Handbook of Clinical Neurology, Vol.

92 (3rd series)
Stroke, Part I
M. Fisher, Editor
# 2009 Elsevier B.V. All rights reserved

Chapter 2

The ischemic cascade and mediators of ischemic injury

MATTHIAS ENDRES1 *, ULRICH DIRNAGL1, AND MICHAEL A. MOSKOWITZ 2

1
Charite´-Universita€tsmedizin Berlin, Berlin, Germany
2
Massachusetts General Hospital and Harvard Medical School, Charlestown, MA, USA

Cerebrovascular disease ranks second as the cause of that cell death evolves in a temporal and spatial conti-
death worldwide; mortality over the first year after first nuum (Astrup et al., 1981). Almost immediately after
stroke is approximately 20%. The economic and social vessel occlusion, an ischemic core is defined that is
burdens of stroke, however, are not consequences of destined to die irrespective of therapeutic interven-
mortality; they are imposed by the large majority of tions unless blood flow is rapidly restored. Surround-
stroke patients who survive but are physically and men- ing this core lies the penumbra, which is functionally
tally disabled by stroke-induced brain damage (Stroke silent but metabolically still active and hence salvage-
Progress Review Group, 2002). able. Early reperfusion is the major target of most
Stroke is a heterogeneous disease and refers to an experimental interventions, to render cells in the pen-
umbrella of conditions. In the Western world, ischemic umbra resistant to cell death. Importantly, however,
stroke comprises 80–85%, while primary intracerebral the penumbra is dynamic: indeed, over time the ‘core
hemorrhages, subarachnoidal hemorrhages, and sinus grows at the cost of the penumbra’ and previously
thrombosis account for the remaining 15–20%. By con- viable brain becomes infarcted tissue (Ginsberg et al.,
trast, hemorrhagic stroke may comprise up to 50% of all 1999) (Fig. 2.1).
strokes in Asia. Ischemic stroke is caused by a transient
or permanent reduction of blood flow restricted to the 2.2. Active cell death mechanisms
territory of a cerebral artery—typically by embolic or
thrombotic occlusion. In any case, ischemic stroke Although the exact timing and cellular pathways are
eventually results in the death and dysfunction of brain incompletely understood it is commonly believed that
cells. Central to our current understanding of stroke mechanisms actively promoting cell death are trig-
pathophysiology is evidence derived from animal mod- gered after stroke (Lo et al., 2005). Cell death occurs
els. Mimicking the condition of ischemic stroke are by a necrotic pathway characterized by ischemic or
models of focal cerebral ischemia. The most frequently edematous cell changes, by an apoptotic pathway with
used species are rodents such as mice and rats but also a number of morphological (e.g., apoptotic bodies, bleb-
other mammals, including dogs, cats, and sheep as well bing), biochemical (e.g., DNA laddering), pharma-
as primates (Lo et al., 2003). cological, and molecular characteristics (e.g., activation
of caspases), or by autophagocytosis. A number of dif-
2.1. Temporal and spatial events after stroke: ferent stages of the cell death process and major patho-
the concept of an ischemic penumbra physiological pathways have emerged from the
literature (Dirnagl et al., 1999; Lo et al., 2005). First
Not all brain cells die immediately after an ischemic comes an induction stage with energy failure, increase
stroke. Ever since Astrup first introduced the concept of intracellular calcium, and release of excitatory
of an ischemic penumbra, i.e. a perilesional area that amino acids (Lipton et al., 1999). This in turn triggers
surrounds the ischemic core, it is generally accepted the activation of downstream perpetrators of ischemic

*
Correspondence to: Priv.-Doz. Dr M. Endres, Klinik und Poliklinik für Neurologie, Charité-Universitätsmedizin Berlin, Campus
Mitte, Schumannstrasse. 20/21, D-10117 Berlin, Germany. E-mail: matthias.endres@charite.de, Tel: þ49.30.450.560.257,
Fax: þ49.30.450.560.932.
32 M. ENDRES ET AL.
Morphology Biochemistry

Ionic failure
Infarction Anoxic depolarization

PENUMBRA CORE
Glucose utilization ↓
Glutamate release
Glucose utilization
Protein synthesis ↓
Inflammation Acidosis
and Oxygen extraction
Apoptosis
Selective gene expression

Fig. 2.1. The ischemic penumbra. A brain region of low perfusion in which cells have lost their membrane potential terminally
(“core”) is surrounded by an area in which intermediate perfusion prevails (“penumbra”)(from Dirnagl et al., 1999).

damage, including free radical and peroxynitrite pro- 2.3. Excitotoxicity, energy failure, and
duction, calpain, phospholipases, and poly-ADP- ionic imbalance
ribose polymerase activation. Concomitantly, apopto-
tic pathways are initiated. Waves of peri-infarct Brain is one of the most metabolically active organs
depolarization further compromise the energy balance and depends almost exclusively upon oxidative phos-
of ischemic neurons in the penumbra. Inflammation phorylation for energy. It is exquisitely sensitive to
then amplifies tissue damage. Secondary stages of disturbances in oxygen and glucose supply (Siesjo,
cell death may involve long-term changes in macro- 1978; Hansen, 1985; Erenciska and Silver, 1989). Fol-
molecules and other key metabolites. All these events lowing focal ischemia there is a profound deprivation
are potential targets for therapeutic interventions of oxygen and glucose. Available evidence suggests
(Fig. 2.2). that gray and white matter both suffer immediate loss

Energy failure Peri-infarct

Gl
G

Na+ K+
Gl
u
Ca2+ Depolari-
u

zation
Gl

K+ Glu

Na+ Cell Glu

Glutamate swelling
release Ca2+

Mitochondrial damage
Enzyme-
induction
DNA damage

Free radicals
Membrane NO
Apoptosis
degradation

Inflammatory
mediators

Microglial Leukocyte
activation infiltration

Fig. 2.2. Simplified overview of pathophysiological mechanisms in the focally ischemic brain (from Dirnagl et al., 1999).
 THE ISCHEMIC CASCADE AND MEDIATORS OF ISCHEMIC INJURY 33
of function with anoxia (Stys et al., 1992). Within as robust as focal ischemia (Kondo et al., 1997; Fuji-
minutes both neuronal and non-neuronal cells become mura et al., 1998, 1999). Mice overexpressing superox-
depolarized (e.g., anoxic depolarization) and voltage- ide dismutase have reduced injury following focal
dependent calcium channels are activated (Martin ischemia, giving support to the notion that interven-
et al., 1994; Paschen, 2000). Depolarization also tions aimed at enhancing endogenous repair mechan-
induces the release of (predominantly excitatory) neu- ism may be attractive therapeutic targets for ischemic
rotransmitters from presynaptic terminals into the stroke (Kinouchi et al., 1991; Sheng et al., 1999).
synaptic cleft (Zipfel et al., 1999). In particular, bind- In addition to reactive free oxygen species, nitrosa-
ing of glutamate to ionotropic N-methyl-d-aspartate tive stress contributes to tissue damage. Nitric oxide
(NMDA) and a-amino-3-hydroxy-5-methyl-4-isoazole (NO) itself may play both beneficial and deleterious
propionic acid (AMPA) receptors promotes excessive roles during brain ischemia depending upon when
calcium influx. In turn, increased intracellular calcium and where it is produced. NO is synthesized from
levels act as universal second and third messenger L-arginine by several isoforms of NO synthase (NOS;
to trigger an array of downstream phospholipases i.e. NOS I, II, and III). In situations of increased oxida-
and proteases that degrade membrane and proteins tive stress NO reacts with superoxide anions to gener-
essential for cellular integrity (such as actin, spectrin, ate the highly reactive and cytotoxic peroxynitrite,
laminin, etc.; Chen et al., 1998; Furukawa et al., which may damage virtually every cellular component
1997; Endres et al., 1999). Calcium ions and other ions (Beckman and Koppenol, 1997; Iadecola, 1997). Dur-
also enter mitochondria through the mitochondrial per- ing ischemia constitutive, calcium-dependent neuronal
meability transition pore, causing dysfunction and type I NOS (nNOS) may be activated via calcium
mitochondrial swelling. Moreover, sodium and chlor- influx. Within minutes after induction of focal cerebral
ide enter neurons via channels for monovalent ions ischemia cortical NO levels may rise from 10 nmol/L
(e.g., the AMPA receptor) passively followed by to 2 mmol/L (Dalkara and Moskowitz, 1994). In addi-
water. The ensuing intracellular (‘cytotoxic’) edema tion, inducible (type II) NOS (iNOS), which is cal-
may negatively impact perfusion in the peri-infarct cium/calmodulin-independent and is not normally
region. Imbalances of other ions are also important: present in healthy brain tissue, is induced after brain
for example, large amounts of zinc that are stored in ischemia in non-neuronal cells such as microglia and
vesicles of excitatory neurons become released upon leukocytes, but also astrocytes and endothelial cells.
depolarization and contribute to excitotoxic cell death Once expressed, iNOS is constitutively active and
(Frederickson, 1989; Weiss et al., 1993; Sorensen confers high NO output, contributing to secondary
et al., 1998). late-phase tissue damage (Iadecola et al., 1997).
Experiments using knockout mice have helped to clar-
2.4. Oxidative and nitrosative stress ify the respective roles of the different NOS isoforms
in ischemic injury. Animals that lack expression of
As a consequence of ischemia but particularly after nNOS (so called nNOS knockout mice) have 40%
reperfusion, reactive free oxygen species are generated. smaller ischemia lesion volumes (Huang et al.,
These reactive free oxygen species are also key media- 1994). Similarly, pharmacological inhibition of iNOS
tors of tissue damage in other organs subjected to reper- or deletion of the iNOS gene leads to a reduction in
fusion injury, such as heart or kidney. Free radicals ischemic damage (Iadecola et al., 1997). Interestingly,
trigger a vicious cycle in the mitochondria with inhibi- iNOS is not expressed before 24 hours after ischemia
tion of electron transport mechanisms leading to excess onset, and iNOS inhibition reduces tissue damage even
superoxide production and also to activation of mito- when administered as late as 1 day after the occlusion,
chondrial permeability transition (Fiskum et al., 1999; which is consistent with the hypothesis that ischemic
Kroemer and Reed, 2000; Chan, 2001). Additional injury indeed evolves over several days.
sources of reactive free oxygen species are enzymatic The generation of nitric oxide and reactive free oxy-
processes such as the cyclooxygenase-dependent con- gen species is linked to DNA damage and activation
version of arachidonic acid to prostanoids, or the degra- of the nuclear enzyme poly-ADP-ribose polymerase
dation of hypoxanthine. In fact, oxidative stress is (PARP-1). PARP-1 is activated by single-stranded
closely linked to excitotoxicity, energy loss, and ionic DNA nicks (which may be generated by oxidative and
imbalances, and all these events contribute to tissue nitrosative damage) and in turn consumes large amounts
damage. Injury induced by oxidative stress is mitigated of b-nicotinamide adenine dinucleotide (NADþ) to pro-
by endogenous mechanisms involving superoxide duce poly-ADP chains for post-translational modifica-
dismutase or glutathione; however, these scavenging tion of a number of repair enzymes including PARP-1
mechanisms are typically insufficient during an injury itself (Zhang et al., 1994). It has been assumed that the
34 M. ENDRES ET AL.
consumption of NADþ and subsequent depletion of (1) The total number and frequency of peri-infarct
ATP results in cellular energy depletion and subsequent depolarizations correlate with the final lesion volume
cell death (Zhang et al., 1994; Eliasson et al., 1997). and cell loss and it has been shown that each propagat-
Ischemic cell death is suppressed by inhibiting PARP-1 ing depolarization contributes to the immediate growth
activity or by deleting the PARP1 gene, indicating the of the lesion from the core to the periphery (Mies
potential of PARP-1 as a therapeutic target (Eliasson et al., 1993; Back et al., 1994, 1996; Hoehn-Berlage,
et al., 1997; Endres et al., 1997). Consistent with the 1995); (2) therapeutic interventions aimed at reducing
notion that PARP-1 is a downstream effector of NO- peri-infarct depolarizations (such as NMDA and
mediated neurotoxicity are experiments to demonstrate AMPA receptor blockade) provide tissue sparing
that nNOS knockout mice have significantly reduced (Iijima et al., 1992) while (3) induction of additional
levels of poly-ADP-ribose formation in ischemic tissue peri-infarct depolarizations (e.g., by administration of
compared to wild-type controls (Endres et al., 1998a; topical potassium chloride solution) further increases
Dawson and Dawson, 2004). infarct volume. So far, however, convincing evidence
In contrast to nNOS and iNOS, an additional iso- that peri-infarct depolarizations occur after ischemic
form, type III or endothelial NOS (eNOS), plays a stroke in humans is lacking (Back et al., 2000).
protective role during cerebral ischemia (Endres
et al., 2004). Endothelial NO possesses vasodilatory, 2.6. Inflammation and immunity
anti-inflammatory, antithrombotic, and antiprolifera-
tive properties. NO generated by eNOS is crucial in Early after the onset of ischemia, the expression of pro-
vascular function, maintaining cerebral blood flow inflammatory genes, including those encoding NF-kB,
and reducing cerebral infarct volume, especially in hypoxia-inducible factor, and interferon 1b, is triggered
the early stages following cerebral ischemia. Animals (Pirttilä and Kauppinen, 1992; Lindsberg et al., 1996;
lacking eNOS expression develop larger cerebral Iadecola, 1997). A cascade of events is then initiated
infarctions following focal brain ischemia (Huang that includes the expression of adhesion molecules
et al., 1994, 1996). Several modalities that upregulate (such as intercellular and vascular adhesion molecules,
eNOS expression and/or activity have been identified, ICAM and VCAM, as well as selectins), endothelial
including HMG-CoA reductase inhibitors (statins), activation, pro-inflammatory and pro-thrombotic inter-
physical activity, steroid hormones, and nutrients actions between vessel wall and blood constituents pro-
(Endres et al., 1998b; Hafezi-Moghadam et al., 2002; moting thrombogenesis, and microvascular plugging
Endres et al., 2004). They all increase NO bioavail- (Connolly et al., 1996; Stanimirovich and Satoh, 2000;
ability, leading to enhanced cerebral blood flow and Frijns and Kappelle, 2002). Of note, serological markers
protection from ischemic stroke. of vessel inflammation including C-reactive protein and
sICAM have been linked to increased stroke risk (Rid-
2.5. Peri-infarct depolarizations ker et al., 2000; Tanne et al., 2002).
Adhesion molecules expressed by the endothelium
Anoxic depolarization is caused by the loss of energy interact with receptors on neutrophils to promote their
and oxygen in neurons and non-neuronal cells and entry into brain, followed days later by macrophages and
the acute release of glutamate and other excitatory monocytes (Iadecola, 1997). Inflammatory responses,
amino acids. Cells in the core of the lesion remain however, are mediated not only by bone marrow and
depolarized and go on to die. In the peri-infarct zone, blood-derived cells but also by immunocompetent cells
cells may repolarize, however at a high cost of energy resident in the brain such as microglia. These cells consti-
expenditure. Following depolarization neurotransmit- tute up to 20% of the total brain cell number and become
ters and potassium are released into the extracellular activated after brain insults—particularly in the penum-
space and then initiate repetitive waves of so-called bra (Tarozzo et al., 2002). Typically, microglial cells
peri-infarct depolarizations (Nedergaard and Hansen, retract their processes and adapt an ‘ameboid’ morphol-
1993; Back et al., 1994, 1996). In animals, these ogy as they become activated. Similar to leukocytes, acti-
peri-infarct depolarizations occur at a frequency of vated microglia produce a plethora of pro-inflammatory
1–4 events per hour for at least 6–8 hours. They typi- cytokines along with toxic metabolites (Yrjanheikki
cally originate in the core of the lesion and then propa- et al., 1999). Recent evidence demonstrates that bone-
gate to the periphery in a wave-like fashion (Busch marrow-derived mononuclear cells that invade the brain
et al., 1996; Wolf et al., 1997). There are several lines upon ischemic insult may differentiate into microglia
of evidence from ischemic animal models (in mouse, and become part of the surviving brain (Priller et al.,
rat, and cat) that the phenomenon of peri-infarct 2001). Also, macroglia such as astrocytes become acti-
depolarizations in fact contributes to tissue damage: vated and proliferate. These cells also contribute to the
 THE ISCHEMIC CASCADE AND MEDIATORS OF ISCHEMIC INJURY 35
production of pro-inflammatory cytokines and also neu- and Gill, 1996), a number of molecular, biochemical,
roprotective factors including erythropoietin, transform- and pharmacological criteria to define apoptotic cell
ing growth factor (TGF)-b, and metallothionein death are fulfilled in ischemic neuronal death. Both
(Letterio, 2000; Ruscher et al., 2002; Trendelenburg caspase-dependent and independent mechanisms have
et al., 2002). been described. Caspases are aspartate-specific
There is evidence to support the notion that the cysteine proteases, constitutively expressed in brain,
inflammatory reaction following ischemic stroke is including neurons, and are activated by intrinsic and
both protective and destructive. In support of the latter extrinsic stimuli (Yuan and Yanker, 2000). Extrinsic
notion, it was demonstrated that ischemic brain injury pathways include the activation of death receptors,
is reduced either when neutrophil invasion is blocked e.g., Fas and tumor necrosis factor (TNF)-a receptors
by systemic neutropenia or when the expression of (Martin-Villaba et al., 1999). Upon activation, Fas
adhesion molecules is blocked by specific antibodies assembles as part of a death-inducing signaling com-
or via genetic deletion (e.g., using ICAM-1 antibodies plex along with Fas-associated protein with death
or ICAM-1-knockout mice; Connolly et al., 1996). domain (FADD) and pro-caspase-8 (Qiu et al., 2002).
In addition, it has recently become clear that brain Intrinsic activators comprise oxygen radical formation,
ischemia also increases susceptibility to systemic DNA damage, rise in intracellular calcium, and lysoso-
infection by brain-specific mechanisms (Meisel et al., mal protease activation (Nicotera and Lipton, 1999;
2005). Ischemic brain injury disturbs the homeostasis Budd et al., 2000; Salvesen, 2001). Once activated,
between the central nervous system and the immune caspases cleave a number of downstream substrates
system, leading to immunosuppression. In fact, in a that include other (‘executioner’) caspases, DNA
rodent model of focal brain ischemia neuroendocrine- repair enzymes such as PARP, cytoskeletal proteins,
mediated systemic immunosuppression results in the presenilin, huntingtin, and caspase-activated deoxyri-
development of spontaneous systemic bacterial infec- bonuclease (ICAD). The most important executioner
tions (Prass et al., 2003). caspase in the brain is caspase-3, which is activated
early after ischemia, particularly in the peri-infarct
2.7. Apoptosis-like pathways region (Namura et al., 1998; Endres et al., 1998c).
Cytochrome c plays an important role in the activation
Loss of membrane integrity, cell swelling, and organelle of caspases. Once released from the mitochondria
failure, i.e. typical characteristics of necrosis, are the cytochrome c forms an ‘apoptosome’ complex in the
most prominent features of cell death following presence of ATP and caspase-9. Apoptosome-forma-
brain ischemia (Majno and Joris, 1995). There has been tion is suppressed by Bcl-Xl (or BCL2L1) (Green
vigorous scientific debate about the mechanism of and Reed, 1998) while cytosolic Bid, a pro-apoptotic
ischemic cell death with respect to apoptosis versus Bcl-2 family member, facilitates cytochrome c release
necrosis, although most agree that necrotic cell death and apoptosome formation (Wei et al., 2001). Down-
predominates in ischemic stroke. The term ‘apoptosis’ stream markers of apoptosis include the appearance
was introduced into the scientific literature by Kerr in of terminal deoxynucleotidyl transferase-mediated
1972, although the typical morphological hallmarks of dUTP-biotin nick-end labeled (TUNEL) cells and bio-
apoptotic cell death were described by pathologists more chemical evidence of oligonucleosomal DNA frag-
than 100 years ago (Kerr et al., 1972; Majno and Joris, mentation (‘DNA laddering’), which has been
1995). These hallmarks include cytoplasmic shrinkage, demonstrated in stroke models in numerous reports
blebbing, nuclear segmentation, chromatin condensation, (Linnik et al., 1993; MacManus et al., 1993; Tominaga
and formation of apoptotic bodies. In general, the ‘deci- et al., 1993; Li et al., 1995a, b; Endres et al., 1998c).
sion’ for apoptotic versus necrotic cell death depends on Inhibition of caspase, for example using small oligo-
the nature and intensity of the stimulus, the type of cell, peptides that mimic the cleavage site of a caspase sub-
and the stage reached in its developmental life-cycle strate, protects from cerebral ischemia (Hara et al.,
(Bonfoco et al., 1995; Nicotera et al., 2000). In line with 1997; Fink et al., 1998; Endres et al., 1998c). Interest-
this, a number of reports have proposed that necrosis is ingly, following mild insults, caspase-3 activation is
the predominant mechanism that follows acute, perma- delayed until 9 hours, consistent with the treatment
nent vascular occlusion, whereas in milder injury, window for caspase-inhibitors, which is also extended
cell suicide becomes unmasked and death resembles to 9 hours after vessel occlusion (Fink et al., 1998).
apoptosis, particularly within the ischemic penumbra Moreover, animals with reduced expression of cas-
(Du et al., 1996; Endres et al., 1998c). pase-1 or -3 are protected from mild experimental
Although bona fide morphological criteria of apop- ischemia (Friedlander et al., 1997; Schielke et al.,
tosis are apparently lacking (van Lookeren Campagne 1998).
36 M. ENDRES ET AL.
In addition, there is evidence for caspase- 2.8. Stroke-induced endogenous
independent apoptotic mechanisms in cerebral ische- neuroprotection
mia. Apoptosis-inducing factor (AIF) is considered as
a key-signaling molecule in this cascade (Susin et al., Historically, research on the pathophysiology of stroke
1999). Apparently, PARP-1 activation may promote was biased towards investigating mechanisms of tissue
the mitochondrial release of AIF. AIF then re-locates damage, while endogenous protective responses to
to the nucleus, where it promotes chromatin condensa- challenges like substrate deprivation and the secondary
tion. Cell death by AIF is resistant to treatment with effectors of cell death (reviewed above) received little
pan-caspase inhibitors but can be suppressed by neu- attention. Recently, however, it became clear that
tralizing AIF before nuclear translocation, supporting the final outcome of a stroke is the result not only of
the notion that AIF acts independently of caspases. tissue destruction and reorganization or plasticity but
A growing body of evidence indicates that compo- also of endogenous neuroprotection.
nents of the cell cycle machinery become activated Central to this research are models of ischemic pre-
in neurons subjected to cell death stimuli and play spe- conditioning (Dirnagl et al., 2003). Ischemic precondi-
cific roles in their death (Katchanov et al., 2001; tioning refers to a process of protecting cells, tissues,
Herrup and Arendt, 2002; Greene et al., 2004). In this or whole organisms induced by one or more exposures
context it has been suggested that multiple cyclin- to a subliminal noxious stimulus (trigger). Subliminal
dependent kinases (CDKs) may participate in neuronal ischemia (‘ischemic tolerance’), hypoxia, reactive
death and two central hypotheses have been espoused. oxygen free radicals, inflammation, etc. can serve as
The first describes the paradoxical situation by which tolerance-inducing stimuli. Studies in patients experien-
terminally differentiated post-mitotic neurons reacti- cing transient ischemic attacks indicate that ischemic
vate components of the cell cycle machinery, includ- tolerance does also exist in humans. A transient
ing CDKs, which however confers cell death instead ischemic attack, by definition, does not lead to infarc-
of mitosis (Copani et al., 2002). Indeed, activation tion. Subsequent to a transient ischemic attack, brain
and up regulation of a number of cell cycle compo- infarcts that occur are smaller compared to lesions
nents has been reported in several cell death models without preceding transient ischemic attack (Weih
including stroke (Timsit et al., 1999; Osuga et al., et al., 1999; Moncayo et al., 2000). In models of cerebral
2000; Wang et al., 2002; Qiu et al., 2002). Activation ischemia preconditioning, many sensors, transducers,
of the cell cycle machinery may be triggered by and effectors of stroke leading to endogenous neuro-
the stroke-induced downregulation of endogenously protection have been identified (Kirino, 2002; Dirnagl
expressed CDK inhibitors of the INK and KIP family, et al., 2003; Kariko et al., 2004; Trendelenburg
such as p16, p19, p21, or p27 (Katchanov et al., 2001). and Dirnagl, 2005). These include anti-excitotoxic,
A second hypothesis proposes that deregulated CDK 5, anti-inflammatory, and anti-apoptotic mechanisms.
which is not involved in cell cycle progression but Hypoxia-inducible-factor (HIF)-1-dependent signaling
rather implicated in transcription, neuronal function, may serve as a typical example of such an endogenous
and differentiation, induces neuronal damage. Activa- neuroprotective response. Focal cerebral ischemia leads
tion of CDK 5 following brain ischemia may be trig- to local hypoxia and consequently to the nuclear translo-
gered via calcium-mediated activation of the protease cation and DNA-binding of HIF-1, a key sensor of
calpain, which in turn converts p35 into a smaller, hypoxia in almost all mammalian cells. HIF-1, among
more stable and mislocalized p25 form. This then acti- other proteins, including glycolytic enzymes, induces
vates CDK 5 to induce cell death. Indeed, accumu- the expression of erythropoietin in astrocytes. Erythro-
lation of p25 after transient forebrain ischemia poietin, in a paracrine fashion, binds to neuronal ery-
activates CDK 5 and induces ischemic cell death thropoietin receptors. Via the activation of a protein
(Wang et al., 2003). In fact, Park and co-workers have kinase cascade, in particular phosphoinositol-3 kinase,
demonstrated that CDK 5 acts preferentially to regulate the pro-apoptotic protein BAD becomes phosphorylated
excitotoxic damage while cell-cycle-promoting CDKs and thus inactivated (Ruscher et al., 2002). By this
such as CDK 2, 4, and 6 are involved in more delayed mechanism HIF-1-induced erythropoietin expression
types of cell death (Rashidian et al., 2005). However, can salvage neurons challenged by ischemia, particu-
the downstream effectors that mediate neuron cell death larly in the penumbra. Understanding the signaling by
in response to cell cycle activation are largely unknown. which the brain inactivates mechanisms of cell death
Recently, the pro-apoptotic Bcl-2 homologue of BH3 in focal cerebral ischemia may lead us to novel strate-
domain-only molecule, Bim, has been identified as a gies for the treatment of patients with stroke. For
direct target of a neuronal E2F-dependent apoptotic example, erythropoietin, as well as other neuroprotec-
pathway (Biswas et al., 2005). tive hematopoietic cytokines expressed in the brain
 THE ISCHEMIC CASCADE AND MEDIATORS OF ISCHEMIC INJURY 37
(such as granulocyte colony stimulating factor and excitotoxic injury differs from gray matter as there
granulocyte–macrophage colony stimulating factors), are no synapses and vesicular release is minimal.
is presently being tested as a neuroprotectant in clinical Recently, a number of quantitative techniques have
stroke trials (Ratan et al., 2004). been developed to assess white matter injury as well
as to protect white matter in ischemia (Dewar et al.,
2.9. The blood–brain barrier, the 1999). Similar to gray matter, energy deprivation leads
microcirculation, and the neurovascular unit to dysfunction of axons via activation of voltage-
dependent sodium and calcium channels, and consecu-
The newly introduced concept of the ‘neurovascular tive activation of calcium-dependent disruptive
unit’ comprised of cerebral endothelial cells, astro- pathways (Goldberg and Ransom, 2003). Of interest,
cytes, and neurons, along with extracellular matrix, the white matter seems particularly resistant to hypo-
provides an opportunity to look beyond a single cell glycemia with astrocytic intracellular glycogen stores
and focus on cell–cell interaction—particularly after playing a protective role (Wender et al., 2000).
ischemic stroke (Lo et al., 2003, 2005). The integrity Indeed, the astrocyte glycogen content defines the
of the blood–brain barrier depends primarily on the period of hypoglycemia in which cells are resistant
interaction of the extracellular matrix with endothelial to injury. Particularly vulnerable to ischemic damage
cells and astrocytes. Perturbation of the extracellular are oligodendrocytes, the myelin-forming cells of the
matrix following ischemic stroke includes the degrada- central nervous system. These cells express non-
tion of type IV collagen, laminin, and fibronectin. NMDA glutamate receptors and can be damaged
A number of proteases such as cathepsins, plus hepar- and killed by glutamate excitotoxicity via intracellu-
anases, contribute to this, as do plasminogen activator lar calcium overload and subsequent calcium-
and matrix metalloproteinases (MMPs). MMP-2 and dependent cascades, as described above (McDonald
-9 are induced very early after ischemia onset, which et al., 1998). Notably, oligodendrocytic AMPA
correlates with the risk for hemorrhagic transforma- receptors lack the calcium-impermeable GluR2 subu-
tion, edema formation, and the extent of tissue demise nit, which may render these cells particularly vulner-
(Gasche et al., 1999; Heo et al., 1999; Lapchak et al., able to injury (Matute et al., 2002).
2000). Indeed, increased MMP plasma and brain levels
have been demonstrated in stroke patients (Clark et al., 2.11. Concluding remarks
1997; Montaner et al., 2001). In experimental stroke
models, MMP inhibitors reduce ischemic brain damage Cell death following brain ischemia is mediated by a
(Rosenberg and Navratil, 1997; Asahi et al., 2001) and complex interplay of a number of pathophysiologi-
MMP-9 knockout mice are protected against cerebral cally distinct mechanisms. Both blood vessels and
ischemia (Asahi et al., 2001) parenchyma have been implicated and these complex
interactions define the fate of compromised tissues
2.10. White matter and cells. The past two decades have witnessed an
enormous expansion in our understanding of events
In contrast to research on gray matter injury, white that determine the fate of brain cells following vessel
matter injury following stroke has been largely occlusion. In fact, novel death mechanisms have
neglected. Indeed, it may well be that the failure to been implicated to explain the diversity of cell death
develop clinically effective neuroprotective strategies phenotypes resembling necrosis, apoptosis, necropto-
for stroke in part relate to the inattention to white mat- sis, and autophagy in ischemic models. Their rele-
ter injury and the failure to ameliorate ischemic vance to embolic, thrombotic strokes and transient
damage to white matter. Most stroke models are per- ischemic attacks in humans, however, remains to be
formed in rodents, whose brains are a little less than determined. Each cell death mechanism critically
10% white matter by volume (Zhang and Sejnowski, relates in unique ways to energy depletion, deploy-
2000) while the human brain contains more than ment of endogenous proteases, oxygen radical gen-
50% white matter by volume. Hence, protecting gray eration, fluxing of cations, receptor activation, and
matter alone may have considerably less benefit in active cell death programs that determine the fate
humans than in rodents. However, much less is under- of injured parenchymal and vascular cells. Despite
stood about the cellular pathways of ischemic injury in this complexity, there is reason for optimism, as
white matter, which is composed of axons, oligoden- recent developments in molecular and cell biology,
drocytes, astrocytes, and blood vessels. Laminin, plus neuroimaging, are providing tools for better
fibronectin, and chondroitin proteoglycans are matrix translation of discoveries from bench to bedside
proteins that envelop these cells. Furthermore (Endres et al., 2007).
38 M. ENDRES ET AL.
References Dewar D, Yam P, McCulloch J (1999). Drug development
for stroke: importance of protecting cerebral white matter.
Asahi M, Wang X, Mori T, et al. (2001). Effects of matrix Eur J Pharmacol 375: 41–50.
metalloproteinase 9 gene knockout on the proteolysis of Dirnagl U, Iadecola C, Moskowitz MA (1999). Pathobiology
blood–brain barrier and white matter components after of ischaemic stroke. An integrated view. Trends Neurosci
cerebral ischemia. J Neurosci 21: 7724–7732. 22: 391–397.
Back T, Kohno K, Hossmann KA (1994). Cortical negative
Dirnagl U, Simon RP, Hallenbeck JM (2003). Ischemic tol-
DC deflections following middle cerebral artery occlusion
erance and endogenous neuroprotection. Trends Neurosci
and KCl-induced spreading depression: effect on blood
26: 248–254.
flow, tissue oxygenation, and electroencephalogram.
Du C, Hu R, Csernansky CA, et al. (1996). Very delayed
J Cereb Blood Flow Metab 14: 12–19.
infarction after mild focal cerebral ischemia: a role for
Back T, Ginsberg MD, Dietrich DW, et al. (1996). Induction
apoptosis? J Cereb Blood Flow Metab 16: 195–201.
of spreading depression in the ischemic hemisphere fol-
Eliasson MJL, Sampei K, Mandir AS, et al. (1997). Poly
lowing experimental middle cerebral artery occlusion:
(ADP-ribose)polymerase gene disruption renders mice
effect on infarct morphology. J Cereb Blood Flow Metab
resistant to cerebral ischemia. Nat Med 3: 1089–1095.
16: 202–213.
Endres M, Wang Z-Q, Namura S, et al. (1997). Ischemic
Back T, Hirsch JG, Szabo K (2000). Failure to demonstrate
brain injury is mediated by the activation of poly(ADP-
peri-infarct depolarizations by repetitive MR diffusion
ribose)polymerase. J Cereb Blood Flow Metab 17:
imaging in acute human stroke. Stroke 31: 2901–2906.
1143–1151.
Beckman JS, Koppenol WH (1997). Nitric oxide, superox-
Endres M, Scott GS, Namura S, et al. (1998a). Role of
ide, and peroxynitrite: the good, the bad, and ugly. Am J
perosynitrite and neuronal nitric oxide synthase in the
Physiol 271: C1424–C1437.
activation of poly(ADP-ribose)synthetase, against peroxy-
Biswas SC, Liu DX, Greene LA (2005). Bim is a direct tar-
nitrite-induced glial damage and stroke development.
get of a neuronal E2F-dependent apoptotic pathway.
Neurosci Lett 248: 41–44.
J Neurosci 25: 8349–8358.
Bonfoco E, Krainc D, Ankarcrona M, et al. (1995). Apoptosis Endres M, Laufs U, Huang Z, et al. (1998b). Stroke protec-
and necrosis: two distinct events induced, respectively, by tion by 3-hydroxy-3 methylglutaryl (HMG)-CoA reduc-
mild and intense insults with N-methyl-D-aspartate or tase inhibitors mediated by endothelial nitric oxide
nitric oxide/superoxide in cortical cell cultures. Proc Natl synthase. Proc Natl Acad Sci U S A 95: 8880–8885.
Acad Sci U S A 92: 72162–72166. Endres M, Namura S, Shimizu-Sasamata M, et al. (1998c).
Budd SL, Tenneti L, Lishnak T, et al. (2000). Mitochondrial Attenuation of delayed neuronal death after mild focal cer-
and extramitochondrial apoptotic signalling pathways in ebral ischemia in mice by inhibitors of the caspase family.
cerebrocortical neurons. Proc Natl Acad Sci U S A 97: J Cereb Blood Flow Metab 18: 238–247.
6161–6666. Endres M, Fink KB, Zhu J (1999). Neuroprotective effects of
Busch E, Gyngell ML, Eis M, et al. (1996). Potassium- gelsolin during murine stroke. J Clin Invest 103: 347–354.
induced cortical spreading depression during focal cere- Endres M, Laufs U, Liao JK, et al. (2004). Targeting ENOS
bral ischemia in rats: contribution to lesion growth for stroke protection. Trends Neurosci 27: 283–289.
assessed by diffusion-weighted NMR and biochemical Endres M, Engelhardt B, Koistinaho J, et al. (2008). Improv-
imaging. J Cereb Blood Flow Metab 16: 1090–1099. ing outcome after stroke: translational research after
Chan PH (2001). Reactive oxygen radicals in signaling and SAINT-II. Cerebrovasc Dis 25(3): 268–278.
damage in the ischemic brain. J Cereb Blood Flow Metab Erenciska M, Silver IA (1989). ATP and brain function.
21: 2–14. J Cereb Blood Flow Metab 9: 2–19.
Chen J, Nagayama T, Jin K (1998). Induction of caspase-3- Fink K, Zhu J, Namura S, et al. (1998). Prolonged therapeu-
like protease may mediate delayed neuronal death in the tic window for ischemic brain damage caused by delayed
hippocampus after transient cerebral ischemia. J Neurosci caspase activation. J Cereb Blood Flow Metab 18:
18: 4914–4928. 1071–1076.
Clark AW, Krekoski CA, Bou SS, et al. (1997). Increased Fiskum G, Murphy AN, Beal MF (1999). Mitochondria in
gelatinase A (MMP-2) and gelatinase B (MMP-9) activ- neurodegeneration: acute ischemia and chronic neurode-
ities in human brain after focal ischemia. Neurosci Lett generative diseases. J Cereb Blood Flow Metab 19:
238: 53–56. 351–369.
Connolly ES Jr, Winfree CJ, Springer TA, et al. (1996). Cere- Frederickson CJ (1989). Neurobiology of zinc and zinc-
bral protection in homozygous null ICAM-1 mice after mid- containing neurons. Int Rev Neurobiol 31: 145–238.
dle cerebral artery occlusion. Role of neutrophil adhesion in Friedlander RM, Gagliardini V, Hara H, et al. (1997).
the pathogenesis of stroke. J Clin Invest 97: 209–216. Expression of a dominant negative mutant of interleukin-
Dalkara T, Moskowitz MA (1994). The complex role of 1 beta converting enzyme in transgenic mice prevents neu-
nitric oxide in the pathophysiology of focal cerebral ische- ronal death induced by trophic factor withdrawal and
mia. Brain Pathol 4: 49–57. ischemic brain injury. J Exp Med 185: 933–940.
Dawson VL, Dawson T (2004). Deadly conversations: Frijns CJM, Kappelle LJ (2002). Inflammatory cell adhesion
nuclear-mitochondrial cross-talk. J Bioenerg Biomembr molecules in ischemic cerebrovascular disease. Stroke 33:
36: 287–294. 2115–2122.
 THE ISCHEMIC CASCADE AND MEDIATORS OF ISCHEMIC INJURY 39
Fujimura M, Morita-Fujimura Y, Murakami K (1998). Cyto- Iijima T, Mies G, Hossmann KA (1992). Repeated negative
solic redistribution of cytochrome c after transient focal DC deflections in rat cortex following middle cerebral
cerebral ischemia in rats. J Cereb Blood Flow Metab 18: artery occlusion are abolished by MK-801: effect on
1239–1247. volume of ischemic injury. J Cereb Blood Flow Metab
Fujimura M, Morita-Fujimura Y, Kawase M, et al. (1999). 12: 727–733.
Manganese superoxide dismutase mediates the early Kariko K, Weissman D, Welsh FA (2004). Inhibition of
release of mitochondrial cytochrome C and subsequent toll-like receptor and cytokine signaling—a unifying
DNA fragmentation after permanent focal cerebral ische- theme in ischemic tolerance. J Cereb Blood Flow Metab
mia in mice. J Neurosci 19: 3414–3422. 24: 1288–1304.
Furukawa K, Fu W, Li Y (1997). The actin-severing protein Katchanov J, Harms C, Gertz C, et al. (2001). Mild cerebral
gelsolin modulates calcium channel and NMDA receptor ischemia induces loss of cyclin-dependent kinase inhibi-
activities and vulnerability to excitotoxicity in hippocam- tors and activation of cell cycle machinery before delayed
pal neurons. J Neurosci 17: 8178–8186. neuronal cell death. J Neurosci 21: 5045–5053.
Gasche Y, Fujimura M, Morita-Fujimura Y, et al. (1999). Kerr JF, Wyllie AH, Currie AR (1972). Apoptosis: a basic
Early appearance of activated MMP-9 after focal cerebral biological phenomenon with wide-ranging implication in
ischemia in mice. J Cereb Blood Flow Metab 19: tissue kinetics. Br J Cancer 26: 239.
1020–1028. Kinouchi H, et al. (1991). Attenuation of focal cerebral
Ginsberg MD, Belayev L, Zhao W, et al. (1999). The acute ischemic injury in transgenic mice overexpressing CuZn
ischemic penumbra: topography, life span, and therapeutic superoxide dismutase. Proc Natl Acad Sci U S A 88:
response. Acta Neurochir Suppl (Wien) 73: 45–50. 11158–11162.
Goldberg MP, Ransom BR (2003). New light on white mat- Kirino T (2002). Ischemic tolerance. J Cereb Blood Flow
ter. Stroke 34: 330–332. Metab 22: 1283–1296.
Green DR, Reed JC (1998). Mitochondria and apoptosis. Kondo T, et al. (1997). Reduction of CuZn-superoxide dis-
Science 281: 1309–1312. mutase activity exacerbates neuronal cell injury and
Greene LA, Biswas SC, Liu DX (2004). Cell cycle mole- edema formation after transient focal cerebral ischemia.
cules and vertebrate neuron death: E2F at the hub. Cell J Neurosci 17: 4180–4189.
Death Differ 11: 49–60. Kroemer G, Reed JC (2000). Mitochondrial control of cell
Hafezi-Moghadam A, Simoncini T, Yang Z, et al. (2002). death. Nat Med 6: 513–519.
Acute cardiovascular protective effects of corticosteroids Lapchak PA, Chapman DF, Zivin JA (2000). Metalloprotei-
are mediated by non-transcriptional activation of endothe- nase inhibition reduces thrombolytic (tissue plasminogen
lial nitric oxide synthase. Nat Med 8: 473–479. activator)-induced hemorrhage after thromboembolic
Hansen AJ (1985). Effect of anoxia on ion disturbances in stroke. Stroke 31: 3034–3040.
the brain. Physiol Rev 65: 101–148. Letterio JJ (2000). Murine models define the role of TGF-
Hara H, Friedlander RM, Gagliardini V, et al. (1997). Inhibi- beta as a master regulator of immune cell function.
tion of interleukin 1b converting enzyme family proteases Cytokine Growth Factor Rev 11: 81–87.
reduces ischemic and excitotoxic neuronal damage. Proc Li Y, Chopp M, Jiang N, et al. (1995a). In situ detection of
Natl Acad Sci U S A 94: 2007–2012. DNA fragmentation after focal cerebral ischemia in mice.
Heo JH, Lucero J, Abumiya T, et al. (1999). Matrix metal- Mol Brain Res 28: 164–168.
loproteinases increase very early during experimental focal Li Y, Chopp M, Jiang N, et al. (1995b). Induction of DNA
cerebral ischemia. J Cereb Blood Flow Metab 19: 624–633. fragmentation after 10 to 120 min of focal cerebral ische-
Herrup K, Arendt T (2002). Re-expression of cell cycle pro- mia in rats. Stroke 26: 1252–1258.
teins induces neuronal cell death during Alzheimer’s dis- Lindsberg PJ, Carpen O, Paetau A, et al. (1996). Endothelial
ease. J Alzheimers Dis 4: 243–247. ICAM-1 expression associated with inflammatory cell
Hoehn-Berlage M (1995). Diffusion-weighted NMD ima- response in human ischemic stroke. Circulation 94: 939–945.
ging: application to experimental focal cerebral ischemia. Linnik MD, Zobrist RH, Hatfield MD (1993). Evidence sup-
NMR Biomed 8: 345–358. porting a role for programmed cell death in focal cerebral
Huang Z, Huang PL, Panahian N, et al. (1994). Effects of ischemia in rats. Stroke 24: 2002–2009.
cerebral ischemia in mice deficient in neuronal nitric Lipton P (1999). Ischemic cell death in neurons. Physiol Rev
oxide synthase. Science 265: 1883–1885. 79: 1431–1568.
Huang Z, Huang PL, Ma J, et al. (1996). Enlarged infarcts in Lo EH, Dalkara T, Moskowitz MA (2003). Mechanisms,
endothelial nitric oxide synthase knockout mice are atte- challenges, and opportunities in stroke. Nat Rev Neurosci
nuated by nitro-L-arginine. J Cereb Blood Flow Metab 4: 399–415.
16: 981–987. Lo EH, Moskowitz MA, Jacobs TP (2005). Exciting, radical,
Iadecola C (1997). Bright and dark sides of nitric oxide in suicidal. How brain cells die after stroke? Stroke 36:
ischemic brain injury. Trends Neurosci 20: 132–139. 189–192.
Iadecola C, Zhan F, Casey R, et al. (1997). Delayed reduc- McDonald JW, Althomsons SP, Hyra KL, et al. (1998). Oli-
tion of ischemic brain injury and neurological deficits in godendrocytes from forebrain are highly vulnerable to
mice lacking the inducible nitric oxide synthase gene. AMPA/kainate receptor mediated excitotoxicity. Nat
J Neurosci 17: 9157–9164. Med 4: 291–297.
40 M. ENDRES ET AL.
MacManus JP, Buchan AM, Hill IE, et al. (1993). Global Qiu J, Whalen MJ, Lowenstein P, et al. (2002). Upregulation
ischemia can cause DNA fragmentation indicative of of the Fas receptor death-inducing signaling complex after
apoptosis in rat brain. Neurosci Lett 164: 89–92. traumatic brain injury in mice and humans. J Neurosci 22:
Majno G, Joris I (1995). Apoptosis, oncosis, and necrosis. 3504–3511.
An overview of cell death. Am J Pathol 146: 3–15. Rashidian J, Iyirharo G, Aleyasin H, et al. (2005). Multiple
Martin RL, Lloyd HG, Cowan AI (1994). The early events of cyclin-dependent kinases signals are critical mediators of
oxygen and glucose deprivation: setting the scene for neu- ischemic/hypoxic neuronal death in vitro and in vivo. Proc
ronal death? Trends Neurosci 17: 251–257. Natl Acad Sci U S A 102: 14080–14085.
Martin-Villalba A, Herr I, Jeremias I, et al. (1999). CD95 Ratan RR, Siddiq A, Aminova L, et al. (2004). Translation of
ligand (Fas-L/APO-1L) and tumor necrosis factor-related ischemic preconditioning to the patient: prolyl hydroxy-
apoptosis-inducing ligand mediate ischemia-induced lase inhibition and hypoxia inducible factor-1 as novel tar-
apoptosis in neurons. J Neurosci 19: 3809–3817. gets for stroke therapy. Stroke 35: 2687–2689.
Matute C, Alberdi E, Ibarretxe G, et al. (2002). Excitotoxicity Ridker PM, Hennekens CH, Buring JE, et al. (2000). C-reac-
in glial cells. Eur J Pharmacol 447: 239–246. tive protein and other markers of inflammation in the pre-
Meisel C, Schwab JM, Prass K, et al. (2005). Central nervous diction of cardiovascular disease in women. N Engl J Med
system injury-induced immune deficiency syndrome. Nat 342: 836–843.
Rev Neurosci 6: 775–786. Rosenberg GA, Navratil M (1997). Metalloproteinase inhibi-
Mies G, Iijima T, Hossmann KA (1993). Correlation between tion blocks edema in intracerebral hemorrhage in the rat.
peri-infarct DC shift and ischaemic neuronal damage in rat. Neurology 48: 921–927.
Neuroreport 4: 709–711. Ruscher K, Freyer D, Karsch M, et al. (2002). Erythropoietin is
Moncayo J, de Freitas GR, Bogousslavsky J, et al. (2000). a paracrine mediator of ischemic tolerance in the brain: evi-
Do transient ischemic attacks have a neuroprotective dence from an in vitro model. J Neurosci 22: 10291–10301.
effect? Neurology 54: 2089–2094. Salvesen GS (2001). A lysosomal protease enters the death
Montaner J, Alvarez-Sabı́n J, Molina C, et al. (2001). Matrix scene. J Clin Invest 107: 21–22.
metalloproteinase expression after human cardioembolic Schielke GP, Yang GY, Shivers BD, et al. (1998). Reduced
stroke: temporal profile and relation to neurological ischemic brain injury in interleukin-1 beta converting
impairment. Stroke 32: 1759–1766. enzyme-deficient mice. J Cereb Blood Flow Metab 18:
Namura S, Zhu J, Fink K, et al. (1998). Activation and clea- 180–185.
vage of caspase-3 in apoptosis induced by experimental Sheng H, Bart RD, Oury TD, et al. (1999). Mice overexpres-
cerebral ischemia. J Neurosci 18: 3659–3668. sing extracellular superoxide dismutase have increased
Nedergaard M, Hansen AJ (1993). Characterization of corti- resistance to focal cerebral ischemia. Neuroscience 88:
cal depolarizations evoked in focal cerebral ischemia. 185–191.
J Cereb Blood Flow Metab 13: 568–574. Siesjo BK (1978). Brain energy metabolism and catechola-
Nicotera P, Lipton SA (1999). Excitotoxins in neuronal minergic activity in hypoxia, hypercapnia and ischemia.
apoptosis and necrosis. J Cereb Blood Flow Metab 19: J Neuro Trans Suppl 14: 17–22.
583–591. Sorensen JC, Mattsson B, Andreasen A, et al. (1998). Rapid
Nicotera P, Leist M, Fava E, et al. (2000). Energy require- disappearance of zinc positive terminals in focal brain
ment for caspase activation and neuronal cell death. Brain ischemia. Brain Res 812: 265–269.
Pathol 10: 276–282. Stanimirovich D, Satoh K (2000). Inflammatory mediators of
Osuga H, Osuga S, Wang F, et al. (2000). Cyclin-dependent cerebral endothelium: a role in ischemic brain inflamma-
kinases as a therapeutic target for stroke. Proc Natl Acad tion. Brain Pathol 10: 113–126.
Sci U S A 97: 10254–10259. Stroke Progress Review Group (2002). Report of the Stroke
Paschen W (2000). Role of calcium in neuronal injury. Brain Progress Review Group. National Institute of Neurological
Res Bull 53: 409–413. Disorders and Stroke, Bethesda, MD, pp. 1–116.
Pirttilä TR, Kauppinen RA (1992). Recovery of intracellular Stys PK, Waxman SG, Ransom BR (1992). Effects of tem-
pH in cortical brain slices following anoxia studied by perature on evoked electrical activity and anoxic injury in
nuclear magnetic resonance spectroscopy: role of lactate CNS white matter. J Cereb Blood Flow Metab 12: 977–986.
removal, extracellular sodium and sodium/hydrogen Susin SA, Lorenzo HK, Zamzami N, et al. (1999). Molecular
exchange. Neuroscience 47: 155–164. characterization of mitochondrial apoptosis-inducing fac-
Prass K, Meisel C, Hoflich C, et al. (2003). Stroke-induced tor. Nature 397: 441–446.
immunodeficiency promotes spontaneous bacterial infec- Tanne D, Haim M, Boyko V, et al. (2002). Soluble intercellu-
tions and is mediated by sympathetic activation—reversal lar adhesion molecule-1 and risk of future ischemic stroke:
by poststroke T helper cell type 1-like immunostimulation. a nested case-control study from the Bezafibrate Infarction
J Exp Med 198: 725–736. Prevention (BIP) study cohort. Stroke 33: 2182–2186.
Priller J, Fluger A, Wehner T, et al. (2001). Targeting gene- Tarozzo G, Campanella M, Ghiani M, et al. (2002). Expres-
modified hematopoietic cells to the central nervous sys- sion of fractalkine and its receptor, CX3CR1, in response
tem: use of green fluorescent protein uncovers microglial to ischaemia-reperfusion brain injury in the rat. Eur J
engraftment. Nat Med 7: 1356–1361. Neurosci 15: 1663–1668.
 THE ISCHEMIC CASCADE AND MEDIATORS OF ISCHEMIC INJURY 41
Timsit S, Rivera S, Ouaghi P, et al. (1999). Increased cyclin Weiss JH, Hartley DM, Koh JY, et al. (1993). AMPA recep-
D1 in vulnerable neurons in the hippocampus after ischae- tor activation potentiates zinc neurotoxicity. Neuron 10:
mia and epilepsy: a modulator of in vivo programmed cell 43–49.
death? Eur J Neurosci 11: 263–278. Wender R, Brown AM, Fern R, et al. (2000). Astrocytic gly-
Tominaga T, Kure S, Narisawa K, et al. (1993). Endonu- cogen influences axon function and survival during glu-
clease activation following focal ischemic injury in the cose deprivation in central white matter. J Neurosci 20:
rat. Brain Res 608: 21–26. 6804–6810.
Trendelenburg G, Dirnagl U (2005). Neuroprotective role of Wolf T, Lindauer U, Reuter U, et al. (1997). Noninvasive
astrocytes in cerebral ischemia: focus on ischemic precon- near-infrared spectroscopy monitoring of regional cerebral
ditioning. Glia 50: 307–320. blood oxygenation changes during peri-infarct depolariza-
Trendelenburg G, Prass K, Priller J, et al. (2002). Serial ana- tions in focal cerebral ischemia in the rat. J Cereb Blood
lysis of gene expression identifies metallothionein-II as Flow Metab 17: 950–954.
major neuroprotective gene in mouse focal cerebral ische- Yrjanheikki J, Tikka T, Keinanan R, et al. (1999). A tetracy-
mia. J Neurosci 22: 5879–5888. cline derivative, minocycline, reduces inflammation and
Van Lookeren Campagne M, Gill R (1996). Ultrastructural protects against focal cerebral ischemia with a wide
morphological changes are not characteristic of apoptotic therapeutic window. Proc Natl Acad Sci U S A 96:
cell death following focal cerebral ischaemia in the rat. 13496–13500.
Neurosci Lett 213: 111–114. Yuan J, Yanker BY (2000). Apoptosis in the nervous system.
Wang F, Corbett D, Osuga H, et al. (2002). Inhibition of Nature 407: 802–809.
cyclin-dependent kinases improves CA1 neuronal survival Zhang K, Sejnowkoski TJ (2000). A universal scaling law
and behavioral performance after global ischemia in the between gray matter and white matter of cerebral cortex.
rat. J Cereb Blood Flow Metab 22: 171–182. Proc Natl Acad Sci U S A 97: 5621–5626.
Wang J, Liu S, Fu Y, et al. (2003). Cdk5 activation induces Zhang J, Dawson VL, Dawson TM (1994). Nitric oxide acti-
hippocampal CA1 cell death by directly phosphorylating vation of poly(ADP-ribose)synthetase in neurotoxicity.
NMDA receptor. Nat Neurosci 6: 1039–1047. Science 263: 687–689.
Wei MC, Zong WX, Cheng EH, et al. (2001). Proapoptotic Zipfel GJ, Lee JM, Choi DW (1999). Reducing calcium over-
BAX and BAK: a requisite gateway to mitochondrial dys- load in ischemic brain. N Engl J Med 341: 1543–1544.
function and death. Science 292: 727–730.
Weih M, Kallenberg K, Bergk A, et al. (1999). Attenuated
stroke severity after prodromal TIA: a role for ischemic
tolerance in the brain? Stroke 30: 1851–1854.