- Email: [email protected]
Neuronal injury in experimental status epilepticus in the rat: role of hypoxia
Neuroscience Letters 222 (1997) 207–209
Neuronal injury in experimental status epilepticus in the rat: role of hypoxia Masahiro Sasahira a, Roger P. Simon a, David A. Greenberg a , b ,* a Department of Neurology, University of Pittsburgh School of Medicine, S-526 Biomedical Science Tower, Pittsburgh, PA 15213, USA Department of Neurobiology, University of Pittsburgh School of Medicine, S-526 Biomedical Science Tower, Pittsburgh, PA 15213, USA
b
Received 30 September 1996; revised version received 8 January 1997; accepted 10 January 1997
Abstract While it seems axiomatic that hypoxia is a risk factor for neuronal death during prolonged seizures, the classic neuropathologic literature does not confirm such an association. We investigated this issue by inducing status epilepticus in normoxic (PaO2 ≈ 100 mmHg) and hypoxic (PaO2 ≈ 50 mmHg) rats, using heat-shock protein (HSP) expression as an index of early cell injury and acid fuchsin staining to detect cell death. Neither stress protein induction nor neuronal death was increased in the selectively vulnerable CA3c region of hippocampus, or in cerebral cortex, of hypoxic compared to normoxic animals. These data support the concept that moderate hypoxia is not a risk factor for brain injury from status epilepticus. 1997 Elsevier Science Ireland Ltd. Keywords: Heat-shock protein; Hypoxia; Status epilepticus; Seizures; Cell death
Most treatment protocols for status epilepticus emphasize immediate attention to issues of oxygenation [7]. However, the experimental and neuropathologic literature that might support a role for hypoxia in neuronal injury during status epilepticus is inconclusive. Maximal oxygenation neither increases nor decreases neuronal injury during experimental status [13,16]. Whether hypoxia in the absence of ischemia is a risk factor is uncertain [5,9]. Blennow et al. showed that in bicuculline-induced status epilepticus, there was less histopathologic change in cortex, hippocampus, and striatum of animals maintained at PaO2 ≈ 50 mmHg than at PaO2 ≈ 100 mmHg [4]. However, the interpretation of these results is limited by their reliance on staining techniques (hematoxylin and eosin, cresyl violet, and Luxol fast blue) that are comparatively insensitive to early neuronal injury. So¨derfeldt et al. used toluidine blue staining and electron microscopy to study the effect of hypoxia in bicuculline-induced status [16] and found that hypoxia ‘appeared to exaggerate the extent of neuronal alteration’, especially mitochondrial swelling. But only cortical neurons were studied, no statistical ana-
* Corresponding author. Tel.: +1 412 6482623; fax: +1 412 6481239; e-mail: [email protected]
lysis was provided, and electrographic seizure activity was not measured. In addition, because animals were sacrificed immediately after 2 h of status in both of these studies, reversible and irreversible neuronal injury could not be differentiated. We revisited this issue using more sensitive measures of neuronal injury (immunocytochemical detection of heatshock protein (HSP) to assess cell injury and acid-fuchsin staining to quantify cell death [3]) in rats sacrificed 72 h after bicuculline-induced status epilepticus with PaO2 maintained at ~50 or ~100 mmHg. Non-fasted male Sprague–Dawley rats (350–470 g) were anesthetized with 4% halothane, then incubated and ventilated with 70% N2O/27% O2/3% halothane. Atropine (0.25 mg/kg) and pancuronium bromide (4 mg/kg) were administered intravenously via a femoral venous catheter. Arterial blood pressure was monitored continuously by polygraphic recording via a femoral arterial catheter. Arterial blood gases were monitored every 10 min during status. The electroencephalogram (EEG) was recorded continuously from biparietal electrodes attached to the skull with jeweler’s screws. The fraction of inspired O2 (FiO2) was regulated to produce PaO2 ≈ 100 mmHg (n = 13) or 50 mmHg (n = 13). Halothane was discontinued before status was induced.
0304-3940/97/$17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0304-3940 (97 )1 3378-X
208
M. Sasahira et al. / Neuroscience Letters 222 (1997) 207–209
Table 1 Physiologic variables prior to and during 1 h of bicuculline-induced status epilepticus in hypoxic (n = 13) and normoxic (n = 13) rats Baseline
Status epilepticus
Group
T (°C)
pHa
PaCO2*
PaO2**
Duration (min)
Tmax (°C)
pHa
PaCO2
PaO2**
Hypoxic Normoxic
37.1 ± 3.0 36.5 ± 1.4
7.49 ± 0.03 7.45 ± 0.08
31.6 ± 2.3 35.0 ± 4.7
53.5 ± 2.9 105.1 ± 6.2
27.2 ± 3.1 29.9 ± 5.7
37.6 ± 4.9 37.5 ± 3.4
7.37 ± 0.03 7.36 ± 0.08
35.5 ± 1.6 37.7 ± 4.7
49.7 ± 1.9 104.5 ± 5.5
a, Arterial; *P , 0.05, **P , 0.001 (t-test).
Bicuculline was dissolved in 10 N HCl (brought to pH 5.5–5.8 with NaOH) and administered intravenously at 1.2 mg/kg to induce status. Acute pulmonary hypertension and pulmonary edema were prevented by pre-administration of phentolamine (0.1 mg/kg) and by the removal of 4–5 ml of arterial blood, which were then reinfused after the hypertensive surge [4,16]. Bicuculline was re-administered 10 and 20 min after the first injection to prolong status. Electrographic seizure activity was terminated after 1 h by the administration of diazepam (14–15 mg/kg). The EEG patterns were classified into four types, as previously reported [10]: type I, baseline activity; type II, discrete spikes superimposed on a normal baseline; type III, high-voltage spikes and sharp waves separated by at least 1 s of suppressed EEG activity; and type IV, continuous high-voltage, high frequency discharges lasting at least 1 s. The total duration of type IV seizure activity was taken as the duration of status because this electrographic pattern correlates best with histologic brain injury from status epilepticus [10]. After 72 h, animals were sacrificed by intraperitoneal administration of chloral hydrate (400 mg/kg), followed by intracardiac perfusion with 100 ml of normal saline and 400 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were removed and kept in fixative for 24 h at 4°C and 35 and 50 mm sections were then prepared on a vibratome. Neuronal injury was assessed in a blinded manner by counting CA3c hippocampal and cerebral cortical neurons that stained with a monoclonal antibody to the non-constitutive, 72 kDa, HSP72, and neuronal cell death was assessed by counting acid fuchsin-stained neurons [10,14], in predetermined grids on a single section taken at bregma minus 3.3 mm. Within that plane, cortical grids were centered 5.7 mm lateral to the midline and 6.6 mm dorsal to the interaural line. Neither HSP immunoreactivity nor acid-fuchsin staining was observed in control animals not given bicuculline. Data are presented as means ± SD. Statistical analyses were performed using ttests, corrected for multiple comparisons as necessary. Physiologic measures in the normoxic and hypoxic groups are given in Table 1. The only significant differences between the two groups were in PaO2 and, to a less pronounced extent, baseline PaCO2. In particular, there were no differences in temperature or arterial blood pH, each of which has been shown to modify excitotoxic neu-
ronal injury, or in the duration of status. The quantitation of HSP immunoreactivity and acid fuchsin staining of neurons in cortex and hippocampus is shown in Fig. 1. There were no significant differences in HSP72 immunoreactivity or in acid-fuchsin staining between the normoxic and hypoxic groups in either of the brain regions studied. This study shows that cell injury (assessed by stress protein expression) and cell death (measured by acid-fuchsin staining) occur with the same frequency after bicuculline-induced status epilepticus whether animals are normoxic or hypoxic. Induced expression of the stress protein, HSP72, is a reliable and sensitive marker of cell
Fig. 1. Histopathologic sequelae of experimental status epilepticus. (A) HSP72-immunoreactive cells in hippocampal area CA3c (original magnification, 40 × ). (B) Counts of HSP72-immunoreactive and acid fuchsin-staining cells in CA3c sector of hippocampus and in cerebral cortex of normoxic (n = 13) and hypoxic (n = 13) rats. Error bars are SD.
M. Sasahira et al. / Neuroscience Letters 222 (1997) 207–209
injury from a wide range of cerebral insults, including seizures and ischemia [12,14,17] and represents a cellular response to the protein denaturation [2]. Denatured proteins activate heat-shock factors [18], which bind to heatshock elements to initiate the transcription of HSP72 mRNA [19,20]. While HSP72 immunoreactivity is, therefore, a sign of neuronal injury, the eventual fate of these injured neurons is uncertain [8]. In contrast, acid fuchsinstained neurons indicate cell death [3]. Although we examined only a single time point after seizures (72 h), it is unlikely that this obscured cell death occurring at earlier times and no longer apparent owing to phagocytic removal of dead cells. This is because, inter alia, we observed no difference in the total number of (stained plus unstained) cells at 72 h after status between normoxic and hypoxic animals. Our data confirm the proposal of Blennow et al. [5] that hypoxia during status epilepticus is not a risk factor for cell death. In another study, Blennow et al. [4] used conventional light-microscopic stains to identify subacute or chronically injured (pale-staining) cells, but these morphological features may be subtler and more difficult to quantify than HSP72 immunocytochemistry. Our findings are also consistent with data in primates, which showed no relationship between the lowest recorded oxygen concentration in cerebral venous or arterial blood and morphologic brain damage from status [11]. We did not confirm Blennow et al.’s suggestion that hypoxia during status is actually neuroprotective. This may be because our model employs repeated bicuculline injections to produce similar seizure durations in the normoxic and hypoxic groups. In the Blennow studies, status was briefer in the hypoxic animals, and this may have explained the protective effect. An anticonvulsive effect of hypoxia has been demonstrated in awake rats treated with kainic acid [1]. The implications of our findings for the treatment of status epilepticus are unclear. For example, even if moderate hypoxia has no directly deleterious effect on cerebral neurons during status, severe hypoxia leads to centrallyinduced hypotension [6], which could cause ischemic injury to the brain, and status in the setting of ischemic brain injury has a poor prognosis [15]. [1] Amano, S., Obata, T., Hazama, F., Kashiro, N. and Shimada, M., Hypoxia prevents seizures and neuronal damages of the hippocampus induced by kainic acid in rats, Brain Res., 523 (1990) 121–126. [2] Ananthan, J., Goldberg, A. and Voellmy, R., Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat
209
shock genes, Science, 232 (1986) 522–524. [3] Auer, R., Wieloch, T., Olsson, Y. and Siesjo, B., The distribution of hypoglycemic brain damage, Acta Neuropathol. (Berl.), 64 (1984) 177–191. [4] Blennow, G., Brierley, J., Meldrum, B. and Siesjo, B., Epileptic brain damage: the role of systemic factors that modify cerebral energy metabolism, Brain, 101 (1978) 687–700. [5] Blennow, G., Nilsson, B. and Siesjo, B., Influence of reduced oxygen availability on cerebral metabolic changes during bicucullineinduced seizures in rats, J. Cerebral Blood Flow Metab., 5 (1985) 439–445. [6] Cross, C., Rieben, P., Barron, C. and Salisbury, P., Effects of arterial hypoxia on the heart and circulation: an integrative study, Am. J. Physiol., 205 (1963) 963–970. [7] Epilepsy Foundation of America, Treatment of convulsive status epilepticus. Recommendations of the Epilepsy Foundation of America’s working group on status epilepticus, J. Am. Med. Assoc., 270 (1993) 854–859. [8] Gonzalez, M., Shiraishi, K., Hisanaga, K., Sagar, S., Mandabach, M. and Sharp, F., Heat shock proteins as markers of neuronal injury, Mol. Brain Res., 6 (1989) 93–100. [9] Kreisman, N., LaManna, J., Rosenthal, M. and Sick, T., Oxidative metabolic responses with recurrent seizures in rat cerebral cortex: role of systemic factors, Brain Res., 218 (1981) 175–188. [10] Lowenstein, D., Simon, R. and Sharp, F., The pattern of 72-kDa heat shock protein-like immunoreactivity in the rat brain following flurothyl-induced status epilepticus, Brain Res., 531 (1990) 173– 182. [11] Meldrum, B. and Brierley, J., Prolonged epileptic seizures in primates. Ischemic cell change and its relation to ictal physiological events, Arch. Neurol., 28 (1973) 10–17. [12] Shimosaka, S., So, Y. and Simon, R., Distribution of HSP72 induction and neuronal death following limbic seizures, Neurosci. Lett., 138 (1992) 202–206. [13] Siesjo, B., Cell damage in the brain: a speculative synthesis, J. Cerebral Blood Flow Metab., 1 (1981) 155–185. [14] Simon, R., Cho, H., Gwinn, R. and Lowenstein, D., The temporal profile of 72-kDa heat-shock protein expression following global ischemia, J. Neurosci., 11 (1991) 881–889. [15] Snyder, B., Hauser, W., Loewenson, R., Leppik, I., Ramirez, L. and Gumnit, R., Neurologic prognosis after cardiopulmonary arrest: III, Seizure activity, Neurology, 30 (1980) 1292–1297. [16] Soderfeldt, B., Blennow, G., Kalimo, H., Olsson, Y. and Siesjo, B., Influence of systemic factors on experimental epileptic brain injury. Structural changes accompanying bicuculline-induced seizures in rats following manipulations of tissue oxygenation or alpha-tocopherol levels, Acta Neuropathol. (Berl.), 60 (1983) 81–91. [17] Vass, K., Welch, W. and Nowak, T., Localization of 70-kDa stress protein induction in gerbil brain after ischemia, Acta Neuropathol. (Berl.), 77 (1988) 128–135. [18] Wu, C., Wilson, S., Walker, B., David, I., Paisley, T., Zimarino, V. and Ueda, H., Purification and properties of Drosophila heat shock activator protein, Science, 238 (1987) 1247–1253. [19] Xiao, H. and Lis, J., Germline transformation used to define key features of heat-shock response, Science, 239 (1988) 1139–1142. [20] Zimarino, V., Tsai, C. and Wu, C., Complex modes of heat shock factor activation, Mol. Cell Biol., 10 (1990) 752–759.
Neuronal injury in experimental status epilepticus in the rat: role of hypoxia Masahiro Sasahira a, Roger P. Simon a, David A. Greenberg a , b ,* a Department of Neurology, University of Pittsburgh School of Medicine, S-526 Biomedical Science Tower, Pittsburgh, PA 15213, USA Department of Neurobiology, University of Pittsburgh School of Medicine, S-526 Biomedical Science Tower, Pittsburgh, PA 15213, USA
b
Received 30 September 1996; revised version received 8 January 1997; accepted 10 January 1997
Abstract While it seems axiomatic that hypoxia is a risk factor for neuronal death during prolonged seizures, the classic neuropathologic literature does not confirm such an association. We investigated this issue by inducing status epilepticus in normoxic (PaO2 ≈ 100 mmHg) and hypoxic (PaO2 ≈ 50 mmHg) rats, using heat-shock protein (HSP) expression as an index of early cell injury and acid fuchsin staining to detect cell death. Neither stress protein induction nor neuronal death was increased in the selectively vulnerable CA3c region of hippocampus, or in cerebral cortex, of hypoxic compared to normoxic animals. These data support the concept that moderate hypoxia is not a risk factor for brain injury from status epilepticus. 1997 Elsevier Science Ireland Ltd. Keywords: Heat-shock protein; Hypoxia; Status epilepticus; Seizures; Cell death
Most treatment protocols for status epilepticus emphasize immediate attention to issues of oxygenation [7]. However, the experimental and neuropathologic literature that might support a role for hypoxia in neuronal injury during status epilepticus is inconclusive. Maximal oxygenation neither increases nor decreases neuronal injury during experimental status [13,16]. Whether hypoxia in the absence of ischemia is a risk factor is uncertain [5,9]. Blennow et al. showed that in bicuculline-induced status epilepticus, there was less histopathologic change in cortex, hippocampus, and striatum of animals maintained at PaO2 ≈ 50 mmHg than at PaO2 ≈ 100 mmHg [4]. However, the interpretation of these results is limited by their reliance on staining techniques (hematoxylin and eosin, cresyl violet, and Luxol fast blue) that are comparatively insensitive to early neuronal injury. So¨derfeldt et al. used toluidine blue staining and electron microscopy to study the effect of hypoxia in bicuculline-induced status [16] and found that hypoxia ‘appeared to exaggerate the extent of neuronal alteration’, especially mitochondrial swelling. But only cortical neurons were studied, no statistical ana-
* Corresponding author. Tel.: +1 412 6482623; fax: +1 412 6481239; e-mail: [email protected]
lysis was provided, and electrographic seizure activity was not measured. In addition, because animals were sacrificed immediately after 2 h of status in both of these studies, reversible and irreversible neuronal injury could not be differentiated. We revisited this issue using more sensitive measures of neuronal injury (immunocytochemical detection of heatshock protein (HSP) to assess cell injury and acid-fuchsin staining to quantify cell death [3]) in rats sacrificed 72 h after bicuculline-induced status epilepticus with PaO2 maintained at ~50 or ~100 mmHg. Non-fasted male Sprague–Dawley rats (350–470 g) were anesthetized with 4% halothane, then incubated and ventilated with 70% N2O/27% O2/3% halothane. Atropine (0.25 mg/kg) and pancuronium bromide (4 mg/kg) were administered intravenously via a femoral venous catheter. Arterial blood pressure was monitored continuously by polygraphic recording via a femoral arterial catheter. Arterial blood gases were monitored every 10 min during status. The electroencephalogram (EEG) was recorded continuously from biparietal electrodes attached to the skull with jeweler’s screws. The fraction of inspired O2 (FiO2) was regulated to produce PaO2 ≈ 100 mmHg (n = 13) or 50 mmHg (n = 13). Halothane was discontinued before status was induced.
0304-3940/97/$17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0304-3940 (97 )1 3378-X
208
M. Sasahira et al. / Neuroscience Letters 222 (1997) 207–209
Table 1 Physiologic variables prior to and during 1 h of bicuculline-induced status epilepticus in hypoxic (n = 13) and normoxic (n = 13) rats Baseline
Status epilepticus
Group
T (°C)
pHa
PaCO2*
PaO2**
Duration (min)
Tmax (°C)
pHa
PaCO2
PaO2**
Hypoxic Normoxic
37.1 ± 3.0 36.5 ± 1.4
7.49 ± 0.03 7.45 ± 0.08
31.6 ± 2.3 35.0 ± 4.7
53.5 ± 2.9 105.1 ± 6.2
27.2 ± 3.1 29.9 ± 5.7
37.6 ± 4.9 37.5 ± 3.4
7.37 ± 0.03 7.36 ± 0.08
35.5 ± 1.6 37.7 ± 4.7
49.7 ± 1.9 104.5 ± 5.5
a, Arterial; *P , 0.05, **P , 0.001 (t-test).
Bicuculline was dissolved in 10 N HCl (brought to pH 5.5–5.8 with NaOH) and administered intravenously at 1.2 mg/kg to induce status. Acute pulmonary hypertension and pulmonary edema were prevented by pre-administration of phentolamine (0.1 mg/kg) and by the removal of 4–5 ml of arterial blood, which were then reinfused after the hypertensive surge [4,16]. Bicuculline was re-administered 10 and 20 min after the first injection to prolong status. Electrographic seizure activity was terminated after 1 h by the administration of diazepam (14–15 mg/kg). The EEG patterns were classified into four types, as previously reported [10]: type I, baseline activity; type II, discrete spikes superimposed on a normal baseline; type III, high-voltage spikes and sharp waves separated by at least 1 s of suppressed EEG activity; and type IV, continuous high-voltage, high frequency discharges lasting at least 1 s. The total duration of type IV seizure activity was taken as the duration of status because this electrographic pattern correlates best with histologic brain injury from status epilepticus [10]. After 72 h, animals were sacrificed by intraperitoneal administration of chloral hydrate (400 mg/kg), followed by intracardiac perfusion with 100 ml of normal saline and 400 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were removed and kept in fixative for 24 h at 4°C and 35 and 50 mm sections were then prepared on a vibratome. Neuronal injury was assessed in a blinded manner by counting CA3c hippocampal and cerebral cortical neurons that stained with a monoclonal antibody to the non-constitutive, 72 kDa, HSP72, and neuronal cell death was assessed by counting acid fuchsin-stained neurons [10,14], in predetermined grids on a single section taken at bregma minus 3.3 mm. Within that plane, cortical grids were centered 5.7 mm lateral to the midline and 6.6 mm dorsal to the interaural line. Neither HSP immunoreactivity nor acid-fuchsin staining was observed in control animals not given bicuculline. Data are presented as means ± SD. Statistical analyses were performed using ttests, corrected for multiple comparisons as necessary. Physiologic measures in the normoxic and hypoxic groups are given in Table 1. The only significant differences between the two groups were in PaO2 and, to a less pronounced extent, baseline PaCO2. In particular, there were no differences in temperature or arterial blood pH, each of which has been shown to modify excitotoxic neu-
ronal injury, or in the duration of status. The quantitation of HSP immunoreactivity and acid fuchsin staining of neurons in cortex and hippocampus is shown in Fig. 1. There were no significant differences in HSP72 immunoreactivity or in acid-fuchsin staining between the normoxic and hypoxic groups in either of the brain regions studied. This study shows that cell injury (assessed by stress protein expression) and cell death (measured by acid-fuchsin staining) occur with the same frequency after bicuculline-induced status epilepticus whether animals are normoxic or hypoxic. Induced expression of the stress protein, HSP72, is a reliable and sensitive marker of cell
Fig. 1. Histopathologic sequelae of experimental status epilepticus. (A) HSP72-immunoreactive cells in hippocampal area CA3c (original magnification, 40 × ). (B) Counts of HSP72-immunoreactive and acid fuchsin-staining cells in CA3c sector of hippocampus and in cerebral cortex of normoxic (n = 13) and hypoxic (n = 13) rats. Error bars are SD.
M. Sasahira et al. / Neuroscience Letters 222 (1997) 207–209
injury from a wide range of cerebral insults, including seizures and ischemia [12,14,17] and represents a cellular response to the protein denaturation [2]. Denatured proteins activate heat-shock factors [18], which bind to heatshock elements to initiate the transcription of HSP72 mRNA [19,20]. While HSP72 immunoreactivity is, therefore, a sign of neuronal injury, the eventual fate of these injured neurons is uncertain [8]. In contrast, acid fuchsinstained neurons indicate cell death [3]. Although we examined only a single time point after seizures (72 h), it is unlikely that this obscured cell death occurring at earlier times and no longer apparent owing to phagocytic removal of dead cells. This is because, inter alia, we observed no difference in the total number of (stained plus unstained) cells at 72 h after status between normoxic and hypoxic animals. Our data confirm the proposal of Blennow et al. [5] that hypoxia during status epilepticus is not a risk factor for cell death. In another study, Blennow et al. [4] used conventional light-microscopic stains to identify subacute or chronically injured (pale-staining) cells, but these morphological features may be subtler and more difficult to quantify than HSP72 immunocytochemistry. Our findings are also consistent with data in primates, which showed no relationship between the lowest recorded oxygen concentration in cerebral venous or arterial blood and morphologic brain damage from status [11]. We did not confirm Blennow et al.’s suggestion that hypoxia during status is actually neuroprotective. This may be because our model employs repeated bicuculline injections to produce similar seizure durations in the normoxic and hypoxic groups. In the Blennow studies, status was briefer in the hypoxic animals, and this may have explained the protective effect. An anticonvulsive effect of hypoxia has been demonstrated in awake rats treated with kainic acid [1]. The implications of our findings for the treatment of status epilepticus are unclear. For example, even if moderate hypoxia has no directly deleterious effect on cerebral neurons during status, severe hypoxia leads to centrallyinduced hypotension [6], which could cause ischemic injury to the brain, and status in the setting of ischemic brain injury has a poor prognosis [15]. [1] Amano, S., Obata, T., Hazama, F., Kashiro, N. and Shimada, M., Hypoxia prevents seizures and neuronal damages of the hippocampus induced by kainic acid in rats, Brain Res., 523 (1990) 121–126. [2] Ananthan, J., Goldberg, A. and Voellmy, R., Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat
209
shock genes, Science, 232 (1986) 522–524. [3] Auer, R., Wieloch, T., Olsson, Y. and Siesjo, B., The distribution of hypoglycemic brain damage, Acta Neuropathol. (Berl.), 64 (1984) 177–191. [4] Blennow, G., Brierley, J., Meldrum, B. and Siesjo, B., Epileptic brain damage: the role of systemic factors that modify cerebral energy metabolism, Brain, 101 (1978) 687–700. [5] Blennow, G., Nilsson, B. and Siesjo, B., Influence of reduced oxygen availability on cerebral metabolic changes during bicucullineinduced seizures in rats, J. Cerebral Blood Flow Metab., 5 (1985) 439–445. [6] Cross, C., Rieben, P., Barron, C. and Salisbury, P., Effects of arterial hypoxia on the heart and circulation: an integrative study, Am. J. Physiol., 205 (1963) 963–970. [7] Epilepsy Foundation of America, Treatment of convulsive status epilepticus. Recommendations of the Epilepsy Foundation of America’s working group on status epilepticus, J. Am. Med. Assoc., 270 (1993) 854–859. [8] Gonzalez, M., Shiraishi, K., Hisanaga, K., Sagar, S., Mandabach, M. and Sharp, F., Heat shock proteins as markers of neuronal injury, Mol. Brain Res., 6 (1989) 93–100. [9] Kreisman, N., LaManna, J., Rosenthal, M. and Sick, T., Oxidative metabolic responses with recurrent seizures in rat cerebral cortex: role of systemic factors, Brain Res., 218 (1981) 175–188. [10] Lowenstein, D., Simon, R. and Sharp, F., The pattern of 72-kDa heat shock protein-like immunoreactivity in the rat brain following flurothyl-induced status epilepticus, Brain Res., 531 (1990) 173– 182. [11] Meldrum, B. and Brierley, J., Prolonged epileptic seizures in primates. Ischemic cell change and its relation to ictal physiological events, Arch. Neurol., 28 (1973) 10–17. [12] Shimosaka, S., So, Y. and Simon, R., Distribution of HSP72 induction and neuronal death following limbic seizures, Neurosci. Lett., 138 (1992) 202–206. [13] Siesjo, B., Cell damage in the brain: a speculative synthesis, J. Cerebral Blood Flow Metab., 1 (1981) 155–185. [14] Simon, R., Cho, H., Gwinn, R. and Lowenstein, D., The temporal profile of 72-kDa heat-shock protein expression following global ischemia, J. Neurosci., 11 (1991) 881–889. [15] Snyder, B., Hauser, W., Loewenson, R., Leppik, I., Ramirez, L. and Gumnit, R., Neurologic prognosis after cardiopulmonary arrest: III, Seizure activity, Neurology, 30 (1980) 1292–1297. [16] Soderfeldt, B., Blennow, G., Kalimo, H., Olsson, Y. and Siesjo, B., Influence of systemic factors on experimental epileptic brain injury. Structural changes accompanying bicuculline-induced seizures in rats following manipulations of tissue oxygenation or alpha-tocopherol levels, Acta Neuropathol. (Berl.), 60 (1983) 81–91. [17] Vass, K., Welch, W. and Nowak, T., Localization of 70-kDa stress protein induction in gerbil brain after ischemia, Acta Neuropathol. (Berl.), 77 (1988) 128–135. [18] Wu, C., Wilson, S., Walker, B., David, I., Paisley, T., Zimarino, V. and Ueda, H., Purification and properties of Drosophila heat shock activator protein, Science, 238 (1987) 1247–1253. [19] Xiao, H. and Lis, J., Germline transformation used to define key features of heat-shock response, Science, 239 (1988) 1139–1142. [20] Zimarino, V., Tsai, C. and Wu, C., Complex modes of heat shock factor activation, Mol. Cell Biol., 10 (1990) 752–759.