XVII. European Stroke Conference
Nice, France

Oral Session:

Experimental studies I
Date: Thursday 15 May 2008  
Time: 10:45 - 10:55 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

01
Nuclear factor-kB and postischemic inflammation are suppressed in CD36-null mice after middle cerebral artery occlusion
A. Kunz    T. Abe    K. Hochrainer    M. Shimamura    J. Anrather    G. Racchumi    P. Zhou    C. Iadecola              
 

1 Division of Neurobiology, Weill-Cornell Medical College, New York, NY, U.S.A. 2 Dept. of Neurology, University Hospital Carl Gustav Carus, Technical University Dresden, Dresden, Germany

GERMANY

Background: CD36, a scavenger receptor involved e.g. in inflammatory signaling (J Clin Invest 108: 785, 2001), also contributes to ischemic injury through yet unidentified mechanisms (J Neurosci 25: 2504, 2005). We investigated whether CD36 participates in the mechanisms of postischemic inflammation and thereby contributes to ischemic damage. Methods: The middle cerebral artery was transiently occluded in wild type (WT) or CD36 null mice (KO) and 72hrs later, injury volume, mRNA expression of the nuclear factor kB (NFkB)-dependent inflammatory genes iNOS, ELAM, ICAM, and nox2, and neutrophil infiltration were analyzed. Postischemic NFkB activation was examined by electrophoretic mobility shift essay. Results: In KO, injury volume was reduced (-62+/-5%; p<0.05; n=6/group) and inflammatory genes were markedly attenuated (iNOS: -74+/-3%; ELAM: -80+/-11%; ICAM: -70+/-3%; nox2: -76+/-4%; p<0.05; n=5/group). Also, the number of infiltrating neutrophils was reduced (KO: 281+/-93; WT: 1938+/-296; p<0.05; n=5/group). Consistent with the finding of suppressed inflammatory response in KO, antinflammatory treatment with the iNOS inhibitor aminoguanidine reduced injury in WT (-45+/-13%; n=6), but not in KO (p>0.05 from vehicle; n=6). In contrast, the neuroprotective agent NS398 reduced injury both in WT (-40+/-15%) and KO (-59+/-4%; p<0.05; n=6/group). Activation of NFkB, a transcription factor that coordinates postischemic inflammation, was attenuated in KO, but not in WT or in WT treated with NS398. In contrast to focal ischemia, neuroinflammation induced by intracerebroventricular injection of interleukin (IL)1β did not attenuate inflammatory gene expression, neutrophil infiltration or NFkB activitation in KO (p>0.05; n=5/group). Discussion: The data unveil an unrecognized role of CD36 in activating NFkB and triggering postischemic inflammatory response. Moreover, CD36 is specifically involved in the mechanisms of postischemic inflammation, but not in IL1ß-induced inflammation. CD36 inhibition may be promising to counteract the deleterious effects of postischemic inflammation.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 10:55 - 11:05 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

02
Endogenous, anti-inflammatory effects of LXR receptors in experimental stroke
J.R.Morales    I. Ballesteros    J.G.Zarruk    F. Nombela    J. Vivancos    A. Castrillo    I. Lizasoain    O. Hurtado    M.A.Moro       
 

Faculty of Medicine, Universidad Complutense

SPAIN

Background: Recent work has identified Liver X receptors alpha (LXRalpha) and beta (LXRbeta) as potent anti-inflammatory molecules in macrophages and other immune cells. We have explored possible role and anti-inflammatory mechanisms of endogenous LXR receptors in acute stroke using LXRalpha,beta-/- mice. Methods: Focal permanent ischemia was induced by middle cerebral artery occlusion (MCAO) in wild type controls and Nr1h3-/-Nr1h2-/- double mutant (LXRalpha,beta -/-) mice on a Sv129/C57BL/6 background, obtained through a collaboration with Drs. David Mangelsdorf and Peter Tontonoz. Infarct size was determined by 2,3,5-triphenyltetrazolium chloride (TTC) staining. Neurological assesment was performed as previously described (Caso et al. Circulation 2007). mRNA levels of inflammatory mediators were studied by quantitative RT-PCR as described (Pereira et al., JCBFM 2006). Results: Infarct volume was greater in mice lacking both LXR isoforms when compared to control mice. Moreover, when neurological tests were applied to this sets of mice, lower scores demonstrated a protective action in WT, whereas LXRalpha,beta-/- mice showed poor neurological status. WT animals also showed a lower expression of brain mRNA levels of several inflammatory mediators than their LXRalpha,beta-/- littermates 8 h after MCAO. Discussion: These data show that an endogenous LXR activatory pathway during experimental stroke mediates a potent and natural protection due to anti-inflammatory mechanisms, which could be caused by physiological LXR agonists such as oxysterols. This evidence also suggests that levels of endogenous LXR agonists might serve as prognostic markers in stroke patients.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:05 - 11:15 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

03
Rosiglitazone induces a switch to an anti-inflammatory status through 5-lipoxigenase in experimental stroke
M. Sobrado    M.P.Pereira    O. Hurtado    F. Nombela    E. Escolar    V.G.Romera    I. Lizasoain    J. Vivancos    M.A.Moro       
 

Faculty of Medicine, Universidad Complutense

SPAIN

Background: we have recently shown that the PPARgmma agonist rosiglitazone (RSG) induces 5-lipoxigenase (5-LO) expression when administered after experimental stroke (Pereira et al. JCBFM 2006; ISC 2007), which plays a neuroprotective effect, since inhibition of 5-LO abolishes rosiglitazone neuroprotection. Since 5-LO plays a key role in inflammatory responses, participating in the synthesis of both inflammatory and anti-inflammatory compounds, we have explored the mechanism underlying 5-LO-mediated neuroprotective after RSG administration. Methods: Focal permanent ischemia was induced by middle cerebral artery occlusion (MCAO) in Fischer rats. Levels of the 5-LO products leukotriene B4 (LTB4) and lipoxin A4 (LXA4), as well as quantitation of PPARgamma activation were assessed by EIA following manufacturers’ instructions. Infarct size was determined by 2,3,5-triphenyltetrazolium chloride (TTC) staining. RSG, T0070907 (PPARgamma antagonist) and/or LXA4 were administered 10 min after MCAO, and BWA4C (selective 5-LO inhibitor) was administered 3 h after experimental ischemia. Results: RSG induced 5-LO expression, as well as the synthesis of LXA4, an effect that was inhibited by the selective 5-LO inhibitor BWA4C, indicating that RSG-induced 5-LO causes the synthesis of this eicosanoid. RSG also inhibited the synthesis of LTB4 induced by MCAO. LXA4 caused neuroprotection by reducing infarct size and neurological deficits when administered icv after MCAO. Interestingly, this effect was partly mediated by PPARgama, since it was inhibited by T0070907 co-administration. Discussion: RSG activates a neuroprotective axis by inhibiting the synthesis of the pro-inflammatory LTB4 and promoting the synthesis of the anti-inflammatory LXA4, which exerts neuroprotection by a partly PPARgamma-dependent mechanism.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:15 - 11:25 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

04
Magnetic Resonance Molecular Imaging of Adhesion Molecule Expression in Acute Stroke In Wild-Type and P-Selectin-Deficient Mice.
A.Y.Jin    J. Kaur    U.I.Tuor    D. Rushforth    J. Petersen    R.N.Muller    S. Boutry    P.A.Barber              
 

University of Calgary

CANADA

Introduction: Endothelial activation following transient cerebral ischemia is partly mediated by the expression of the adhesion molecules P- and E-selectin. We hypothesized that magnetic resonance molecular imaging using selectin-specific Gd-DTPA-B(sLeX)A would demonstrate the differential expression of P- and E-selectin between the ischemic and contralateral hemispheres in wild-type and P-selectin-deficient mice following acute stroke. Methods: Focal cerebral ischemia under normothermia was induced by transient occlusion of the left middle cerebral artery (i.e. >70% decrease in cerebral blood flow for 30 minutes) in male C57 Black 6 wild-type (WT) (n=15) and P-selectin knockout (KO) (n=11) mice, followed by 24 hours reperfusion and injection of contrast – either Gd-DTPA-B(sLeX)A or the non-targeted agent Gd-DTPA. MRI was done at 9.4 T. From serial T1 maps, the pre-post contrast difference in T1 in both the ischemic cortex (∆T1 Stroke side) and the control side cortex (∆T1 Control side) was evaluated at each time point. The contrast effect at each time point was evaluated as: ∆T1 Stroke side - ∆T1 Control side. E- and P-selectin were quantified by ELISA in the ischemic and control cortex and protein determination was done by modified Lowry assay. Results: In WT mice, both Gd-DTPA-B(sLeX)A and Gd-DTPA accumulated similarly in the stroke cortex, possibly due to blood brain barrier (BBB) compromise. In KO mice, there was greater stroke cortex accumulation of Gd-DTPA-B(sLeX)A that was not observed with Gd-DTPA. E- and P-selectin expression in WT mice were increased preferentially in the stroke side. In KO mice, only E-selectin expression was increased in the stroke side. Discussion: Interhemispheric T1 changes with Gd-DTPA-B(sLeX)A following transient focal cerebral ischemia appeared to reflect the interhemispheric differential expression of E-selectin in KO mice. P-selectin may play a role in BBB compromise following acute stroke. Increased expression of E-selectin in P-selectin-deficient mice following acute stroke is a novel finding.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:25 - 11:35 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

05
Oxygen Therapy Reduces Secondary Hemorrhage after Thrombolysis in Experimental Thromboembolic Stroke
L. Sun    C. Mueller    S. Heiland    C. Sommer    W. Zhou    R. Veltkamp                            
 

Ruprecht-Karls-University Heidelberg, University Mainz

GERMANY

Background: Secondary hemorrhage is an important complication of thrombolytic therapy in acute ischemic stroke. Hyperbaric oxygen (HBO) and normobaric hyperoxia (NBO) have been shown to protect the brain parenchyma as well as the brain microvascular permeability. This study investigated the effect of both oxygen therapies on hemorrhagic transformation after thrombolysis. Methods: Wistar rats (n= 60) were subjected to thromboembolic middle cerebral artery occlusion (MCAO). Animals breathed either air, 100% O2 at ambient pressure (NBO) or 100% O2 at 3 bar (HBO) initiated 60 min after MCAO for 1 h. Rt-PA (9mg/kg) was administered intravenously to animals after oxygen therapy. Serial MR-imaging was performed 20 min, 2.5 h and 24 h after MCAO. Hemoglobin concentration was quantified by spectrophotometry 24 h after MCAO. Bleeding was also measured semiquantitatively on trichrom stained sections and on T2* weighted MR images. Results: Severity of ischemia did not differ among groups. HBO and NBO reduced blood-brain barrier permeability on postcontrast T1-weighted (T1w) images. Mean abnormal enhancing volumes were 68.8+/- 44.3 mm3 in air, 43.0+/- 23.8 mm3 in NBO and 45.6+/- 23.6 mm3 in HBO at 2,5 h; and 81.6+/- 60.3 mm3, 49.7+/- 38.5 mm3 and 43.1+/- 20.7 mm3 at 24 h after MCAO. HBO- treated rats had significantly lower hemoglobin content (13.4+/- 1.1 µg) than rats treated with NBO (15.8+/- 2.6 µg) and air (19.2+/- 5.1 µg, p< 0.05, ANOVA). HBO and NBO treatment decreased histologic hemorrhage in the ischemic hemisphere compared to air (hemorrhagic scale: 3.4+/- 1.5 in HBO, 3.7+/-1.3 in NBO und 5+/- 1 in air group, p< 0.05, ANOVA) Conclusions: HBO and NBO reduce brain microvascular permeability and hemorrhagic transformation after thrombolysis in thromboembolic cerebral ischemia.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:35 - 11:45 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

06
Combination of hyperbaric oxygenation (HBO) and selective JNK-inhibition by D-JNK-I-1 as a potent neuroprotective strategy in a focal ischemia/reperfusion injury model in rats.
C.C.Eschenfelder    J.R.Liu    Y. Zhao    J. Meyne    G. Deuschl    A. Koch    T. Herdegen                     
 

University Hospital Schleswig-Holstein, Campus Kiel

GERMANY

Background: In pervious studies, HBO was proven as a potent neuroprotective treatment in animal models of acute stroke by reducing apoptotic cell death. Furthermore, c-Jun N-terminal kinase (JNK) is activated in ischemic neurons downstream of the activated glutamate receptors and thus mediating neuronal cell death. The highly specific peptide JKN-inhibitor D-JNK-I-1 (XG-102) has been shown to be neuroprotective in acute stroke in mammals, acting by intracellular inhibition of JNK-kinase. We thus hypothesized that combination of both treatments might enhance the neuroprotective effect of each regimen. Infarct volumetry and clinical outcome were the primary endpoints. Methods: male SD-rats (200-220g) were randomized in 5 groups (n=6, respectively). The suture model induced transient MCAO for 90 min. Any treatment started 3h following MCAO. HBO was performed in a pressure chamber with 100% O2, 3 ATA for 1h at 3h following MCAO. D-JNK-I-1 was administered intraperitoneally 3h following MCAO (11mg/kg) . Controls: Sham (no MCAO), MCAO alone. Treatment groups: MCAO+HBO; MCAO+D-JNK-I-1; MCAO+HBO+D-JNK-I-1. Clinical outcome was used using Garcia and Bederson score. Infarct volume was assessed histologically 24h following MCAO. Results: Data are mean±SD. Total infact volume in controls was: Sham (0±0); MCAO (308mm3±63,4). MCAO+XG-102 (169mm3±84,6); MCAO+HBO (106mm3±82,4); MCAO+XG-102+HBO (66mm3±38,8). Garcia- and Bederson-scores showed significant improvement in all treatment groups compared to control (MCAO), however, the best results were seen in MCAO+XG-102+HBO group (12,8 vs. 7 points). Discussion: We conclude from the presented data that combination of HBO and D-JNK-I-1 (XG-102) treatment in acute transient stroke in rats augments the neuroprotective effect of each therapy alone and thus might be a promising way for further treatment studies in humans.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:45 - 11:55 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

07
Hypothermia of 34°C prolongs the therapeutic time window for thrombolysis after experimental thromboembolic stroke
R.A.Kollmar    T. Blank    S. Schwab                                                 
 

University of Erlangen-Nürnberg

GERMANY

Hypothermia is neuroprotective in experimental stroke and may prolong the so far limited therapeutic time window for thrombolysis by rt-PA. Therefore, hypothermia of 34°C and effects on delayed thrombolysis including reperfusion-associated injury were investigated in a model of thromboembolic stroke (TE). Methods - Male Wistar rats (n=48) were subjected to TE and equally split in 4 groups: Normothermia (37°C) was used for the control group (C) and the thrombolysis group (T) which received rt-PA 90 min after TE. Hypothermia (34°C) was performed from 1.5 to 5 hours after TE in group H and in the combination group that received rt-PA in addition (HT). After 24 hours infarct size, brain edema and neuroscore were assessed. Protein markers for inflammation, adhesion and blood brain barrier (BBB) disruption were determined. MRI-measurements investigated infarct evolution and blood flow parameters. Results – The infarct volume and brain swelling were smaller in the group H compared to group C, T, and HT (p<0.05 to p<0.01). rt-PA treatment in group T resulted in larger infarct and brain swelling than all others. Additional hypothermic therapy in group HT reduced these parameters compared to group T (p<0.05). Moreover, the neuroscore was better for group H vs. C and T. Animals of group HT performed better than in group T (p<0.05). Lower serum concentration of sICAM-1, TIMP-1 and MMP-2 were shown for hypothermic groups (H, TH) compared to group T. Conclusion – Hypothermia reduced side-effects of rt-PA treatment and prolonged the therapeutic time window for thrombolysis in our model of TE.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 11:55 - 12:05 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

08
Intravenous administration of 99mTc-HMPAO-labelled human mesenchymal stem cells after stroke: in vivo imaging and biodistribution
O. Detante    A. Moisan    J. Dimastromatteo    M.J.Richard    L. Riou    E. Grillon    E. Barbier    M. Hommel    C. Ghezzi    C. Rémy
 

INSERM Grenoble Institut Neurosciences

FRANCE

Stem cell administration is a promising therapy for stroke. Human bone marrow mesenchymal stem cells (hMSC) offer the advantage of not originating from a tumoural source. Intravenous (IV) cell injection is more secure than a local graft after a stroke for clinical trials. No study assessed the biodistribution of IV injected hMSC after stroke. Our aim was to evaluate this distribution in rats. METHODS: Rats are allocated into 2 groups: 1) group underwent a right cerebral ischemia by middle cerebral artery occlusion (MCAo, n=9); 2) control group (n=9) without lesion. One week after MCAo, both groups received an IV injection of ~3.4x106 hMSC labelled by 99mTc-HMPAO for 2 nuclear imaging (0-2h / 18-20h after cell injection) and isolated organ counting (2h / 20h). Histology was performed. RESULTS: By imaging, we observed a trend towards a higher activity in the brain after ischemia compared with control group. By organ counting, at 20h after injection, we showed an increase of right brain activity in MCAo group (6.5±0.9 vs 3.6±1.2 % of injected dose (ID)/g). We estimate the proportion of hMSC having migrated to the damaged brain to be 1/10,000 injected cells (2-fold higher than in control rats). In MCAo group, 2h after injection, the right brain activity was higher than in the left one (11.6±2.8 vs 9.8±1.7 %ID/g). An initial hMSC trapping in the lungs was followed by an important decrease of pulmonary activity between 2 and 20h after injection in both groups. An activity increase was showed in the spleen in contrast to the decrease in other organs. By histology, we identified hMSC in spleen, liver, lung and brain. These results suggest that IV injected hMSC can colonize an ischemic lesion in the brain despite an initial cell entrapment in the lungs.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 12:05 - 12:15 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

09
Autologous bone marrow administration after 24 hours reduces behavioral deficits and lesion size in a novel large animal model of stroke
J. Boltze    H. Barthel    A. Förschler    B. Nitzsche    C.M.Boltze    A. Reischauer    A. Hoffmann    O. Sabri    F. Emmrich    U. Gille
 

Fraunhofer Institute for Cell Therapy and Immunology (IZI)

GERMANY

Introduction Efficiency of experimental bone marrow (BM) cell therapies of stroke has been shown in rodents. However, transfer to clinical application requires close-to-practice large animal models. We evaluated benefit of autologous BM cell transplantation in a novel sheep model of focal cerebral ischemia allowing control of lesion size and subsequent functional deficits. Material & Methods 30 adult rams were subjected to permanent middle cerebral artery occlusion (MCAO) for stroke induction. 24 hours after MCAO, 15 animals received 4.0 to 5.1x10E6 autologous mononuclear BM cells per kilogram bodyweight. 15 sheep served as controls. Functional outcome was continuously observed by behavioral phenotyping. Lesion size development was monitored by MRI and PET performed at days 1, 14 and 42 before brains were removed for further histological investigation. Results In BM cell treated animals, functional improvement was enhanced as compared to control animals (p<0.01) while control animals suffered from moderate to severe motor and sensory dysfunctions like ataxia, absent startle reflexes and spatial hemineglect for the entire observation period. MRI investigations showed similar lesion size in both groups at day 1 (p=0.59), but reduction of lesion size and hemispherical atrophy in cell treated rams 42 days upon MCAO (p<0.01). These findings could be confirmed by 15O-water- and 18F-Desoxyglucose PET (p < 0.05). No tumor formation was observed upon BM cell administration. Conclusion Autologous BM marrow administration 24 hours following stroke is safe and effective in sheep. BM cell administration might be used as a novel treatment option in fighting stroke in upcoming clinical trials.

Experimental studies I
Date: Thursday 15 May 2008  
Time: 12:15 - 12:25 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

10
Differential Modulation of the Cellular Immune System after Large Hemispheric versus Cortical Murine Ischemic Stroke
A. Liesz    S. Hagmann    A. Hug    C. Zschoche    J. Adamek    T. Giese    R. Veltkamp                     
 

University Heidelberg

GERMANY

Background: Cerebral ischemia induces systemic immunodepression which increases the susceptibility to infections but may also limit autoimmune responses against the injured brain. Because infarct size may be an important determinator of the systemic immune response, we systematically compared cellular immune parameters in 2 models of focal ischemia. Methods: In C57Bl/6 mice (n=160), occlusion of the middle cerebral artery (MCAo) was induced either by coagulation of the distal MCA causing small cortical infarcts or by intraluminal filament for 90min. The respective sham operation served as control. Differential leucocyte counts were performed in blood, spleen, lymph node, and thymus. Lymphocyte subpopulations were further characterized by flow cytometry 24h, 3d, and 7d after MCAo. Results: In large infarcts, a marked decrease of leucocyte counts (e.g. 58% in lymph node) and most lymphocyte subsets was observed in all organs already 24h after stroke with no differences between left- and right-sided ischemia. In contrast, small infarcts in the coagulation model led to no significant changes in differential blood count and alterations only in specific lymphocyte populations. B cells show a higher vulnerability after stroke than T cells. Remarkably, the regulatory T cell population remained stable in both models. Mature thymic cells were uneffected while immature CD4+CD8+ cells were depleted by over 90% 3d after stroke. Discussion: Infarct size is a major determinator of post-stroke systemic cellular immune modulation. Cerebral ischemia has a differential impact on different lymphocyte subpopulations. Whether this reflects a biologically meaningful adaptive response remains to be shown. Supported by grants from MWK Baden-Württemberg and the Else-Kröner-Fresenius Stiftung

Experimental studies I
Date: Thursday 15 May 2008  
Time: 12:25 - 12:35 -  Room: Euterpe
Chair: M. Fisher, USA and O. Lindvall, Sweden

11
DIFFERENT INJURY MECHANISMS ARE ACTIVATED AFTER CEREBRAL ISCHEMIA DEPENDING ON THE OCCURRENCE OF REPERFUSION.
M. Gutiérrez    I. Ayuso    M. Salinas    E. Díez Tejedor    J. Masjuan    M. Alonso de Leciñana                            
 

Neurology and*Biochemistry-Investigat Ramón y Cajal Hosp. **Cerebrovasc Research, La Paz Hosp

SPAIN

Background and aims Different injury mechanisms might be responsible of ischemic cerebral damage depending on the occurrence of reperfusion. We study activation of calpain, implicated in necrosis, and caspase-3 in apoptosis, in a model of permanent or transient focal cerebral ischemia. Materials and methods Long- Evans rats were subjected to permanent middle cerebral artery occlusion for 5min (n=4), 30min (n=4), 1h (n=5), 3h (n=4), 72h (n=6), to transient occlusion for 1h followed by different reperfusion periods: 2h (n=3), 23h (n=3) and 72h (n=3), or to different duration of ischemia 1h (n=3), 3h (n=3), 24 h (n=3) followed by reperfusion up to 72 hours of survival. Samples were obtained from the cortex at the infarct core, the penumbra and from symmetrical areas in the contralateral hemisphere. Inmunoblots for espectrin degradation in its specific fragments of 120 KDa by caspase and of 145 KDa by calpain as well as cleavage of eIF4G were used to investigate activation of apoptosis and necrosis respectively. Results After 1 hour of permanent ischemia maximal calpain activation was observed at the infarct core while it was lower at the penumbra. There was no caspase activation. Reperfusion did not avoid calpain activation, but produced caspase activation that was greater in the core than in the penumbra. Conclusions Calpain activation causes necrotic death in permanent or transient focal cerebral ischemia, while caspase activation (i.e. apoptosis) is only present if reperfusion occurs. This indicates that reperfusion triggers additional injury mechanisms that should be specifically inhibited to ensure effective neuroprotection after reperfusion.