2

Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in Medicinepresented at Uppsala University in 2002.

ABSTRACT

Lennmyr, F. 2002. Signal Transduction in Focal Cerebral Ischemia – Experimental Studies onVEGF, MAPK and Src family kinases. Acta Universitatis Upsaliensis. ComprehensiveSummaries of Uppsala Dissertations from the Faculty of Medicine 1133. 55pp. Uppsala.ISBN 91-554-5267-1.

Cerebral ischemia elicits a wide range of events, including complex activation of variousintracellular signaling pathways. This study aims to investigate the expression of vascularendothelial growth factor (VEGF) and the activation pattern of mitogen-activated proteinkinases (MAPK) in reponse to focal cerebral ischemia. Furtermore, the functional roles of thep38 MAPK and the Src family kinases (SFKs) are investigated with specific signaltransduction inhibitors in the rat in vivo.VEGF was found upregulated in several cell types including neurons, glia and vascular cellsafter both permanent and transient cerebral ischemia. VEGF-receptor 1 (VEGFR1) wasexpressed in a similar manner, while VEGFR2 expression was more restricted and confined toendothelial cells and glia.The main MAPK pathways, including extracellular-regulated kinase(ERK), c-jun N-terminal kinase (JNK) and p38, were differentially activated by cerebralischemia. ERK activation was present in blood vessels, suggesting a potential role inneovascularization. JNK was also activated in blood vessels in the infarcted hemisphere,possibly reflecting an interaction with ERK, whereas p38 activity was absent in vessels. Inneurons, ERK was activated in cortical cells up to days of survival, while no substantial JNKor p38 activation was seen in ischemic neurons. Invading macrophages showed distinctactivation of p38 and to some extent also JNK but not ERK. Glia showed activation of allMAPK to a variable extent.Pretreatment with the p38-inhibitor SB203580 before transient cerebral ischemia (ischemia-reperfusion) was investigated with magnetic resonance imaging (MRI). The experiment groupsuffered worse infarcts and blood-brain barrier (BBB) damage than controls, which contraststo previous studies. The results might be attributed to interference with protective effects ofthe vehicle or with preconditioning mechanisms. The SFK-inhibitor PP2 significantly reducedinfarct size after cerebral ischemia-reperfusion, which is consistent with previously reportedeffects in permanent ischemia. Due to the multifunctional role of SFKs, it cannot be easilyconcluded in exactly what cellular context(s) SFKs are of importance to cerebral ischemia.In conclusion, the VEGF and MAPK systems of extra- and intracellular signaling areactivated in focal cerebral ischemia. Manipulation of p38 as well as SFK in vivo can influencethe course of transient cerebral ischemia, which may be of significance to the understandingof the pathology of cerebral ischemia and to the development of therapeutic strategies.

Fredrik Lennmyr, Department of Medical Sciences, Uppsala University, University Hospital,SE-751 85 Uppsala, Sweden

© Fredrik Lennmyr 2002

ISSN 0282-7476ISBN 91-554-5267-1

Printed in Sweden by Akademitryck, Edsbruk 2002

3

4

ORIGINAL PAPERS

This doctoral thesis is based on the following papers

I Fredrik Lennmyr, Khaled Ahmad Ata, Keiko Funa, Yngve Olsson,

Andreas Terént (1998) Expression of Vascular Endothelial Growth Factor

and its Receptors (Flt-1 and Flk-1) following Permanent and Transient

Occlusion of the Middle Cerebral Artery in the Rat. J Neuropathol Exp

Neurol 57(9):874-82

II Fredrik Lennmyr, Pär Gerwins, Khaled Ahmad Ata, Sofia Karlsson,

Andreas Terént (2002) Activation of Mitogen-activated Protein Kinases

in Experimental Cerebral Ischemia. Acta Neurologica Scandinavica (in

press)

III Fredrik Lennmyr, Anders Ericsson, Pär Gerwins, Håkan Ahlström,

Andreas Terént (2002) Increased brain injury and vascular leakage after

pretreatment with p38-inhibitor SB203580 in transient ischemia.

Manuscript

IV Fredrik Lennmyr, Anders Ericsson, Pär Gerwins, Susanne Akterin, Håkan

Ahlström, Andreas Terént (2002) Src family kinase-inhibitor PP2 reduces

focal ischemic brain injury. Submitted

5

TABLE OF CONTENTS

ORIGINAL PAPERS ______________________________________________________________________ 4

ABBREVIATIONS________________________________________________________________________ 6

INTRODUCTION ________________________________________________________________________ 7

CEREBRAL ISCHEMIA _____________________________________________________________________ 7VASCULAR ENDOTHELIAL GROWTH FACTOR (VEGF)_____________________________________________ 8

General aspects ______________________________________________________________________ 8VEGF in cerebral ischemia _____________________________________________________________ 9

MITOGEN-ACTIVATED PROTEIN KINASES (MAPK) _______________________________________________ 9General aspects ______________________________________________________________________ 9MAPK in cerebral ischemia ____________________________________________________________ 10

SRC FAMILY KINASES ____________________________________________________________________ 14General aspects _____________________________________________________________________ 14Src family kinases in cerebral ischemia ___________________________________________________ 16

AIMS __________________________________________________________________________________ 17

METHODS _____________________________________________________________________________ 18

MODEL OF CEREBRAL ISCHEMIA____________________________________________________________ 18IMMUNOHISTOCHEMISTRY (PAPER I AND II) __________________________________________________ 20SIGNAL TRANSDUCTION INHIBITORS (PAPER III AND IV) _________________________________________ 21

SB203580 – a p38-inhibitor ____________________________________________________________ 21PP2 – an SFK-inhibitor _______________________________________________________________ 21

MAGNETIC RESONANCE IMAGING (PAPER III AND IV) ___________________________________________ 22General comments ___________________________________________________________________ 22Diffusion-weighted imaging ____________________________________________________________ 23T2-weighted imaging _________________________________________________________________ 23T1-weighted dynamic contrast enhanced imaging ___________________________________________ 23

RESULTS AND DISCUSSION _____________________________________________________________ 26

PAPER I ______________________________________________________________________________ 26VEGF expression in different cell types ___________________________________________________ 26VEGF-receptor findings _______________________________________________________________ 28

PAPER II______________________________________________________________________________ 29Vascular activation of ERK and JNK _____________________________________________________ 29Neuronal ERK activation ______________________________________________________________ 31Activation of JNK and p38 _____________________________________________________________ 32From descriptive to functional studies ____________________________________________________ 33

PAPER III _____________________________________________________________________________ 33Brief discussion of the background_______________________________________________________ 33Evaluating MCAO with DWI ___________________________________________________________ 34Augmented ischemic injury by SB203580__________________________________________________ 36The role of p38 in cerebral ischemia _____________________________________________________ 37

PAPER IV _____________________________________________________________________________ 38Reduced infarct size with PP2 treatment – a matter of vascular permeability? _____________________ 38Possible involvement of SFKs in other injury mechanisms _____________________________________ 39

CONCLUSIONS_________________________________________________________________________ 41

SAMMANFATTNING PÅ SVENSKA _______________________________________________________ 42

ACKNOWLEDGMENTS _________________________________________________________________ 43

REFERENCES __________________________________________________________________________ 45

6

ABBREVIATIONS

ADC Apparent diffusion coefficient

BBB Blood-brain barrier

FGF Fibroblast growth factor

DMSO Dimethylsulfoxide

DWI Diffusion-weighted imaging

EGF Endothelial growth factor

ERK Extracellular-regulated kinase

IL-1 Interleukin 1

JNK c-jun-N-terminal kinase

MAPK Mitogen-activated protein kinase

MCAO Middle cerebral artery occlusion

MRI Magnetic resonance imaging

NGF Nerve growth factor

NMDA N-methyl-D-aspartate

PDGF Platelet derived growth factor

PMCAO Permanent middle cerebral artery occlusion

ROS Reactive oxygen species

SAPK Stress-activated protein kinase, synonymous to JNK

SH-domain Src homology domain

TGF-ß Transforming growth factor beta

TMCAO Transient middle cerebral artery occlusion

TNF-α Tumor necrosis factor alpha

T1, T2WI T1, T2-weighted imaging

VEGF Vascular endothelial growth factor

VEGFR1, 2 Vascular endothelial growth factor receptor 1, 2

VP Vascular permeability

7

INTRODUCTION

Cerebral ischemia

Cerebral ischemia plays a central role in various types of brain injury, either as the

primary cause (e.g. ischemic stroke, cardiac arrest) or as an aspect of other diseases

(e.g. hemorrhage, malignancy, infection, surgery or trauma). This thesis will

concentrate on primary focal cerebral ischemia.

Stroke is one of the leading causes to global mortality and importantly also to long-

and short-term morbidity with associated individual suffering and considerable costs

(Bonita R et al. 1994). The stroke incidence may be reduced by primary and secondary

prevention, where research has been relatively successful in developing effective drugs

targeting e.g. blood pressure, platelets and the coagulation system (Hankey GJ and

Warlow CP 1999). However, the attempts to design treatment for cerebral ischemia in

the acute stage has been less successful, and today the medical therapy is directed

towards optimized physiology and in selected cases also thrombolysis (Hacke W et al.

1995).

Even though reperfusion, spontaneous or therapeutic, offers a possible salvage to the

ischemic tissue, it is likely to be associated with specific injury mechanisms (Jean WC

et al. 1998). In fear of malignant brain edema due to reperfusion of large ischemic

regions, patients with widespread ischemic changes have been excluded from some

clinical studies on thrombolysis (Hacke W et al. 1995), although the circumstances

leading to brain edema are still not completely understood. According to Igor Klatzo in

1967, the mechanisms for this brain edema may be divided into a cytotoxic (cellular)

and a vasogenic part (Klatzo I 1967). The former is commonly considered to

correspond to the acute swelling of astrocytes while the latter depends on a loss of

blood-brain barrier (BBB) integrity for plasma constituents, undermining the

physiological colloid-osmotic mechanisms.

In experimental stroke models, several drugs have been effective in reducing ischemic

brain injury, but the translation of these results into clinical practice has thus far been

8

disappointing. The reasons for this have been debated, and common arguments include

poor validity of animal models for human stroke, heterogeneity among included

subjects and inappropriate therapeutic time windows. Nevertheless, it must be

concluded we currently lack sufficient knowledge to provide effective medical

treatment for cerebral ischemia. Therefore, the underlying mechanisms need further

investigations in order to understand the pathology, find new therapeutic possibilities

and to be able to adequately design clinical studies.

Vascular endothelial growth factor (VEGF)

General aspects

VEGF is an approximately 45-kD protein, whose effects on endothelial proliferation

were described in 1989 (Ferrara N and Henzel WJ 1989). The existence of a protein

with ability to increase vascular permeability had been recognized a few years earlier,

and this protein was denominated vascular permeability factor (VPF) (Senger DR et al.

1986). Soon after the characterization of VEGF, it was discovered that it was identical

to VPF. Both acronyms can be found in the literature, however in this thesis,

VEGF/VPF will be referred to as VEGF.

Structurally, VEGF is a homodimeric protein with different isoforms, differing in size

and kinetic behavior. Evidence indicates that the different kinetic properties not only

affect the distribution of the ligand, but also modulate the ligand-receptor binding

(Gitay-Goren H et al. 1996). The isoforms are derived through alternative splicing of

the VEGF mRNA (Neufeld G et al. 1996), and even if human and rodent VEGF are

nearly homologous, the rodent isoforms are generally 1 amino acid shorter, and there

is no known analogue to the largest human isoform (VEGF206). The characteristics of

the main isoforms are summarized in table 1.

VEGF has been described as the main direct-acting endothelial mitogen along with the

angiopoietin family (Ferrara N and Alitalo K 1999). The principal receptor types for

VEGF are VEGFR1 (or Flt-1) and VEGFR2 (or KDR in humans, Flk-1 in rodents),

which both are tyrosine kinase receptors of approximately 180 and 200-220 kDa,

9

mediating different cellular effects. Most cellular effects of VEGF, including

proliferation and vascular permeability, have been associated with VEGFR2, although

there is evidence for biological roles for VEGFR1 as well (Clauss M et al. 1996;

Shibuya M et al. 1999).

_____________________________________________________________________

Table 1. VEGF isoforms.

Human Rodent Exons Kinetics

VEGF121 VEGF120 1-5, 8 Soluble

VEGF165 VEGF164 1-5, 7, 8 Soluble/heparin-bound

VEGF189 VEGF 188 1-5, 6a, 7, 8 Heparin-bound

VEGF206 not known 1-5, 6ab, 7, 8 Heparin-bound

_____________________________________________________________________

VEGF in cerebral ischemia

Brain ischemia upregulates several growth factors/cytokines, such as IL-1 (Minami M

et al. 1992), PDGF (Iihara K et al. 1994; Krupinski J et al. 1997), TGF-ß (Krupinski J

et al. 1996; Ata KA et al. 1999) and TNF-α (Liu T et al. 1994), which to a variable

extent have been characterized as mediators in cerebral ischemia. VEGF, with its dual

properties (mitogen and VP), might be of particular interest, since vascular leakage

with consequent vasogenic edema is a commonly accepted injury mechanism in

cerebral ischemia. On the other hand, re-establishment of the circulation through

angiogenesis might benefit the recovery of the injured brain. There is evidence of

angiogenesis in human brain infarcts (Krupinski J et al. 1994), and VEGF has been

proposed as a favorable mediator in this process (Marti HJ et al. 2000). Thus, there

may be alternate roles for VEGF in cerebral ischemia, which motivated an

investigation with immunohistochemistry at the ligand-receptor level (Paper I).

Mitogen-activated protein kinases (MAPK)

General aspects

Mitogen-activated protein kinases (MAPK) is an evolutionary conserved family of

signal transduction proteins, which regulate differentiation, proliferation and

10

survival/apoptosis. The principal MAPK pathways include ERK (Boulton TG et al.

1990), JNK (Kyriakis JM and Avruch J 1990) and p38 (Brewster JL et al. 1993; Han J

et al. 1994), which all share a common pattern of activation, but differ in activating

stimuli and downstream substrates. Characteristically, the MAPKs are activated by the

MEK (MAPKK) family, which in turn is activated by MEKK (MAPKKK), as

illustrated in figure 1. MAPK are activated by phosphorylation of a conserved motif

(Thr-X-Tyr) in the activation loop (T-loop), and the MAPK themselves exert Ser/Thr

kinase activity on downstream substrates, which include both nuclear and cytosolic

structures. These are recognized by the MAPK through defined structural properties,

especially the proline residue located immediately C-terminally (+1) from the Ser/Thr

position of known substrates of all 3 pathways. For review, see Schaeffer and Weber

(Schaeffer HJ and Weber MJ 1999).

ERK typically reponds to growth factor stimulation (Boulton TG et al. 1991), and

participates in proliferation and differentiation signaling, depending on the

circumstances (Marshall CJ 1995). The ERK pathway is also involved in promoting

cell survival, thereby counteracting JNK and p38, which are implicated in apoptosis

(Xia Z et al. 1995) and inflammation (Lee JC et al. 1994), and respond to

environmental stress and inflammatory cytokines (Raingeaud J et al. 1995).

MAPK in cerebral ischemia

Ischemia provides many putative stimuli for MAPK activation such as stress,

cytokines and growth factors. Several studies have addressed the functions of MAPK

in neuronal cells, e.g. PC12-cells, in vitro and the accumulated body of information is

substantial. ERK may mediate either proliferation or differentiation dependent on

temporal activation aspects (Marshall CJ 1995). ERK also promotes cell survival in

contrast to JNK and p38 which have been associated with apoptosis in PC12-cells (Xia

Z et al. 1995). On the other hand, ERK activation may not be critical for cell survival

(Creedon DJ et al. 1996), and there is evidence that JNK, not ERK, signals

proliferation upon prolactin stimulation in PC12-cells (Cheng Y et al. 2000).

11

The complexity of MAPK signaling, as exemplified above, indicates a need to

investigate the distribution of activated MAPK in vivo, in order to improve the

understanding of these pathways. There are few descriptive studies on the distribution

of MAPK in cerebral ischemia, and those were carried out with different animal

12

models, time windows and evaluation methods. It has been shown that p38 is activated

in microglia days after global ischemic insult, however no ERK or JNK activation

could be demonstrated at this stage (Walton KM et al. 1998). In a similar model of

global ischemia, all 3 MAPK pathways were found activated after hours in the

hippocampus (Shioda S et al. 1998). JNK has also been suggested as a mediator of

apoptosis in transient focal ischemia, although the cellular localization was not

simultaneously determined (Herdegen T et al. 1998). However, it has been suggested

by others that JNK is activated in neurons and glia after cerebral ischemia (Hayashi T

et al. 2000; Wu DC et al. 2000). ERK activation has been reported in the cortical

neurons after ischemia (Alessandrini A et al. 1999; Irving EA et al. 2000), and ERK

has been suggested to mediate ischemic brain damage (Rundén E et al. 1998;

Alessandrini A et al. 1999), which contrasts to its supposed role in promoting cell

survival (Xia Z et al. 1995).

Summarizing available data on MAPK activation in cerebral ischemia (table 2), some

aspects still appear incomplete from a descriptive point of view. The studies rarely

include simultaneous observations of both permanent and transient ischemia and the

observation times generally comprise early or late stages, rather than spanning over a

longer period. Although the MAPK pathways theoretically may be involved in the

ischemic injury process by regulation of the BBB, activation in vascular cells has not

been described. Regarding the suggested roles of JNK and p38, particularly JNK has

been poorly substantiated from a descriptive perspective. It therefore seemed relevant

to perform an immunohistochemical study of the distribution of the activated MAPK

in transient and permanent ischemia over a relatively extended period of time (6 hours-

1 week) (Paper II).

13

_____________________________________________________________________

Table 2. Survey over descriptive studies on MAPK in cerebral ischemia.

First author Model ERK JNK p38 Time Comment

Walton –98 global no no µ-glia 4d

Shioda –98 global + (N) + (N) + (N) 3h-1w Hippocampal cell extracts

Herdegen -98 TMCAO + 3h-5d Cell type not determined

Alessandrini -99 TMCAO + (N) 1d-3d Cortical neurons; PD protective

Sugino –00 global + (N) + + 15min-3d Nerve fibers; SB protective

Irving –00 T/PMCAO + (N, G) + (N, G) 6h-1d ERK oligo-, p38 astroglia

Wu –00 PMCAO + (N, G) + (N, G) + (N, G) 5min-1d Astroglia

Hayashi –00 TMCAO + (N) 1h-3d JNK-1

Legos –01 PMCAO + 15min-5d SB239063 protective

N=neurons, G=glia, PD=PD98059, SB=SB203580, min=minutes, h=hours, d=days, w=week

_____________________________________________________________________

The understanding of the function of different MAPK pathways in cerebral ischemia

have been facilitated by the use of signal inhibitors administered in vivo, but the

findings are diverse and the understanding is still limited. ERK activation can be

experimentally inhibited by adminstration of the MEK-inhibitor PD98059, which has

been demonstrated to reduce transient focal ischemic injury (Alessandrini A et al.

1999), whereas in global ischemia it was ineffective in reducing delayed neuronal

death (DND) (Sugino T et al. 2000). Interestingly, the p38-inhibitor SB203580 was

found ineffective in reducing focal injury, but effective in limiting DND (Alessandrini

A et al. 1999; Sugino T et al. 2000). Other studies demonstrate that inhibition of p38

by novel agents may reduce brain injury after focal permanent ischemia (Barone FC et

al. 2001; Legos JJ et al. 2001). The functional role of JNK is even more uncertain, due

to lack of specific inhibitors. However, it has been suggested that administration of

pituitary adenylate cyclase-activating polypeptide (PACAP) may protect hippocampal

cells from JNK-mediated apoptosis through increased IL-6 levels (Shioda S et al.

1998).

In endothelial cells, ERK and p38 may respond to VEGF stimulation (Rousseau S et

al. 1997; Pedram A et al. 1998; Gupta K et al. 1999), and p38 has been suggested as a

14

point of convergence for both VEGF- and TNF-α-induced VP (Clauss M et al. 2001),

presumably by mediating cytoskeletal changes (Kato Y et al. 2001). Both VEGF and

TNF-α have been implicated as mediators of blood-brain barrier breakdown in

cerebral ischemia (Yang GY et al. 1999; Zhang ZG et al. 2000). Based on the theories

of p38 as a mediator of VEGF-induced VP, and the possible influence of p38 on

infarct size, an experiment was designed where the p38-inhibitor SB203580 was given

prior to transient MCAO. The ischemic injury was evaluated with serial MRI in order

to determine infarct size and vascular leakage (Paper III).

Src family kinases

General aspects

The Src family protein tyrosine kinases (SFK) participate in multiple intracellular

pathways, and the details and biological context of this family was outlined in a review

by Thomas and Brugge (Thomas SM and Brugge JS 1997). Briefly, the prototype

member of this family, Src (v-Src), was identified in the oncogenic Rous Sarcoma

Virus in the late 1970’s, and it was subsequently characterized as a protein tyrosine

kinase. There are at least eleven known SFK members, of which some are ubiquitously

expressed (e.g. Src, Fyn, Yes) and some are more tissue-specific (e.g. Hck, Lyn, Lck).

SFKs consist of six functional subunits, largely sharing homology (Src homology, SH-

domains), except for the unique domain, which is situated between the SH3 and the N-

terminal SH4 domains, and distinguishes the members. SFKs lack surface receptor

properties, but can link to the cell membrane and other structures to interact with e.g.

integrin receptors, receptor tyrosine kinases and G-protein coupled receptors, and the

downstream targets include structural elements of the cell, receptors and nuclear

motifs. The SFKs themselves are regulated by phosphorylation of specific tyrosine

residues, which alters the catalytic activity and the affinity for substrates.

In the brain, Src is involved in the regulation of NMDA-receptor channel (Yu XM et

al. 1997), and Fyn regulates hippocampal function as well as myelination in the mouse

(Umemori H et al. 1994; Miyakawa T et al. 1996). Evidence also suggest that SFK

15

control neuronal cell structure in terms of NGF- and Ca2+-induced neurite outgrowth in

vitro (Rusanescu G et al. 1995). Examples of the interaction between SFKs and other

structures are schematically illustrated in figure 2.

16

Src family kinases in cerebral ischemia

Among the wide range of identified injury mechanisms in cerebral ischemia, the

multifunctional SFKs may theoretically be integrated in several pathways, exemplified

by inflammatory damage (Lowell CA et al. 1996; Mocsai A et al. 2000), reactive

oxygen species (Nakamura K et al. 1993) and NMDA-receptor-mediated Ca2+ toxicity

(Yu XM et al. 1997). Furthermore, SFKs may participate in the regulation of cell

survival and apoptosis (Thomas SM and Brugge JS 1997). Although injurious effects

of SFKs have not been established after transient cerebral ischemia, there is evidence

of Src and the NMDA-receptor interaction after global ischemia (Pei L et al. 2000; Pei

L et al. 2000; Liu Y et al. 2001). However, there is one study on permanent focal

ischemia, where Src-deficient or SFK-inhibitor-treated mice were less susceptible to

ischemia (Paul R et al. 2001). The same investigators had previously described that

Src mediates VEGF-induced VP (Eliceiri BP et al. 1999), and analogously, this was

proposed to be the mechanism behind Src-mediated ischemic brain injury (Paul R et

al. 2001). Based on those findings, it seemed appropriate to evaluate the therapeutic

potential of SFK inhibition in a model of transient focal ischemia with MRI, with

respect to both infarct size and vascular leakage (Paper IV).

17

AIMS

Paper I

-To investigate the cellular sources and temporal pattern of VEGF and VEGF-receptor

expression in the time window of 1 to 3 days after transient and permenant MCAO.

Paper II

-To examine the cellular pattern of MAPK activation 6 hours to 1 week after transient

and permanent MCAO.

Paper III

-To measure the effects on infarct size and vascular leakage of the MAPK p38-

inhibitor SB203580 in transient MCAO up to 4 days of survival.

Paper IV

-To explore the effects of the SFK-inhibitor PP2 on infarct size and vascular leakage 1

day after transient MCAO.

18

METHODS

Model of cerebral ischemia

There are several models of cerebral ischemia using various animals. Rats and mice

are probably the most widespread model animals for focal cerebral ischemia, although

there is a number of studies that use rabbits, monkeys and other animals as well.

Presumably due to the development of transgenic models, mice are becoming

increasingly popular, however these animals may require more delicate microsurgery

because of their small size.

In the rat, two common methods to produce middle cerebral artery occlusion (MCAO)

can be distinguished. In the method described by Tamura et al in 1981, the middle

cerebral artery (MCA) is occluded through a subtemporal craniotomy (Tamura A et al.

1981). Another approach, which was described by Koizumi et al in 1986 (Koizumi J et

al. 1986) and by Longa et al in 1989 (Longa EZ et al. 1989), is to occlude the MCA

with an intraluminal filament inserted through the carotid vessels. The latter technique,

which was used throughout this work, is illustrated in figure 3. The experimental

designs for the different papers are shown in table 3.

_____________________________________________________________________

Table 3. Experimental designs.

Cerebral ischemia Observation times Evaluation

Paper I TMCAO (2 h), PMCAO 1 - 3 days Immunohistochemistry

Paper II TMCAO (2 h), PMCAO 6 h – 1 week Immunohistochemistry

Paper III TMCAO (90 min) onset – 4 days MRI

Paper IV TMCAO (90 min) 1 day MRI, TTC, neurology

TTC=triphenyl-tetrazolium-chloride staining_____________________________________________________________________

19

Briefly, the left carotid region was exposed through a cervical midline incision, after

which the common carotid artery was separated from the vagal nerve (figure 3).

Therafter, the external carotid artery (ECA) and its small branches were coagulated

and divided. After closing the pterygopalatine artery (PPA) branch of the internal

carotid artery (ICA) with a silk ligature, the bifurcation was clamped and a 3/0

filament (Ethilon, Eticon) was inserted through the ECA. The filament was gently

advanced into the ICA until the feeling of slight resistance, indicating position at the

origin of the MCA. The filament was secured with a silk ligature during MCAO, or

coagulated at the place of insertion in the ECA in case of permanent MCAO. In paper

I and II, the filament was removed after 2 hours of MCAO and the ligature of the PPA

was left in place. In paper III and IV, the duration of the MCAO was 90 min, and the

20

PPA ligature was untied before wound closure. For anesthesia, intraperitoneal fentanyl

and fluanisone (HypnormTM Janssen) was used together with midazolam (Dormicum

® Roche).

Immunohistochemistry (Paper I and II)

Standard procedures for immunohistochemistry were used in paper I and II. The rats

were perfusion-fixed with cold formaldehyde solution. Samples were taken from

anatomically defined rostrocaudal levels, embedded into paraffin, cut into 5µm thin

sections and treated with different methods to expose antigens (enzymes, microwave

treatment). Commercially available antibodies with specificity as described by the

manufacturer were used to label the desired antigens. Both monoclonal (phospho-

ERK, phospho-JNK) and polyclonal antibodies were used. For some (VEGF, Flt-1)

but not all (Flk-1, phospho-p38) polyclonal antibodies, blocking peptides were also

available and used to show specificity.

The primary antibody was labeled with a biotinylated secondary antibody of

appropriate origin. Using avidin-biotin-peroxidase complex and development with

diaminobenzidine in diluted hydrogen peroxide solution, the molecular complex

becomes brown-colored and possible to localize histologically in the light microscope.

Background staining was reduced as much as possible by minimizing the antibody

concentrations, and various blocking procedures were carried out in the different

staining protocols to eliminate unspecific immunoreaction. The degree of

immunostaining was classified by semi-quantitation, where each staining was assigned

a discrete value ranging from 0 (no reactive cells) to 3 (high frequency of reactive

cells). The observations were made in a blinded manner by different investigators, and

discrepancies were discussed for homogeneity.

In paper I, the frequencies of immunopositive cells (neurons, glia and vascular cells)

were categorized as described above. The findings were also topographically divided

into infarct core, border zone as visualized with microtubuli-associated protein 2

(MAP-2) immunostaining, and contralateral hemisphere. The MCAO groups were

then compared with controls and tested for statistical significance using the non-

21

parametric Mann-Whitney U-test. In similarity, the assessment of immunostaining in

paper II was also based on the relative frequencies of positive cells, however overall

assessments were formed for the different groups and no statistical analysis was

performed.

Signal transduction inhibitors (Paper III and IV)

SB203580 – a p38-inhibitor

For experimental inhibition of p38 there are several known inhibitors. These share the

pyridylimidazole structure exemplified by 4-(4-Fluorophenyl)-2-(4-

methylsulfinylphenyl)-5-(4-pyridyl)1H-imidazole (SB203580), and preferentially

inhibit the p38 α and ß but not the other isoforms. The inhibitors bind to the ATP-

binding site of p38, and the selectivity seems to depend on the Thr106 residue present

in the affected but not in the other isoforms (Wilson KP et al. 1997). SB203580 (figure

4) was selected on basis of experiences in experimental in vivo applications in other

studies (Alessandrini A et al. 1999; Sugino T et al. 2000), where the drug was injected

into the lateral brain ventricle prior to ischemia as in paper III. The substance was

purchased from a commercial source, whose documentation contradicted simultaneous

ERK and JNK inhibition up to 100µM, which was the concentration used in paper III.

The potency of the p38-inhibitors is commonly expressed in terms of suppressive

effects on cytokine induction, and although more potent inhibitors than SB203580

have been developed during recent years, the access to those is yet limited.

PP2 – an SFK-inhibitor

There are two common and structurally related SFK-inhibitors, PP1 and PP2, of which

we used PP2, 4-Amino-5-(4-chlorophenyl)-7-( t-butyl)pyrazolo[3,4-d]pyrimidine, in

paper IV. The chemical structure is shown in figure 4. PP2 inhibits multiple SFK

members, and was originally characterized in a study on T-cells (Hanke JH et al.

1996). The selectivity of PP1 and PP2 is likely to depend on a specific residue

corresponding to Thr338 of Src. This structure that can be found also in other kinases,

indicating possible overlapping effects on other pathways. However, the IC50-values

suggest that the effects of PP1 on e.g. p38 are moderate compared with the effects on

22

SFKs (Liu Y et al. 1999). The intraperitoneal administration route and dosage used in

paper IV were similar to another study in mice by Paul et al (Paul R et al. 2001).

O

CH3

F

N

N

NHS

SB203580

Cl

NN

N

NH2

CH3H C3

H C3

PP2

Figure 4. Inhibitors of p38 (SB203580) and Src family kinases (PP2).

Magnetic resonance imaging (Paper III and IV)

General comments

MRI has become a routine method in imaging the central nervous system in humans.

Basically, a strong magnetic field is used to weakly magnetize the tissue, and by using

radio-frequency (RF) pulses and magnetic field gradients, the nuclear moments (spins)

can be altered and recorded, allowing anatomical information to be derived. In medical

applications, the nuclear spins commonly refer to protons (hydrogen atom nuclei). The

relaxation times T1 and T2 reflect the microdynamics of water in tissue, and the

standard procedures in MRI include T1- and T2-weighted imaging (T1WI, T2WI), as

well as diffusion-weighted imaging (DWI) sequences. T1WI is sensitive to the

remagnetization of the tissue, e.g.after a RF-pulse, while T2WI is sensitive to the

decline in magnetization, viewed from a perspective angular to the magnetic field

direction. DWI is sensitive to changes in the refocusing of the magnetic moments,

which occurs in after reversing the magnetic moments by 180° after the initial RF

pulse in a spin echo sequence with additional sensitizing field gradients. These

23

changes depend on the mobility of the protons (water diffusion), and processes that

limit this mobility, e.g. acute ischemia, consequently influence the signal. Quantitative

information can be deducted from the DWI by constructing maps of the “apparent”

diffusion coefficient (ADC). Furthermore, the signal intensity can be manipulated by

administration of contrast agents, which are commonly used to measure perfusion, but

also to give anatomical contrast enhancement and to detect dynamic changes.

MRI applications were used in paper III and IV to evaluate the ischemic lesions. In

paper III, the rats were scanned with a panel of different MRI sequences in connection

to the MCAO procedure, and after 1 and 4 days. In paper IV, the scans were

performed after 1 day of survival. All MRI was performed with isoflurane anesthesia.

Diffusion-weighted imaging

Among the standard sequences, DWI is considered to be the most sensitive to cerebral

ischemia, by means of lowered ADC in the immediate phase after onset. Reperfusion

normalizes the ADC if the duration is limited to 30 min in the rat (Li F et al. 2000;

Neumann-Haefelin T et al. 2000). Partial return of the ADC is observed after

prolonged ischemia, however 2.5 hours duration seems too long for this phenomenon

to occur (Neumann-Haefelin T et al. 2000). Based on these facts, DWI was performed

to identify the ischemic lesion during MCAO and to ascertain reperfusion after

filament removal in paper III. DWI was repeated on subsequent scans on day 1 and 4

in paper III, and was also included in the panel of sequences used in paper IV.

T2-weighted imaging

Initially, the ischemic lesion can not be detected by T2WI, however ischemic changes

start to develop on T2WI within the first hours after ischemic onset, and after 1 day,

the infarcts appear as hyperintensities on T2WI. In contrast to the fluctuating changes

on DWI, the T2WI-changes are commonly accepted to reflect permanent damage.

T2WI was included in all panels used in paper III and IV.

T1-weighted dynamic contrast enhanced imaging

To investigate the vascular leakage over the blood-brain barrier, a MRI contrast agent

consisting of a gadolinium chelate was used in a dynamic scanning procedure of

24

T1WI. At each observation, an initial scan was performed after which the rat was

injected with contrast agent and then scanned 7-fold post-contrast, producing a series

of 8 images for each anatomical, rostro-caudal level.

In normal brain tissue, little or no gadolinium agent escapes from the plasma

compartment. However, in case or disrupted integrity of the cerebral blood vessels,

contrast passage can be observed as a gradual increase in signal intensity in the leaking

areas. Increases in signal intensity can readily be detected by comparing the last image

with the pre-contrast, or preferably, the first post-contrast images. In paper IV, such

comparisons revealed areas, anatomically corresponding to ischemic changes on DWI

or T2WI, where intensities had increased more than 3 SD compared with the

presumably normal tissue. Measurement of the signal intensity across time in such

regions revealed gradually increasing values, consistent with vascular leakage.

To better visualize the vascular leakage, the increase in intensity was fitted pixel-wise

to a function of vascular leakage from a 2-compartment model, where the circulating

plasma constitutes 1 of the compartments (Tofts PS et al. 1999). Under the assumption

that the leakage is limited by the permeability surface area, a coefficient (Kps),

positively correlated to the leakage, was derived. These Kps-values were mapped

anatomically to produce images reflecting the vascular leakage. The procedure is

illustrated in figure 5. The Kps-maps were evaluated with semi-quantitation of the

leakage extent. Each slice was assigned a value of 0 (normal), 1 (small) or 2 (large

leakage), and the values were summarized for each rat and observation time point to

form a score (ranging from 0 to 24 for 12 slices).

25

26

RESULTS AND DISCUSSION

Paper I

VEGF expression in different cell types

The study demonstrates widespread expression of VEGF after transient and permanent

MCAO. Several cell types were able to express VEGF in response to the insult,

however the most striking reactivity could be attributed neuronal cell bodies in the

infarct border region. In addition, prominent vascular reaction could be observed

selectively in the media of leptomeningeal arteries, but also in the vascular walls and

in the endothelium of penetrating vessels. Different investigators have reached various

conclusions on which may be the main source of VEGF in cerebral ischemia. Kovács

et al studied permanent MCAO and found the initial expression in macrophages while

neurons showed increased reactivity after days of survival (Kovács Z et al. 1996).

Conversely, Hayashi et al observed neuronal and pial expression after transient

MCAO, which is consistent with the findings in paper I (Hayashi T et al. 1997). Cobbs

et al showed major reaction in glia compared with neurons and vessels after transient

MCAO (Cobbs CS et al. 1998). However, these findings are contradicted by Plate et

al, in whose report significant expression was restricted to macrophages, while

neuronal or glia expression could not be reproduced after permanent MCAO (Plate KH

et al. 1999).

Regarding the reaction in endothelial cells, it has been questioned whether this reflects

ligand expression or binding (Kovács Z et al. 1996), and indeed, studies with in situ

hybridization have not been able to reproduce the VEGF reaction for endothelial cells

at the gene level (Cobbs CS et al. 1998; Plate KH et al. 1999). However, since VEGF

induction is regulated not only on transcriptional level (Stein I et al. 1998), negative

findings at the mRNA level need to be interpreted with precaution. Remarkable

reaction in large pial and in parenchymal vessels, as seen in paper I, have been noted

also after cerebral cold injury, where VEGF was suggested to play a role in BBB

breakdown (Nag S et al. 1997). Vascular presence of VEGF in response to cerebral

27

stab injury have also been reported, however that paper focused on the expression in

glia (Papavassiliou E et al. 1997).

After the publication of paper I, the effects of VEGF on the BBB and brain edema

have been reported with various results in the literature. Topical application of VEGF

reduced infarct size and edema after transient focal ischemia (Hayashi T et al. 1998),

and treatment with an intraventricular VEGF antibody also supported a protective role

for VEGF in a similar model (Bao WL et al. 1999). In contrast, another VEGF-

antagonist proved effective in reducing infarct size and edema (van Bruggen N et al.

1999), and the same group demonstrated that administration of recombinant VEGF

worsened BBB disruption and infarct size when given early after onset, while late

administration successfully enhanced angiogenesis and reduced ischemic damage

(Zhang ZG et al. 2000). Evidently, the importance of VEGF to ischemic brain injury

cannot easily be deducted from available studies. There might be variation in effects

between different cell types, the responsiveness may be different to endogenous levels

compared with exogenously administered VEGF , and the timing of manipulation may

be critical to outcome as previously suggested (Zhang ZG et al. 2000). At what tissue

concentration levels VEGF may function in the ischemic brain is virtually unknown.

Possibly, there is an optimal concentration level, at which trophic effects of VEGF

dominate without significant development of vasogenic edema. This view, if correct,

makes the prospect of VEGF therapy very intricate and points out a need for further

diagnostic advances to individualize intervention.

Surprisingly, the VEGF reaction in paper I was not restricted to the ischemic

hemisphere, which contrasts to the first description of VEGF expression after

permanent focal ischemia (Kovács Z et al. 1996). This contralateral appearance was

seen after 1 day and normalized after 3 days, and the underlying mechanism is elusive.

Possibly, signals through the commissure tracts initiate contralateral induction of

VEGF, but paracrine or mechanical mechanisms are imaginable as well.

28

VEGF-receptor findings

After transient and permanent MCAO, upregulation of VEGFR1 could be seen in

various cells, compared with the constitutive expression. VEGFR1-reactivity in

ipsilateral endothelial cells was more evident after transient than permanent MCAO,

however neuronal and glia expression was seen in both designs. In line with the

observations in paper I, VEGFR1 expression has been reported by Kovács et al after

permanent ischemia (Kovács Z et al. 1996). Surprisingly, immunostaining for

VEGFR1 occurred also to a variable extent in the non-ischemic hemisphere, and the

rationale for this is mysterious. Since few biological functions, except for chemotaxis

(Barleon B et al. 1996; Clauss M et al. 1996), have been attributed to VEGFR1, this

finding is left as observational.

There was no widespread upregulation by number of cells for VEGFR2 after MCAO,

however regional immunoreaction was quite distinct compared with controls. The

frequencies of endothelial reaction were comparable to the basal level, which was

unexpected since VEGFR2 accounts for most of the biological VEGF effects

described in cultured cells (Shibuya M et al. 1999). Possibly, the more disseminated

expression of VEGFR1 reported here and by Kovács et al (Kovács Z et al. 1996)

reflects a more significant role in cerebral ischemia than in other models. In contrast to

the prevailing paradigm at the time, VEGFR2 was seen not only in endothelial cells

but also in glia cells mainly in the ipsilateral cortex. Subsequent studies by different

groups have demonstrated that VEGFR2 mediates proliferation in peripheral Schwann

cells (Sondell M et al. 1999; Schratzberger P et al. 2000) and pericytes (Yamagishi S

et al. 1999). Moreover, it has been shown that neurotrophic effects of VEGF are

mediated by VEGFR2 (Sondell M et al. 1999; Sondell M et al. 2000), which may

explain the findings in paper I. Functional studies will be required to investigate

VEGFR-mediated growth effects on the ischemic brain in vivo.

The literature on VEGF-induced VP is less comprehensive than on the mitogenic

effects of VEGF. In assays, VEGF has extremely potent effects on VP compared with

other well-known mediators (Collins PD et al. 1993). The reason for this vascular

leakage is not entirely clear, but it may be functional in the sense that fibrin is allowed

29

to form outside the blood vessels, thus providing a stroma for newly formed vessels

(Dvorak HF et al. 1995). The exact mechanisms for VP increase in response to VEGF

have not been clearly established, but at the level of endothelial cells leakage can

principally exist between or across individual cells. VEGFR2 is considered to mediate

VP (Shibuya M et al. 1999), although it has been shown that disruption of endothelial

cell-cell connections may involve also VEGFR1 (Lindner V and Reidy MA 1996).

Endothelial gap junction communication is reduced through VEGFR2 and Src-MAPK

pathways, however these effects may not be entirely specific to VEGF (Suarez S and

Ballmer-Hofer K 2001). Recently, a novel vesiculo-vacuolar organelle (VVO) has

been discovered to transport macromolecules through the endothelial cell in a

VEGFR2-dependent manner (Dvorak AM and Feng D 2001). This VVO may origin

from fusioning of plasma-membrane invagionations, caveolae (Vasile E et al. 1999),

whose formation may be regulated by p38 and SFKs (Volonte D et al. 2001). It has

been shown that VEGF-induced VP is mediated through Src (Eliceiri BP et al. 1999),

and this mechanism might be of relevance to ischemic brain injury (Paul R et al.

2001). Among the MAPK pathways, especially p38 has been suggested to mediate VP

in the context of tumor metastasis seeding (Kato Y et al. 2001). Since strong candidate

mediators of vascular leakage such as VEGF and TNF-α (Yang GY et al. 1999) have

been shown to converge on p38 (Clauss M et al. 2001), the question is raised whether

p38 may regulate vascular effects of VEGF in the ischemic brain. Evidently, it may be

difficult to isolate the effects mediated by VEGFR1 and 2. Regardless of the activities

at the ligand-receptor level of VEGF signaling, the effects may be addressed by

investigating the intracellular signaling cascades downstream of the receptors.

Paper II

Vascular activation of ERK and JNK

In resemblance with the distribution of VEGF described in paper I, there was a

bilateral vascular reaction to activated ERK. With time, this activation increased and

became more prominent in the lesion hemisphere with microvascular engagement,

consistent with neovascularization. JNK was activated in a different pattern, with

distinct activation in blood vessels within the ipsilateral parenchyma, but practically no

30

signs of contralateral reaction. This activation was most apparent in larger arterial

vessels at early survival times, but was seen also in microvessels at a later stage. Even

though it is difficult to draw any firm conclusions on basis of these MAPK

observations, some features can be distinguished and commented.

The contralateral findings suggest that other mechanisms than direct ischemia

regulates ERK activation in the present model, e.g. VEGF or other factors which

appear contralaterally. With respect to the literature, it is conceivable that JNK is

differently regulated, possibly by environmental stress, which could explain the

localization to the ischemic hemisphere. It has been demonstrated in endothelial cells,

that both ERK and JNK are involved in VEGF-induced endothelial proliferation, and it

is possible that JNK, at least partially, lies downstream to MEK1-ERK in this context

(Pedram A et al. 1998). Transferring this idea to the present ischemia model, the

observed activation of both pathways in the population of perifocal microvessels

support a similar interplay in vivo.

Whether changes in vascular integrity are dependent on these pathways in cerebral

ischemia is currently unknown. Interest has been directed towards p38, rather than

ERK and JNK, as a mediator of VP, since p38 is involved in the regulation of cell

structure (Rousseau S et al. 1997) and thus potentially affects vascular integrity

(Clauss M et al. 2001; Kato Y et al. 2001). With reservation for the earliest period up

to 6 hours of survival, which was not covered in paper II, no vascular reaction to

activated p38 could be observed at any time point, although reactivity was present in

other cell types. This suggests that p38 may not be significantly involved in vascular

events at the observed time points after focal ischemic insult. The role of p38, if any,

in endothelial proliferation has remained unclear compared with ERK and JNK, but

interestingly, evidence has been presented that p38 may provide negative feedback

signals to the Ras pathway (Chen G et al. 2000). In such case, it would be logical to

expect little p38 activation in presence of active proliferation.

31

Neuronal ERK activation

ERK activation in neurons and glia was more frequent on, however not entirely

restricted to, the ischemic hemisphere, which resembles the results of studies by others

(Irving EA et al. 2000; Sugino T et al. 2000; Wu DC et al. 2000). However, the

functional importance of ERK remains unclear. There is experimental evidence that

inhibition of the ERK pathway at the MEK level limits brain damage in focal ischemia

(Alessandrini A et al. 1999), however failure to observe such effects after global

ischemia has been reported as well (Sugino T et al. 2000). Besides using different

ischemia models, those reports targeted different neuronal populations (cortex,

hippocampus), which may be one factor behind the discrepant effects. From a

theoretical perspective, predicting the effects of signaling through the ERK pathway in

cerebral ischemia is complicated.

Based on the view of ERK as a putative cell survival factor (Xia Z et al. 1995), it

seems paradoxical to expect injurious effects through this pathway. However, since

ERK is involved in the physiological response to NMDA-receptor-induced memory

functions (Xia Z et al. 1996), it may fit into the theories of excitotoxicity through

excessive glutamate receptor activation. Exposure of hippocampal slice cultures to a

phosphatase inhibitor resulted in sustained ERK activation and non-apoptotic neuronal

death (Rundén E et al. 1998). This cell death was ameliorated by inhibition of MEK1,

indicating a role for the MEK1-ERK pathway in neuronal injury. Nevertheless,

populations of phospho-ERK-positive, appreciably preserved neurons could be seen up

to 1 and 3 days of survival in paper II and it cannot be ruled out that ERK participates

in the survival mechanisms in these cells (figure 6).

32

Figure 6. Phospho-ERK immunostaining. Cortical neurons showing ERK activation(brown colored cells) in the infarct border (down left) after 1 day of PMCAO.Pyknotic cells in the infarcted area (up right).

Activation of JNK and p38

JNK has been pointed out as a possible mediator of neuronal cell death (Herdegen T et

al. 1998; Hayashi T et al. 2000), but in addition to such effects, it has been

demonstrated that JNK may mediate growth signals in neuronal cells as well, e.g.

when PC12-cells are exposed to prolactin (Cheng Y et al. 2000). With exception for

some isolated, non-ischemic nuclei after PMCAO, we did not find any significant JNK

activation in neurons, which impedes further speculations. Conversely, distinct

reaction was seen in certain glia cells of suspected astrocyte origin and also in cortical

cells neighboring neurons, possibly satellite cells. Interestingly, it is known that JNK

mediates NGF-induced apoptosis in cultured oligodendrocytes (Casaccia-Bonnefil P et

al. 1996), which contrasts to the effects on PC12-cells, where NGF-withdrawal

33

activates JNK and p38 (Xia Z et al. 1996). These seemingly contradictory findings

may result from the different cell types used, and there may also be variations between

the effects of the different JNK isoforms (Harrington AW et al. 2002).

Regarding p38, major activation was seen in macrophages within the infarct at later

survival times. Neuronal activation was observed in nuclei around the 3rd ventricle, but

otherwise sparse. Activation was seen in glia cells of astrocyte-like appearance in the

ischemic parenchyma, but the activation in these cell types was diminutive compared

with the reaction in macrophages. Other studies suggest that p38 is rapidly and

transiently activated (Legos JJ et al. 2001) in neurons and astrocytes after ischemia

(Wu DC et al. 2000), which together with the present findings indicate that p38 may

be involved in different phases of the injury process.

From descriptive to functional studies

This descriptive study demonstrates that the main MAPK pathways are differentially

activated after focal cerebral ischemia with largely similar patterns after permanent

and 2 hours transient ischemia. This gives an indication on which of these pathways

are turned on or not in different cell types, but to achieve functional information,

experimental manipulation of these pathways need to be undertaken in vivo.

Paper III

Brief discussion of the background

This study aimed to investigate the role of p38 in focal cerebral ischemia. It has been

suggested that p38 may mediate VEGF-induced VP (Clauss M et al. 2001), however

no signs of vascular activation of p38 could be detected between 6 hours and 1 week

of survival in paper II, although macrophage activation was obvious. As pointed out in

the discussion of paper II, other reports have demonstrated that p38 is activated in the

initial phase of cerebral ischemia, reaching its peak within the first hour of ischemia

(Wu DC et al. 2000; Legos JJ et al. 2001), preferentially in neuronal cells and glia (Wu

DC et al. 2000). Administration of p38-inhibitors has resulted in various effects on the

ischemic injury (Alessandrini A et al. 1999; Sugino T et al. 2000; Barone FC et al.

34

2001), and the reason for this is unclear. A speculation could be that the early p38

activation affects the course of the ischemic injury, and that there might be effects on

the VP and thereby on the BBB as well. Based on these considerations, an

interventional study using a conventional p38-inhibitor, SB203580, a pyridylimidazole

derivative, was designed. The regimen was similar to a previous study with the same

inhibitor in global ischemia, in which neuroprotective effects were found (Sugino T et

al. 2000).

Evaluating MCAO with DWI

The SB203580-treated and the control groups showed no initial difference in lesion

size as determined by ADC below 65%. It has been demonstrated that the degree of

ADC lowering is related to the severity of the ischemia (Hoehn-Berlage M et al.

1995), but it is also true that even severe ADC decreases cannot adequately predict

irreversible damage (Fiehler J et al. 2002). The threshold used (65%) represents a

severe but not extreme decrease with reference to the literature in the field. In

comparison, it has been demonstrated that energy depletion corresponds to a decrease

of ADC to approximately 77%, while tissue acidosis is present at 90% in the rat

(Hoehn-Berlage M et al. 1995). When setting the threshold used in paper III, different

levels were tested, resulting in minor differences between volumes calculated on the

65% and 75% levels. However, at the higher level, there was more noise from the non-

ischemic tissue, which lead to the threshold of 65%. It is reasonable to assume that the

similarity in volumes of decreased ADC rules out any systematic differences in

MCAO between the groups. Likewise, the conformity between groupd in DWI change

after filament withdrawal supports the presence of successful reperfusion in the two

groups (Li F et al. 2000; Neumann-Haefelin T et al. 2000).

The filament model of MCAO has been subjected to criticism due to limited

reproducibility, and different methods to prepare the filament have been proposed to

improve the model (Belayev L et al. 1996). The use of laser-doppler measurement of

the cerebral blood flow has been suggested for adequate monitoring of the surgical

procedure, since it can detect both misplacement of the filament and subarachnoid

hemorrhages (SAH) (Schmid-Elsaesser R et al. 1998). It is evident that such

35

complications may seriously disturb the experiments if not detected, but the extent of

these problems may be discussed. In the latter study, it was clearly demonstrated that

filament misplacement was associated with premature reperfusion and SAH with lack

of reperfusion on filament removal (Schmid-Elsaesser R et al. 1998). The rate of SAH

was about 40%, which is considerably higher than what we have been able to detect by

macroscopic inspection, and what has been reported in by others (Li F et al. 1999).

In order to validate the MCAO model in our hands, we collected the ADC data from

all rats (n=27), including pilot experiments, that had undergone DWI scanning prior to

and immediately after filament withdrawal. In 25 rats (93%), lowered ADC was

present in the ipsilateral MCA territory, indicating successful MCAO. In 21 (84%) of

the rats with MCAO, the ADC was partially normalized after removal of the filament,

indicating reperfusion. Of the remaining 4 rats (16%), all had macroscopic hemorrhage

on post-mortem examination, of which 3 rats failed to show ADC recovery and 1 died

before the second DWI scan (but was included in the analysis since the intention was

to scan before and after suture removal). These data suggest that in absence of

macroscopic hemorrhage, the incidence of significant SAH is likely to be low in our

hands. The different ADC patterns are displayed in figure 7. The rats without visible

36

ischemic changes on DWI (7% of total) showed no neurological motor deficit on

recovery from anesthesia. Even though it may be of interest to measure blood flow

with laser-doppler during the surgical procedure, this prolongs the operation, which

may add variation to the experiment as well. The relations between DWI changes,

hemorrhage and neurological deficits as described above support the relevance in

using initial motor deficit and absence of macroscopic SAH as inclusion criteria in this

model.

Augmented ischemic injury by SB203580

In the panel of MRI sequences used, T2WI most adequately reflects the definitive

ischemic injury after 1 day, although not visible from the beginning of the ischemia.

Although no significant differences in DWI lesions could be detected initially, the

lesion volumes on T2WI were significantly larger in the SB203580-treated group than

in controls after 1 day (300±95mm3 vs. 126±75 mm3; p<0.01). This pattern was still

present on 4 days of observation, however the lesions appeared more patchy with

consequent difficulties to delineate the lesion, which was the reason for choosing not

to calculate the infarct volumes after 4 days. The increased infarct size in SB203580-

treated rats on T2WI was accompanied by increased leakage of contrast agent, as

measured semi-quantitatively from the Kps-maps. Interestingly, the difference

between groups in vascular leakage also evolved across time (median leakage score

after 1 day 18.5; range 15-21 vs. 6.5; 4-17; p<0.05; after 4 days 11; 6-15 vs. 3.5; 1-9;

p<0.05), without significant difference initially after reperfusion.

Taken together, the findings demonstrate that pretreatment with the p38-inhibitor did

not significantly influence the initial extent of the ischemic lesion as detected by DWI

or vascular leakage, suggesting that other mechanisms than VP may be critical at that

stage. However, the course of the ischemic injury was affected negatively over the

observation period, with SB203580-treated rats suffering worse infarcts and leakage,

indicating that the drug administration had interfered with the capability of the brain to

endure the ischemic insult.

37

The role of p38 in cerebral ischemia

The detrimental effects of SB203580 in paper III lead to speculations on a protective

function of p38 in cerebral ischemia. In that sense, the findings in paper III adds to

previous reports, where negative and neutral effects have been attributed to p38.

Neuronal survival was promoted by SB203580 in global ischemia (Sugino T et al.

2000), while it was ineffective compared with the MEK-inhibitor PD98059 in

permanent ischemia (Alessandrini A et al. 1999). Another group has published

evidence that a second-generation compound, related to SB203580, was effective in

reducing permanent ischemic injury (Barone FC et al. 2001). In that material, the

administration of the inhibitor was prolonged for several hours after ischemia, i.e.

another time window than in paper III, and perhaps affecting the inflammatory

reaction.

A speculative explanation to the findings in paper III could be that instead of

preventing damage, the p38-inhibitor may have inhibited a protective mechanism such

as ischemic preconditioning from the laceration produced at drug injection, or the

putative beneficial effects of the vehicle (DMSO 3%). In the heart, it has been shown

that ischemic preconditioning is p38-dependent with MAPK-activated protein kinase 2

and small heat shock proteins as downstream targets (Cohen MV et al. 2000; Joyeux

M et al. 2000; Zhao TC et al. 2001). The presence of such preconditioning

mechanisms in the brain is largely unexplored, but it has been demonstrated that

intraventricular injection of thrombin both activates MAPK and induces tolerance to a

subsequent cerebral insult (Xi G et al. 1999; Xi G et al. 2000; Xi G et al. 2001). It is

known that DMSO may have intrinsic neuroprotective properties (Shimizu S et al.

1997; Greiner C et al. 2000), however the interaction with MAPK pathways is not

known. It is possible that cerebral preconditioning mechanisms involve p38, analogous

to the in the heart, however it is conceivable that it may contribute to the ischemic

damage in a later phase. Further research will be required to elucidate these aspects of

the p38 pathway.

38

Paper IV

Reduced infarct size with PP2 treatment – a matter of vascular permeability?

Administration of the SFK-inhibitor PP2 reduced infarct size by approximately 50%

compared with controls after 1 day, as measured with T2WI (120±47 mm3 vs.

313±110 mm3; p<0.05) and tetrazolium staining (173±44 mm3 vs. 375±55 mm3;

p<0.05). Neurological score according to Bederson (Bederson JB et al. 1986) was also

significantly better in the PP2-group compared with controls (median neurological

score 1; range 0-2 vs. 2.5; 1-3; p<0.05.), however no significant effects were

demonstrated on vascular leakage (median score 11; range 6-13 vs. 17; 10-20; n.s.).

The infarct reduction found is consistent with previous observations by Paul et al in

permanent ischemia, where mice were treated with the unspecific SFK-inhibitor PP1

(Paul R et al. 2001). That study also included transgenic mice, and it was concluded

that Src-, but not Fyn-deficiency, rendered tolerance to ischemia. The authors

proposed VEGF-induced VP to be the injury mechanism responsive to SFK inhibition,

and this issue was discussed in the paper I section. As pointed out there, the role of

VEGF in cerebral ischemia is far from uncomplicated. One of the main technical

difficulties is to determine the causal relationship between infarct and leakage, and it is

reasonable to assume that the extent of vascular leakage may depend on the infarct per

se, even if the leakage may be injurious as well.

Paul et al demonstrate that Fyn-deficient mice were not protected against brain

ischemia, and based on developmental aspects of Fyn-NMDA-receptor interactions,

they basically discard the idea that any protective effects of Src-deficiency and PP1

could be related to other mechanisms than VP (Paul R et al. 2001). When scrutinized,

that study provides little evidence of substantial connection between SFKs and

endogenous VEGF levels. Although it was emphasized that vascular leakage was

diminished in the PP1-treated or Src-deficient mice, the causal relationship was not

established. The mentioned (but not demonstrated) inhibition by PP1 of exogenous

VEGF-induced VP represents an artificial situation of uncertain relevance to the

natural injury process. As mentioned above, it was demonstrated in knockout-mice

that VP in response to VEGF was mediated by Src and Yes, but not by Fyn (Eliceiri

BP et al. 1999). Based on those findings, the hypothesis by Paul et al implicates that

39

the ischemic tolerance should be comparable in Src- and Yes-deficient mice.

Regrettably, Paul et al do not address this potentially enlightening issue (Eliceiri BP et

al. 1999; Paul R et al. 2001).

Possible involvement of SFKs in other injury mechanisms

NMDA-receptor-mediated excitotoxicity is an extensively studied and commonly

accepted mechanism of ischemic brain injury. Despite lack of evidence of any

functional importance to infarct development, it is possible that SFKs interact with the

NMDA-receptor in response to cerebral ischemia. There is substantial evidence of

physiological SFK-NMDA-receptor interactions (Ali DW and Salter MW 2001),

where SFKs regulate neuronal differentiation and Ca2+ influx through the NMDA-

receptor channel through tyrosine phosphorylation (Yu XM et al. 1997). SFKs can be

engaged by other pathways in order to modulate the NMDA-receptor, e.g. a route

involving G-protein coupled receptors, protein kinase C (PKC) and cell-adhesion

kinase-ß (CAKß) (Lu WY et al. 1999; Ali DW and Salter MW 2001). SFK-CAKß

interaction may also provide a link to the MAPK pathways (Dikic I et al. 1996), which

are discussed above.

It has been shown that transient global ischemia increases the degree of tyrosine

phosphorylation of the NMDA-receptor in the hippocampus with a peak at 6 hours

after ischemia, and this phenomenon was partially reversed by ketamine, a NMDA-

receptor antagonist (Pei L et al. 2000). In addition, Src co-precipitated with the

phosphorylated NMDA-receptor subunit NR2B (Pei L et al. 2000), and in the same

model, the Src kinase activity was increased by ischemia and partially inhibited by

ketamine (Pei L et al. 2000). Similar observations including both NR2A and B have

been reported by other groups (Takagi N et al. 1999; Cheung HH et al. 2000; Liu Y et

al. 2001), which support the idea of an interaction in vivo.

The role of leukocytes as contributors to ischemic brain injury has been debated

(Emerich DF et al. 2002), however it has been demonstrated that neutrophils may be

involved in transient cerebral ischemia (Matsuo Y et al. 1994). SFKs are likely to

participate in such mechanisms, since they regulate functions characteristic to

neutrophils, such as degranulation and respiratory burst (Lowell CA et al. 1996;

40

Mocsai A et al. 2000). It may also be that SFKs mediate effects of reactive oxygen

species (ROS), since studies demonstrate that SFKs are activated upon ROS

stimulation in PC12- and endothelial cells, leading to MAPK activation (Jope RS et al.

2000; Yoshizumi M et al. 2000). Although there is no evidence for a pathogenic role

of SFKs in the examples of brain injury modalities above, at least inflammatory and

ROS damage (Cao X and Phillis JW 1994) have been attributed a therapeutic window

resembling what was described by Paul et al (Paul R et al. 2001). The peak in NMDA-

receptor-induced Src activity after global ischemia appears after 6 hours, indicating

that this pathway may be available for intervention in the same window as SFKs.

Another possibility could be that the SFKs influence rheology, based on evidence of

SFK involvement in platelet function (Grondin P et al. 1991; Horvath AR et al. 1992).

For example, Src activation has been shown in response to von Willebrand factor

stimulation of platelets (Jackson SP et al. 1994) and the anchoring of the platelet

integrin GP IIb/IIIa into fibrin structures is likely to be Src-dependent (Schoenwaelder

SM et al. 1994). Platelet inhibition through GPIIb/IIIa has been approved for treatment

of acute coronary syndromes and also attributed a therapeutic potential in acute

ischemic stroke (Bogousslavsky J and Leclerc JR 2001; Junghans U et al. 2002). In

paper IV, the cerebral blood flow (CBF) was not measured, but in the report by Paul et

al, the reduced injury by PP1was accompanied by improved CBF, indicating a possible

confounding factor (Paul R et al. 2001). Thus, there are several potential injury

modalities that might explain the protective effects of SFK-inhibition, demonstrated in

paper IV and by Paul et al (Paul R et al. 2001), and determining which pathway(s) may

be critical in this respect will require thorough research. Nevertheless, the findings in

paper IV and by Paul et al suggest a considerable therapeutic potential with an

attractive therapeutic window, which encourages further investigations within this

field.

41

CONCLUSIONS

Based on this doctoral thesis it is concluded that:

√ VEGF and VEGF-receptors are upregulated after focal cerebral ischemia

√ VEGF and VEGFR1 after cerebral ischemia is present in various cell types,

while VEGFR2 is chiefly restricted to endothelial cells and glia

√ MAPKs are differentially activated in response to focal cerebral ischemia

√ ERK and JNK pathways may be associated with post-ischemic

neovascularization

√ p38 activation is associated with the inflammatory response to cerebral

ischemia

√ Pretreatment with the p38-inhibitor SB203580 may aggravate brain injury

after transient ischemia

√ Inhibition of SFK has the potential to reduce injury after transient focal

cerebral ischemia.

42

SAMMANFATTNING PÅ SVENSKA

Detta arbete avhandlar hjärnans signaleringsmekanismer vid akut nedsatt blodcirkulation, somvid blodpropp i hjärnan (ischemisk stroke). Den modell som använts i delarbetena I – IVinnefattar avstängning av den mellersta hjärnartären (a cerebri media) hos råtta, vilketkärlanatomiskt relativt väl avspeglar den vanligaste stroketypen hos människa. Metodenmedger studier av såväl permanent som temporär avstängning (ischemi), där det sistnämndainnebär återkomst av blodcirkulation (reperfusion).

De mekanismer som studeras är kommunikation dels mellan celler (extracellulärt) i form avtillväxtfaktorn VEGF och dess cellytereceptorer, och dels inom cellen (intracellulärt),representerat av s k MAP-kinaser (MAPK) och Src-familjen av kinaser (SFK). Dessaintracellulära signalmolekyler fungerar som enzymer i intracellulära signaleringskaskader ochkatalyserar biokemiska reaktioner där en fosfatgrupp kopplas till utvalda proteiner. Genomdenna typ av reaktioner reglerar cellen bl a celldelning, utmognad och programmerad celldöd(apoptos).

Med hjälp av antikroppar, framtagna för att känna igen specifika strukturer, kan man studeraförekomsten av t ex ett protein i vävnadssnitt från den skadade hjärnan (s k immunhistokemi).I delarbete I och II används denna metod för att studera förekomst (expression) av VEGF ochdess receptorer samt aktivering av tre olika MAPK (betecknade ERK, JNK och p38) efterpermanent och temporär ischemi

Vissa signalvägar kan hämmas (inhiberas) experimentellt med hjälp av specifika inhibitorer,vilket tillämpats i stor utsträckning främst i in vitro-forskning, dvs ej i levande organismer(in vivo). Effekterna på skadestorlek och kärlfunktion studeras i delarbete III – IV in vivogenom tillförsel av p38- och SFK-inhibitorer i anslutning till ischemin. För att utvärderaskadan användes magnetisk resonanstomografi (MRT), en metod som lämpar sig väl för attundersöka skador i nervsystemet.

I delarbete I kunde en ökning av VEGF och dess receptorer påvisas i olika celltyper, vilkettolkas som att hjärnan aktiverar detta system som en del av svaret på skadan. VEGF kan habetydelse för kärlnybildning och kärlfunktion efter stroke, men även effekter på olika typer avnervceller. De MAPK som studerades i delarbete II visade olika mönster av aktivering, främstöverensstämmande med en roll för ERK och JNK i hjärnans blodkärl och för p38 iinflammatoriska celler efter ischemi.

I delarbete III sågs en ökad skadeutbredning och ökat läckage från hjärnans blodkärl efterbehandling med p38-hämmare, vilket innebär att p38 skulle kunna uppbära en gynnsam roll ivissa skeden av skadeprocessen. Detta fynd kompletterar rön från andra forskargrupper, därp38 har antagits bidra till ischemisk hjärnskada. I delarbete IV kunde minskad skada ses hosdjur som fått SFK-hämmare, vilket stöder möjligheten att SFK bidrar till skadeutvecklingen.SFK är dock inblandade många signalvägar, vilket gör resultaten svårtolkade ur ettmekanistiskt perspektiv. Man kan dra slutsatsen att dessa signalvägar sannolikt är involveradei skadeutvecklingen vid hjärnischemi och att ytterligare forskning kommer att behövas för attnå mer fullständig kunskap på området.

43

ACKNOWLEDGMENTS

Jag vill framföra mitt varma tack till:

Andreas Terént, min handledare för goda råd, outsinligt stöd och entusiasm underdoktorandtiden och för att ha visat mig stort förtroende och vänskap i alla upptänkligasituationer genom åren

Yngve Olsson, som introducerade mig för immunhistokemi, för medförfattarskap ochgoda råd inom neuropatologi samt för välbehövlig uppmuntran i stunder av trötthetoch tvivel i början av arbetet

Pär Gerwins, för handledning inom signaltransduktion och för idérikedom ochinspirerande diskussioner

Anders Ericsson, för handledning inom MRT, för gott humör trots udda försökstiderdag som natt samt för tips om bra jazzplattor

Sverker Ljunghall och Kjell Öberg, f d och nuvarande prefekter vid Institutionenför Medicinska Vetenskaper för möjligheten att påbörja och slutföra detta arbete

Lars Wiklund and Torsten Gordh, för stöd från Anestesi- och intensivvårdskliniken,min kliniska hemvist

Khaled Ahmad Ata, för medförfattarskap, enastående immunhistokemi samt förutvecklande diskussioner om forskning, religion, politik och mycket annat

Håkan Ahlström, för medförfattarskap och för den välkomnande miljön på MRT-sektionen

Keiko Funa, för medförfattarskap, diskussioner och för att ha visat vägen in iimmunhistokemins detaljer

Sofia Karlsson och Susanne Akterin, mina projektstudenter och medförfattare, föridogt och stringent arbete med immunhistokemin i delarbete II respektive hjälp medpilotförsöken till delarbete IV

Ulla Svensson, som hållit ordning på doktoranderna på labbet, troget ställt upp i allalägen med teknisk assistans och stor erfarenhet

Ann-Christine Syvänen, Gisela Barbany m fl i Molekylärmedicingruppen, LenaClaesson-Welsh, Emma Rennel m fl på Rudbecklaboratoriet samt Anders Larssonoch David Carlander på Klinisk Kemi för möjligheten att gästarbeta i respektivelaboratorier

44

Bengt Fellström, för välkomnande generositet i början av arbetet när strokemodellenskulle etableras på labbet

Jan Melin, Lilija Zezina och Sven-Olof och Elisabet Granstam, mina kollegor pålabbet för vänskap, diskussioner och bataljer

Lars Johansson, för initierad hjälp med diverse MRT-frågor, många skratt och för attmed sin närvaro fått Västsverige att kännas lite mer närbeläget, och Atle Bjørnerud,som bidragit med kunskap, hjälp och hemgjord mjukvara för bearbetning av MRT-bilder

Fredrik Claussen, Niklas Marklund, Jonas Isaksson m fl i Neurokirurgensforskningsgrupp för diskussioner och lån av utrustning

Gunilla Tibbling, som lärde mig grunderna i immunhistokemi, och MadeleineJarild, som bidragit med förstklassigt arbete med vävnadssnittning

Forskare, doktorander och teknisk personal på Centrum för Klinisk MedicinskForskning

Kollegor på Anestesi- och intensivvårdskliniken för diskussioner och gemenskap

Kollegor och klinikledningar inom Skaraborgs Sjukhus för förståelse och visatintresse för min forskning under AT i Lidköping och under tiden på Anestesikliniken iSkövde

Solveig Karlsson, för ovärderlig administrativ och praktisk hjälp med allehanda saker

Mina gamla kursare och kollegor Filip Sköldberg, Maria Holstad, Eva Penno, ErikBackhaus, Jakob Johansson, Marie Andersson-Olerud m fl, för vetenskapliga,kliniska och allmänt sociala ljuspunkter, och för uppdateringar av sociogrammen

Peder Olofsson, min vän och kollega för reflektioner och spekulationer om samtidaföreteelser som jobb, liv och annat

Mamma och Pappa för att ni alltid funnits där för mig

Johanna och Edvin, min egen familj, för all kärlek och tålamod.

Finansiellt stöd har erhållits från Hjärt-Lungfonden (75015), KungligaVetenskapssamhället och Svenska Läkaresällskapet, Erik, Karin och Gösta SelandersStiftelse.

Astra Arcus (London) bidrog med handledning kring etablerandet av försöksmodelleni Uppsala.

45

REFERENCES

Alessandrini A, Namura S, Moskowitz MA and Bonventre JV (1999). “MEK1protein kinase inhibition protects against damage resulting from focal cerebralischemia.” Proc Natl Acad Sci U S A 96 (22): 12866-9.

Ali DW and Salter MW (2001). “NMDA receptor regulation by Src kinase signallingin excitatory synaptic transmission and plasticity.” Curr Opin Neurobiol 11 (3): 336-42.

Ata KA, Lennmyr F, Funa K, Olsson Y and Terent A (1999). “Expression oftransforming growth factor-beta1, 2, 3 isoforms and type I and II receptors in acutefocal cerebral ischemia: an immunohistochemical study in rat after transient andpermanent occlusion of middle cerebral artery.” Acta Neuropathol (Berl) 97 (5): 447-55.

Bao WL, Lu SD, Wang H and Sun FY (1999). “Intraventricular vascular endothelialgrowth factor antibody increases infarct volume following transient cerebralischemia.” Chung Kuo Yao Li Hsueh Pao 20 (4): 313-8.

Barleon B, Sozzani S, Zhou D, Weich HA, Mantovani A and Marme D (1996).“Migration of human monocytes in response to vascular endothelial growth factor(VEGF) is mediated via the VEGF receptor flt-1.” Blood 87 (8): 3336-43.

Barone FC, Irving EA, Ray AM, Lee JC, Kassis S, Kumar S, et al. (2001). “SB239063, a Second-Generation p38 Mitogen-Activated Protein Kinase Inhibitor,Reduces Brain Injury and Neurological Deficits in Cerebral Focal Ischemia.” JPharmacol Exp Ther 296 (2): 312-321.

Bederson JB, Pitts LH, Tsuji M, Nishimura MC, Davis RL and Bartkowski H(1986). “Rat middle cerebral artery occlusion: evaluation of the model anddevelopment of a neurologic examination.” Stroke 17 (3): 472-6.

Belayev L, Alonso OF, Busto R, Zhao W and Ginsberg MD (1996). “MiddleCerebral Artery Occlusion in the Rat by Intraluminal Suture - Neurological andPathological Evaluation of an Improved Model.” Stroke 27: 1616-23.

Bogousslavsky J and Leclerc JR (2001). “Platelet glycoprotein IIb/IIIa antagonistsfor acute ischemic stroke.” Neurology 57 (5): S53-7.

Bonita R, Beaglehole R and Asplund K (1994). “The worldwide problem of stroke.”Curr Opin Neurol 7 (1): 5-10.

Boulton TG, Nye SH, Robbins DJ, Ip NY, Radziejewska E, Morgenbesser SD, etal. (1991). “ERKs: a family of protein-serine/threonine kinases that are activated andtyrosine phosphorylated in response to insulin and NGF.” Cell 65 (4): 663-75.

46

Boulton TG, Yancopoulos GD, Gregory JS, Slaughter C, Moomaw C, Hsu J, et al.(1990). “An insulin-stimulated protein kinase similar to yeast kinases involved in cellcycle control.” Science 249 (4964): 64-7.

Brewster JL, de Valoir T, Dwyer ND, Winter E and Gustin MC (1993). “Anosmosensing signal transduction pathway in yeast.” Science 259 (5102): 1760-3.

Cao X and Phillis JW (1994). “alpha-Phenyl-tert-butyl-nitrone reduces corticalinfarct and edema in rats subjected to focal ischemia.” Brain Res 644 (2): 267-72.

Casaccia-Bonnefil P, Carter BD, Dobrowsky RT and Chao MV (1996). “Death ofoligodendrocytes mediated by the interaction of nerve growth factor with its receptorp75.” Nature 383 (6602): 716-9.

Chen G, Hitomi M, Han J and Stacey DW (2000). “The p38 pathway providesnegative feedback for ras proliferative signaling.” J Biol Chem 275 (50): 38973-80.

Cheng Y, Zhizhin I, Perlman RL and Mangoura D (2000). “Prolactin-induced cellproliferation in PC12 cells depends on JNK but not ERK activation.” J Biol Chem 275(30): 23326-32.

Cheung HH, Takagi N, Teves L, Logan R, Wallace MC and Gurd JW (2000).“Altered association of protein tyrosine kinases with postsynaptic densities aftertransient cerebral ischemia in the rat brain.” J Cereb Blood Flow Metab 20 (3): 505-12.

Clauss M, Sunderkotter C, Sveinbjornsson B, Hippenstiel S, Willuweit A, MarinoM, et al. (2001). “A permissive role for tumor necrosis factor in vascular endothelialgrowth factor-induced vascular permeability.” Blood 97 (5): 1321-9.

Clauss M, Weich H, Breier G, Knies U, Rockl W, Waltenberger J, et al. (1996).“The vascular endothelial growth factor receptor Flt-1 mediates biological activities.Implications for a functional role of placenta growth factor in monocyte activation andchemotaxis.” J Biol Chem 271 (30): 17629-34.

Cobbs CS, Chen J, Greenberg DA and Graham SH (1998). “Vascular endothelialgrowth factor expression in transient focal cerebral ischemia in the rat.” Neurosci Lett249 (2-3): 79-82.

Cohen MV, Baines CP and Downey JM (2000). “Ischemic preconditioning: fromadenosine receptor of KATP channel.” Annu Rev Physiol 62: 79-109.

Collins PD, Connolly DT and Williams TJ (1993). “Characterization of the increasein vascular permeability induced by vascular permeability factor in vivo.” Br JPharmacol 109 (1): 195-9.

47

Creedon DJ, Johnson EM and Lawrence JC (1996). “Mitogen-activated proteinkinase-independent pathways mediate the effects of nerve growth factor and cAMP onneuronal survival.” J Biol Chem 271 (34): 20713-8.

Dikic I, Tokiwa G, Lev S, Courtneidge SA and Schlessinger J (1996). “A role forPyk2 and Src in linking G-protein-coupled receptors with MAP kinase activation.”Nature 383 (6600): 547-50.

Dvorak AM and Feng D (2001). “The vesiculo-vacuolar organelle (VVO). A newendothelial cell permeability organelle.” J Histochem Cytochem 49 (4): 419-32.

Dvorak HF, Brown LF, Detmar M and Dvorak AM (1995). “Vascular permeabilityfactor/vascular endothelial growth factor, microvascular hyperpermeability, andangiogenesis.” Am J Pathol 146 (5): 1029-39.

Eliceiri BP, Paul R, Schwartzberg PL, Hood JD, Leng J and Cheresh DA (1999).“Selective requirement for Src kinases during VEGF-induced angiogenesis andvascular permeability.” Mol Cell 4 (6): 915-24.

Emerich DF, Dean RL, 3rd and Bartus RT (2002). “The role of leukocytesfollowing cerebral ischemia: pathogenic variable or bystander reaction to emerginginfarct?” Exp Neurol 173 (1): 168-81.

Ferrara N and Alitalo K (1999). “Clinical applications of angiogenic growth factorsand their inhibitors.” Nat Med 5 (12): 1359-64.

Ferrara N and Henzel WJ (1989). “Pituitary follicular cells secrete a novel heparin-binding growth factor specific for vascular endothelial cells.” Biochem Biophys ResCommun 161 (2): 851-8.

Fiehler J, Foth M, Kucinski T, Knab R, von Bezold M, Weiller C, et al. (2002).“Severe ADC decreases do not predict irreversible tissue damage in humans.” Stroke33 (1): 79-86.

Gitay-Goren H, Cohen T, Tessler S, Soker S, Gengrinovitch S, Rockwell P, et al.(1996). “Selective binding of VEGF121 to one of the three vascular endothelialgrowth factor receptors of vascular endothelial cells.” J Biol Chem 271 (10): 5519-23.

Greiner C, Schmidinger A, Hulsmann S, Moskopp D, Wolfer J, Kohling R, et al.(2000). “Acute protective effect of nimodipine and dimethyl sulfoxide against hypoxicand ischemic damage in brain slices.” Brain Res 887 (2): 316-22.

Grondin P, Plantavid M, Sultan C, Breton M, Mauco G and Chap H (1991).“Interaction of pp60c-src, phospholipase C, inositol-lipid, and diacyglycerol kinaseswith the cytoskeletons of thrombin-stimulated platelets.” J Biol Chem 266 (24):15705-9.

48

Gupta K, Kshirsagar S, Li W, Gui L, Ramakrishnan S, Gupta P, et al. (1999).“VEGF prevents apoptosis of human microvascular endothelial cells via opposingeffects on MAPK/ERK and SAPK/JNK signaling.” Exp Cell Res 247 (2): 495-504.

Hacke W, Kaste M, Fieschi D, Toni D, Lesaffre E, von Kummer R, et al. (1995).“Intravenous Thrombolysis With Recombinant Tissue Plasminogen Activator forAcute Hemispheric Stroke - The European Cooperative Acute Stroke Study(ECASS).” JAMA 274: 1017-25.

Han J, Lee JD, Bibbs L and Ulevitch RJ (1994). “A MAP kinase targeted byendotoxin and hyperosmolarity in mammalian cells.” Science 265 (5173): 808-11.

Hanke JH, Gardner JP, Dow RL, Changelian PS, Brissette WH, Weringer EJ, etal. (1996). “Discovery of a novel, potent, and Src family-selective tyrosine kinaseinhibitor. Study of Lck- and FynT-dependent T cell activation.” J Biol Chem 271 (2):695-701.

Hankey GJ and Warlow CP (1999). “Treatment and secondary prevention of stroke:evidence, costs, and effects on individuals and populations.” Lancet 354 (9188): 1457-63.

Harrington AW, Kim JY and Yoon SO (2002). “Activation of Rac GTPase by p75is necessary for c-jun N-terminal kinase-mediated apoptosis.” J Neurosci 22 (1): 156-66.

Hayashi T, Abe K and Itoyama Y (1998). “Reduction of ischemic damage byapplication of vascular endothelial growth factor in rat brain after transient ischemia.”J Cereb Blood Flow Metab 18 (8): 887-95.

Hayashi T, Abe K, Suzuki H and Itoyama Y (1997). “Rapid Induction of VascularEndothelial Growth Factor Gene Expression After Transient Middle Cerebral ArteryOcclusion in Rats.” Stroke 28: 2039-44.

Hayashi T, Sakai K, Sasaki C, Zhang WR, Warita H and Abe K (2000). “c-Jun N-terminal kinase (JNK) and JNK interacting protein response in rat brain after transientmiddle cerebral artery occlusion.” Neurosci Lett 284 (3): 195-9.

Herdegen T, Claret FX, Kallunki T, Martin-Villalba A, Winter C, Hunter T, etal. (1998). “Lasting N-terminal phosphorylation of c-Jun and activation of c-Jun N-terminal kinases after neuronal injury.” J Neurosci 18 (14): 5124-35.

Hoehn-Berlage M, Norris DG, Kohno K, Mies G, Leibfritz D and Hossmann KA(1995). “Evolution of regional changes in apparent diffusion coefficient during focalischemia of rat brain: the relationship of quantitative diffusion NMR imaging toreduction in cerebral blood flow and metabolic disturbances.” J Cereb Blood FlowMetab 15 (6): 1002-11.

49

Horvath AR, Muszbek L and Kellie S (1992). “Translocation of pp60c-src to thecytoskeleton during platelet aggregation.” Embo J 11 (3): 855-61.

Iihara K, Sasahara M, Hashimoto N, Uemura Y, Kikuchi H and Hazama F(1994). “Ischemia induces the expression of the platelet-derived growth factor-B chainin neurons and brain macrophages in vivo.” J Cereb Blood Flow Metab 14 (5): 818-24.

Irving EA, Barone FC, Reith AD, Hadingham SJ and Parsons AA (2000).“Differential activation of MAPK/ERK and p38/SAPK in neurones and glia followingfocal cerebral ischaemia in the rat.” Brain Res Mol Brain Res 77 (1): 65-75.

Jackson SP, Schoenwaelder SM, Yuan Y, Rabinowitz I, Salem HH and MitchellCA (1994). “Adhesion receptor activation of phosphatidylinositol 3-kinase. vonWillebrand factor stimulates the cytoskeletal association and activation ofphosphatidylinositol 3-kinase and pp60c-src in human platelets.” J Biol Chem 269(43): 27093-9.

Jean WC, Spellman SR, Nussbaum ES and Low WC (1998). “Reperfusion injuryafter focal cerebral ischemia: the role of inflammation and the therapeutic horizon.”Neurosurgery 43 (6): 1382-96.

Jope RS, Zhang L and Song L (2000). “Peroxynitrite modulates the activation of p38and extracellular regulated kinases in PC12 cells.” Arch Biochem Biophys 376 (2):365-70.

Joyeux M, Boumendjel A, Carroll R, Ribuot C, Godin-Ribuot D and Yellon DM(2000). “SB 203580, a mitogen-activated protein kinase inhibitor, abolishes resistanceto myocardial infarction induced by heat stress.” Cardiovasc Drugs Ther 14 (3): 337-43.

Junghans U, Seitz RJ, Ritzl A, Wittsack HJ, Fink GR, Freund HJ, et al. (2002).“Ischemic brain tissue salvaged from infarction by the GP IIb/IIIa platelet antagonisttirofiban.” Neurology 58 (3): 474-6.

Kato Y, Lewalle JM, Baba Y, Tsukuda M, Sakai N, Baba M, et al. (2001).“Induction of sparc by vegf in human vascular endothelial cells.” Biochem BiophysRes Commun 287 (2): 422-6.

Klatzo I (1967). “Presidental address. Neuropathological aspects of brain edema.” JNeuropathol Exp Neurol 26 (1): 1-14.

Koizumi J, Yoshida Y, Nakazawa T and Ooneda G (1986). “Experimental studiesof ischemic brain edema, I: a new experimental model of cerebral embolism in rats inwhich recirculation can be introduced in the ischemic area.” Jpn J Stroke 8: 1-8.

Kovács Z, Ikezaki K, Samoto K, Inamura T and Fukui M (1996). “VEGF and flt -Expression Time Kinetics in Rat Brain Infarct.” Stroke 27: 1865-73.

50

Krupinski J, Issa R, Bujny T, Slevin M, Kumar P, Kumar S, et al. (1997). “Aputative role for platelet-derived growth factor in angiogenesis and neuroprotectionafter ischemic stroke in humans.” Stroke 28 (3): 564-73.

Krupinski J, Kaluza J, Kumar P, Kumar S and Wang JM (1994). “Role ofangiogenesis in patients with cerebral ischemic stroke.” Stroke 25 (9): 1794-8.

Krupinski J, Kumar P, Kumar S and Kaluza J (1996). “Increased Expression ofTGF-ß1 in Brain Tissue After Ischemic Stroke in Humans.” Stroke 27: 852-57.

Kyriakis JM and Avruch J (1990). “pp54 microtubule-associated protein 2 kinase. Anovel serine/threonine protein kinase regulated by phosphorylation and stimulated bypoly-L- lysine.” J Biol Chem 265 (28): 17355-63.

Lee JC, Laydon JT, McDonnell PC, Gallagher TF, Kumar S, Green D, et al.(1994). “A protein kinase involved in the regulation of inflammatory cytokinebiosynthesis.” Nature 372 (6508): 739-46.

Legos JJ, Erhardt JA, White RF, Lenhard SC, Chandra S, Parsons AA, et al.(2001). “SB 239063, a novel p38 inhibitor, attenuates early neuronal injury followingischemia.” Brain Res 892 (1): 70-7.

Li F, Omae T and Fisher M (1999). “Spontaneous hyperthermia and its mechanismin the intraluminal suture middle cerebral artery occlusion model of rats.” Stroke 30(11): 2464-70.

Li F, Silva MD, Liu KF, Helmer KG, Omae T, Fenstermacher JD, et al. (2000).“Secondary decline in apparent diffusion coefficient and neurological outcomes after ashort period of focal brain ischemia in rats.” Ann Neurol 48 (2): 236-44.

Lindner V and Reidy MA (1996). “Expression of VEGF receptors in arteries afterendothelial injury and lack of increased endothelial regrowth in response to VEGF.”Arterioscler Thromb Vasc Biol 16 (11): 1399-405.

Liu T, Clark RK, McDonnell PC, Young PR, White RF, Barone FC, et al. (1994).“Tumor necrosis factor-alpha expression in ischemic neurons.” Stroke 25 (7): 1481-8.

Liu Y, Bishop A, Witucki L, Kraybill B, Shimizu E, Tsien J, et al. (1999).“Structural basis for selective inhibition of Src family kinases by PP1.” Chem Biol 6(9): 671-8.

Liu Y, Zhang G, Gao C and Hou X (2001). “NMDA receptor activation results intyrosine phosphorylation of NMDA receptor subunit 2A(NR2A) and interaction ofPyk2 and Src with NR2A after transient cerebral ischemia and reperfusion.” Brain Res909 (1-2): 51-8.

Longa EZ, Weinstein PR, Carlson S and Cummins R (1989). “Reversible middlecerebral artery occlusion without craniectomy in rats.” Stroke 20 (1): 84-91.

51

Lowell CA, Fumagalli L and Berton G (1996). “Deficiency of Src family kinasesp59/61hck and p58c-fgr results in defective adhesion-dependent neutrophil functions.”J Cell Biol 133 (4): 895-910.

Lu WY, Xiong ZG, Lei S, Orser BA, Dudek E, Browning MD, et al. (1999). “G-protein-coupled receptors act via protein kinase C and Src to regulate NMDAreceptors.” Nat Neurosci 2 (4): 331-8.

Marshall CJ (1995). “Specificity of receptor tyrosine kinase signaling: transientversus sustained extracellular signal-regulated kinase activation.” Cell 80 (2): 179-85.

Marti HJ, Bernaudin M, Bellail A, Schoch H, Euler M, Petit E, et al. (2000).“Hypoxia-induced vascular endothelial growth factor expression precedesneovascularization after cerebral ischemia.” Am J Pathol 156 (3): 965-76.

Matsuo Y, Onodera H, Shiga Y, Nakamura M, Ninomiya M, Kihara T, et al.(1994). “Correlation between myeloperoxidase-quantified neutrophil accumulationand ischemic brain injury in the rat. Effects of neutrophil depletion.” Stroke 25 (7):1469-75.

Minami M, Kuraishi Y, Yabuuchi K, Yamazaki A and Satoh M (1992). “Inductionof interleukin-1 beta mRNA in rat brain after transient forebrain ischemia.” JNeurochem 58 (1): 390-2.

Miyakawa T, Yagi T, Kagiyama A and Niki H (1996). “Radial maze performance,open-field and elevated plus-maze behaviors in Fyn-kinase deficient mice: furtherevidence for increased fearfulness.” Brain Res Mol Brain Res 37 (1-2): 145-50.

Mocsai A, Jakus Z, Vantus T, Berton G, Lowell CA and Ligeti E (2000). “Kinasepathways in chemoattractant-induced degranulation of neutrophils: the role of p38mitogen-activated protein kinase activated by Src family kinases.” J Immunol 164 (8):4321-31.

Nag S, Takahashi JL and Kilty DW (1997). “Role of Vascular Endothelial GrowthFactor in Blood-Brain Barrier Breakdown and Angiogenesis in Brain Trauma.” JNeuropathol Exp Neurol 56 (8): 912-21.

Nakamura K, Hori T, Sato N, Sugie K, Kawakami T and Yodoi J (1993). “Redoxregulation of a src family protein tyrosine kinase p56lck in T cells.” Oncogene 8 (11):3133-9.

Neufeld G, Cohen T, Gitay-Goren H, Poltorak Z, Tessler S, Sharon R, et al.(1996). “Similarities and differences between the vascular endothelial growth factor(VEGF) splice variants.” Cancer Metastasis Rev 15 (2): 153-8.

Neumann-Haefelin T, Kastrup A, de Crespigny A, Yenari MA, Ringer T, SunGH, et al. (2000). “Serial MRI after transient focal cerebral ischemia in rats: dynamics

52

of tissue injury, blood-brain barrier damage, and edema formation.” Stroke 31 (8):1965-72.

Papavassiliou E, Gogate N, Proescholdt M, Heiss JD, Walbridge S, Edwards NA,et al. (1997). “Vascular endothelial growth factor (vascular permeability factor)expression in injured rat brain.” J Neurosci Res 49 (4): 451-60.

Paul R, Zhang ZG, Eliceiri BP, Jiang Q, Boccia AD, Zhang RL, et al. (2001). “Srcdeficiency or blockade of Src activity in mice provides cerebral protection followingstroke.” Nat Med 7 (2): 222-7.

Pedram A, Razandi M and Levin ER (1998). “Extracellular signal-regulated proteinkinase/Jun kinase cross-talk underlies vascular endothelial cell growth factor-inducedendothelial cell proliferation.” J Biol Chem 273 (41): 26722-8.

Pei L, Li Y, Yan JZ, Zhang GY, Cui ZC and Zhu ZM (2000). “Changes andmechanisms of protein-tyrosine kinase and protein-tyrosine phosphatase activities afterbrain ischemia/reperfusion.” Acta Pharmacol Sin 21 (8): 715-20.

Pei L, Li Y, Zhang GY, Cui ZC and Zhu ZM (2000). “Mechanisms of regulation oftyrosine phosphorylation of NMDA receptor subunit 2B after cerebralischemia/reperfusion.” Acta Pharmacol Sin 21 (8): 695-700.

Plate KH, Beck H, Danner S, Allegrini PR and Wiessner C (1999). “Cell typespecific upregulation of vascular endothelial growth factor in an MCA-occlusionmodel of cerebral infarct.” J Neuropathol Exp Neurol 58 (6): 654-66.

Raingeaud J, Gupta S, Rogers JS, Dickens M, Han J, Ulevitch RJ, et al. (1995).“Pro-inflammatory cytokines and environmental stress cause p38 mitogen- activatedprotein kinase activation by dual phosphorylation on tyrosine and threonine.” J BiolChem 270 (13): 7420-6.

Rousseau S, Houle F, Landry J and Huot J (1997). “p38 MAP kinase activation byvascular endothelial growth factor mediates actin reorganization and cell migration inhuman endothelial cells.” Oncogene 15 (18): 2169-77.

Rundén E, Seglen PO, Haug FM, Ottersen OP, Wieloch T, Shamloo M, et al.(1998). “Regional selective neuronal degeneration after protein phosphatase inhibitionin hippocampal slice cultures: evidence for a MAP kinase- dependent mechanism.” JNeurosci 18 (18): 7296-305.

Rusanescu G, Qi H, Thomas SM, Brugge JS and Halegoua S (1995). “Calciuminflux induces neurite growth through a Src-Ras signaling cassette.” Neuron 15 (6):1415-25.

Schaeffer HJ and Weber MJ (1999). “Mitogen-activated protein kinases: specificmessages from ubiquitous messengers.” Mol Cell Biol 19 (4): 2435-44.

53

Schmid-Elsaesser R, Zausinger S, Hungerhuber E, Baethmann A, Reulen HJ andGarcia JH (1998). “A critical reevaluation of the intraluminal thread model of focalcerebral ischemia : evidence of inadvertent premature reperfusion and subarachnoidhemorrhage in rats by laser-doppler flowmetry.” Stroke 29 (10): 2162-70.

Schoenwaelder SM, Jackson SP, Yuan Y, Teasdale MS, Salem HH and MitchellCA (1994). “Tyrosine kinases regulate the cytoskeletal attachment of integrin alphaIIb beta 3 (platelet glycoprotein IIb/IIIa) and the cellular retraction of fibrin polymers.”J Biol Chem 269 (51): 32479-87.

Schratzberger P, Schratzberger G, Silver M, Curry C, Kearney M, Magner M, etal. (2000). “Favorable effect of VEGF gene transfer on ischemic peripheralneuropathy.” Nat Med 6 (4): 405-13.

Senger DR, Perruzzi CA, Feder J and Dvorak HF (1986). “A highly conservedvascular permeability factor secreted by a variety of human and rodent tumor celllines.” Cancer Res 46 (11): 5629-32.

Shibuya M, Ito N and Claesson-Welsh L (1999). “Structure and function of vascularendothelial growth factor receptor-1 and -2.” Curr Top Microbiol Immunol 237: 59-83.

Shimizu S, Simon RP and Graham SH (1997). “Dimethylsulfoxide (DMSO)treatment reduces infarction volume after permanent focal cerebral ischemia in rats.”Neurosci Lett 239 (2-3): 125-7.

Shioda S, Ozawa H, Dohi K, Mizushima H, Matsumoto K, Nakajo S, et al. (1998).“PACAP protects hippocampal neurons against apoptosis: involvement of JNK/SAPKsignaling pathway.” Ann N Y Acad Sci 865: 111-7.

Sondell M, Lundborg G and Kanje M (1999). “Vascular endothelial growth factorhas neurotrophic activity and stimulates axonal outgrowth, enhancing cell survival andSchwann cell proliferation in the peripheral nervous system.” J Neurosci 19 (14):5731-40.

Sondell M, Sundler F and Kanje M (2000). “Vascular endothelial growth factor is aneurotrophic factor which stimulates axonal outgrowth through the flk-1 receptor.”Eur J Neurosci 12 (12): 4243-54.

Stein I, Itin A, Einat P, Skaliter R, Grossman Z and Keshet E (1998). “Translationof vascular endothelial growth factor mRNA by internal ribosome entry: implicationsfor translation under hypoxia.” Mol Cell Biol 18 (6): 3112-9.

Suarez S and Ballmer-Hofer K (2001). “VEGF transiently disrupts gap junctionalcommunication in endothelial cells.” J Cell Sci 114 (Pt 6): 1229-35.

54

Sugino T, Nozaki K, Takagi Y, Hattori I, Hashimoto N, Moriguchi T, et al.(2000). “Activation of mitogen-activated protein kinases after transient forebrainischemia in gerbil hippocampus.” J Neurosci 20 (12): 4506-14.

Takagi N, Cheung HH, Bissoon N, Teves L, Wallace MC and Gurd JW (1999).“The effect of transient global ischemia on the interaction of Src and Fyn with the N-methyl-D-aspartate receptor and postsynaptic densities: possible involvement of Srchomology 2 domains.” J Cereb Blood Flow Metab 19 (8): 880-8.

Tamura A, Graham DI, McCulloch J and Teasdale GM (1981). “Focal cerebralischaemia in the rat: 1. Description of technique and early neuropathologicalconsequences following middle cerebral artery occlusion.” J Cereb Blood Flow Metab1 (1): 53-60.

Thomas SM and Brugge JS (1997). “Cellular functions regulated by Src familykinases.” Annu Rev Cell Dev Biol 13: 513-609.

Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp MV, et al.(1999). “Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols.” J MagnReson Imaging 10 (3): 223-32.

Umemori H, Sato S, Yagi T, Aizawa S and Yamamoto T (1994). “Initial events ofmyelination involve Fyn tyrosine kinase signalling.” Nature 367 (6463): 572-76.

van Bruggen N, Thibodeaux H, Palmer JT, Lee WP, Fu L, Cairns B, et al. (1999).“VEGF antagonism reduces edema formation and tissue damage afterischemia/reperfusion injury in the mouse brain.” J Clin Invest 104 (11): 1613-20.

Vasile E, Qu H, Dvorak HF and Dvorak AM (1999). “Caveolae and vesiculo-vacuolar organelles in bovine capillary endothelial cells cultured with VPF/VEGF onfloating Matrigel-collagen gels.” J Histochem Cytochem 47 (2): 159-67.

Volonte D, Galbiati F, Pestell RG and Lisanti MP (2001). “Cellular stress inducesthe tyrosine phosphorylation of caveolin-1 (Tyr(14)) via activation of p38 mitogen-activated protein kinase and c- Src kinase. Evidence for caveolae, the actincytoskeleton, and focal adhesions as mechanical sensors of osmotic stress.” J BiolChem 276 (11): 8094-103.

Walton KM, DiRocco R, Bartlett BA, Koury E, Marcy VR, Jarvis B, et al. (1998).“Activation of p38MAPK in microglia after ischemia.” J Neurochem 70 (4): 1764-7.

Wilson KP, McCaffrey PG, Hsiao K, Pazhanisamy S, Galullo V, Bemis GW, et al.(1997). “The structural basis for the specificity of pyridinylimidazole inhibitors of p38MAP kinase.” Chem Biol 4 (6): 423-31.

55

Wu DC, Ye W, Che XM and Yang GY (2000). “Activation of mitogen-activatedprotein kinases after permanent cerebral artery occlusion in mouse brain.” J CerebBlood Flow Metab 20 (9): 1320-30.

Xi G, Hua Y, Keep RF, Duong HK and Hoff JT (2001). “Activation of p44/42mitogen activated protein kinases in thrombin- induced brain tolerance.” Brain Res895 (1-2): 153-9.

Xi G, Keep RF, Hua Y and Hoff JT (2000). “Thrombin preconditioning, heat shockproteins and thrombin-induced brain edema.” Acta Neurochir Suppl 76: 511-5.

Xi G, Keep RF, Hua Y, Xiang J and Hoff JT (1999). “Attenuation of thrombin-induced brain edema by cerebral thrombin preconditioning.” Stroke 30 (6): 1247-55.

Xia Z, Dickens M, Raingeaud J, Davis RJ and Greenberg ME (1995). “Opposingeffects of ERK and JNK-p38 MAP kinases on apoptosis.” Science 270 (5240): 1326-31.

Xia Z, Dudek H, Miranti CK and Greenberg ME (1996). “Calcium influx via theNMDA receptor induces immediate early gene transcription by a MAP kinase/ERK-dependent mechanism.” J Neurosci 16 (17): 5425-36.

Yamagishi S, Yonekura H, Yamamoto Y, Fujimori H, Sakurai S, Tanaka N, et al.(1999). “Vascular endothelial growth factor acts as a pericyte mitogen under hypoxicconditions.” Lab Invest 79 (4): 501-9.

Yang GY, Gong C, Qin Z, Liu XH and Lorris Betz A (1999). “Tumor necrosisfactor alpha expression produces increased blood-brain barrier permeability followingtemporary focal cerebral ischemia in mice.” Brain Res Mol Brain Res 69 (1): 135-43.

Yoshizumi M, Abe J, Haendeler J, Huang Q and Berk BC (2000). “Src and Casmediate JNK activation but not ERK1/2 and p38 kinases by reactive oxygen species.”J Biol Chem 275 (16): 11706-12.

Yu XM, Askalan R, Keil GJ, 2nd and Salter MW (1997). “NMDA channelregulation by channel-associated protein tyrosine kinase Src.” Science 275 (5300):674-8.

Zhang ZG, Zhang L, Jiang Q, Zhang R, Davies K, Powers C, et al. (2000). “VEGFenhances angiogenesis and promotes blood-brain barrier leakage in the ischemicbrain.” J Clin Invest 106 (7): 829-38.

Zhao TC, Hines DS and Kukreja RC (2001). “Adenosine-induced latepreconditioning in mouse hearts: role of p38 MAP kinase and mitochondrial K(ATP)channels.” Am J Physiol Heart Circ Physiol 280 (3): H1278-85.