Pulsinelli (Eds.), Cerebrovascular Diseases, Raven
Press, New York, 1985, pp. 187-198. 37 Somjen, G.G. and Aitken, P.G., The ionic and metabolic responses associated with neuronal depression of Leao's type in cerebral cortex and in hippocampal formation, An. Acad. Brasil. Cienc., 56 (1984) 495-504. 38 Somjen, G.G., and Balestrino, M., Chlorpromazine improves resistance of hippocampal slices to hypoxic damage, Soc. Neurosci. Abstr., 11 (1985) 374. 39 Symon, L., Flow thresholds in brain ischaemia and the effects of drugs, Br. J. Anaesth., 57 (1985) 34-43. 40 Thomas, G.E., Levitsky, S. and Feinberg, H., Chlorpromazine inhibits loss of contractile function, compliance and ATP in ischemic rabbit heart, J. Mol. Cell. Cardiol., 15 (1983) 621-628. 41 Wode-Helgodt, N., Borg, S., Fyro, B. and Sedvall, G., Clinical effects and drug concentrations in plasma and cerebrospinal fluid in psychotic patients treated with fixed doses of chlorpromazine, Acta Psychiat. Scand., 58 (1978) 149-173.
219
BRE 12074
Chlorpromazine Protects Brain Tissue in Hypoxia by Delaying Spreading Depression-Mediated Calcium Influx MAURIZIO BALESTRINO and GEORGE G. SOMJEN
Departmentof Physiology, Duke University Medical Center, Durham, NC27710 (U.S.A.) (Accepted 18 March 1986)
Key words: Hypoxia - - Cerebral hypoxia - - Cerebrovascular disorder - - Chlorpromazine - - Spreading depression - - Calcium
We have investigated the possible protective effect of chlorpromazine in hypoxia of brain tissue, using rat hippocampal slices maintained at 35-36 °C. The recovery of synaptic transmission along the Schaffer collaterals to the CA1 pathway after 9 min hypoxia was compared in chlorpromazine-treated and in control slices. Recovery upon reoxygenation was the exception in control slices, while it was observed in approximately 50 and 100% of slices treated with 7 and 70/~M chlorpromazine, respectively. Chlorpromazine also significantly delayed the occurrence of the hypoxia-induced spreading depression (SD). Recovery took place when SD occurred late during hypoxia, not when it occurred early. In those slices in which 7~tM chlorpromazine afforded no protection, SD occurred as early as it did in control slices. In further experiments, we deliberately induced SD during hypoxia in 70/~M-treated slices by topically applying a drop of high-K+ artificial cerebrospinal fluid (ACSF). Recovery was not observed when SD was induced early, but it was observed when it was induced near the end of the hypoxic period. Slices exposed to the same period of hypoxia in Ca2+-free ACSF recovered synaptic transmission (even without chlorpromazine treatment) despite early induction of SD. We conclude that: (1) chlorpromazine protects brain tissue from hypoxia-induced irreversible loss of synaptic transmission; (2) it does so by delaying the occurrence of SD, and hence shortening the time spent in the SD-induced depolarized state; and (3) the harm done by SD in hypoxia is related to the influx of Ca 2+ into neurons. INTRODUCTION Knowledge of drugs that can p r e v e n t the d a m a g e induced by hypoxia in the central nervous system could lead to new t h e r a p e u t i c a l a p p r o a c h e s , and also to a better understanding of the mechanisms involved in such damage. A m o n g the theories to explain neuronal d a m a g e due to hypoxia, one of the most widely consid e r e d 23'31'35 is the intracellular accumulation of Ca 2+. Despite a large a m o u n t of indirect evidence, the 'Ca 2+ hypothesis' is, so far, u n p r o v e n 19'36. A l s o , the possible role of the hypoxia-induced spreading depression ('anoxic d e p o l a r i z a t i o n ' ) in m e d i a t i n g both intracellular Ca 2÷ accumulation and hypoxic d a m a g e remains hypothetical 19. W e decided to study possible protection by chlorpromazine because it has been r e p o r t e d that this drug mitigates ischemic d a m a g e in liver 16 and heart 18,4°.
W e tested its possible protective effect against brain tissue hypoxia using in vitro h i p p o c a m p a l slices. In this p r e p a r a t i o n ' p u r e ' hypoxia can be induced, without the systemic, vascular and o t h e r cerebral effects that occur in in vivo models. W e found that chlorpromazine protects brain slices from hypoxia-induced irreversible loss of synaptic transmisson; an analysis of this protection suggests that spreading depression is a m a j o r determinant of hypoxic damage. W e also found additional evidence that spreading depression-induced damage is linked to the massive influx of Ca 2÷ that it determines. Portions of this w o r k have been r e p o r t e d in abstract form 9,38. MATERIALS AND METHODS W e used 180-250 g female S p r a g u e - D a w l e y rats
Correspondence: M. Balestrino, Department of Physiology, Box 3709-N, Duke University Medical Center, Durham, NC 27710, U.S.A. 0006-8993/86/$03.50 ~ 1986 Elsevier Science Publishers B.V. (Biomedical Division)
220 from Charles Rivers. Slices were prepared and maintained as described earlier 11'17. Briefly, 400#m transverse slices were cut from one hippocampus and maintained at 35-36 °C in an interface chamber. Artificial cerebrospinal fluid (ACSF) used for perfusion contained (in mM): NaC1, 130; KCI, 3.5; NaH2PO 4, 1.25; N a H C O 3, 24; CaC12, 1.2; MgSO4, 1.2; glucose 10. This medium was bubbled with 95% 02-5% CO2 to reach a pH of 7.4. The same mixture was used for the gas phase. Potentials evoked by stimulation of the Schaffer collaterals with a tungsten microcathode were recorded from both stratum (st.) radiatum and st. pyramidale of the CA1 region with glass micropipettes, 1-6 Mf2 resistance, filled with 150 mM NaCI solution. Population spike amplitude (measured as described by Aitken I and expressed as a percentage of prehypoxic value) was used to evaluate recovery of synaptic transmission. Ground-referenced extracellular voltage was also continuously recorded from the electrode in st. pyramidale, and sometimes also from the electrode in st. radiatum, using a DCcoupled amplifier. Hypoxia always consisted of switching the gas phase to 95% N2-5% CO2 for 9 min. To investigate a possible protective effect of chlorpromazine, 3 groups of slices were studied. The first group served as a control, the other two were perfused with ACSF containing 7 or 70/~M chlorpromazine (chlorpromazine hydrochloride, Sigma). 'Treated' slices were taken from rats that had received an i.p. injection of 30 mg/kg chlorpromazine 30 min before sacrifice 18'4°. Rats for the control group received an equivalent volume of i.p. saline. The following protocol was used. For every experiment, 3-8 slices were cut from the brain of each rat and placed in the same recording chamber. The number of slices cut from every rat, as well as rats' weights, were not statistically different in the 3 groups. After 90 min of incubation, viability of all slices was checked. Occasional slices that did not generate a population spike of at least 1 mV were discarded. One slice from the batch was then selected for continuous testing of evoked and sustained potentials. Evoked responses were tested with a stimulus strength adjusted to evoke, at the beginning of the experiment, a population spike of about 75% maximal amplitude. Hypoxia was induced after a 30 min control period, that is, 120 min following slice preparation. The period of hypoxia lasted 9 min. Re-
sponses were sampled 5 and 2 min before the beginning of hypoxia, every minute during hypoxia, and 5, 10, 15 and 20 min after the beginning of reoxygenation. Follow-up after hypoxia was 1 h (if postsynaptic potentials recovered within this period) or 2 h (if no recovery was observed). After this time, the other slices in the batch (that had been exposed in the same chamber to the same degree of hypoxia) were checked again for the presence of synaptic transmission. Absence of postsynaptic responses after hypoxia was always confirmed by increasing stimulation intensity. To deliberately induce spreading depression at a desired time, we used a broken glass micropipette 4 to deliver a drop of high-K + ACSF to st. pyramidale of CA1. Preliminary experiments with fully oxygenated slices showed that this method reliably induces a typical, fully reversible episode of spreading depression. High-K + ACSF was identical in composition to the ACSF used for slices perfusion (including the appropriate addition of chlorpromazine or the appropriate changes in [Ca 2÷] and [Mg2+]), except for complete replacement of NaC1 with KCI (see ref. 28). When Ca2+-free perfusion was desired, we allowed 30 min after changing perfusion medium for equilibration before inducing hypoxia. In Ca2+-free ACSF Mg 2+ concentration was raised to 6 mM (see ref. 2). We switched back to perfusion with normal ACSF when we started reoxygenation. RESULTS
Electric responses during hypoxia and reoxygenation The responses evoked by stimulation of Schaffer collaterals consisted of a presynaptic volley6, a focal postsynaptic potential (fEPSP) 3° and a compound action potential (or 'population spike') 5. They are shown in Figs. 1A and 4A. Postsynaptic responses were abolished in treated as well as control slices within 3 min after oxygen withdrawal. The fEPSP disappeared later than the population spike, and the time of onset of synaptic block was not different in control and chlorpromazine treated slices. The presynaptic volley proved more resistant to hypoxia, disappearing only 3-7 min after oxygen withdrawal, often coincident with the occurrence of spreading depression (see below). One to 3 min after oxygen withdrawal there was a slow, negative shift of potential in
221
both st. p y r a m i d a l e and st. radiatum. This negative shift could reach 2 - 5 m V amplitude, and was followed by a sudden large negative shift of potential (Fig. 2), again occurring at the same time in both strata. This is the h a l l m a r k of the hypoxia-induced spreading depression 15,26'27.
C Chlorpromozine;SD by high [K*]
,~ Control
I - -
N2
I
9' Hypolio
Protective effect o f chlorpromazine Since stimulation during h y p o x i a m a y m a k e a difference in the establishment of hypoxic d a m a g e , for statistical analysis we s e p a r a t e d continuously monBEFORE HYPOXIA
DURING HYPOXIA
AFTER HYPOXIA
A I0 msec SA~ Control
Pv~ EPSP
EPSP ImVl~
SA
lOmse¢ SA PS
SA
I N2
1
N2
02
D Coz~- tree ACSF with Glulomole;SD by high [K ÷]
E] Chlorpromozine
U p o n reoxygenation, the presynaptic volley always recovered within 1 min. Postsynaptic responses either r e a p p e a r e d within 10 min after reoxygenation, or they did not r e a p p e a r at all (Figs. 1 and 4). The focal E P S P recovered earlier than the p o p u l a t i o n spike. Extracellular potential r e t u r n e d rapidly toward its control level in all slices, regardless of recovery of synaptic function, sometimes slightly undershooting, sometimes slightly overshooting it (Fig. 2). These observations are consistent with previous reports 19,29.
I
02
I 02
I N2
I O?
Fig. 2. DC-coupled tracings showing extracellular ground-referenced voltage from st. pyramidale. Negative deflection is downward in all tracings. Each time bar represents 9 rain. N 2, time of induction of hypoxia; 0 2, beginning of reoxygenation. A: control slice. B: slice treated with chlorpromazine as in Fig. 1. C: same as in B, but spreading depression provoked 4 min after beginning of hypoxia by a small drop of high-K+ medium. D: slice perfused with chlorpromazine-free, Ca2+-free, 6 mM [Mg2÷] medium, 16#M glutamate added to the medium, 0.1 Hz antidromic stimulation during hypoxia; spreading depression induced as in C 3 min after beginning of hypoxia.
itored slices (in which responses were s a m p l e d during hypoxia, see Methods) from those in which the responses were checked only before and after hypoxia. A few control slices in the latter group recovered, while no control slice did so in the former group. H o w e v e r , results from these two groups largely agreed. They are r e p o r t e d in Table I. A s it can be seen, recovery, virtually absent in the control groups, was seen in many of the t r e a t e d slices and the difference was statistically significant. Also, the n u m b e r of recovered slices was d o s e - d e p e n d e n t . T h e TABLE I
Effects of chlorpromazine on recovery
CPZ
lOmse¢
i V
-U
Slices were classified as 'recovered' or 'not recovered' according to whether or not focal EPSP and population spike reappeared after hypoxia. Hypoxia time was 9 min in all experiments. Monitoring of slices and chlorpromazine treatment are detailed in the text. Probability shown are for X2-tests. CPZ, chlorpromazine.
Slices (recovered~totalnumber tested) Control Fig. 1. Potentials evoked in CA1 by stimulation of Schaffer collaterals before, during and after hypoxia. Negative deflection downward in all tracings. A: control slice. B: slice dissected from rat injected with chlorpromazine, superfused by ACSF containing 70 MM of the drug. In each condition, upper tracings are from stratum (st.) radiatum, lower tracings are from st. pyramidale. SA, stimulus artifact; PV, presynaptic volley; EPSP, focal EPSP; PS, population spike.
Monitored continuously 0/8 Monitored only before and after hypoxia 7/36
CPZ (7ltM)
CPZ (70ktM)
5/8
5/5
P < 0.01
17/33
25/27
P < 0.01
222 0
0
540-
0
o
-1-
-100
"~ 440-
o c~ cf')
OlD E
~ 340-
000
-200 ~'
o•
g
o
D
~ ec
•
D D
E2~
D
~" 240-
t00 co
DD
E o
D °
~
C3
E
D 140
I ~ l
t__J
Slices
Slices o:C0ntr01s
100
N01 Recovered
Recovered •
-
CPZ 7~M
©= CPZ 70 #M
Fig. 3. Relationship between latency of spreading depression from beginning of hypoxia and recovery of synaptic transmission. Left-side scale represents time (in seconds) between beginning of hypoxia and spreading depression (SD). Right side scale represents time between onset of SD and reoxygenation. Left column of data points represents slices that recovered synaptic transmission upon reoxygenation. Right column represents slices that did not. It should be noted that one slice in the '7 ~M' group is not represented in this figure, since no DC-coupied recording was made in that experiment.
potential that signals the onset of the hypoxia-induced spreading depression-like response 15,2°'26"2v. Spreading depression could be studied only in the continuously monitored slices. Latency between induction of hypoxia and the negative shift of potential was (in seconds, mean _+ S.D.) 237 _ 43 in the control group, 319 +_ 42 in the 7 ~M chlorpromazine group, and 418 _ 99 in the 70/*M chlorpromazine group (P < 0.01, one-way analysis of variance). Examples of DC-coupled tracings from these experiments are shown in Fig. 2, A and B. Chlorpromazine protection was clearly associated with relatively longer latencies of spreading depression. As shown in Fig. 3, early occurrence of spreading depression resulted in no recovery even with chlorpromazine treatment. Also, no slice showed irreversible loss of synaptic transmission if spreading depression occurred after 325 s from the beginning of hypoxia. Since the duration of hypoxia was constant, the time from the beginning of hypoxia to spreading depression was inversely correlated to the time from onset of spreading depression to reoxygenation (Fig. 3, right side scale).
The role of spreading depression We next tried to determine whether the improved BEFORE HYPOXIA
amplitude of the recovered population spike could be compared to the control amplitude in the continuously monitored slices only, since in those that were checked before and after hypoxia the recording site and the distance between the electrodes did not remain constant. In the continuously monitored slices the mean _+ S.D. of the recovered spike was 114 + 39% of the prehypoxic value with 7/~M chlorpromazine (not recovered slices excluded from computation); and 115 _+ 103% in slices treated with 70/tM. Examples of waveforms are shown in Fig. 1. However, there was variability among slices and recovery was incomplete in a number of instances. Protection may have been partial in these, chlorpromazine not preventing the loss of some neurons. Spontaneous changes or 'drift' in population spike amplitude may also have played a role in this variability. Another effect of chlorpromazine was a delay of the sudden, large negative shift of the extracellular
DURING HYPOXIA
AFTER HYPOXIA
ImVL SA-pv / E
CPZ SD ot 4'
SA
SA PV
ImvLEPSP a SA SA PS
B
(2o2-* flee ACSF SO at 4 '
Fig. 4. A: slice treated with chlorpromazine as in Fig. IB, but spreading depression (SD) provoked 4 min after beginning of hypoxia by a small drop of high-K* medium. B: slice subjected to hypoxia in Ca2+-free, 6 mM [Mg2+], 16#M glutamate medium; no chlorpromazine used, 0.1 Hz antidromic stimulation during hypoxia; SD provoked 3 rain afIer beginning of hypoxia as in A. Recordings labeled as in Fig. 1.
223 survival rate of chlorpromazine-treated slices actually depends on the delay of the onset of spreading depression, or whether protection against hypoxia and against spreading depression depends on two different actions of the drug. We deliberately provoked spreading depression at various times during hypoxia in slices treated, as above, with 70/~M chlorpromazine (see Methods for the technique used). A sample record is shown in Fig. 2C. In 5 out of 5 slices in which spreading depression was provoked 4 min after oxygen withdrawal (i.e. at the same time when, on the average, it would occur in untreated slices, see above), recovery was not observed in spite of the presence of chlorpromazine (example in Fig. 4A). However, 3 out of 3 slices in which spreading depression was induced at 7 min after withdrawal of oxygen (2 min before reoxygenation) did recover. This difference is statistically significant (P < 0.05, Ze-test)i In the latter, 'recovered' group, posthypoxic (mean _+ S.D.) population spike was 68 + 34% of prehypoxic value.
synaptic transmission upon reoxygenation (4 slices from 4 rats). In 4 additional slices, similarly treated, we provoked spreading depression at 3 min hypoxia using high-K + ACSF (example in Fig. 2D) and all 4 slices recovered (example in Fig. 4B). In two of the slices incubated in Ca2+-free ACSF we observed spontaneous paroxysmal firing during the early phases of hypoxia 1°. In summary, a total of 6 out of 6 slices in which spreading depression was provoked either after 3 or 4 min hypoxia in Ca2+-free medium did show recovery (mean + S.D. of population spike amplitude being 85+67% of the prehypoxic value). This outcome is significantly different (P < 0.01, Z~-test) from the 5 out of 5 slices (see preceding section) in which recovery failed when spreading depression was provoked in the same way and after an equal duration of hypoxia in Ca2+-containing medium. In addition, 4 other slices recovered in Ca2+-free medium despite the spontaneous occurrence of spreading depression within 4 min of the onset of hypoxia.
Mechanism of damage by spreadingdepression
DISCUSSION
As suggested in the literature ~9 increase in intracellular free Ca 2÷ may be the reason why spreading depression is harmful in hypoxia. We tested this hypothesis by inducing hypoxia in Ca2+-free ACSF. [Mg2+]o was raised to 6 mM to maintain normal electric excitability of neurons 2. Four minutes after induction of hypoxia (see above) spreading depression was provoked by the local application of high-K + ACSF. We observed recovery in both of two such experiments. Still, it was possible that other effects of Ca2+-free ACSF, especially its effect on quantal release of neurotransmitters 24, including glutamate, were at work and somehow accounted for at least part of this protection. In fact, synaptic transmission 32 and hypoxia-induced glutamate release 13,33 have been suggested to play a role in hypoxic damage. In 4 additional experiments, we therefore added 16 ~M glutamate (maximal average concentration in ischemic brain, as calculated from ref. 13) to the Ca2÷-free, high-Mg 2+ ACSF. We also stimulated antidromically CA1 pyramidal cells at 0.1 Hz during hypoxia, to cause periodic cell firing in spite of blocked synaptic transmission. Slices so treated underwent spreading depression without provocation after approximately 4 rain hypoxia, yet recovered
Our data indicate that chlorpromazine can protect brain tissue against hypoxic damage. Whether or not the mechanism of protection is similar to that operating in other organs ~6'~s'4°, remains to be seen. In brain tissue, synaptic recovery was clearly correlated with the delay in the onset of the negative shift in sustained potential during hypoxia. This is often referred to as 'anoxic depolarization' (refs. 15, 19). However, there is little doubt that its mechanisms are similar to those of Leao's 'spreading depression' (refs. 15, 19, 25-27). That chlorpromazine can delay the hypoxia-induced spreading depression has already been noted in 1957 by Benesova et al. 12, but these authors did not investigate the effect of the drug on functional recovery. Indirect evidence linking spreading depression to irreversible damage is, however, already available in the literature. In 1957 Bures and Buresova TM demonstrated that cold delayed the occurrence of the spreading depression-like 'terminal depolarization'. Since Hirsch had earlier shown 21 that cooling also protects brain tissue from ischemic damage, Bures and Buresova speculated that any therapeutic intervention able to delay terminal depolarization would
224 improve the chances of recovery of brain function after ischemia. Moreover, it has been found in this laboratory 34 that brain slices submerged in ACSF, incubated at 29 °C and made hypoxic for as long as 45 min do not show spreading depression and do recover synaptic transmission after reoxygenation (these slices show a state of posthypoxic hyperexcitability). Others have reported that in ischemic brain necrotic phenomena take place only in those areas where ion derangements typical of spreading depression are observed during deprivation of blood, and not in those areas (notably the 'ischemic penumbra') where these derangements are not observed 7,39. While all these data agree with the notion of a causal relationship between spreading depression and irreversible damage, they still do not prove it. By deliberately inducing spreading depression at various times during hypoxia we could show a true causal relationship: when we induced it early enough after withdrawal of oxygen, chlorpromazine no longer afforded any protection. Protection in our slices was due to the delay in spreading depression. It is of course well known that in normally oxygenated brain tissue spreading depression can be induced, both in vivo and in vitro, without causing irreversible damage 15'37. It therefore seems that spreading depression interacts with oxygen deprivation in bringing about irreversible damage. Clues to the mechanisms involved in this interaction may be found by noting that this damage can be modulated by both the timing of spreading depression during hypoxia and the presence of extracellular Ca 2+. As to the former, spreading depression was well tolerated when it occurred late during a fixed period of hypoxia, not when it occurred early. In principle, the critical variable could be either the time between onset of hypoxia and of spreading depression, or the time between the onset of spreading depression and of reoxygenation. The latter is more likely: the critical factor for irreversible damage may be the length of time spent in the spreading depression-induced depolarized state 19. Two observations seem to support this view. Symon 39 reported that brain damage in vivo correlated with the time during which the tissue had been exposed to flow rate lower than that at which 'significant increases in potassium activity are first seen during progressive brain ischemia'. These 'significant increases in potassium activity' are proba-
bly those that are recorded by others as spreading depression 2°. Also, Lipton and Kass have found (personal communication) that superfusion of brain slices with 60 mM [K +] medium for 5 min results in irreversible loss of synaptic transmisson, while after superfusion for 3 min some slices recover, some do not. Superfusion with 60 mM [K +] is likely to have caused spreading depression. If we assume so, the time spent in the spreading depression-induced depolarized state by the slices that did not recover was remarkably similar in Lipton and Kass's and our experiments. The question may be raised, whether the small drops of high-K + ACSF used to provoke spreading depression could have, in themselves, been the cause of irreversible damage to neurons. This was almost certainly not the case, since synaptic transmission always recovered when such droplets were applied to fully oxygenated slices, or to hypoxic slices in Ca 2+free medium, or to hypoxic slices in chlorpromazinecontaining ACSF 2 min before reoxygenation. As to the second of the above-mentioned modulating factors, hypoxic spreading depression requires the presence of extracellular C a 2+ to exert its harmful effect. Neither decreased cell firing in Ca2+-free ACSF nor a lack or decrease of Ca2+-dependent glutamate release in the course of hypoxia could explain this protection: we observed recovery in Ca2+-free medium even from slices that were treated with high doses of extracellular glutamate and were forced to fire during hypoxia by antidromic stimulation, as well as from those that showed hypoxia-induced paroxysmal firing. Our data therefore support the view that spreading depression is harmful in hypoxia because it allows Ca 2+ to cross the cell membrane and to accumulate in the intracellular space 19. Kass and Lipton 23 also explained the protection they observed in Ca 2÷free ACSF by a lack of intracellular accumulation of this ion. The recently reported 22 protection by high[Mg2+] ACSF has also been explained in terms of Ca 2+ antagonism by this ion. These notions may lead to a clinical application in brain protection during hypoxic-ischemic states. From this point of view, it must be noted that it is not yet certain how loss of synaptic transmission in vitro relates to neuron damage in vivo; nor whether the recovery of synaptic function reliably predicts longterm survival in the post-hypoxic period. However,
225 our data seem strong enough to encourage further exploration of the clinical potential of chlorprom-
spreading depression) may prove useful in favoring postischemic recovery in intact brains.
azine and related drugs. Concentrations of chlorpromazine of 10/.tM and more (a certain a m o u n t being b o u n d by proteins 3) have b e e n reported in the ce-
ACKNOWLEDGEMENTS
rebrospinal fluid of some patients treated with the drug for psychiatric conditions8'41. P r e t r e a t m e n t with chlorpromazine (or other drugs capable of delaying
REFERENCES 1 Aitken, P.G., Kainic acid and penicillin: differential effects on excitatory and inhibitory interactions in the CA1 region of the hippocampal slice, Brain Research, 325 (1985) 261-269. 2 Aitken, P.G. and Somjen, G.G., The sources of extracellular potassium accumulation in the CA1 region of hippocampal slices, Brain Research, 369 (1986) 163-167. 3 Alfredsson, G. and Sedvall, G., Protein binding of chlorpromazine in cerebrospinal fluid and serum, Int. Pharrnacopsychiat., 15 (1980) 261-269. 4 Alger, B.E., Dhanjal, S.S., Dingledine, R., Garthwaite, J., Henderson, G. et al., Brain slice methods. In R. Dingledine (Ed.), Brain slices, Plenum Press, New York, 1984, pp. 381-437. 5 Andersen, P., Bliss, T.V.P. and Skrede, K.K., Unit analysis of hippocampal population spikes, Exp. Brain Res., 13 (1971) 208-221. 6 Andersen, P., Silfvenius, H., Sundberg, S.H., Sveen, O. and Wigstrom, H., Functional characteristics of unmyelinated fibers in the hippocampal cortex, Brain Research, 144 (1978) 11-18. 7 Astrup, J., Siesjo, B.K. and Symon, L., Thresholds in cereberal ischemia; the ischemic penumbra, Stroke, 12 (1981) 723-725. 8 Axelsson, S., Jonsson, S. and Nordgren, L., Cerebrospinal fluid levels of chlorpromazine and its metabolites in schizophrenia, Arch. Psychiat. Nervenkr., 221 (1975) 167-170. 9 Balestrino, M., Mechanism of protection by chlorpromazine: the role of spreading depression in hypoxic brain damage, Abstr. 30th Congress of IUPS, Vancouver, Canada, in press. 10 Balestrino, M. and Aitken, P.G., Paroxysmal firing in hippocampal slices during the early phase of hypoxia, Soc. Neurosci. Abstr., 11 (1985) 433. 11 Balestrino, M., Aitken, P.G. and Somjen, G.G., The effects of moderate changes of extracellular K+ and Ca2+ on synaptic and neural function in the CA1 region of the hippocampal slice, Brain Research, in press. 12 Benesova, O., Buresova, O. and Bures, J., Die Wirkung des Chlorpromazins und der Glykamie auf das elektrophysiologisch kontrollierte Uberleben der Hirnrinde bei verschiedenen Korpertemperaturen, Arch. Exp. Pathol. Pharmakol., 231 (1957) 550-561. 13 Benveniste, H., Drejer, J., Schousboe, A. and Diemer, N.H., Elevation 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.
This work was supported by U S P H S G r a n t NS 18670.
14 Bures, J. and Buresova, O., Die anoxische Terminaldepolarisation als Indicator der Vulnerabilitat der Grosshirnrinde bei Anoxie und Ischamie, Pfliigers Arch., 264 (1957) 325-334. 15 Bures, J., Buresova, O. and Krivanek, J., The Mechanisms and the Applications of Leao's Spreading Depression of Electroencephalographic Activity, Academia, Prague and Academic Press, New York, 1974. 16 Chien, K.R., Abrams, J., Pfau, R.G. and Farber, J.L., Prevention by chlorpromazine of ischemic liver cell death, Am. J. Pathol., 88 (1977) 539-558. 17 Dingledine, R. and Somjen, G.G., Calcium dependance of synaptic transmission in the hippocampal slice, Brain Research, 207 (1981) 218-222. 18 Edoute, Y., Van Der Merwe, E.L., Sanan, D., Kotze, J.C.N., Van Niekerk, I. and Lochner, A., Normothermic ischaemic cardiac arrest and reperfusion of the isolated working heart: effect of chlorpromazine on functional, metabolic and morphological recovery, J. Mol. Cell. Cardiol., 15 (1983) 603-620. 19 Hansen, A.J., Effect of anoxia on ion distribution in the brain, Physiol. Rev., 65 (1985) 101-148. 20 Hansen, A.J. and Zeuthen, T., Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex, Acta Physiol. Scand., 113 (1981) 437-445. 21 Hirsch, H., Erholung und Wiederbelebung des Warmblutergehirns bei Normo- und Hypothermie, Proc. 20th Int. Physiol. Congr., Brussels, 1956, p. 426. 22 Kass, I.S. and Cottrell, J.E., Effects of blocking calcium entry with nimodipine and magnesium in anoxic hippocampal slices, Soc. Neurosci. Abstr., 11 (1985) 375. 23 Kass, I. and Lipton, P., Mechanisms involved in irreversible anoxic damage to the in vitro rat hippocampal slice, J. Physiol. (London), 332 (1982) 459-472. 24 Katz, B. and Miledi, R., The timing of calcium action during neuromuscular transmission, J. Physiol. (London), 189 (1967) 535-544. 25 Leao, A.A.P., Spreading depression of activity in the cerebral cortex, J. Neurophysiol., 7 (1944) 359-390. 26 Leao, A.A.P., Further observations on the spreading depression of activity in the cerebral cortex, J. Neurophysiol., 10 (1947) 409-414. 27 Leao, A.A.P., The slow voltage variation of cortical spreading depression of activity, Electroencephalogr. Clin. Neurophysiol., 3 (1951) 315-321. 28 Leao, A.A.P. and Morison, R.S., Propagation of spreading cortical depression, J. Neurophysiol., 8 (1945) 33-45. 29 Lipton, P. and Whittingham, T.S., The effect of hypoxia on evoked potentials in the in vitro hippocampus, J. Physiol. (London), 287 (1979) 427-438.
226 30 Lomo, T., Patterns of activation in a monosynaptic cortical pathway: the perforant path input to the dentate area of the hippocampal formation, Exp. Brain Res., 12 (1971) 18-45. 31 Meldrum, B.S., Simon, R.P., Griffiths, T. and Evans, M.C., Hyperexcitability and brain protection, Clin. Neuropharmacol., Suppl. 1,7 (1984) 500-501. 32 Rothman, S.M., Synaptic activity mediates death of hypoxic neurons, Science, 220 (1983) 536-537. 33 Rothman, S., Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death, J. Neurosci., 4 (1984) 1884-1891. 34 Schiff, S.J. and Somjen, G.G., Hyperexcitability following moderate hypoxia in hippocampal tissue slices, Brain Research, 337 (1985) 337-340. 35 Siesjo, B.K., Cell damage in the brain: a speculative synthesis, J. Cereb. Blood Flow Metabol., 1 (1981) 155-185. 36 Siesjo, B.K. and Wieloch, T., Molecular mechanisms of ischemic brain damage: Ca2+-related events. In F. Plum and W. Pulsinelli (Eds.), Cerebrovascular Diseases, Raven
Press, New York, 1985, pp. 187-198. 37 Somjen, G.G. and Aitken, P.G., The ionic and metabolic responses associated with neuronal depression of Leao's type in cerebral cortex and in hippocampal formation, An. Acad. Brasil. Cienc., 56 (1984) 495-504. 38 Somjen, G.G., and Balestrino, M., Chlorpromazine improves resistance of hippocampal slices to hypoxic damage, Soc. Neurosci. Abstr., 11 (1985) 374. 39 Symon, L., Flow thresholds in brain ischaemia and the effects of drugs, Br. J. Anaesth., 57 (1985) 34-43. 40 Thomas, G.E., Levitsky, S. and Feinberg, H., Chlorpromazine inhibits loss of contractile function, compliance and ATP in ischemic rabbit heart, J. Mol. Cell. Cardiol., 15 (1983) 621-628. 41 Wode-Helgodt, N., Borg, S., Fyro, B. and Sedvall, G., Clinical effects and drug concentrations in plasma and cerebrospinal fluid in psychotic patients treated with fixed doses of chlorpromazine, Acta Psychiat. Scand., 58 (1978) 149-173.