Chronically Epileptic Human and Rat Neocortex Display a Similar Resistance Against Spreading Depolarization In Vitro
Background and Purpose—Experimental and clinical evidence suggests that prolonged spreading depolarizations (SDs) are a promising target for therapeutic intervention in stroke because they recruit tissue at risk into necrosis by protracted intracellular calcium surge and massive glutamate release. Unfortunately, unlike SDs in healthy tissue, they are resistant to drugs such as N-methyl-d-aspartate-receptor antagonists. This drug resistance of SD in low perfusion areas may be due to the gradual rise of extracellular potassium before SD onset. Brain slices from patients undergoing surgery for intractable epilepsy allow for screening of drugs, targeting pharmacoresistant SDs under elevated potassium in human tissue. However, network changes associated with epilepsy may interfere with tissue susceptibility to SD. This could distort the results of pharmacological tests.
Methods—We investigated the threshold for SD, induced by a gradual rise of potassium, in neocortex slices of patients with intractable epilepsy and of chronically epileptic rats as well as age-matched and younger control rats using combined extracellular potassium/field recordings and intrinsic optical imaging.
Results—Both age and epilepsy significantly increased the potassium threshold, which was similarly high in epileptic rat and human slices (23.6±2.4 mmol/L versus 22.3±2.8 mmol/L).
Conclusions—Our results suggest that chronic epilepsy confers resistance against SD. This should be considered when human tissue is used for screening of neuroprotective drugs. The finding of similar potassium thresholds for SD in epileptic human and rat neocortex challenges previous speculations that the resistance of the human brain against SD is markedly higher than that of the rodent brain.
Several hundred Phase II and III clinical trials on presumed neuroprotective agents for stroke and brain trauma have failed in the last decades (see Washington University Internet Stroke Center, www.strokecenter.org). Obviously, novel approaches to the translation from “bench” to “bedside” have to be adopted. A new roadmap for neuroprotection was suggested by Donnan in 2007,1 the first 4 steps being: (1) better proof of efficacy in animal models; (2) in vitro efficacy in human tissue; (3) in vivo studies of the distribution of neuroprotectants in the human brain; and (4) efficacy in novel human models of acute neuronal injury to prove/disprove the concept. Only after Step 4, novel neuroprotective strategies should enter a Phase II clinical trial. We focused on the second step: “in vitro efficacy in human tissue.” Brain slices from patients with pharmacoresistant epilepsy provide the most physiological preparation of human tissue currently available for studying neuroprotectants in vitro. Unlike human cell cultures, they allow investigating neuroprotective effects on highly complex networks such as neocortex.
In neocortex, the term spreading depolarization (SD) describes abrupt near-complete breakdown of the electrochemical gradients across the neuronal membranes, which leads biochemically and morphologically to the cytotoxic edema.2 SD is observed extracellularly as a large negative slow field potential shift. Experimental evidence suggests that prolonged SDs recruit tissue at risk into necrosis by massive glutamate release and protracted rise of intracellular calcium.2 In the clinic, clusters of SDs were found in patients with delayed ischemic stroke after subarachnoid hemorrhage and in the course of malignant hemispheric stroke.3,4 Preliminary clinical findings suggest that very prolonged SDs are associated with lesion progression in a similar fashion to that in animals.5
The cooperation of several ion channels mediates sodium and calcium inward fluxes that lead up to SD.6,7 Whereas N-methyl-d-aspartate-receptor antagonists can prevent SD in healthy brain tissue,8 their potency to suppress SD is greatly reduced under hypoxia, sodium pump inhibition, or artificial increase of extracellular potassium ([K+]o).9,10 Preliminary evidence suggests that such differences in susceptibility to N-methyl-d-aspartate-receptor antagonists also apply to SDs in the human brain.11,12
Ischemia causes a gradual rise in [K+]o before SD onset, presumably due to decline of sodium pump activity and activation of adenosine 5′-triphosphate- and calcium-sensitive potassium channels.13 This gradual rise of baseline [K+]o could be responsible, at least partially, for drug resistance of SDs under energy depletion.9 Hence, it has been suggested that the high potassium model in human brain slices from patients with intractable epilepsy allows for drug screening targeting drug-resistant SDs in human tissue. However, network changes associated with epilepsy may influence SD susceptibility. Therefore, we investigated effects of age and chronic epilepsy on the potassium threshold for SD in rat neocortical slices and compared the thresholds between chronically epileptic human and rat neocortex. We also investigated whether the effect of epilepsy could be due to changes of GABAergic mechanisms in epileptic tissue.
Materials and Methods
The study was approved by the local ethics committee. Written informed consent was obtained from each patient. For detailed description of the methods, see the online Supplemental Methods section (http://stroke.ahajournals.org).
Bicuculline methiodide was purchased from Sigma.
Chronic epilepsy was triggered by status epilepticus at 12 to 14 weeks due to intraperitoneal injection of pilocarpine as described previously (see online supplemental section).14
Neocortical slices (500 μm) from human temporal or frontal lobe resectates (Supplemental Table I) and horizontal temporohippocampal slices (400 μm thick) from male Wistar rats (Charles River Laboratories, Sulzfeld, Germany; 10-month-old chronically epileptic rats, age-matched controls, and 8-week-old young controls) were prepared as described previously.9 The slices were stored in a humidified, carbogenated interface-type chamber perfused with prewarmed (36°C) artificial cerebrospinal fluid (aCSF) containing (in mM) 129 NaCl, 3 KCl, 1.8 MgSO4, 1.6 CaCl2, 1.25 NaH2PO4, 21 NaHCO3, 10 glucose, pH 7.4, osmolality 303 mOsm/kg.
Extracellular Recordings and Stimulation
Combined field potential/potassium-sensitive microelectrodes were connected to a custom-made differential amplifier and recorded the direct current (DC) shift and [K+]o in layer III of human neocortex or temporal and ectorhinal rat neocortex. Signals were filtered at 1 kHz (field potential) and 3 Hz ([K+]o), then sampled at a rate of 5 kHz and 10 Hz, respectively, by a CED 1401 (Cambridge Electronic Design Limited, Cambridge, UK) and stored on a PC. Slice viability was tested by recording responses to orthodromic bipolar stimulation in layer V. Single or paired stimuli (0.1 ms, 1 to 10 V, 50-ms interval) were delivered using a stimulus isolator in constant voltage mode (ISO Flex; AMPI Instruments, Jerusalem, Israel) controlled by a Master-8 (AMPI Instruments). Only slices showing maximal amplitudes of the population spike ≥1 mV were used.
Intrinsic Optical Imaging
Intrinsic optical signals were obtained with a CCD camera at the same time as transilluminating slices with white light to record SD propagation.9
Data analysis was performed with Spike2 (Version 6; Cambridge Electronic Design Limited, Cambridge, UK) and MATLAB. Data are given as mean value±SD. The groups were compared using 1-way analysis of variance with Fisher least significant difference post hoc test unless otherwise stated. P≤0.05 was considered statistically significant.
Potassium Threshold for SD
The SD threshold in response to stepwise rises of the potassium concentration in the aCSF ([K+]aCSF) was compared among brain slices of chronically epileptic human neocortex (EpiHum, n=10), chronically epileptic rats (EpiRats, n=7), age-matched (old) control rats (OCoRats, n=9), and young control rats (YCoRats, n=8). [K+]aCSF was increased initially to 17.5 mmol/L and increased further by 2.5 mmol/L every 30 minutes until the first SD occurred. The [K+]aCSF threshold for SD was significantly lower in the YCoRats compared with those of the other groups (Table). Moreover, OCoRats showed a significantly lower threshold than EpiRats and the thresholds were similar in neocortex of EpiRat and EpiHum tissue. Linear regression of the pooled data of all groups demonstrated that the time until the first SD correlated significantly with [K+]o as measured in the tissue immediately before the first SD (R=0.508, P=0.004). This indicates a difference in tissue tolerance toward higher levels of [K+]o without SD generation rather than different rates of potassium buffering between the groups. Peak-to-peak amplitude of the DC shift was significantly larger in the EpiHum group, whereas it was significantly smaller in EpiRat neocortex (Table). SD propagation velocity in YCoRats was significantly higher than that in the other groups (Table; Figures 1 and 2). No significant differences in peak concentrations of [K+]o and SD duration (at 50% of the maximal DC amplitude) were observed.
SD Frequency Under Elevated Baseline [K+]aCSF
Perfusion with 25 mmol/L [K+]aCSF for 45 minutes elicited multiple SDs in all groups. Average SD incidence was significantly higher in YCoRats than in the other groups. SD incidence in OCoRats was higher than in the epileptic groups: YCoRats: 14.4±5.0 SDs (n=8); OCoRats: 8.3±4.1 SDs (n=8); EpiRats: 4.0±2.1 SDs (n=6); EpiHum: 2.0±1.5 SDs (n=6); YCoRats versus EpiRats and EpiHum: P<0.001; YCoRats versus OCoRats: P=0.003; OCoRats versus EpiRats and EpiHum: P=0.044 and P=0.005, respectively. Incidence in the 2 epileptic groups was similar. SDs often merged in clusters of OCoRats and YCoRats as reported previously (Figure 3A; 5 of 8 YCoRats and 5 of 9 OCoRats).15 This phenomenon was not observed in the epileptic groups (Figure 3B). The DC shift was significantly smaller in EpiRats compared with the other groups: ±9.0±3.2 mV, P<0.05; YCoRats: −13.4±3.1 mV; OCoRats: −14.5±2.7 mV; EpiHum: −13.6±3.5 mV. Differences in SD velocities and DC durations did not reach statistical significance between groups.
aCSF containing bicuculline (50 μmol/L) induced spontaneous SDs in 4 of 7 YCoRats and 4 of 9 OCoRats. Moreover, interictal-like activity and seizure-like events were triggered in 5 and 1 YCoRats, respectively, and 5 and 2 OCoRats, respectively. Neither SDs nor epileptiform activities were induced by bicuculline in the epileptic tissues (n=6 EpiRats, n=5 EpiHum). Artificial rise of [K+]aCSF always terminated epileptiform activity in the control groups. In presence of bicuculline, SD number per 45 minutes significantly increased in slices of OCoRats (8.3±4.1 versus 13.4±4.3 SDs [n=8], t test, P=0.03) as well as EpiRats (4.0±2.1 versus 7.7±2.0 SDs [n=6], P=0.011) and EpiHum tissue (2.0±1.5 versus 4.4±0.9 SDs [n=5], P=0.014). There was no significant effect in YCoRats (14.4±5.0 versus 15.7±4.9 SDs [n=7]). Bicuculline did not affect SD duration and DC amplitude. Only in YCoRats, SD propagation velocity was significantly faster in presence of bicuculline (11.8±4.6 mm/min versus 6.0±1.5, P=0.015).
In YCoRats, we possibly missed the enhancing effect of bicuculline on SD frequency because the SD frequency was already very high at 25 mmol/L [K+]aCSF even in the absence of bicuculline. Therefore, we tested whether bicuculline would influence the potassium threshold for SDs in YCoRats starting with [K+]aCSF at 12.5 mmol/L and increasing it by 2.5 mmol/L every 30 minutes. Bicuculline lowered significantly the threshold from 17.8±2.1 to 14.4±3.1 mmol/L (n=8, paired t test, P=0.01). We then tested thresholds in presence of bicuculline in 8 patients to verify that bicuculline would also lower the potassium threshold in epileptic human neocortex. Consistently, we found a significant decrease from 22.5±3.0 to 19.7±1.6 mmol/L (n=8, paired t test, P=0.015).
In the current study, we investigated 2 different measures of tissue susceptibility to SD in vitro: (1) the potassium threshold for SD; and (2) the SD frequency under elevated [K+]aCSF. The latter approach was also used previously in vivo.16 Both age and epilepsy increased resistance against SD in rats; resistance was not different between chronically epileptic rat and human neocortex. Slices from young control rats showed faster SD propagation similar to a previous study in gerbils.17 Bicuculline increased SD susceptibility in all groups and propagation velocity in young control rats.
The terms epileptiform activity and SD describe the 2 fundamental spectra of pathological network disturbances in the cortex.18 Whereas epileptic ictaform activity is characterized by a sustained shift of the membrane potential to a level above activation but below inactivation threshold for the action potential-generating channels, the sustained depolarization of SD is above the inactivation threshold.19 Therefore, epileptic ictaform activity is characterized by highly frequent, synchronous firing of neurons, whereas SD is characterized by depression of firing.19
Similar factors can induce experimentally SD and epileptiform activity, for example, high potassium, low magnesium, or bicuculline.18 Consistently, SD co-occurs typically in models for acute status epilepticus.20,21 Similar observations also exist in patients with acute neuronal injury in whom SDs were significantly more abundant than epileptic activities, but SDs always co-occurred with epileptic ictaform activity whenever epileptic ictaform activity was detected.22 In contrast, resistance against SD seems to increase in models for chronic epilepsy such as the pentylenetetrazole model or that of prolonged blood brain–barrier disruption in rats.23,24 Thus, the present finding of increased resistance against SD in chronically epileptic rats after pilocarpine-induced status epilepticus is consistent with previous observations in other models. This secondary effect of epileptic network changes could be related to alterations in potassium buffering: ictaform activity is associated with moderate rise of [K+]o to the ceiling level of approximately 12 mmol/L. Repeated stimulus-induced rises of [K+]o to this level seem to boost up mechanisms for potassium reuptake,25 which should counteract SD ignition.18 However, our findings suggest that age and epilepsy increased the tissue's ability to tolerate higher [K+]o rather than buffer it more efficiently.
By inhibiting GABAA receptors with bicuculline, we investigated whether altered GABAergic transmission is responsible for differences in SD threshold. GABAA receptors are ligand-gated ion channels concerned mainly with passing of chloride ions across the cell membrane. Moreover, bicuculline inhibits small-conductance calcium-activated potassium channels (SK channels), which mediate slow afterhyperpolarization after the action potential in many neurons.26 Above 5 μmol/L, bicuculline induces epileptiform activity in healthy neocortex.27 Such epileptiform activity was inhibited by the high potassium medium in the present study. Excess [K+]o forces chloride uptake and this may decrease the proepileptic effect of inhibiting GABAA receptors, which are dependent on the chloride equilibrium potential.28
A role of GABAergic transmission for SD ignition was suggested previously.27 Those authors observed that subepileptic doses of bicuculline triggered SD in neocortical slices from healthy rats in contrast to slices from rats and patients with epilepsy in a similar fashion to the present study. Experimental evidence suggests that GABAergic transmission is disturbed in chronically epileptic human tissue.29 However, in the present study, bicuculline increased the susceptibility to SD under high [K+]aCSF in a similar fashion in all groups. This suggests that differences in susceptibility to SD between groups were not mediated by altered GABAergic transmission or by differences in SK channel function. Further studies should determine why chronic epilepsy increases resistance against SD and whether this protects against ischemic damage. To our knowledge, this is unknown but experimental evidence suggests that status epilepticus early in life in fact modifies infarct volume later in life. However, this effect seems complex because it depended on the inducer of status epilepticus.30
It has been proposed that SDs are a target for therapeutic intervention in the ischemic penumbra because their blockade may inhibit lesion progression.31 Previous studies reported that the resistance against SD is significantly higher in human epileptic tissue compared with that of healthy rats.9,32 We confirmed this here. Nevertheless, this was often related to species differences rather than structural and functional alterations of epileptic tissues because earlier reports had suggested that the gyrencephalic cortex of cats and monkeys is more resistant against SD than the lissencephalic cortex of rodents.33 However, we found that rat and human chronically epileptic tissues were similarly resistant against SD. Consistently, recent clinical trials suggested that patients with hemorrhagic and ischemic stroke or brain trauma display SDs in vivo in a similar fashion to rodents.3,4,34 Thus, species differences may be overestimated.35 The present data indicate that epileptic network changes could be more important than species differences for the SD threshold, and effects of neuroprotectants could be overestimated in epileptic tissue. It is unlikely that the results would have been different when the patients had not been pretreated with antiepileptic drugs because experimental evidence suggests that antiepileptic drugs further increase the SD threshold.16
In conclusion, we recommend that results gained in human epileptic tissue are compared with those in both nonepileptic and epileptic tissues of old rodents. With this modification, preclinical studies according to step 2 of Donnan's new roadmap for acute neuroprotection in stroke should reach higher validity.1
Sources of Funding
Supported by grants from the Deutsche Forschungsgemeinschaft (DFG SFB Tr3 D10), the Bundesministerium für Bildung und Forschung (Center for Stroke Research Berlin, 01 EO 0801), the Bernstein Center for Computational Neuroscience Berlin 01GQ1001C B2, and the Kompetenznetz Schlaganfall (J.P.D.).
Costantino Iadecola, MD, was the Guest Editor for this paper.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.621581/-/DC1.
- Received March 28, 2011.
- Revision received April 22, 2011.
- Accepted April 27, 2011.
- © 2011 American Heart Association, Inc.
- Donnan GA
- Dreier JP,
- Woitzik J,
- Fabricius M,
- Bhatia R,
- Major S,
- Drenckhahn C,
- et al
- Oliveira-Ferreira AI,
- Milakara D,
- Alam M,
- Jorks D,
- Major S,
- Hartings JA,
- et al
- Petzold GC,
- Windmüller O,
- Haack S,
- Major S,
- Buchheim K,
- Megow D,
- et al
- Dreier JP,
- Major S,
- Manning A,
- Woitzik J,
- Drenckhahn C,
- Steinbrink J,
- et al
- Sakowitz OW,
- Kiening KL,
- Krajewski KL,
- Sarrafzadeh AS,
- Fabricius M,
- Strong AJ,
- et al
- Erdemli G,
- Xu YZ,
- Krnjevic K
- Kann O,
- Kovács R,
- Njunting M,
- Behrens CJ,
- Otáhal J,
- Lehmann TN,
- et al
- Somjen GG
- Kager H,
- Wadman WJ,
- Somjen GG
- Mody I,
- Lambert JD,
- Heinemann U
- Fabricius M,
- Fuhr S,
- Bhatia R,
- Boutelle M,
- Hashemi P,
- Strong AJ,
- et al