Directory UMM :Data Elmu:jurnal:B:Brain Research:Vol882.Issue1-2.2000:

www.elsevier.com / locate / bres

Research report

Plasminogen activation in experimental permanent focal cerebral

ischemia

a b b a

Thomas Pfefferkorn , Christoph Wiessner , Peter R. Allegrini , Brigitte Staufer ,

a a a a

Milan R. Vosko , Martin Liebetrau , Gundula Bueltemeier , Christian U.A. Kloss ,

a ,

_*

Gerhard F. Hamann

a

¨ ¨

Department of Neurology, Klinikum Grosshadern, Ludwig-Maximilians-Universitat Munchen, Marchioninistr. 15, 81377 Munich, Germany

b

Novartis Pharma Inc., Nervous System Research, Neurodegeneration, Basel, Switzerland Accepted 25 July 2000

Abstract

Background: Previous experimental work using in situ zymography has shown very early increased plasminogen activation in ischemic regions after 3 h of ischemia with and without reperfusion. The objective of the present study was to evaluate the time course and extent of plasminogen activation in long-term permanent focal cerebral ischemia. Material and methods: The middle cerebral artery in male Fisher rats was irreversibly occluded by electrocoagulation. Duration of ischemia was 48, 72, and 168 h. Occlusion was controlled in vivo by MRI at day 2. Plasminogen activation was detected by in situ zymography of 10 mm cryosections with an overlay containing plasminogen and the plasmin substrate caseine. Areas of plasminogen activation were compared to structural lesions (immuno-histochemical loss of microtubule-associated protein 2; MAP 2). Results: Compared to controls, increased plasminogen activation was observed in the basal ganglia and the cortex of the ischemic hemisphere after 48, 72, and 168 h (affected area of basal ganglia: 44.5621.9, 70.162.3 and 66.662.8%, respectively; affected area of cortex: 63.469.8, 67.760.7 and 64.063.7%, respectively). The duration of ischemia had no significant influence on the extent of plasminogen activation. Areas of increased plasminogen activation significantly overlapped with and exceeded areas of MAP 2 loss (P,0.005). Discussion: Permanent focal cerebral ischemia leads to increased plasminogen activation in ischemic regions. This plasminogen activation remains elevated at persistent levels over days. It may contribute to extracellular matrix (ECM) disruption, secondary hemorrhage, and brain edema in subacute stages of ischemic stroke.  2000 Elsevier Science B.V. All rights reserved.

Theme: Disorders of the nervous system

Topic: Ischemia

Keywords: Plasminogen activation; Cerebral ischemia; In situ zymography

1. Introduction increased rate of intracerebral hemorrhage. In an European

study, the rate of symptomatic intracerebral hemorrhage in Intravenous thrombolysis with tissue type plasminogen the t-PA-treated group was 19.8% compared to 6.5% in the activator (t-PA) was shown to improve outcome in acute placebo group [7].

ischemic stroke, when therapy was started in the first 3 h Possible mechanisms of plasmin-mediated intracranial after symptom onset [25]. However, the positive effect of bleeding include extracellular matrix (ECM) degradation early recanalisation may be in part outweighed by an with consecutive vessel leakage, extravasation, and petechial or symptomatic hemorrhage [6]. The ECM components laminin, fibronectin, and collagen IV were shown to be reduced in ischemic regions after focal *Corresponding author. Tel.:149-89-7095-3670; fax: 1

49-89-7095-cerebral ischemia and reperfusion [8]. Using antibodies 3677.

against these ECM components and hemoglobin,

extrava-E-mail address: hamann@brain.nefo.med.uni-muenchen.de (G.F.

Hamann). sation of blood components could be demonstrated around

leaking vessels by immunofluorescence double staining Experiments were performed in male Fisher 344 rats

techniques [9]. (Iffa Credo, France) weighing 220–300 g. Animals were

Plasmin was shown to directly degrade ECM com- housed under standard conditions with free access to rat ponents such as laminin and fibronectin [14,15]. Besides chow and tap water before and after surgery.

direct plasmin effects, activation of matrix metalloprotein- Irreversible occlusion of the right middle cerebral artery ases (MMPs) by plasmin may play a crucial role in was performed as described previously [22,24]. Briefly, ischemia-related ECM degradation. Pro-MMP 9 was animals were anesthetized with 2% isoflurane in a 70 / 30 shown to be converted to its active form (MMP 9, (v / v) nitrous oxide / oxygen mixture and, using an oper-gelatinase B) in the presence of plasmin [17]. In accord- ating microscope (Wild), the right MCA was exposed by a ance with this study, pro-MMP 9 activation was reduced in subtemporal craniectomy. The artery and its lenticulo-plasminogen deficient knock-out mice [13]. In a cell line striate branches were then occluded by bipolar electro-experiment, pro-MMP 2 activation was inhibited by plas- coagulation. Afterwards, retracted soft tissues were re-min inhibitors, suggesting a role of plasre-min in the conver- placed, wounds were sutured, anesthesia was discontinued, sion of pro-MMP 2 to its active form (MMP 2, gelatinase and the rats were put back into their cages. Body

tempera-A) as well [2]. ture was maintained at 378C by means of a rectal probe

The role of MMPs in focal cerebral ischemia has been connected to a heating pad (CMA 150, Carnegie Medicine) addressed in several studies. In experimental focal cerebral during surgery and until animals regained consciousness. ischemia and reperfusion in primates, MMP 2 was in- Thereafter, rectal temperature was checked frequently creased after 1 h of middle cerebral artery occlusion (every 10–15 min) during the following 2 h and, if (MCAO) and remained elevated for at least 7 days [10]. A necessary, it was corrected to 378C using a heating pad similar increase was observed for MMP 9 in subjects with which was placed under the cage. After this, animals were hemorrhagic transformation [10]. A clinical study in returned to their home cages and allowed free access to humans revealed increased MMP 2 and 9 levels in the food and water.

cerebrospinal fluid of patients with acute stroke. These In order to verify successful MCAO and to visualize the levels correlated with the extent of peri-infarct edema [3]. lesioned territory all brains were investigated by NMR In a post-mortem study, MMP 2 and 9 were increased in imaging after 48 h of permanent ischemia. MRI inves-infarcted brain regions 2 days and several months after tigations were performed on a Bruker DBX 47 / 30 (4.7 T) stroke [5]. These studies suggest detrimental effects of equipped with an actively shielded gradient insert (max. MMP 2 and 9 in cerebral ischemia such as brain edema 100 mT / m). RF pulses were transmitted and the MRI 1

and secondary hemorrhage. signal was acquired with a homogeneous birdcage H

Little is known about plasminogen activation in focal resonator (inner diameter, 3.5 cm). The lesion volume was cerebral ischemia. Investigations in mice showed increased determined by means of T weighted quantitative in vivo2 plasminogen activation in infarcted regions after 2 h of MRI [1]. Rats were anesthetized with 1–1.5% isoflurane focal cerebral ischemia [27]. In previous experiments in delivered via a face mask and positioned with their heads rats, plasminogen activation was increased in regions of in the resonator. Each animal was then subjected to one structural damage after 3 h of focal ischemia and different imaging cycle, in which 13 continuous T2 weighted reperfusion intervals [19]. The time course of plasminogen coronal slices of the brain with a thickness of 1.2 mm were activation in subacute stages (days 2–7) of focal perma- taken using a RARE sequence (optimized parameters: nent cerebral ischemia has not yet been investigated. repetition time 3000 ms; effective echo time 66 ms; spatial

2

However, it may play a role in subacute hemorrhage and resolution in plane (156mm) ). The total measuring time edema formation after ischemic stroke. was 5 min.

To investigate plasminogen activation in long-term The animals survived for 48, 72, and 168 h (groups 1, 2, ischemia we used a rat model with permanent MCAO for 3; n54, 5, and 5). Thereafter, animals of group 2 and 3

2–7 days. (animals of group 1 remained anesthetized after MRI) were

anesthetized again and perfused with 50 ml ice-cold physiological saline, containing 1% bovine serum albumin,

2. Materials and methods 4 IE / ml heparin, and 36 mg / ml sodium nitroprusside at 5

ml / min. The brains were removed and frozen in OCT

2.1. Animal model embedding medium (Sakura Finetec) on dry ice and stored

at2808C until further processing. All animal experiments were performed after ethical

approval by the responsible government institution (the 2.2. Preparation of cryostat sections ¨

‘Kantonales Veterinaramt’, Basel, Switzerland, reference

adjacent sections were examined for MAP 2 immuno- acetone and chloroform, washed in phosphate buffer histochemistry and zymographic plasminogen activation. solution (PBS), and incubated with blotto (50 g non-fat dry milk, 10 ml horse serum, 0.3 mmol / l sodium azide 2.3. Zymographic detection of plasminogen activation diluted in Tris–saline stock (38.5 mmol / l Tris, 150 mmol / l NaCl)) to reduce unspecific binding. After repeated The method was described previously [21]. After thaw- washing in PBS, the sections were incubated with the ing and drying at room temperature, sections were incu- secondary biotinylated antibody (horse, anti-mouse, Vector bated with an overlay consisting of commercial non-fat Laboratories, dilution 1:200, incubation period 2 h at milk powder (containing the plasmin substrate casein), 378C). This was followed by blocking endogenous per-agarose, and human plasminogen. The overlay was pre- oxidase with H O for 20 min. Thereafter sections were2 2 pared in several steps. Three different solutions were incubated with the ABC-Complex (Vectastain-Elite-Kit, prepared. Solutions 1 (100 ml PBS124.6 mg MgSO₄1 Vector Laboratories, incubation 30 min at 378C). Per-13.2 mg CaCl ) and 2 (10 ml dH O11.6 g commercial2 2 oxidase signal was developed with AEC (AEC-Kit, non-fat milk powder) were separately heated to 508C, Biomeda). Nuclei were counterstained with Mayer’s hema-solution 3 (10 ml dH O1250 mg agarose) to 708C. The2 toxylin (Sigma), and tissue was blued in saturated sodium solutions were mixed (0.75 ml of solution 1, 0.70 ml of bicarbonate.

solution 2, 0.50 ml of solution 3), and the mixture cooled

to 378C. Then 100 ml of plasminogen (1.5 mg / ml) were 2.5. Morphometric analysis added. This mixture was evenly spread over the heated

sections (378C) and immediately covered by a coverslip. To evaluate areas of visible plasminogen activation and After incubation at 378C for 22 h the sections were decreased MAP 2 antigenicity, a morphometric program immediately inspected for zones of overlay lysis and (Optimas 6.5, Media Cybernetics, MD) was used. Areas of digitally scanned for later computer analysis. The incuba- plasmin-mediated lysis and MAP 2 loss were measured tion period of 22 h was chosen, since plasmin-mediated and expressed as percent of the total ischemic hemisphere. lysis had then reached a maximum. Longer incubation Affected areas were also evaluated separately for the basal periods up to 36 h showed no further expansion of lytic ganglia and the cortex (Fig. 1). To minimize the error of regions. To exclude unspecific overlay lysis, control spatial resolution, adjacent brain sections were evaluated sections of all groups were incubated with an overlay for zymography and MAP 2 immunohistochemistry.

Cor-lacking plasminogen. responding MRI images were selected from the individual

MRI series to illustrate overlapping of MRI changes with 2.4. Immunohistochemistry overlay lysis and MAP 2 loss. This selection was guided

by a rat brain atlas [18]. For MAP 2 detection, a monoclonal mouse antibody

(Boehringer Mannheim, dilution 1:800, incubation for 2 h 2.6. Statistical analysis at 378C, followed by 24 h at 48C) was used. Before

incubation, sections were fixated in a 1:1 mixture of For statistical evaluation, data were presented as means

and standard errors of mean (S.E.M.). To analyse the time minogen, limited areas of ‘background’ overlay lysis were course of overlay lysis and MAP 2 loss non-parametric seen in the non-ischemic (control) hemisphere. This lysis testing (Kruskal–Wallis, Mann–Whitney) was performed. uniformly occurred in two different patterns: first, small, To compare areas of overlay lysis with areas of MAP 2 pin-point-shaped regions with no preference for distinct loss in individual sections the Wilcoxon matched-pairs anatomical structures; secondly, a circular rim of lysis signed-ranks test was used. following pial structures around the section (Fig. 2). In all individual animals, this ‘background’ lysis affected less than 20% of the cortex and less than 10% of the basal

3. Results ganglia of the control hemisphere.

3.1. Lesions detected by MRI 3.3. Plasminogen activation in ischemic regions

In all animals, MRI showed clearly delineated lesions in Additional extensive plasminogen activation beyond the ischemic hemisphere after 48 h of MCAO (Fig. 2). ‘background’ lysis was observed in the basal ganglia and While the cortex was consistently affected in all animals, cortical regions of the ischemic hemisphere as large the basal ganglia were spared in one animal of group 1 (48 coherent zones of overlay lysis (Fig. 2). After 48, 72, and

h of ischemia). 168 h of ischemia, the affected area in the basal ganglia

was 44.5621.9, 70.162.3 and 66.662.8%, respectively 3.2. Plasminogen activation in controls (Fig. 3). The affected area in the cortex measured 63.469.8, 67.760.7 and 64.063.7%, respectively (Fig. 3). Sections without plasminogen in the overlay showed no No significant differences in the time course of permanent zones of lysis (Fig. 1). Using overlays containing plas- ischemia from 48 to 168 h were observed. The relatively

Fig. 3. Areas of MAP 2 loss (left column) and plasminogen activation (right column), presented as percent of total basal ganglia and cortex area of the ischemic hemisphere6standard error of mean. The area of overlap between MAP 2 loss and plasminogen activation is indicated by hatched lines. In the area of MAP 2 loss this overlap is almost complete and significantly higher than expected if areas were randomly distributed (Wilcoxon; P,0.005). The area of plasminogen activation exceeds the area of structural injury (Wilcoxon; *indicates significance at P,0.05 for single groups; lack of significance in the other groups may be due to low animal number; for all groups combined: P,0.005). Duration of ischemia (h) has no influence on the size of affected areas.

low mean with a high standard error of mean in the extent with ischemic intervals between 2 and 7 days, up to 70% of basal ganglia affection after 48 h (44.5621.9%) is of the respective areas showed increased plasminogen caused by exceptional basal ganglia sparing in one animal. activation, which may suggest an expansion of the areas of plasminogen activation between 3 and 48 h. However, this 3.4. Plasminogen activation and MAP 2 demarcation point remains speculative, since different rat strains (Wis-tar versus Fisher rats) and occlusion techniques (intralumi-Areas showing MAP 2 loss were almost completely nal thread versus electrocoagulation) were used in the two affected by increased plasminogen activation (Figs. 2 and studies.

3). Comparing individual sections, this colocalisation was In the present study, one animal of group 1 (48 h of significantly higher than expected if areas were randomly ischemia) showed no major overlay lysis in the basal distributed (Wilcoxon; P,0.005). On the other hand, ganglia. This lack of plasminogen activation corresponded increased plasminogen activation was not merely restricted to basically normal MAP 2 immunohistochemistry and to zones with detectable structural injury (Fig. 2). In the lack of major MRI changes in the basal ganglia. It is basal ganglia and the cortex of the ischemic hemisphere, known from the used rat model, that, regarding MRI the area of increased plasminogen activation exceeded the alterations due to ischemia, the basal ganglia can be spared area of MAP 2 loss (Wilcoxon; P,0.005). Separated for in rare cases [24]. The lack of basal ganglia affection in different time points, this excess was significant for the one animal explains the low mean and high variability for basal ganglia at 72 h and for the cortex after 72 and 168 h area quantification of plasminogen activation and MAP 2

(Wilcoxon; P,0.05; Fig. 3). loss in group 1 (Fig. 3).

In all groups, areas with structural injury (loss of MAP 2 antigenicity) constantly showed increased plasminogen

4. Discussion activation. Beyond that, the area of increased plasminogen

activation exceeded the area of structural injury. This Our results demonstrate increased plasminogen activa- exceeding plasminogen activation was mainly located tion in the basal ganglia and the cortex following perma- around the border zone of the structural injury and may nent focal cerebral ischemia. This increase was limited to therefore simply represent diffusion of proteolytic agents the ischemic hemisphere. Comparing different time inter- (plasmin, plasminogen activators) into the casein overlay. vals of permanent ischemia (48, 72, 168 h), no significant In some cases, however, areas of increased plasminogen differences in the extent of plasminogen activation could activation in the ischemic hemisphere had no contact to

be detected. and were remote from areas of structural injury (as

aggravation of structural injury after focal ischemia may be It may also be indirectly influenced by plasminogen considered. One previous study supports this interpreta- activator inhibitors (plasminogen activator inhibitors 1 and tion: Wang and co-workers found a decrease in infarct size 2; PAI 1 and 2). Supposing stable PA activity, decreased after experimental focal ischemia in t-PA-deficient mice, PAI activity should lead to increased overall plasminogen suggesting negative effects of plasminogen activation on activation. Shatos et al. found decreased levels of PAI 1 structural integrity [27]. However, these findings could not after exogenous exposure to t-PA in human cell line be confirmed by intervention experiments with exogenous experiments, suggesting synergistic effects of increased PA t-PA [11,12,16]. Supposing a negative effect of plas- and decreased PAI activity [23]. Regardless of the exact minogen activation on structural integrity, an expansion of mechanisms and the enzymes involved, a shift towards the structural injury into areas with excess plasminogen increased plasmin activity is likely to affect ECM integrity. activation over time should be expected. In our experi- Our experiments demonstrate this ongoing and stable ments, however, the area of MAP 2 loss and the excess imbalance at subacute stages of permanent ischemia. area of increased plasminogen activation compared to Potential beneficial effects (maintenance of microvascular MAP 2 loss remained stable over the time course of flow, capillary remodeling) may be outweighed by brain

permanent ischemia. edema and cerebral hemorrhage due to PPS-mediated

The areas of increased plasminogen activation were ECM disruption. detected by histological in situ zymography. Using an

overlay lacking plasminogen, no lysis was seen. Overlay

lysis in the presence of plasminogen depended on the _{Acknowledgements} presence of brain tissue. Therefore, overlay lysis

unques-tionably represented endogenous plasminogen activator- _{This study was supported by the German Research} mediated proteolysis. Two limitations of the method have _{Foundation (DFG; No. Ha-2078 / 5-1). The authors thank} to be mentioned: first, in situ zymography does not reveal _{Mrs. Judy Benson for language editing of the manuscript.} information about the degree of activation. Second, in situ _{The authors also thank Dr. Tobias Back for initiation of the} zymography does not allow to differentiate between differ- _{cooperation between Novartis Pharma Inc., Basel,} Switzer-ent PAs. From other studies, limited data regarding the _{land and the Department of Neurology, Klinikum} Gross-roles of different PAs in plasminogen activation following _{hadern, Munich, Germany.}

focal cerebral ischemia are available. A possible role of urokinase type plasminogen activator (uPA) in plas-minogen activation following focal cerebral ischemia was

References

suggested by a study of Rosenberg et al., that found an increase of uPA in the ischemic hemisphere of

sponta-[1] P.R. Allegrini, D. Sauer, Application of magnetic resonance imaging neously hypertensive rats after 12 and 24 h and 5 days of

to the measurement of neurodegeneration in the rat brain: MRI data permanent MCAO but a decrease of t-PA after 12 and 24 h _{correlate strongly with histology and enzymatic analysis, Magn.} [20]. These results have not yet been confirmed in nor- Res. Imag. 10 (1992) 773–778.

[2] E.N. Baramova, K. Bajou, A. Remacle, C. L’Hoir, H.W. Krell, U.H. motensive rats.

Weidle, A. Noel, J.M. Foidart, Involvement of PA / plasmin system As stated in Section 1, increased plasminogen activation

in the processing of MMP-9 and in the second step of pro-to plasmin may effect the ECM in two ways. First, plasmin

MMP-2 activation, FEBS Lett. 405 (1997) 157–162.

directly degrades ECM components. Secondly, it is in- _{[3] C. Braker, S. Shapiro, N. Lahat, E. Sobel, S. Honigman, A. Miller,} volved in the activation of pro-MMP 2 and 9 to their active Matrix metalloproteinases (MMPs) and IL-8 CSF levels are corre-forms, the collagen IV degrading gelatinases A and B. lated with brain edema following stroke (Abstract), J. Neurol. 246

(Suppl. 1) (1999) 10. Besides its ECM disrupting properties with consecutive

[4] Z. Chen, S. Strickland, Neuronal death in the hippocampus is microvascular injury, the PPS may also be involved in

promoted by plasmin-catalyzed degradation of laminin, Cell 91 neuronal damage. This was demonstrated in experiments _{(1997) 917–925.}

with knock-out mice, where t-PA deficiency protected [5] A.W. Clark, C.A. Krekoski, S. Bou, K.R. Chapman, R.E. Edwards, against neuronal damage provoked by excitatory amino Increased gelatinase A (MMP-2) and gelatinase B (MMP-9) ac-tivities in human brain after focal ischemia, Neurosci. Lett. 238 acids [26] and focal ischemia [27]. As in PPS-mediated

(1997) 53–56. microvascular injury, this neuronal damage also seems to

[6] G.J. del Zoppo, R. von Kummer, G.F. Hamann, Ischaemic damage depend on ECM disruption. In experiments with t-PA _{of brain microvessels: inherent risk for thrombolytic treatment in} deficient mice, the blocking of the ECM component stroke, J. Neurol. Neurosurg. Psychiatry 65 (1998) 1–9.

laminin with specific antibodies restored excitotoxic neuro- [7] W. Hacke, M. Kaste, C. Fieschi et al., Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric nal sensitivity. This suggests that the disruption of the

stroke. The European Cooperative Acute Stroke Study (ECASS), J. neuron ECM interaction (by PPS-mediated proteolysis or

Am. Med. Assoc. 274 (1995) 1017–1025.

blocking of laminin) substantially contributes to t-PA- _{[8] G.F. Hamann, Y. Okada, R. Fitridge, G.J. del Zoppo, Microvascular} mediated neuronal vulnerability [4]. basal lamina antigens disappear during cerebral ischemia and

¨

[9] G.F. Hamann, Y. Okada, G.J. del Zoppo, Hemorrhagic transforma- [19] T. Pfefferkorn, B. Staufer, M. Liebetrau, G. Bultemeier, M.R. Vosko, tion and microvascular integrity during focal cerebral ischemia / C. Zimmermann, G.F. Hamann, Plasminogen activation in focal reperfusion, J. Cereb. Blood Flow Metab. 16 (1996) 1373–1378. cerebral ischemia and reperfusion, J. Cereb. Blood Flow Metab. 20 [10] J. Hoe Heo, J. Lucero, T. Abumiya, J.A. Koziol, B.R. Copelands, (2000) 337–342.

G.J. del Zoppo, Matrix metalloproteinases increase very early during [20] G.A. Rosenberg, M. Navratil, F. Barone, G. Feuerstein, Proteolytic experimental focal cerebral ischemia, J. Cereb. Blood Flow Metab. cascade enzymes increase in focal cerebral ischemia in rat, J. Cereb.

19 (1999) 624–633. Blood Flow Metab. 16 (1996) 360–366.

[11] E. Kilic, D.M. Hermann, K.A. Hossmann, Recombinant tissue [21] A. Sappino, R. Madani, J. Huarte, D. Belin, J.Z. Kiss, A. Wohlwend, plasminogen activator reduces infarct size after reversible thread J. Vassalli, Extracellular proteolysis in the adult murine brain, J. occlusion of middle cerebral artery in mice, Neuroreport 10 (1999) Clin. Invest. 92 (1993) 679–685.

107–111. [22] D. Sauer, P.R. Allegrini, A. Cosenti, A. Pataki, A. Amacker, G.E. [12] G.M. Klein, H. Li, A.M. Buchan, Tissue plasminogen activator does Fagg, Characterization of the cerebroprotective effect of the com-not increase neuronal damage in rat models of global and focal petitive NMDA receptor antagonist CGP 40116 in a rat model of ischemia, Neurology 52 (1999) 1381–1384. focal cerebral ischemia: an in vivo magnetic resonance study, J. [13] H.R. Lijnen, J. Silence, G. Lemmens, L. Frederix, D. Collen, Cereb. Blood Flow Metab. 13 (1993) 595–602.

Regulation of gelatinase activity in mice with targeted inactivation [23] M.A. Shatos, J.M. Doherty, P.L. Penar, B.E. Sobel, Suppression of of components of the plasminogen / plasmin system, Thromb. Haem- plasminogen activator inhibitor-1 release from human cerebral ost. 79 (1998) 1171–1176. endothelium by plasminogen activators. A factor potentially pre-[14] L.A. Liotta, R.H. Goldfarb, R. Brundage, G.P. Siegal, V. Terranova, disposing to intracranial bleeding, Circulation 94 (1996) 636–642. S. Garbisa, Effect of plasminogen activator (urokinase), plasmin, [24] A. Tamura, D.I. Graham, J. McCulloch, G.M. Teasdale, Focal and thrombin on glycoprotein and collagenous components of basal cerebral ischemia in the rat: 1. Description of technique and early membrane, Cancer Res. 41 (1981) 4629–4636. neuropathological consequences following middle cerebral artery [15] L.A. Liotta, R.H. Goldfarb, V.P. Terranova, Cleavage of laminin by occlusion, J. Cereb. Blood Flow Metab. 1 (1981) 53–60.

thrombin and plasmin: alpha thrombin selectively cleaves the beta [25] The National Institutes of Neurological Disorders and Stroke rtPA chain of laminin, Thromb. Res. 21 (1981) 663–673. Stroke Study Group, Tissue plasminogen activator for acute is-[16] W. Meng, X. Wang, M. Asahi, T. Kano, K. Asahi, R.H. Ackermann, chemic stroke, New Engl. J. Med. 333 (1995) 1581–1587.

E.H. Lo, Effects of tissue type plasminogen activator in embolic [26] S.E. Tsirka, A. Gualandris, D.G. Amaral, S. Strickland, Excitotoxin-versus mechanical models of focal cerebral ischemia in rats, J. induced neuronal degeneration and seizure are mediated by tissue Cereb. Blood Flow Metab. 19 (1999) 1316–1321. plasminogen activator, Nature 377 (1995) 340–344.

[17] Y. Okada, Y. Gonoji, K. Naka, K. Tomita, I. Nakanishi, K. Iwata, K. [27] Y.F. Wang, S.E. Tsirka, S. Strickland, P.E. Stieg, S.G. Soiano, S.A. Yamahita, T. Hayakawa, Matrix metalloproteinase 9 (92-kDa Lipton, Tissue plasminogen activator (tPA) increases neuronal gelatinase / type IV collagenase) from HT 1080 human fibrosarcoma damage after focal cerebral ischemia in wild-type and tPA-deficient cells. Purification and activation of the precursor and enzymic mice, Nat. Med. 4 (1998) 228–231.

properties, J. Biol. Chem. 267 (1992) 21712–21719.