www.elsevier.com / locate / bres
Helene Nallet , Eric T. MacKenzie, Simon Roussel
University of Caen, CNRS-UMR 6551, Cyceron, Boulevard Henri Becquerel, BP 5229, 14074 Caen Cedex, France Accepted 25 July 2000
Abstract
Transient ischaemic depolarizations (IDs) are thought to play a key role in the pathogenesis of focal cerebral ischaemia. Most transient IDs are akin to spreading depression (SD), although a negative DC shift is not observed in half the cases. The other IDs may represent transient anoxic depolarizations. Using cortical DC and blood flow recordings, following middle cerebral artery occlusion in rats, we show here that: (i) these later depolarizations do indeed represent transient anoxic depolarizations; (ii) SD-like IDs, DC and haemodynamic parameters are similar to those of SDs when blood flow remains close to base line and; (iii) when blood flow decreases, the hyperaemia associated with SD-like IDs is largely reduced and there is an increasing proportion of cortical sites which fail to display a DC shift. These data demonstrate the coexistence of two mechanisms of IDs, and yield new information as to the flow-dependence of DC and haemodynamic correlates of SD-like IDs, the pathophysiological significance of which remains to be determined. 2000 Elsevier Science B.V. All rights reserved.
Theme: Disorders of the nervous system
Topic: Ischaemia
Keywords: Transient ischaemic depolarization; Spreading depression; Anoxic depolarization; Cerebral blood flow; Middle cerebral artery occlusion; Rat
1. Introduction such IDs are transient anoxic depolarizations (ADs) and are thus associated with a reversible decrease in cerebral In the first hours following middle cerebral artery blood flow.
occlusion (MCAo), the transient ischaemic depolarizations The second type of event, with a duration of less than 7 (IDs) that occur intermittently in the penumbra are thought min (short IDs), are akin to SD, a phenomenon originally
˜
to be involved in the final infarction. We have previously described by Leao [15] as a reversible and transient demonstrated [18] that different types of IDs coexist in the negative shift of the direct current (DC) which spreads penumbra. While their characteristics (time of occurrence, over the cortex at a rate of 2–5 mm / min. One hypothesis, amplitude, duration and propagation) have been described that would explain the deleterious role of SD-like IDs, is in detail, their relationships with the underlying haemo- that repolarization requires extra energy which cannot be dynamic status remain unclear. supplied due to the limited haemodynamic reserve in the About 10% of IDs, the duration of which is more than 7 penumbral tissue. While SD is coupled to a compensatory min (long IDs), occur essentially in the first hour following rise in CBF and does not cause damage in an intact animal, ischaemia and have characteristics markedly different from the presence or not of a hyperaemia associated with SD-those of spreading depression (SD) and therefore cannot be like IDs remains to be clarified [1,10].
assimilated to the phenomenon of SD. We hypothesize that In addition, we have previously showed that, with ongoing ischaemia, some cortical sites intermittently fail to display a DC shift and are entirely replaced by suppressed electrocorticographic activity, while an ID wave is ob-*Corresponding author. Tel.: 133-2-3147-0282; fax: 1
33-2-3147-served on the other electrodes [18]. We have referred to 0222.
E-mail address: [email protected] (H. Nallet). these events as ECoG silences. The relationship between 0006-8993 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved.
the proportion of ECoG silences and the residual blood 37.58C. The cortical potential and CBF were recorded for flow is also unknown. 20 min prior to MCAo or stimulation, and over the Accordingly, the aim of our present study was to clarify following 4 to 5 h. Signals were digitized using LAB-the haemodynamic correlates of LAB-the various alterations in TECH software (Laboratory Technologies Corporation, ionic homeostasis that follow MCAo. This characterisation UK). During the recordings, the concentration of halothane should enable the determination as to whether long IDs was reduced to 0.8–1%.
represent AD and to clarify the relationship between the
residual blood flow and the hyperaemia associated with 2.4. Middle cerebral artery occlusion either IDs or the proportion of ECoG silences.
In the ischaemia group, permanent focal cerebral is-chaemia was induced by a remote controlled method of
2. Materials and methods intraluminal MCAo [20]. A cylinder of thermofusible adhesive (0.3832 mm), attached to a nylon thread (diam-Adult male Sprague–Dawley rats weighing 308630 g eter 0.22 mm), was advanced from the lumen of the were assigned to two experimental groups: a MCAo or external carotid artery (ECA) into the internal carotid ‘ischaemia’ group (n516), and an electrically-induced SD artery, 5 mm beyond the surface of the skull. A catheter or ‘control’ group (n54). Experimental procedures were was placed through the portion of the thread remaining performed according to the appropriate European direc- outside the ECA stump, inserted into the stump and tives and French national legislation. secured. Subsequently, the MCAo was achieved by
ad-vancing the thread by a further 4 mm. 2.1. General preparation
2.5. Electrically-induced spreading depression Anaesthesia with halothane (4%) was induced and
maintained (1.5–2%) during surgery in a mixture of In control animals, an additional right frontal O :N O (30:70) with the animal breathing spontaneously2 2 craniotomy (diameter 2 mm) with the dura mater intact (n511, ischaemia group) or intubated and artificially was performed 2 mm rostral to the bregma and 2 mm ventilated (n55 for ischaemia group, n54 for control lateral to the sagittal suture. Two insulated Ag /AgCl wires, group). A catheter was inserted into the femoral artery for were placed 2 mm apart between the calvarium and the arterial pressure monitoring and to obtain samples for gas meningeal surface. Trains of stimuli (2 ms pulses, 8 mA, at analysis. Rectal temperature was kept close to 37.58C with 80 Hz for 10 s) were then delivered every hour for 5 h. a feed-back controlled heating pad.
2.6. Data and statistical analyses 2.2. Electrophysiological recordings
Recordings were analysed using DADISP (DSP De-The animal’s head was fixed in a stereotaxic frame and velopment, UK) and Statview (Abacus Concepts, USA). the skull was exposed. A right frontoparietal craniotomy (8 Blood flow was expressed as percentage of base line. mm34 mm) was performed and the dura mater was left Blood flow changes (hyperaemia) were expressed in intact, laterally between the level of the lambda and the arbitrary units; 1 unit equals 1% of base line. Electro-bregma, so as to insert (2 mm apart) four glass microelec- physiological and CBF data were analysed by ANOVA. trodes (E1 to E4; tip diameter, 8mm) filled with physiolog- Each LDF recording was analysed in relation to the ical saline, 200mm deep into the cortex, within the MCA electrophysiological data arising from the closest electrode. territory (stereotaxic coordinates: all electrodes 3 mm Repeated-measures ANOVA followed by Student’s t-test lateral to the sagittal suture, E1 1 mm posterior to bregma, and a Bonferroni correction were used to compare physio-E4 7 mm posterior to the bregma, E2 and E3 in between). logical parameters between the groups. Contingency tables
2
Ag /AgCl wires were inserted into the microelectrodes and were analysed by the x test. Values are given as connected to an amplifier. An Ag /AgCl reference disk mean6S.E.M. and P,0.05 was accepted as significant. electrode was inserted under the skin of the neck.
2.3. Cerebral blood flow recordings 3. Results
Cerebral blood flow (CBF) was measured continuously 3.1. Physiological variables by laser doppler flowmetry (LDF). One or two LDF probes
animals was observed only for PaCO (46.42 61.4 mmHg with the post-occlusion time (Fig. 1). To test the hypoth-and 38.861.4 mmHg, respectively; P50.005) and, as esis that ECoG silences could be due to refractoriness, we would be expected, for pHa (7.3460.01 and 7.3960.01, calculated the average lag time between an ECoG silence respectively; P50.04). The other systemic parameters and the previous event. An ECoG silence occurred on were: PaO2513065 mmHg; MAP59262 mmHg. average 22624 min after the preceding event wave.
3.2. Ischaemia group 3.2.2. CBF changes associated with long IDs
Most long IDs (70.6% of long IDs) were associated with 3.2.1. CBF and DC /ECoG recordings a decrease in CBF to 18.165.7% of base line values and, Immediately following MCAo, the residual CBF ranged at the time of repolarization, with a recovery of CBF from 100 to 9.6% of the base line values. A terminal AD towards pre-ischaemic values (73.6614.7% of base line occurred in only one animal with a CBF of 15% of control values) (Fig. 2A).
values. In all other animals, IDs affecting sequentially all These long IDs were recorded as the first electrophysio-or some of the electrodes (98 series of events, each logical event and occurred on all of the four electrodes. No observed at four locations, corresponding to 392 ischaemic long IDs were observed twice in the same animal. The events) were noted at irregular intervals, were reversible duration of the remaining long IDs was 9.860.9 min, a and were of variable duration. Based on the criteria value slightly above our threshold and these events were previously validated in our model [18], two populations of therefore considered SD-like depolarizations (duration of IDs were separated according to their duration: long IDs SD-like IDs to include these later events: 2.360.1 min; (4.3% of events; duration: 23.463.6 min; amplitude: amplitude: 16.660.5 mV).
24.262.8 mV) and SD-like depolarizations (55.4% of
events; duration: 2.160.1 min; amplitude: 16.660.5 mV). 3.2.3. CBF changes associated with SD-like IDs and The remaining events (40.3%) were ECoG silences. The ECoG silences
total number of recorded events decreased progressively SD-like depolarizations and ECoG silences occurred
Fig. 2. Examples of CBF changes associated with electrophysiological ischaemic events in four different animals. The LDF probe and electrode are approximately 1–2 mm apart. (A) Transient decrease in CBF associated with a long ID; (B) Hyperaemia associated with SD-like ID at normal CBF values; (C) Hyperaemia associated with a SD-like ID and ECoG silence (second event) at a residual flow value of 80% of base line; (D) Insignificant variations in CBF associated with two SD-like depolarizations at a residual flow value of 60% of base line.
when the residual CBF was greater than 25.5% of pre- 3.3. Control group ischaemic values and were associated with either no
change or a variable increase in CBF (Fig. 2B, C, D). The In the control group, a deflection of the DC potential data were analysed as a function of the residual CBF, (duration: 1.0860.03 min; amplitude: 22.460.5 mV), was binned by intervals of 10%: only the electrophysiological observed nearly universally after stimulation (99.4%). SDs data arising from the electrode closest to each laser- were associated with a hyperaemia which was also depen-doppler probe was taken into consideration. It must be dent on the pre-event blood flow (which spontaneously noted that we cover a wide range of residual CBF because ranged from 60 to 120% of base line in this group): in this the residual CBF varies from animal to animal. In each rat, case, the greater the pre-event flow, the lesser the hy-residual flow was similar at the two LDF probe locations. peraemia (Fig. 3). The amplitude and duration of SDs were The average amplitude of the hyperaemia associated with not dependent on the pre-event flow.
SD-like depolarizations was a function of the residual
CBF: the greater the residual CBF, the greater the hy- 3.4. Comparison between ischaemic and control groups peraemia (Fig. 3). The percentage of ECoG silences
Fig. 3. Histogram of the average hyperaemia (mean6S.E.M.;DCBF in arbitrary units: 1 unit equals 1% of base line) associated with SD-like events (short IDs and ECoG silences) and electrically-induced SDs at different residual or pre-event CBF values. In the ischaemia group, the hyperaemia is significantly dependent on the residual CBF (ANOVA, P,0.0001, F513.952, 9 degrees of freedom by Fischer PLSD post-hoc tests (not shown) were performed to compare the amplitude of hyperaemia associated with SD-like depolarization, between each pair of residual CBF bins (45 comparisons): 30 comparisons were significant and 15 were non-significant. In the stimulation group, the hyperaemia is significantly dependent on the pre-event CBF (ANOVA,
P50.0004; F56.383, 5 degrees of freedom; Fischer PLSD post-hoc tests, not shown: 8 significant and 7 non-significant comparisons). Comparison of the two groups is performed by two-way ANOVA (group; pre-event flow from 60 to 110% of base line): there is an interaction between group and pre-event flow effects (P,0.0001). *P,0.002: significant difference between the two groups (Student’s t-test with Bonferonni correction).
flow effects (P,0.0001) and the amplitude was signifi- profile of extracellular potassium increase [19], their cantly different between the two groups at flows ranging duration [5] or their time of occurrence [5], some IDs have from 60 to 90% of base line (Fig. 3). With the electro- been postulated to correspond to an anoxic depolarization. physiological data (amplitude and duration), there was no These hypotheses were based on investigations in which interaction between the group and the pre-event flow blood flow was not measured. As in the present study, long effects (P.0.52) but a significant group effect (P,0.02). IDs occurred shortly after the initial fall in blood flow and repolarization followed the recovery of blood flow, these long IDs can be assimilated to anoxic depolarizations. The
4. Discussion initial decrease in CBF may be accompanied by an
Fig. 4. Percentage of ECoG silences relative to the total number of ischaemic events (ECoG silences and SD-like IDs) at different residual CBF values.
2
The distribution of ECoG silences was significantly different from that of short IDs for the entire range of residual flow values (x test, P50.0103).
MCAo, both the electrophysiological characteristics and absence of depolarization is an adaptive mechanism which the associated hyperaemia of IDs are similar to those of may spare intracellular ATP concentrations. Based on this
1
SDs. However, when blood flow is decreased, these hypothesis, the opening of K -ATP channels by a de-characteristics no longer match those of SDs. ficiency in ATP [11] would explain both the suppression of With respect to their electrophysiological characteristics, the potassium efflux and its dependence on blood flow, it would appear that ECoG silences are more frequent the since the hyperpolarization due to the opening of
ATP-1
lower the residual blood flow. The basis of this phenom- dependent K channels would counteract the depolariza-enon is not known. ECoG silences occur locally during an tion and since ATP concentrations are expected to decrease SD-like wave (which can be seen as a depolarization on at lower rates of flow.
duration of SD-like IDs has been largely overestimated in Acknowledgements
the literature.
´ `
Helene Nallet received a bursary from the University of Spreading depression has been nearly universally
associ-Caen and the studies were carried out within the IFR 47 ated with a hyperaemia and our data confirm that this
‘Neuro-imagerie Fonctionnelle’. The investigations were increase in blood flow is inversely proportional to the base
supported by grants from the CNRS and the University of line flow [14]. Under ischaemic conditions, our data
Caen. The authors thank Isabelle and Pascal Duchatelle for confirm that, at low residual rates of flow, the hyperaemia
their help with the microelectrode preparation. becomes negligible or absent [1]. This finding is in
agreement with the idea that the deleterious effects of SD-like IDs are due to a reduced haemodynamic reserve in
References
the penumbra and, therefore, to a mismatch between energy requirements and availability. This hypothesis is
[1] T. Back, K.A. Hossman, Cortical negative DC deflections following challenged, however, by data which show that, in the
middle cerebral artery occlusion and KCl-induced spreading depres-absence of an ischaemic focus but with a reduced blood
sion: effect on blood flow, tissue oxygenation, and electroence-flow, SDs are not deleterious even when the total period of phalogram, J. Cereb. Blood Flow Metab. 14 (1994) 12–19. depolarization lasts up to 70 min [8]. It has been proposed [2] S.M. Bowyer, Y.C. Okada, N. Papuashvili, J.E. Moran, G.L.
Barkley, K.M.A. Welch, N. Tepley, Analysis of MEG signals of that an even more reduced blood flow or the presence of an
spreading depression with propagation constrained to a rectangular ischaemic focus is necessary for SDs to induce damage [8].
cortical strip. I. Lissencephalic rabbit model, Brain Res. 843 (1999) Another hypothesis can be invoked. We have here demon- 71–78.
strated that the hyperaemia is proportional to residual [3] S.M. Bowyer, N. Tepley, N. Papuashvili, S. Kato, G.L. Barkley, blood flow. When blood flow is in the physiological range K.M.A. Welch, Y.C. Okada, Analysis of MEG signals of cortical spreading depression with propagation constrained to a rectangular or only slightly reduced, the hyperaemia is similar to that
strip. II. Gyrencephalic swine model, Brain Res. 843 (1999) 79–86. seen following SD. Since a wave of depolarization induces
[4] D.M. Colonna, W. Meng, D.D. Deal, M. Gowda, D.W. Busija, an increase in blood flow only in those tissues in which Neuronal NO promotes cerebral cortical hyperemia during cortical perfusion is not severely reduced, i.e. in the periphery of spreading depression in rabbits, Am. J. Physiol. 41 (1997) H1315– the ischaemic focus, a steal phenomenon could occur. H1322.
[5] R.M. Dijkhuizen, J.P. Beekwilder, B.H. Van der Worp, J.W.B. Van Although we did not observe a decrease in blood flow
der Sprenkel, K.A.F. Tulleken, K. Nicolay, Correlation between during the SD-like IDs seen at low residual rates of flow,
tissue depolarizations and damage in focal ischemic rat brain, Brain one might envisage that this phenomenon exists in the Res. 840 (1999) 194–205.
vicinity of the ischaemic focus, where a further minor [6] K.J. Escott, J.S. Beech, K.K. Haga, S.C.R. Williams, B.S. Meldrum, decrease in flow may be sufficient to attain the threshold P.M.W. Bath, Cerebroprotective effect of the nitric oxide synthase inhibitors, 1-(2-trifluoromethylphenyl) imidazol and 7-nitro indazol, for membrane failure. In addition, this later hypothesis
after transient focal cerebral ischaemia in the rat, J. Cereb. Blood would explain the neuroprotective effects of the inhibition
Flow Metab. 18 (1998) 281–287.
of neuronal nitric oxide synthase [6], a pharmacodynamic [7] A.R. Gardner-Medwin, N. van Bruggen, S.R. Williams, R.G. Ahier, effect which is reported to block the hyperaemia associated Magnetic resonance imaging of propagating waves of spreading with SDs [4] and would accordingly be expected to depression in the anaesthetised rat, J. Cereb. Blood Flow Metab. 14
(1994) 7–11. increase brain damage rather than to decrease it.
¨ ¨
[8] G. Gido, T. Kristian, B.K. Siesjo, Induced spreading depression in In conclusion, two mechanisms of transient IDs coexist:
energy-compromised neocortical tissue: calcium transients and one is a decrease in CBF followed by a recovery, which histopathological correlates, Neurobiol. Dis. 1 (1994) 31–41. induces transient anoxic depolarization; the second is [9] R. Gill, P. Andine, L. Hillered, L. Persson, H. Hagberg, The effect´ spreading depression. SD-like IDs propagate to all cortical of MK-801 on cortical spreading depression in the penumbral zone following focal ischaemia in the rat, J. Cereb. Blood Flow Metab. 12 recording sites no matter the level of blood flow. With a
(1992) 371–379. normal flow, SD-like ID electrophysiological and
haemo-[10] T. Iijima, G. Mies, K.A. Hossman, Repeated negative DC deflec-dynamic effects are identical to those of SDs. With reduced tions in rat cortex following middle cerebral artery occlusion are flow, the hyperaemia associated with SD-like IDs is less abolished by MK-801: effect on volume of ischemic injury, J. prominent and a large proportion of the waves fail to Cereb. Blood Flow Metab. 12 (1992) 727–733.
[11] C. Jiang, F.J. Sigworth, G.G. Haddad, Oxygen deprivation activates display the depolarization in the near entirety of the 1
an ATP-inhibitable K channel in substantia nigra neurons, J. recording sites; the pathophysiological significance of this
Neurosci. 14 (1994) 5590–5602.
phenomenon remains to be determined. These findings [12] P.E. Kunkler, R.P. Kraig, Calcium waves precede electrophysiologi-should be useful for the interpretation of magnetic reso- cal changes of spreading depression in hippocampal organ cultures, nance images following cerebral ischaemia in both ex- J. Neurosci. 18 (1998) 3416–3425.
[13] L.L. Latour, Y. Hasegawa, J.E. Formato, M. Fisher, C.H. Sotak, perimental and clinical studies. Finally, our current
ob-Spreading waves of decreased diffusion coefficient after cortical servations open new research perspectives towards the
stimulation in the rat brain, Magn. Reson. Med 32 (1994) 189–198. understanding of the mechanisms and roles of the altera- [14] M. Lauritzen, Regional cerebral blood flow during cortical spreading tions of the characteristics of SD secondary to a reduced depression in rat brain: increased reactive hyperperfusion in low
[15] A.P. Leao, Spreading depression of activity in the cerebral cortex, J. [18] H. Nallet, E.T. MacKenzie, S. Roussel, The nature of penumbral Neurophysiol. 7 (1944) 359–390. depolarizations following focal cerebral ischaemia in the rat, Brain [16] A. Mancuso, N. Derugin, Y. Ono, K. Hara, F.R. Sharp, P.R. Res. 842 (1999) 148–158.
Weinstein, Transient MRI-detected water apparent diffusion coeffi- [19] M. Nedergaard, A.J. Hansen, Characterization of cortical depolariza-cient reduction correlates with c-fos mRNA but not hsp70 mRNA tions evoked in focal cerebral ischemia, J. Cereb. Blood Flow induction during focal cerebral ischemia in rats, Brain Res. 839 Metab. 13 (1993) 568–574.
