- Email: [email protected]
Prior ischemic stress protects against experimental stroke
Neuroscience Leiters, 163 (1993) 135 137 © 1993 Elsevier Ireland Ltd. All rights reserved 0304-3940/93l$ 06.00
135
NSL 09950
Prior ischemic stress protects against experimental stroke Roger P. Simon*, Masaki Niiro, Ryder Gwinn University qf Cali[brnia, San Francisco Medical Center, Department o]Neurology, 3rd & Parnassus, SFGH 4M62, 1001 Potrero A venue, San Francisco, CA 94143-0870, USA (Received 28 May 1993; Revised version received 14 July 1993; Accepted 3 August 19931 Key words: Stress protein; Heat shock protein: Ischemic tolerance We studied the possible role of prior ischemic stress as a protective mechanism against cerebral infarction in rats. Two brief periods of global cerebral ischemia~ separated by 24 h, did not cause cell death in brain, but did produce neuronal stress, as demonstrated by induction of the nonconstitutive 72 kDa heat shock protein (HSP72). Forty-eight hours later, animals subjected to prior ischemia had smaller infarct from permanent middle cerebral artery occlusion than did sham-operated controls. These findings support an association between ischemia-induced stress, HSP72 induction, and attenuation of injury from subsequent focal cerebral ischemia.
Heat shock proteins (HSP) have been associated with the induction of tolerance to multiple stresses, in both in vivo and in vitro systems [2]. Recent in vivo data suggest that heat shock proteins induced either by prior hyperthermia [4, 8] or brief periods of global ischemia [7] result in tolerance of the hippocampal pyramidal neurons to subsequent global ischemia. To determine the possible relevance of this phenomenon to stroke, we studied the effect of brief periods of global ischemia, sufficient to induce HSP72, on infarct size resulting from subsequent permanent middle cerebral artery (MCA) occlusion in the rat. Male Sprague-Dawley rats weighing 325400 g were intubated and mechanically ventilated with 1 3% halothane anesthesia in a nitrous oxide/oxygen gas mixture (70%/30%) for all operations. The femoral artery was catheterized for monitoring mean arterial pressure and collecting blood gas and blood glucose samples. During surgery body temperature was maintained at 37 + 1.0°C by a rectal probe and homeothermic blanket (Harvard). Global ischemia was induced by four-vessel occlusion (permanent bilateral vertebral with transient bilateral carotid ligation; 4VO) producing ischemia for six minutes during which EEG isoelectricity was confirmed. Mean arterial blood pressure was over 80 mmHg before clamping of common carotid arteries. Twenty-four hours after the first episode of global ischemia, a second 4VO was produced by clamping the common carotid arteries for 6 *Corresponding author. Fax: (1) (415) 476-5582.
min (n = 13). Blood pressure and isoelectric EEG requirements were identical to those above. Forty-eight hours after the second 4VO, four animals were sacrificed for immunocytochemical staining of HSP72 [15] and nine animals underwent focal ischemia produced by permanent left middle cerebral artery occlusion (MCAO) via a base of the skull approach [14]. Core temperature was maintained during surgery at 37 + I°C with a self-regulating heating pad. Core temperature in this model is not different from brain temperature as assessed by temporalis muscle temperature [3, 16]. Brain temperature must be reduced to approximately 24°C to affect infarct size in this model [10, 11]. Cerebral blood flow (CBF) was not directly measured in these experiments as we have shown, in this model, that raising CBF to 100-350% of control does not alter infarct size [17]. Control rats (n = 9) received electrocautery of the vertebral arteries, ligation of the external carotid arteries and placement of 3-0 silk threads under the common carotid arteries. Twenty-four hours after the first sham operation, a second sham 4VO was done by exposing the common carotid arteries and placing threads under each of them. Forty-eight hours after the second sham 4VO, MCAO was done as described above. Forty-eight hours after MCAO, all animals were reanesthetized with chloral hydrate (400 mg/kg), and killed by decapitation, brains were removed and 1.4 mm coronal sections cut and stained with 2% 2,3,5-triphenyl tetrazolium chloride (TTC) [1]. Infarct area was obtained by image analysis of six sec-
136
tions corresponding to 12.2, 10.8, 9.4, S.(), 6.0, and 5.2 mm anterior to the interaural line The total infarct volume was then determined by summing the infarct areas t"o1" each of the six sections, multiplied by the section thickness, and expressed as a percentage of the conlralateral hemispheric volume [18]. The percent hemispheric infarct in the double 4VO group in which HSP had been induced (n = 9) was compared with that in the double sham 4VO group (n = 9) using one-way analysis of variance (a = 0.05). There were no statistically signilicant differences in the plasma glucose concentrations, blood gas values oz" systemic blood pressures of animals between test groups, or between the various measurement times using repeated measures ANOVA (a = 0.05). Global ischemia (double 4VO) induced robust expression of HSP throughout the cortex (Fig. 1), as described previously [15], and confirmed by Northern and Western blotting [6, 19]. With this brief duration of global ischemia only rare cell death (acid-fuchsin stained neurons) is seen [15].
D i s t a n c e anterior to interaural line ( m m )
70 e-,
p •
60
=E
50
e-
12.2
10.8
9.4
8.0
6.6
5.2
t
t
"T
I
]
O Sham • 4VO
}--
/;/+\,
~yj. t~~~ t 30 ? 40
v
i._
10
I 1
= 2
= 3
I 4
= 5
I 6
Section n u m b e r Fig. 2. Infarct area in brain sections from rats subjected to two 6-min episodes of global ischemia (4VO pre-treatment) or no global ischemia (sham treatment) prior to MCA occlusion. Bars indicate standard errors. Volumetric calculations from the six serial sections of each brain
(area under curves) revealeda significantoveralldifferencebetween the two groups one way ANOVA, a = 0.05.
Fig. 1. A coronal section of rat brain 24 h following the second period of global ischemia (see text). Immunocytochemical staining with a monoclonal antibody against HSP72 demonstrates widespread induction of the protein throughout the dorsolateral cortical mantle (solid arrows). Ventral cortex and basal structures did not exhibit HSP72-1ike immunoreactivity (open arrow).
The infarction following permanent MCA occlusion was compared between brains in which prior double 4VO had induced HSP and in controls without prior 4VO. Whole brain infarct volume was smaller (33.8 + 9.2%, n = 9) in the animals with previous global ischemia than in sham-operated controls (45.3 + 9.0%, n = 9) (Fig. 2). Brief periods of global ischemia are associated with neuronal stress, as documented by robust expression of the major nonconstitutive heat shock protein, HSP72 [15]. The timing and duration of global ischemia in this study were designed to induce HSP but not neuronal cell death. Under these conditions, rats subjected to subsequent MCAO have smaller infarcts than control animals. These experiments suggest that factors induced by ischemia (perhaps heat shock proteins) protect against subsequent ischemic injury. We have shown the upregulation by global ischemia of other potential candidates as well, which include calcium binding proteins and other stress proteins, such as the glucose-regulated proteins [6], which might also protect the brain during stroke. Although no direct proof of a protective role of HSP has been demonstrated in this study or in others of ischemia, the heat shock proteins are clearly protective in other states of induced cellular stress. Microinjection of antibodies against heat shock proteins eliminates induci-
137
ble heat tolerance in fibroblasts in vitro [12]. Of particular relevance to ischemia, prior heat shock inhibits glutamate toxicity, in cortical cultures; this effect requires new protein synthesis as it is blocked by inhibitors of protein or R N A synthesis [13]. In global ischemia with reperfusion, prior hyperthermia [4, 8] or sublethal global ischemia [7] reduced the vulnerability or CA I pyramidal neurons [7] to ischemia and induced HSR The studies reported here suggest that this phenomenon of ischemic tolerance demonstrated previously in CA I neurons in global ischemia applies to a broader population of neurons, as well as to focal ischemia. The reduction of infarct volume in these experiments was characterized by a narrowing of the radius, i.e. the protected areas were at the edges of the infarct. HSP expression occurs around the rim of the infarct in control animals [5], presumably reflecting some degree of ischemic stress in neurons that nonetheless survive. By induction of HSPs prior to local arterial occlusion, this protected zone is extended. Accordingly, stress protein induction could be another factor in addition to glutamate antagonists [16] or free radical scavengers [9] that defines the extent of the penumbral region in focal infarction. Supported by NIH Program Project Grant NS 14543. 1 Bederson. J.A., et al., Evaluation of 2,3,5-triphenyltetrazolium chloride as a stain for detection and quantification of experimental cerebral infarction in rats, Stroke, 17 (1986) 1304 1307. 2 Brown, I.R.. Induction of heat shock (stress) genes in the mammalian brain by' hyperthermia and other traumatic events: a current perspective, .1. Neurosci. Res., 27 (1990} 247 255. 3 Busto, R., el al., Small differences in intraischemic brain temperature critically determine the extent of ischemic neuronal injury, J. Cereb. Blood Flow Metab., 7 (1987) 729 738. 4 Chopp, M., el al., Transient hyperthermia protects against subsequent forcbrain ischemic cell damage in the rat. Neurology, 39 (1989) 1396 1398.
5 Gonzalcz, M.F., et al.. Heat shock proteins as markers of neural injury, Mol. Brain Res.. 6 (1~,~89) 93 10(). 6 Gwmn. R.P.. el al., Increased expression of m R N A encoding calbmdin-D28K, the glucose regulated proteins, or the 72 kl)a heatshock protcin in throe models of CNS injury. Soc. Ncurosci. Abstr.. 17(1991) 1083. 7 Kitagawa, K., et al.. "lschemic tolerance" phenomenon Ik)und in the brain. Brain Res.. 528 11990)21 24. 8 Kitagawa, K.. et al., Hyperthcrmia-induced neuronal protection against ischemic injtu'y m gerbil,,. ,1. ('erch Blood F'low Metab.. II 11991)449 452. 9 Mon_,,cr, H., Hartley, D.M. and Choi, D., 21-Aminostcroids attenuale excitotoxic neuronal injury in cortical cell cultures. Neuron, 5 (19901 121 126. 10 Morika,aa. E.. c t a l . , Modcratc celcbral h,.pothcrmia fails to ;imelioratc focal ischemic injury. J. ('crcb. Blood |:lov, Mctab.. 12 (1992) 380 38'-). 11 Onesti. S.T.. ct al,, Transient hypothcrmia reduces tk~cal ischcmic injury m the rat, Neurosurgcr}, 3 (1991) 369 373. 12 Riabm~ol, K.T., Mizzen. L.A. and Welch. W.J., ttcat shock is lethal to libroblasts microinjected with antibodies against hsp70. Scicncc. (1988) 433 43(~. 13 Rordorf. G.. Koroshctz, W.,l. and Bonvenlre. ,I.V., tteat shock proleers cultures neurons from glutamate toxicity, Neuron, 7 (1991) 11143 1(i51. 14 Shiraishi. K. and Simon, R.P.. A model of proximal MCA occhlsion i n t h c r a t . Ncurosci. Methods, 31)(1989) 169 174. 15 Simon. R.P.. el al.. The temporal profile of 72-kDa hcat-shock protcin expression follo',~ing global ischemia, J. Neurosci., I 1 (1991) 881 889. 16 Simon. R.R and Shiraishi, K.. 3:-Methyl-I>aspartate antagonist reduces stroke size and regional glucose metabolism. Ann. Neurol.. 27 (1991)) 606 611. 17 Simon. R.R and Gwinn, R.. Brain acidosis reduced by hypcrcarbic xcntilation protects against focal ischemia. Neurology. 43 (1993) A338. 18 Swanson. R.A.. el al.. A semi-automated method Ik~r mcasurmg brain inlhrct volume, J. Cereb. Blood Flow Metabol., I(t (19891 29O 293. 19 Va~,s, K., Welch. W.T. and Nowak. T.J.. Localization of 7()kl) stress protein induction in gerbil brain aftcr isehemia, Acta Neurol. Palhol.,7711988) 128 135.
135
NSL 09950
Prior ischemic stress protects against experimental stroke Roger P. Simon*, Masaki Niiro, Ryder Gwinn University qf Cali[brnia, San Francisco Medical Center, Department o]Neurology, 3rd & Parnassus, SFGH 4M62, 1001 Potrero A venue, San Francisco, CA 94143-0870, USA (Received 28 May 1993; Revised version received 14 July 1993; Accepted 3 August 19931 Key words: Stress protein; Heat shock protein: Ischemic tolerance We studied the possible role of prior ischemic stress as a protective mechanism against cerebral infarction in rats. Two brief periods of global cerebral ischemia~ separated by 24 h, did not cause cell death in brain, but did produce neuronal stress, as demonstrated by induction of the nonconstitutive 72 kDa heat shock protein (HSP72). Forty-eight hours later, animals subjected to prior ischemia had smaller infarct from permanent middle cerebral artery occlusion than did sham-operated controls. These findings support an association between ischemia-induced stress, HSP72 induction, and attenuation of injury from subsequent focal cerebral ischemia.
Heat shock proteins (HSP) have been associated with the induction of tolerance to multiple stresses, in both in vivo and in vitro systems [2]. Recent in vivo data suggest that heat shock proteins induced either by prior hyperthermia [4, 8] or brief periods of global ischemia [7] result in tolerance of the hippocampal pyramidal neurons to subsequent global ischemia. To determine the possible relevance of this phenomenon to stroke, we studied the effect of brief periods of global ischemia, sufficient to induce HSP72, on infarct size resulting from subsequent permanent middle cerebral artery (MCA) occlusion in the rat. Male Sprague-Dawley rats weighing 325400 g were intubated and mechanically ventilated with 1 3% halothane anesthesia in a nitrous oxide/oxygen gas mixture (70%/30%) for all operations. The femoral artery was catheterized for monitoring mean arterial pressure and collecting blood gas and blood glucose samples. During surgery body temperature was maintained at 37 + 1.0°C by a rectal probe and homeothermic blanket (Harvard). Global ischemia was induced by four-vessel occlusion (permanent bilateral vertebral with transient bilateral carotid ligation; 4VO) producing ischemia for six minutes during which EEG isoelectricity was confirmed. Mean arterial blood pressure was over 80 mmHg before clamping of common carotid arteries. Twenty-four hours after the first episode of global ischemia, a second 4VO was produced by clamping the common carotid arteries for 6 *Corresponding author. Fax: (1) (415) 476-5582.
min (n = 13). Blood pressure and isoelectric EEG requirements were identical to those above. Forty-eight hours after the second 4VO, four animals were sacrificed for immunocytochemical staining of HSP72 [15] and nine animals underwent focal ischemia produced by permanent left middle cerebral artery occlusion (MCAO) via a base of the skull approach [14]. Core temperature was maintained during surgery at 37 + I°C with a self-regulating heating pad. Core temperature in this model is not different from brain temperature as assessed by temporalis muscle temperature [3, 16]. Brain temperature must be reduced to approximately 24°C to affect infarct size in this model [10, 11]. Cerebral blood flow (CBF) was not directly measured in these experiments as we have shown, in this model, that raising CBF to 100-350% of control does not alter infarct size [17]. Control rats (n = 9) received electrocautery of the vertebral arteries, ligation of the external carotid arteries and placement of 3-0 silk threads under the common carotid arteries. Twenty-four hours after the first sham operation, a second sham 4VO was done by exposing the common carotid arteries and placing threads under each of them. Forty-eight hours after the second sham 4VO, MCAO was done as described above. Forty-eight hours after MCAO, all animals were reanesthetized with chloral hydrate (400 mg/kg), and killed by decapitation, brains were removed and 1.4 mm coronal sections cut and stained with 2% 2,3,5-triphenyl tetrazolium chloride (TTC) [1]. Infarct area was obtained by image analysis of six sec-
136
tions corresponding to 12.2, 10.8, 9.4, S.(), 6.0, and 5.2 mm anterior to the interaural line The total infarct volume was then determined by summing the infarct areas t"o1" each of the six sections, multiplied by the section thickness, and expressed as a percentage of the conlralateral hemispheric volume [18]. The percent hemispheric infarct in the double 4VO group in which HSP had been induced (n = 9) was compared with that in the double sham 4VO group (n = 9) using one-way analysis of variance (a = 0.05). There were no statistically signilicant differences in the plasma glucose concentrations, blood gas values oz" systemic blood pressures of animals between test groups, or between the various measurement times using repeated measures ANOVA (a = 0.05). Global ischemia (double 4VO) induced robust expression of HSP throughout the cortex (Fig. 1), as described previously [15], and confirmed by Northern and Western blotting [6, 19]. With this brief duration of global ischemia only rare cell death (acid-fuchsin stained neurons) is seen [15].
D i s t a n c e anterior to interaural line ( m m )
70 e-,
p •
60
=E
50
e-
12.2
10.8
9.4
8.0
6.6
5.2
t
t
"T
I
]
O Sham • 4VO
}--
/;/+\,
~yj. t~~~ t 30 ? 40
v
i._
10
I 1
= 2
= 3
I 4
= 5
I 6
Section n u m b e r Fig. 2. Infarct area in brain sections from rats subjected to two 6-min episodes of global ischemia (4VO pre-treatment) or no global ischemia (sham treatment) prior to MCA occlusion. Bars indicate standard errors. Volumetric calculations from the six serial sections of each brain
(area under curves) revealeda significantoveralldifferencebetween the two groups one way ANOVA, a = 0.05.
Fig. 1. A coronal section of rat brain 24 h following the second period of global ischemia (see text). Immunocytochemical staining with a monoclonal antibody against HSP72 demonstrates widespread induction of the protein throughout the dorsolateral cortical mantle (solid arrows). Ventral cortex and basal structures did not exhibit HSP72-1ike immunoreactivity (open arrow).
The infarction following permanent MCA occlusion was compared between brains in which prior double 4VO had induced HSP and in controls without prior 4VO. Whole brain infarct volume was smaller (33.8 + 9.2%, n = 9) in the animals with previous global ischemia than in sham-operated controls (45.3 + 9.0%, n = 9) (Fig. 2). Brief periods of global ischemia are associated with neuronal stress, as documented by robust expression of the major nonconstitutive heat shock protein, HSP72 [15]. The timing and duration of global ischemia in this study were designed to induce HSP but not neuronal cell death. Under these conditions, rats subjected to subsequent MCAO have smaller infarcts than control animals. These experiments suggest that factors induced by ischemia (perhaps heat shock proteins) protect against subsequent ischemic injury. We have shown the upregulation by global ischemia of other potential candidates as well, which include calcium binding proteins and other stress proteins, such as the glucose-regulated proteins [6], which might also protect the brain during stroke. Although no direct proof of a protective role of HSP has been demonstrated in this study or in others of ischemia, the heat shock proteins are clearly protective in other states of induced cellular stress. Microinjection of antibodies against heat shock proteins eliminates induci-
137
ble heat tolerance in fibroblasts in vitro [12]. Of particular relevance to ischemia, prior heat shock inhibits glutamate toxicity, in cortical cultures; this effect requires new protein synthesis as it is blocked by inhibitors of protein or R N A synthesis [13]. In global ischemia with reperfusion, prior hyperthermia [4, 8] or sublethal global ischemia [7] reduced the vulnerability or CA I pyramidal neurons [7] to ischemia and induced HSR The studies reported here suggest that this phenomenon of ischemic tolerance demonstrated previously in CA I neurons in global ischemia applies to a broader population of neurons, as well as to focal ischemia. The reduction of infarct volume in these experiments was characterized by a narrowing of the radius, i.e. the protected areas were at the edges of the infarct. HSP expression occurs around the rim of the infarct in control animals [5], presumably reflecting some degree of ischemic stress in neurons that nonetheless survive. By induction of HSPs prior to local arterial occlusion, this protected zone is extended. Accordingly, stress protein induction could be another factor in addition to glutamate antagonists [16] or free radical scavengers [9] that defines the extent of the penumbral region in focal infarction. Supported by NIH Program Project Grant NS 14543. 1 Bederson. J.A., et al., Evaluation of 2,3,5-triphenyltetrazolium chloride as a stain for detection and quantification of experimental cerebral infarction in rats, Stroke, 17 (1986) 1304 1307. 2 Brown, I.R.. Induction of heat shock (stress) genes in the mammalian brain by' hyperthermia and other traumatic events: a current perspective, .1. Neurosci. Res., 27 (1990} 247 255. 3 Busto, R., el al., Small differences in intraischemic brain temperature critically determine the extent of ischemic neuronal injury, J. Cereb. Blood Flow Metab., 7 (1987) 729 738. 4 Chopp, M., el al., Transient hyperthermia protects against subsequent forcbrain ischemic cell damage in the rat. Neurology, 39 (1989) 1396 1398.
5 Gonzalcz, M.F., et al.. Heat shock proteins as markers of neural injury, Mol. Brain Res.. 6 (1~,~89) 93 10(). 6 Gwmn. R.P.. el al., Increased expression of m R N A encoding calbmdin-D28K, the glucose regulated proteins, or the 72 kl)a heatshock protcin in throe models of CNS injury. Soc. Ncurosci. Abstr.. 17(1991) 1083. 7 Kitagawa, K., et al.. "lschemic tolerance" phenomenon Ik)und in the brain. Brain Res.. 528 11990)21 24. 8 Kitagawa, K.. et al., Hyperthcrmia-induced neuronal protection against ischemic injtu'y m gerbil,,. ,1. ('erch Blood F'low Metab.. II 11991)449 452. 9 Mon_,,cr, H., Hartley, D.M. and Choi, D., 21-Aminostcroids attenuale excitotoxic neuronal injury in cortical cell cultures. Neuron, 5 (19901 121 126. 10 Morika,aa. E.. c t a l . , Modcratc celcbral h,.pothcrmia fails to ;imelioratc focal ischemic injury. J. ('crcb. Blood |:lov, Mctab.. 12 (1992) 380 38'-). 11 Onesti. S.T.. ct al,, Transient hypothcrmia reduces tk~cal ischcmic injury m the rat, Neurosurgcr}, 3 (1991) 369 373. 12 Riabm~ol, K.T., Mizzen. L.A. and Welch. W.J., ttcat shock is lethal to libroblasts microinjected with antibodies against hsp70. Scicncc. (1988) 433 43(~. 13 Rordorf. G.. Koroshctz, W.,l. and Bonvenlre. ,I.V., tteat shock proleers cultures neurons from glutamate toxicity, Neuron, 7 (1991) 11143 1(i51. 14 Shiraishi. K. and Simon, R.P.. A model of proximal MCA occhlsion i n t h c r a t . Ncurosci. Methods, 31)(1989) 169 174. 15 Simon. R.P.. el al.. The temporal profile of 72-kDa hcat-shock protcin expression follo',~ing global ischemia, J. Neurosci., I 1 (1991) 881 889. 16 Simon. R.R and Shiraishi, K.. 3:-Methyl-I>aspartate antagonist reduces stroke size and regional glucose metabolism. Ann. Neurol.. 27 (1991)) 606 611. 17 Simon. R.R and Gwinn, R.. Brain acidosis reduced by hypcrcarbic xcntilation protects against focal ischemia. Neurology. 43 (1993) A338. 18 Swanson. R.A.. el al.. A semi-automated method Ik~r mcasurmg brain inlhrct volume, J. Cereb. Blood Flow Metabol., I(t (19891 29O 293. 19 Va~,s, K., Welch. W.T. and Nowak. T.J.. Localization of 7()kl) stress protein induction in gerbil brain aftcr isehemia, Acta Neurol. Palhol.,7711988) 128 135.