Abstract

An important but poorly understood event associated with ischemia is anoxic depolarization (AD), a sudden and profound depolarization of neurons and glia in cortical and subcortical gray matter. Leao first measured the AD as a wave of electrical silence moving across the cerebral cortex in 1947 and noted its similarity to spreading depression (SD). SD is harmless when coursing through normoxic cortical tissue as during migraine aura. However for 3–4 h following focal ischemia, the additional metabolic stress arising from recurring SD in the penumbra expands the ischemic core, so SD blockade is potentially beneficial therapeutically. In the present study, we measured intrinsic optical signals (IOSs) to monitor anoxic depolarization in submerged rat neocortical slices during O2/glucose deprivation (OGD). After ~6 min of OGD, the AD was imaged as a focal increase in light transmittance which then propagated across neocortical gray at ~2 mm/min. Although the slice was globally stressed, the AD always initiated focally, sometimes at multiple sites. Its propagation was coincident with a transient negative shift in the extracellular potential, the electrical signature of AD. Acute damage to neocortex (measured as a delayed decrease in LT and as a loss of the evoked field potential) followed only where the AD had propagated, so it is the combined metabolic demands of AD and OGD that acutely damages all layers of the neocortex. Glutamate receptor antagonists (2 mM kynurenate or 25 μM AP-5/10 μM CNQX) did not block AD initiation, slow its propagation or prevent post-AD damage. This study shows that acute ischemic damage is greatly exacerberated by AD during metabolic stress and that glutamate receptor antagonists are not protective. Using this slice model, therapeutically tolerable drugs that block the AD and SD can be investigated.

Introduction

One approach to the future management of stroke is the use of neuroprotectants to limit damage to the brain tissue and improve outcome. Effective therapeutic interventions require an understanding of how neurons and glia respond to the first crucial minutes of ischemia. The initial pathophysiological processes include energy failure, loss of ion homeostasis, depolarization and water influx. Because the prime suspect in the initiation of ischemic damage has been excess glutamate release, much research has focused on glutamate receptor antagonists as neuroprotectants. The therapeutic results have been disappointing and the central role of glutamate in acute ischemic damage is increasingly under scrutiny (Obrenovitch and Urenjak, 1997a, b; Obrenovitch, 1999; Obeidat et al., 2000).

An important process associated with ischemia is anoxic depolarization (AD) which was described originally as a propagating electrical silence following interruption of the cerebral circulation (Leao, 1947). It was described as spreading depression-like, a sudden and profound depolarization to the point where neurons can no longer discharge. Spreading depression (SD) involves loss of ion homeostasis and water influx under normoxic conditions (Somjen et al., 1992) that arises focally and then engulfs neurons and glia at a rate of 2–5 mm/min. The cerebral cortex, where it was first measured as a wave of electrical silence lasting several minutes (Leao, 1944), is particularly susceptible. The cause of SD (and thus its prevention) is not understood and its characteristics depend on the tissue's metabolic status. SD without metabolic compromise (as occurs in migraine aura) causes no discernible damage to intact neocortex (Lauritzen, 1987; Nedergaard and Hansen, 1988) or to neocortical slices (Footit and Newberry, 1998; Anderson et al., 1999). In vivo this ‘normoxic’ version of SD is blocked by N-methyl-d-aspartatic acid (NMDA) receptor antagonists but the AD is not blocked by NMDA or non-NMDA receptor antagonists (Hernandez-Caceres et al., 1987; Marranes et al., 1988; Nedergaard and Hansen, 1988; Lauritzen and Hansen, 1992; Nellgard and Wieloch, 1992). For 3–4 h following focal ischemia onset, recurring SD-like events expand the ischemic core and increase the number of ‘at risk’ neurons in the penumbra (Hossman, 1994, 1996; Nedergaard, 1996; Takano et al., 1996; Irwin and Walz, 1999). Preventing penumbral infarction reduces neurological impairment clinically (Furlan et al., 1996), so inhibiting the initiation and propagation of recurrent SD during ischemia should reduce infarct volume.

Cell swelling during ischemia-like conditions contributes a major component of the measured intrinsic optical signal (IOS) which is associated with changes in light transmittance in brain slices under ischemia-like conditions (Basarsky et al., 1998; Obeidat and Andrew, 1998; Aitken et al., 1999; Andrew et al., 1999; Kreisman et al., 2000). In submerged slices, an initial LT increase represents the propagating depolarizing front, whereas a subsequent LT reduction is attributed to dendritic damage (Andrew et al., 1999; Jarvis et al., 1999; Obeidat et al., 2000). Thus the imaging technique permits an assessment of both AD initiation and the acute neuronal damage left in its wake.

In this study, we optically map the effects of simulated global ischemia by depriving the neocortical slice preparation of O2 and lowering glucose to 1 mM. Our results indicate that the so-called anoxic depolarization that arises during global ischemia initially appears as a multi-focal event, each focus spreading out concentrically over the cortex and exacerbating neuronal damage in gray matter. Preliminary data have been presented by Jarvis and Andrew (Jarvis and Andrew, 1998).

Materials and Methods

Neocortical Slice Preparation

Male Sprague–Dawley rats, 21–30 days old (Charles River, St Constant, Quebec, Canada), were housed in a controlled environment (25°C, 12 h light/dark cycle) and fed Purina lab chow and water ad libitum. A rat was placed in a rodent restrainer and guillotined. The brain was excised and placed in ice cold oxygenated (95% O2/5% CO2) artificial cerebrospinal fluid (aCSF). Coronal slices (350–400 μm) of neocortex and underlying striatum or hippocampus were cut using a vibrating microtome (Leica VT100S). Slices were incubated (2–6 h) at 30°C in aCSF before transfer to the recording chamber. A slice was submerged in oxygenated aCSF flowing at a rate of 1–2 ml/min at 37.5°C.

Solutions

The aCSF contained (in mM): NaCl 120, KCl 3.3, NaHCO3 26, MgSO4 1.3, NaH2PO4 1.2, d-glucose 11, CaCl2 1.8. The pH was 7.3–7.4 and the osmolarity was 292–295 mOsm. O2/glucose deprivation (OGD) which simulates ischemia in vitro, was induced by reducing aCSF glucose from 11 to 1 mM and gassing the aCSF with 95% N2/5% CO2. NMDA (100 μM) or ouabain (100 μM) were added to the aCSF as required. Furosemide (5 mM), kynurenic acid (2 mM), dl-2-amino-5-phosphonovaleric acid (AP-5, 50 μM) or (6-cyano-7-nitroquinozaline-2,3-(1H,4H)-dione (CNQX, 10 μM), were added to the aCSF as required. The 50:50 racemic stock of AP-5 made up to 50 μM contained 25 μM of the active d-isomer. All chemicals were purchased from the Sigma Chemical Co.

Imaging Intrinsic Optical Signals (IOSs)

IOSs are generated by changes in light scattering or absorbance within living tissue. In terms of optics, the simplest paradigm is to image change in light transmitted by a submerged brain slice, thereby avoiding complexities associated with measuring reflectance and the tissue/air interface [reviewed by Jarvis et al. (Jarvis et al., 1999)]. A neocortical slice was placed in an imaging chamber with a coverslip as the base. The slice was superfused with flowing aCSF, transilluminated with a broadband halogen light source, and viewed with a 1.25 objective on an inverted microscope (Zeiss Axiovert TV 100). Video images were collected using a charge coupled device (CCD, Cohu) connected to an image processing board (DT 3155, Translation) in a PC controlled by Axon Imaging Workbench software (Axon Instruments). The CCD was set at maximum gain and low black level. The gamma level was set to 1.0 so that CCD output was linear with respect to changes in light intensity. With appropriate filters, the IOS signal comprised the far red to near infrared spectrum (690–1000 nm). Video frames were acquired at 1.3–8 s intervals with 32–256 frames averaged per single image. The transmittance value (T) of the first image (Tcont) was subtracted from each subsequent image (Texpt) of the series, so the difference image (TexpTcont) revealed areas where LT changed over time. To visualize these areas better, light transmittance changes (ΔLT) were pseudocolored. Data were also quantified and graphically displayed by averaging the digital intensities from selected zones of interest (ZOI). Since various regions differed in Tcont, the data were normalized as follows. The change in light transmittance = 

\[{\bigtriangleup}\mathit{LT}\ =\ \frac{(\mathit{T}_{exp}\ {\mbox{--}}\ \mathit{T}_{cont}){/}\mathit{gain}}{\mathit{T}_{cont}}{\times}\ 100\ =\ \frac{{\bigtriangleup}\mathit{T}}{\mathit{T}}\%\]
where the CCD is set at maximum gain and the software gain is set at 1.0. These values were plotted over time.

There are several biophysical changes when cells take up water and swell which act to reduce light scattering (thereby elevating LT). However, when dendritic beading accompanies swelling (as during excitotoxicity or O2/glucose deprivation) the beading becomes the most important biophysical factor, scattering light even as the tissue continues to swell (Jarvis et al., 1999).

Electrophysiology

Extracellular recording of evoked field potentials served as an indicator of synaptic function and slice viability. Simultaneous extracellular recording and IOS imaging during OGD was performed to correlate the onset of the negative shift and LT changes resulting from OGD. The recording micropipette was placed in layer II/III of the neocortical slice and a concentric bipolar stimulating electrode was placed in the immediately underlying V/VI layers. A current pulse (0.1 ms duration; 0.25 Hz) was applied to produce a near-maximal amplitude population spike. Digitized data were plotted using pCLAMP software (Axon Instruments). The recording pipette also served to measure a negative DC shift in the extracellular potential induced by OGD. The DC shift is the electrical signature of the AD and SD.

Histology

For histological analyses, control slices were maintained in the imaging chamber for 15 min at 37.5°C then fixed in Bouin's fluid. Experimental slices were exposed to OGD for 10 min at 37.5°C which evoked an AD episode. After 15 min they were placed in fixative. Following fixation for 24 h in Bouin's fluid, slices were stored in 70% ethanol and processed for paraffin embedding. Sections (7 μm) were stained with hematoxylin/ eosin and photographed using a 40× objective.

Results

Simulated Ischemia Induces a Propagating Anoxic Depolarization in Neocortical Slices

Within 10 min, OGD at 37.5°C induced AD in all neocortical slices tested with a mean latency of 4:26 ± 1:24 (min:s ± SD, n = 32). AD was first detected as one or two focal increases in LT in layers II/III. Each focus usually first appeared as a sphere ~0.5 mm in diameter that migrated out along adjacent cortex of all six layers. The ignition site could be in lateral, central or midline neocortex. Changes in LT were plotted from several zones of interest across the neocortex. AD propagated as a front of elevated LT in gray but not white matter at a rate of 1.50 ± 0.83 mm/min (n = 32). A dramatic decrease in LT then followed in the wake of the elevated LT front. Figure 1A shows migration of the AD front (initiated out of frame) first in one hemisphere and then as an independent event contralaterally. The elevated LT front propagates across the gray matter (Fig. 2A), activating all neocortical layers of a single column concurrently (Fig. 2B).

Bath application of the Na+/K+-ATPase inhibitor ouabain for 2 min produced a response virtually identical to that induced by OGD for 10 min. Following bath exposure to 100 μM ouabain at 37.5°C, SD developed at 2:40 ± 00:50 (n = 6), which was 5 min earlier than that produced by OGD. Otherwise, migration and the irreversible LT reduction appeared comparable to OGD (Fig. 1A,B).

Simultaneous IOS imaging (Fig. 3A) and extracellular recording (Fig. 3B) were performed to correlate changes in extracellular potential with changes in LT during OGD. A recording electrode in layer II/III monitored the extracellular potential (Fig. 3B) as well as the population spike (Fig. 3D) evoked by a stimulating electrode positioned in underlying layers V/VI. A sudden negative shift of 5–10 mV (Fig. 3B) correlated temporally and spatially with the passage of the elevated LT front past the electrode tip (Fig. 3A,C; n = 5). The negative shift started to return to baseline during OGD exposure, usually reaching baseline within 2–3 min but often not returning completely. Significantly, the evoked field potential recorded in layers II/III prior to OGD was abolished following AD and showed no recovery over the following 30 min (Fig. 3D).

The rate that aCSF flows over the slice affects the onset of AD. Most data reported here involved a flow rate of 1–2 ml/min. The exception is Figure 2C where the rate was increased to 4 ml/min. In this case, AD elicted by OGD was consistently delayed by 3–6 min and damage (judged by the degree of LT reduction) was reduced in ten of ten slices tested. It is possible that the faster superfusion rate increases the washout of accumulating extra-cellular K+. This in turn would increase the latency to AD onset. This delay in onset reduces the post-AD period under OGD exposure, meaning less time exposed to the combined metabolic stress of AD and OGD. Thus the LT reduction (denoting damage) is lessened at the faster flow rate.

Histology/Temperature/Furosemide

Histological sections of control tissue (n = 4) maintained at 37.5°C for 15 min (no AD) were compared with experimental slices (n = 4) that supported AD in response to OGD for 10 min at 37.5°C. In contrast to control tissue (Fig. 4A), neocortical neurons that generated the AD displayed nuclear and cytoplasmic swelling (Fig. 4B). In addition, the large primary dendrites that could be discerned running perpendicular to the cortical layers in control slices (Fig. 4A, arrowheads) were not apparent in experimental slices where the neuropil instead appeared mottled (Fig. 4B), suggesting altered structure of these cortical neurons, which were also unable to generate an evoked field potential.

Anoxic depolarization induced by OGD persisted in neocortical slices maintained at a lowered temperature of 32°C (n = 5). The onset time of AD at 32°C was delayed by a mean duration of 3 min (n = 5), but this difference was not statistically significant at P < 0.01 (Fig. 6A). The propagation rate was also not affected (Fig. 6B). Neocortical slices maintained at 22°C did not display an AD following 10 min of OGD (n = 5) and so no large change in LT was observed.

To provide evidence for a possible glial component to the intrinsic optical signal, we pretreated slices with furosemide. This is thought to block astrocyte swelling caused by K+ uptake during neuronal activity by interfering with the Na+/K+/2Cl cotransporter (Walz, 1987). In the present study, however, the presence of 5 mM furosemide did not alter the onset time (Fig. 6A) nor propagation rate (Fig. 6B) of OGD-induced AD (n = 7). An AD event also initiated and propagated within the underlying striatum (n = 4) or hippocampus (n = 2) in the presence of furosemide and was comparable to that in control tissue (not shown).

Glutamate Receptors and SD Induced by OGD

To test the potential role of glutamate receptors in the OGD response, the non-specific glutamate antagonist kynurenate (2 mM) was applied 15–40 min before, during and after OGD at 37.5°C. Treatment with kynurenate did not prevent the initiation and propagation of AD induced by 10 min of OGD in nine of nine neocortical slices tested (Fig. 5A). Moreover, kynurenate did not significantly alter AD onset time (Fig. 6A) nor propagation rate (Fig. 6B). The negative shift recorded extracellularly in layers II/III was similar in waveform to slices without kynurenate (n = 6, not shown). In support of these findings, treatment of slices with a combination of an NMDA and a non-NMDA receptor antagonist, AP5 (d-isomer, 25 μM) and CNQX (10 μM) respectively, did not prevent OGD-induced AD in five of five slices tested (Fig. 5). Neither the time to onset (Fig. 5A) nor the propagation rate (Fig. 5B) were altered by AP-5/CNQX treatment. Likewise, treatment with 50 μM of the d-isomer of AP-5 alone was ineffective in five of five slices tested (not shown).

AD could also be generated by OGD in the underlying hippo-campus (Fig. 5B). The latency to onset in the CA1 region was 7:45 ± 1.27 (n = 7), an average of 3 min after AD onset in neocortex. The onset time was in keeping with that reported in isolated hippocampal slices which we previously termed ‘ischemic SD’ (Obeidat and Andrew, 1998) and so was independent of AD in the overlying neocortex. The propagation rate in the hippocampus ranged between 0.6 and 2 mm/min. In the hippocampus (n = 6), OGD-induced AD was delayed in the presence of kynurenate, in agreement with previous observations (Obeidat et al., 2000) and persisted in hippocampal slices in the presence of AP-5/CNQX (n = 3). Most importantly, in both neocortex and hippocampus, treatment with glutamate receptor antagonists did not affect the subsequent reduction in LT which represents neuronal damage (Fig. 5A).

If extracellular glutamate accumulation has a role in OGD-induced AD, then glutamate receptor agonists should elicit AD in a pattern similar to OGD. However, the application of 100 μM NMDA at 37.5°C (n = 8, Fig. 7A, B) first produced a generalized, not focal, elevation in LT which developed more slowly than AD onset induced by OGD (Figs 1A, 2A). However, there then followed a sudden spreading optical signal reduction (Fig. 7A, B) coincident with a negative shift in layers II/III (Fig. 7C). A generalized elevation in LT was also observed in the hippocampus, but no spreading event was observed in eight of eight hippocampal slices, as previously reported (Jarvis et al., 1999).

A role for voltage-dependent Na+ channels was tested using tetrodotoxin (TTX). The presence of 1 μM TTX neither prevented the initiation nor slowed the propagation of OGD-induced AD in four of four neocortical slices. However, it did delay the onset of the AD by an average of 4 min (P < 0.005; Fig. 6). The presence of 1 μM TTX + 2 mM kynurenate significantly (P < 0.005) delayed the initiation of the AD (induced by 20 min OGD) by ~8 min (n = 7; Fig. 6). Development of the negative optical signal observed following AD proceeded normally. Thus it appears that blocking voltage-dependent Na+ channels and glutamate-gated channels together can delay the onset of the AD, but not prevent it.

Discussion

Imaging Reveals that the AD and OGD Together Exacerbate Neuronal Damage

In animal models of ischemia, the extent of necrotic brain injury is quantified by determining infarct size or neuronal loss sampled histologically hours or days post-insult (Corbett and Nurse, 1998). Recurrent SD, also termed peri-infarct depolarization (PID), contributes to this damage in the penumbra (Nedergaard, 1996; Hossman 1996; Strong et al., 1999). Mapping intrinsic optical signals (IOSs) reveals real-time responses to glutamate agonists or to simulated ischemia in brain slices (Andrew et al., 1996; Obeidat and Andrew, 1998; Polischuk et al., 1998; Aitken et al., 1999; Jarvis et al., 1999). IOS imaging detects cell swelling, clearly demarcating the ignition site and migration front of AD across live submerged slices of hippo-campus or neocortex (Basarsky et al., 1998; Obeidat and Andrew, 1998). The present study suggests that during global ischemia, the AD initiates at multiple sites and propagates outwards like the ripples from several stones thrown in a pond. In their wake the ripples leave further energetically compromised gray matter.

SD involves a sudden and massive increase in membrane permeability causing influx of Na+, Ca2+, Cl and water and efflux of K+ (Somjen et al., 1992). The resultant cell swelling decreases extracellular space (Nicholson and Kraig, 1981; Hansen, 1985). Under normoxic conditions, ion redistribution returns to normal within minutes with no neuronal or glial damage. However, the AD generated during ischemic-like conditions (Leao, 1947) exacerbates neuronal damage (Nedergaard and Hansen, 1993; Mies et al., 1994; Back et al., 1996), presumably by further elevating the metabolic load in the vulnerable penumbra (Obrenovitch, 1995).

In the present study, submerged neocortical slices were imaged at low magnification in response to metabolic compromise by O2/glucose deprivation (OGD) or by ouabain (Balestrino, 1995; Balestrino et al., 1999), both simulating a global ischemia. Either treatment evokes an increase in LT that arises focally at one or several sites and propagates into adjacent cortical tissue. The AD propagation rate of 1.5–2.0 mm/min is in the lower range of those reported for SD in intact cortex (Nedergaard, 1996).

In the hippocampal slice (Obeidat et al., 2000), OGD induced a propagating increase in LT (cell swelling) followed by a rapid decrease in LT over several minutes. This decrease might represent cellular shrinkage because decreased LT is reversibly evoked by hyperosmotic saline (Andrew and MacVicar, 1994; Andrew et al., 1997). However, the more likely cause is dendritic ‘beading’ which develops even as the tissue continues to swell and is indicative of neuronal damage. The necklace-like conformation of hundreds of dendritic processes is highly efficient at scattering light, such that bead formation over several minutes dramatically reduces light transmittance (Polischuk et al., 1998; Jarvis et al., 1999; Obeidat et al., 2000). Where dendrites are lacking (as in hippocampal cell body layers), LT continues to increase. In contrast, the high density of dendrites in all neocortical layers leads to a decreased LT across gray matter once enough beading overwhelms the initial increase in LT generated by cell swelling. Dendritic beading has been observed in cultured neurons following O2/glucose deprivation (Park et al., 1996) and in vivo following ischemia (Hori and Carpenter, 1994). The optical sequence evoked by ouabain is indistinguishable from that evoked by OGD, consistent with previous observations in hippocampus (Obeidat and Andrew, 1998). In both cases there is loss of Na+/K+ pump function causing increases in [K+]o, which appears to be a critical step in the induction of the AD.

IOS imaging and simultaneous recording of the extracellular potential in layers II/III reveal a temporal and spacial correlation between signals. The negative voltage shift recorded extracellularly, which is the electrophysiological signature of the AD, arises as the LT front passes the site of the recording electrode. The negative shift may return to near baseline, but this does not indicate physiological recovery because the evoked field potential is permanently lost following the AD. Moreover, an irreversible decrease in LT develops following the AD. In addition, histological analysis reveals swelling of neuronal nuclei and cell bodies in all cortical layers. The neuropil displays a mottled appearance and indistinct primary dendrites. None of these indicators of neuronal damage are observed in OGD-exposed tissue unless the AD is generated. In contrast, SD can be generated repetitively in our slices each time extracellular K+ is briefly increased and there are no signs of tissue damage (Anderson et al., 1999), so SD without OGD appears innocuous.

AD Resists Pharmacological Blockade

Bath application of the glutamate receptor agonist NMDA produced a general LT increase in neocortex and hippocampus (particulary in the CA1 region). Unlike the AD, the signal developed slowly and uniformly over the slice, did not propagate and was blocked by AP-5. An SD-like negative shift developed just at the time that the LT increase peaked and abruptly reversed, which was the point at which we conjecture light scattering by beaded dendrites began to overwhelm the ‘swelling’ signal. Thus the NMDA-evoked signal sequence had some elements of the AD, but it appeared to engage the entire slice almost simultaneously.

NMDA receptor antagonists block SD under normoxic conditions in vivo (Hernandez-Caceres et al., 1987; Marranes et al., 1988; Nedergaard and Hansen, 1988; Lauritzen and Hansen, 1992; Nellgard and Wieloch, 1992). These same studies show that glutamate receptors do not play a major role in the AD induced by ischemia. A recent study (Rossi et al., 1999) showed that the AD could be blocked by a combination of NMDAR and non-NMDAR antagonists, but used slices from 12-day-old rats where NMDAR density is higher than adults. MK801 is ~10-fold more potent as a neuroprotectant against NMDA- and OGD- mediated neuronal injury in immature rodents than in adults (McDonald and Johnston, 1990). Our study supports in vivo work cited above by showing that NMDA receptor antagonism does not block the AD in slices or offer protection from acute post-AD damage. This is unlike normoxic SD in sister brain slices (Anderson et al., 1999) where NMDA receptor antagonists are effective blockers. This indicates a contribution by these receptors in generating the milder SD. However, the fact that glutamate begins to accumulate only after the anoxic depolarization (Obrenovitch and Urenjak, 1997a; Obrenovitch, 1999) further argues against a role for glutamate in AD initiation and propagation under metabolically compromised conditions.

While both neurons and glia depolarize during AD, their relative contribution to the optical signal is still an open question. A previous study using furosemide (to interfere with the Na+/K+/2Clco-transporter) blocked weak IOSs generated by synaptic stimulation (MacVicar and Hochman, 1991). Swelling by cultured glial cells can be blocked by furosemide (Walz and Hertz, 1984; Kempski et al., 1991). Furosemide has been reported to reduce the duration of normoxic SD induced by KCl application in cat neocortex in vivo (Read et al., 1997). However, in the present study the initiation, propagation and post-AD damage were each furosemide-insensitive, so a glial contribution to the swelling signal was not demonstrated.

Slight reductions in temperature can lessen neuronal damage associated with ischemia (Chen et al., 1993; Dietrich et al., 1997). In the present study, OGD-induced AD persisted in neocortical slices maintained at a lowered temperature of 32°C but not at 22°C, probably because metabolic demand was reduced enough to avoid AD induction. We suggest that reduced temperature can be neuroprotective because recurrent SD in the penumbra is suppressed. Hypothermia reduces the propensity of cortical tissue to propagate SD in the rat (Takaoka et al., 1996) and in rat hippocampal slices (Obeidat et al., 2000). Lowered temperature suppresses AD and reduces infarct size following middle cerebral artery occlusion (Chen et al., 1993; Colbourne et al., 1997; Corbett et al., 2000).

In the present study TTX did not block AD, as also found by Taylor et al. (Taylor et al., 1999) in half of their slices exposed to OGD. In the other half, it was delayed. We found that treatment with glutamate antagonists (kynurenate or AP-5/CNQX) had no effect. Combining TTX and kynurenate delayed AD onset by 8 min, so blockade of both glutamate receptors and voltage-sensitive Na+ channels can delay, but not prevent, the onset of AD. This supports findings by Aitken et al. (Aitken et al., 1988). Other slice studies show that NMDA receptor antagonists are ineffective in blocking hypoxic SD (Jing et al., 1993). Clearly, an important reason why acute stroke damage is difficult to prevent (other than by restoring blood supply or lowering temperature) is because the AD and recurrent SD are so resistant to pharmacological blockade. However, we have recently found that sigma receptor ligands such as dextromethorphan block ‘ischemic’ SD (i.e. AD) or normoxic SD in neocortical slice preparations (Anderson et al., 2000). The findings presented here suggest that an ideal stroke treatment would uncouple the AD from ischemia yet be clinically tolerable. Sigma receptor ligands may prove useful in this regard.

To conclude, this study indicates that AD is an important contributor to neuronal damage immediately following the onset of ischemia. Glutamate receptor antagonists appear to be of little benefit in vivo or in cortical slices during and following the initial ischemic period. The inhibition of AD or recurrent SD in the penumbra could prove to be an effective therapeutic strategy in the treatment of stroke if clinically tolerable drugs could be taken prophylactically or introduced during the 3 h period following stroke when peri-infarct depolarizations recur.

Notes

This work was supported by the Heart and Stroke Foundation of Ontario (Grant B-4003).

Address correspondence to R. David Andrew, Department of Anatomy and Cell Biology, Queen's University, Kingston, Ontario, Canada K7L 3N6. Email: andrewd@post.queensu.ca.

Figure 1.

(A) O2/glucose deprivation (OGD) between 0:00 and 10:00 min (37.5°C), induces anoxic depolarization (AD) in the coronal neocortical slice. AD is imaged as a propagating front of elevated light transmittance (LT) across all six layers of gray matter (blue/yellow pseudocoloring). A decrease in LT (magenta pseudocoloring) follows only in the wake of AD. Black and white images are taken before and after the experiment under identical illumination. Temporal changes in LT within regions of interest across the neocortex (a–d) and within the different cortical layers (I–VI) are plotted in Figure 2A and B. (B) Exposure to 100 μM ouabain from time 0:00 to 2:00 min (37.5°C) induces AD indistinguishable from that produced by OGD (A), but with earlier onset. White matter (wm) is outlined by dotted lines.

(A) O2/glucose deprivation (OGD) between 0:00 and 10:00 min (37.5°C), induces anoxic depolarization (AD) in the coronal neocortical slice. AD is imaged as a propagating front of elevated light transmittance (LT) across all six layers of gray matter (blue/yellow pseudocoloring). A decrease in LT (magenta pseudocoloring) follows only in the wake of AD. Black and white images are taken before and after the experiment under identical illumination. Temporal changes in LT within regions of interest across the neocortex (a–d) and within the different cortical layers (I–VI) are plotted in Figure 2A and B. (B) Exposure to 100 μM ouabain from time 0:00 to 2:00 min (37.5°C) induces AD indistinguishable from that produced by OGD (A), but with earlier onset. White matter (wm) is outlined by dotted lines.

Figure 3.

(A) IOS imaging during OGD (10 min at 37°C) reveals a front of elevated LT propagating across the neocortex. A recording electrode in layer II/III (white dot) monitors the extracellular potential evoked by stimulation in adjacent layers V/VI (*). (B) A negative shift in the extracellular potential occurs at the same time as the LT wave sweeps by the recording electrode (white dot) as shown in C. The extracellular potential returns to near baseline within 2–3 min. (D) The evoked population spike recorded prior to OGD (stim. 1) is abolished during OGD and does not recover after a 20 min wash in regular aCSF (stim. 2).

Figure 3.

(A) IOS imaging during OGD (10 min at 37°C) reveals a front of elevated LT propagating across the neocortex. A recording electrode in layer II/III (white dot) monitors the extracellular potential evoked by stimulation in adjacent layers V/VI (*). (B) A negative shift in the extracellular potential occurs at the same time as the LT wave sweeps by the recording electrode (white dot) as shown in C. The extracellular potential returns to near baseline within 2–3 min. (D) The evoked population spike recorded prior to OGD (stim. 1) is abolished during OGD and does not recover after a 20 min wash in regular aCSF (stim. 2).

Figure 4.

Paraffin sections stained with hematoxylin and eosin comparing (A) control neocortex maintained at 37°C displaying no AD, with (B) O2/glucose deprived neocortex (10 min at 37.5°C) that supported the AD. In B there is nuclear and cytoplasmic swelling of neurons compared to control. Neuronal processes are evident among the cells in control tissue, running parallel to the cortical layers (arrowheads in A). They are not obvious in O2/glucose deprived tissue which displays a mottled appearance.

Figure 4.

Paraffin sections stained with hematoxylin and eosin comparing (A) control neocortex maintained at 37°C displaying no AD, with (B) O2/glucose deprived neocortex (10 min at 37.5°C) that supported the AD. In B there is nuclear and cytoplasmic swelling of neurons compared to control. Neuronal processes are evident among the cells in control tissue, running parallel to the cortical layers (arrowheads in A). They are not obvious in O2/glucose deprived tissue which displays a mottled appearance.

Figure 5.

OGD-induced anoxic depolarization in neocortical gray is not altered by glutamate receptor antagonists. (A) Bath application of 2 mM kynurenate 15 min before, during and after OGD does not affect AD onset or propagation, nor is post-AD damage (magenta pseudocoloring) avoided. (B) AP-5 (25 μM) and CNQX (10 μM) applied 15 min before, during and after OGD does not affect AD generation or migration across neocortex and there is no post-SD damage (magenta). Likewise, AD which initiates later in the CA1 region (arrow, 7:22) is not prevented. Note that subsequent damage in the CA1 region in B (14:09) is not seen in the slice of A (15:00) where hippocampal AD does not arise.

Figure 5.

OGD-induced anoxic depolarization in neocortical gray is not altered by glutamate receptor antagonists. (A) Bath application of 2 mM kynurenate 15 min before, during and after OGD does not affect AD onset or propagation, nor is post-AD damage (magenta pseudocoloring) avoided. (B) AP-5 (25 μM) and CNQX (10 μM) applied 15 min before, during and after OGD does not affect AD generation or migration across neocortex and there is no post-SD damage (magenta). Likewise, AD which initiates later in the CA1 region (arrow, 7:22) is not prevented. Note that subsequent damage in the CA1 region in B (14:09) is not seen in the slice of A (15:00) where hippocampal AD does not arise.

Figure 6.

Histogram summarizes treatments tested for their effects on (A) the time to onset of OGD-induced AD and (B) the AD propagation rate. AD properties induced by OGD alone (black bars) are compared with various treatments in addition to OGD (shaded bars). Statistically significant differences (P < 0.005) are indicated by an asterisk. Lowering the temperature to 22°C was the only treatment that prevented the AD.

Figure 6.

Histogram summarizes treatments tested for their effects on (A) the time to onset of OGD-induced AD and (B) the AD propagation rate. AD properties induced by OGD alone (black bars) are compared with various treatments in addition to OGD (shaded bars). Statistically significant differences (P < 0.005) are indicated by an asterisk. Lowering the temperature to 22°C was the only treatment that prevented the AD.

Figure 7.

Response to bath application of 100 μM NMDA. (A) IOS imaging reveals a generalized increase in LT that develops in all gray matter regions simultaneously starting at 3:20. This is followed by a sudden reversal in LT starting at around 3:35. (B) Normalized LT changes from two areas of interest show the response to NMDA over time. Note that the signals rise and fall at the same time, unlike the propagating AD in Figure 2A. (C) The sudden reversal in the optical signal at the recording site (rec) is coincident with a negative shift in the extracellular potential.

Response to bath application of 100 μM NMDA. (A) IOS imaging reveals a generalized increase in LT that develops in all gray matter regions simultaneously starting at 3:20. This is followed by a sudden reversal in LT starting at around 3:35. (B) Normalized LT changes from two areas of interest show the response to NMDA over time. Note that the signals rise and fall at the same time, unlike the propagating AD in Figure 2A. (C) The sudden reversal in the optical signal at the recording site (rec) is coincident with a negative shift in the extracellular potential.

References

Aitken PG, Balestrino M, Somjen GG (
1988
) NMDA antagonists: lack of protective effects against hypoxic damage in CA1 region of hippocampal slices.
Neurosci Lett
 
89
:
187
–192.
Aitken PG, Tombaugh GC, Turner DA, Somjen GG (
1998
) Similar propagation of SD and hypoxic SD-like depolarization in rat hippocampus recorded optically and electrically.
J Neurophysiol
 
80
:
1514
–1521.
Aitken PG, Fayuk D, Somjen GG, Turner DA (
1999
) Use of intrinsic optical signals to monitor physiological changes in brain tissue slices.
Methods: A Companion to Methods in Enzymology
 
18
:
91
–103.
Anderson, TA, Jarvis CR, Andrew RD (
1999
) Imaging repetitive spreading depression in submerged neocortical slices.
Soc Neurosci Abstr
 
25
:
740
.
Anderson TA, Biedermann AJ, Andrew RD (
2000
) Sigma receptor ligands block spreading depression in rat neocortical slices.
Soc Neurosci Abstr
 
26
:
282
.15.
Andrew RD, MacVicar BA (
1994
) Imaging cell volume changes and neuronal excitation in the hippocampal slice.
Neuroscience
 
62
:
371
–383.
Andrew RD, Adams JR, Polischuk TM (
1996
) Imaging NMDA- and kainate- induced intrinsic optical signals from hippocampal slice.
J Neurophysiol
 
76
:
2707
–2717.
Andrew RD, Lobinowich ME, Osehobo EP (
1997
) Evidence against volume regulation by cortical brain cells during acute osmotic stress.
Exp Neurol
 
142
:
300
–312.
Andrew RD, Jarvis CR, Obeidat AS (
1999
) Potential sources of intrinsic optical signals imaged in live brain slices.
Methods: A Companion to Methods in Enzymology
 
18
:
185
–196.
Back T, Ginsberg MD, Dietrich WD, Watson BD (
1996
) Induction of spreading depression in the ischemic hemisphere following experimental middle cerebral artery occlusion: effect on infarct morphology.
J Cereb Blood Flow Metab
 
16
:
202
–213.
Balestrino M (1995) Studies on anoxic depolarization. In: Brain slices in basic and clinical research (Shurr A, Rigor BM, eds), pp. 273–293. London: CRC Press.
Balestrino M, Young J, Aitken P (
1999
) Block of (Na+, K+) ATPase with ouabain induces spreading depression-like depolarization in hippocampal slices.
Brain Res
 
838
:
37
–44.
Basarsky TA, Duffy SN, Andrew RD, MacVicar BA (
1998
) Imaging spreading depression and associated intracellular calcium waves in brain slices.
J Neurosci
 
18
:
7189
–7199.
Chen Q, Chopp M, Bodzin G, Chen H (
1993
) Temperature modulation of cerebral depolarization during focal cerebral ischemia in rats: correlation with ischemic injury.
J Cereb Blood Flow Metab
 
13
:
389
–394.
Colbourne F, Sutherland G, Corbett D (
1997
) Postischemic hypothermia: a critical appraisal with implications for clinical treatment.
Mol Neurobiol
 
14
:
171
–201.
Corbett D, Nurse S (
1998
) The problem of assessing effective neuroprotection in experimental cerebral ischemia.
Prog Neurobiol
 
54
:
531
–548.
Corbett D, Hamilton M, Colbourne F (
2000
) Persistent neuroprotection with prolonged postischemic hypothermia in adult rats subjected to transient middle cerebral artery acclusion.
Exp Neurol
 
163
:
200
–206.
Dietrich WD, Busto R, Globus MYT, Ginsberg, MD (
1997
) Brain damage and temperature: cellular and molecular mechanisms.
Adv Neurol
 
71
:
177
–197.
Footit DR, Newberry NR (
1998
) Cortical spreading depression induces an LTP-like effect in rat neocortex in vivo.
Brain Res
 
781
:
339
–342.
Furlan M, Marchal G, Viader F, Derlon JM, Baron JC (
1996
) Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra.
Ann Neurol
 
40
:
216
–226.
Hansen AJ (
1985
) Effect of anoxia on ion distribution in the brain.
Physiol Rev
 
65
:
101
–148.
Hernandez-Caceres J, Maclas-Gonzales K, Brozek G, Bures J (
1987
) Systemic ketamine blocks cortical spreading depression but does not delay the onset of terminal anoxic depression.
Brain Res
 
437
:
360
–364.
Hori N, Carpenter DO (
1994
) Functional and morphological changes induced by transient in vivo ischemia.
Exp Neurol
 
129
:
279
–289.
Hossmann KA (
1994
) Glutamate-mediated injury in focal cerebral ischemia: the excitotoxic hypothesis revised.
Brain Pathol
 
4
:
23
–26.
Hossmann KA (
1996
) Periinfarct depolarizations.
Cerebrovasc Brain Metab Rev
 
8
:
195
–208.
Irwin A, Walz W (1999) Spreading depression waves as mediators of secondary Injury and protective mechanisms. In Cerebral ischemia: molecular and cellular pathophysiology (Walz W, ed.). Totow, NJ: Humana Press.
Jarvis CR, Andrew RD (
1998
) Spreading depression mediates acute neuronal damage induced by oxygen/glucose deprivation in neocortical slices.
Soc Neurosci Abstr
 
24
:
982
.
Jarvis CR, Lilge L, Vipond GJ, Andrew RD (
1999
) Interpretation of intrinsic optical signals and calcein fluorescence during acute excitotoxic insult in the hippocampal slice.
NeuroImage
 
10
:
357
–372.
Jing J, Aitken PG, Somjen GG (
1993
) Role of calcium channels in spreading depression in rat hippocampal slices.
Brain Res
 
604
:
251
–259.
Kempski O, von Rosen S, Weigt H, Staub F, Peters J, Baethman A (
1991
) Glial ion transport and volume control.
Ann NY Acad Sci
 
633
:
306
–317.
Kreisman NR, Soliman S, Gozal D (
2000
) Regional differences in hypoxic depolarization and swelling in hippocampal slices.
J Neurophysiol
 
83
:
1031
–1038.
Lauritzen M (
1987
) Cerebral blood flow in migraine and cortical spreading depression.
Acta Neurol Scand Suppl
 
76
:
9
–40.
Lauritzen M, Hansen AJ (
1992
) The effect of glutamate receptor blockade on anoxic depolarisation and cortical spreading depression.
J Cereb Blood Flow Metab
 
12
:
223
–229.
Leao APP (
1944
) Spreading depression of activity in the cerebral cortex.
J Neurophysiol
 
7
:
359
–390.
Leao APP (
1947
) Further observations on the spreading depression of activity in the cerebral cortex.
J Neurophysiol
 
10
:
409
–414.
MacVicar BA, Hochman D (
1991
) Imaging of synaptically evoked intrinsic optical signals in hippocampal slices.
J Neurosci
 
11
:
1458
–1469.
Marrannes R, Willems R, DePrins E (
1988
) Evidence for a role of the NMDA receptor in cortical spreading depression in the rat.
Brain Res
 
457
:
226
–240.
McDonald JR, Johnston MV (
1990
) Physiological and pathophysiological roles of excitatory amino acids during central neurons system development.
Brain Res Rev
 
15
:
41
–70.
Mies G, Kohno K, Hossmann KA (
1994
) Prevention of periinfarct direct current shifts with glutamate antagonist NBQX following occlusion of the middle cerebral artery in the rat.
J Cereb Blood Flow Metab
 
14
:
802
–807.
Nedergaard M (
1996
) Spreading depression as a contributor to ischemic brain damage.
Adv Neurobiol
 
71
:
75
–84.
Nedergaard M, Hansen AJ (
1988
) Spreading depression is not associated with neuronal injury in the normal brain.
Brain Res
 
449
:
395
–398.
Nedergaard M, Hansen AJ (
1993
) Characterization of cortical depolarizations evoked in focal cerebral ischemia.
J Cereb Blood Flow Metab
 
13
:
568
–574.
Nellgard B, Wieloch T (
1992
) NMDA-receptor blockers but not NBQX, an AMPA-receptor antagonist, inhibit spreading depression in the rat brain.
Acta Physiol Scand
 
146
:
497
–503.
Nicholson C, Kraig RP (1981) The behavior of extracellular ions during spreading depression. In: The application of ion-selective electrodes (Zeuthen T, ed.). Amsterdam: Elsevier.
Obeidat, A, Andrew RD (
1998
) Spreading depression determines acute cellular damage in the hippocampal slice during oxygen/glucose deprivation.
Eur J Neurosci
 
10
:
3451
–3461.
Obeidat A, Jarvis CR, Andrew RD (
2000
) Glutamate does not mediate acute neuronal damage following spreading depression induced by O2/glucose deprivation in the hippocampal slice.
J Cereb Blood Flow Metab
 
20
:
412
–422.
Obrenovitch TP (
1995
) The ischemic penumbra: twenty years on.
Cerebrovasc Brain Metab Rev
 
7
:
297
–323.
Obrenovitch TP (
1999
) High extracellular glutamate and neuronal death in neurological disorders: cause, contribution, or consequence.
Ann NY Acad Sci
 
890
:
273
–286.
Obrenovitch TP, Urenjak J (
1997
) Altered glutamatergic transmission in neurological disorders: from high extracellular glutamate to excessive synaptic efficacy.
Prog Neurobiol
 
51
:
39
–87.
Obrenovitch TP, Urenjak J (
1997
) Is high extracellular glutamate the key to excitotoxicity in traumatic brain injury?
J Neurotrauma
 
14
:
677
–697.
Park JS, Bateman MC, Goldberg MP (
1996
) Rapid alterations in dendrite morphology during sublethal hypoxia or glutamate receptor activation.
Neurobiol Disease
 
3
:
215
–227.
Polischuk TM, Jarvis CR, Andrew RD (
1998
) Intrinsic optical signaling denoting neuronal damage in response to acute excitotoxic insult in the hippocampal slice.
Neurobiol Disease
 
4
:
423
–437.
Read SJ, Smith MI, Benham CD, Hunter AJ, Parsons AA (
1997
) Furosemide inhibits regenerative cortical spreading depression in anaesthetized cats.
Cephalalgia
 
17
:
826
–832.
Rossi DJ, Oshima T, Attwell D (
1999
) Glutamate release in severe brain ischemia is mainly by reversed uptake.
Nature
 
403
:
316
–321.
Strong AJ, Smith SE, Whittington DJ, Meldrum, BS, Parsons AA, Krupinski J, Hunter AJ, Patel S (
1999
) Factors influencing the frequency of fluorescence transients as markers of peri-infarct depolarizations in focal cerebral ischemia.
Stroke
 
31
:
214
–221.
Somjen GG, Aitken PG, Czeh GL, Herreras O, Jing J, Young JN (
1992
) Mechanisms of spreading depression: a review of recent findings and a hypothesis.
Can J Physiol Pharmacol
 
70
:
S248
–S254.
Takano K, Latour LL, Formato JE, Carano RA, Helmer KG, Hasegawa Y, Sotak CH, Fisher M (
1996
) The role of spreading depression in focal ischemia evaluated by diffusion mapping.
Ann Neurol
 
39
:
308
–318.
Takaoka S, Pearlstein RD, Warner DS (
1996
) Hypothermia reduces the propensity of cortical tissue to propagate direct current depolarization in the rat.
Neurosci Lett
 
218
:
25
–28.
Taylor CP, Weber ML, Gaughan CL, Lehning EJ, LoPachin RM (
1999
) Oxygen/glucose deprivation in hippocampal slices: altered intraneuronal elemental composition predicts structural and functional damage.
J Neurosci
 
19
:
619
–629.
Verity MA (
1991
) Use and abuse of tissue cultures in neurotoxicity studies.
Neurotoxicology
 
12
:
457
–459.
Walz W (
1987
) Swelling and K+ uptake in cultured astrocytes.
Can J Physiol Pharmacol
 
65
:
1051
–1057.
Walz W, Hertz L (
1984
) Intense furosemide-sensitive potassium accumulation in astrocytes in the presence of pathologically high extracellular potassium levels.
J Cereb Blood Flow Metab
 
4
:
301
–304.
Weber ML, Taylor CP (
1994
) Damage from oxygen and glucose deprivation in hippocampal slices is prevented by tetrodotoxin, lidocaine and phenytoin without blockade of action potentials.
Brain Res
 
664
:
167
–177.