INeurological Disorders – Epidemiology, Clinical Overview,and Model Systems

Neuroprotection. Models, Mechanisms and Therapies. Edited by Mathias B�hrCopyright c 2004 WILEY-VCH Verlag GmbH & Co. KGaA, WeinheimISBN 3-527-30816-4

1Stroke

Andreas Meisel, Konstantin Prass, Tilo Wolf, Ulrich Dirnagl

Abstract

The current understanding of stroke pathophysiology is basically founded on ex-perimental models. Among these, in vitro models with primary cultures of cerebralcells permit stroke pathophysiology to be examined on a molecular level, while an-imal models (mainly mouse and rat) are used to evaluate medical intervention.There is a highly complex sequence of events leading to the eventual ischemic cer-ebral damage that follows a well-defined spatio-temporal pattern. Overwhelmingexcitotoxicity leads to early necrotic cell death in what is to become the core ofthe infarction, while the tissue damage in the surrounding zone called penumbrahappens on a longer time scale. The excitotoxic or inflammatory mechanisms aremilder, bearing the biochemical hallmarks of apoptosis. The brain cells, challengedby such a large-scale assault, activate endogenous protective programs. These havebeen studied by experimentally inducing ischemic tolerance (i. e. ischemic precon-ditioning). Importantly, cerebral ischemia not only affects the brain, but alsoimpacts other systems. For example, stroke induces dramatic immunodepressionthrough over-activation of the sympathetic nervous system. As a result, severebacterial infections such as pneumonia occur. The complex signaling cascadesnot only decide over cell survival in the brain and the neurological deficit, butalso over mortality after stroke from extracerebral complications. Their ability togovern not only the maturation of the eventual infarction but also the immunesystem make them a promising target for intervention and the development ofneuroprotective drugs.

1.1Introduction

In the US more than 600,000 people per year suffer from strokes, and amongacutely hospitalized neurological patients stroke patients make up the largestshare – about 50%. Currently, there are about 4 million stroke survivors [1] in

Neuroprotection. Models, Mechanisms and Therapies. Edited by Mathias B�hrCopyright c 2004 WILEY-VCH Verlag GmbH & Co. KGaA, WeinheimISBN 3-527-30816-4

the US. Mortality from stroke is an estimated 25%, which makes it the third majorcause of death in industrialized countries. Due to the high rate of severe perma-nent disability, stroke is a burden not only for the affected patients and their fa-milies, but also for national economies: In the US, it is estimated that the annualcosts caused by strokes ranges between 30 and 40 billon dollars [2, 3]. In the UK,the cost of one individual patient amounts to £ 30,000 over 5 years [4].The term stroke accommodates a variety of different conditions. About 85% of

all strokes are caused by cerebral ischemia due to vessel occlusion. Primary cere-bral bleeding is rare in comparison –15%. Of the ischemic strokes, 75% arecaused by emboli, of either arterial or cardial origin, while microvascular occlusion,i. e. hyalinosis or in-situ thrombosis is responsible for 20% of cases. Hemodynamicischemia, caused by stenoses of brain-supplying arteries, account for less than 5%of ischemic strokes [5, 6].The prospective risk of suffering an ischemic stroke is partially a function of

social and behavioral (nutrition, tobacco use, stress) factors and longstanding dis-orders like hypertension, diabetes, disorders of cholesterol and lipid metabolism,and obesity. Atherosclerosis is not only the main underlying condition in ischemicstroke, but also in coronary heart disease and peripheral vascular disease. More-over, the genetic background of affected individuals is coming under increasingscientific scrutiny, as might be expected for familial stroke conditions likeCADASIL (cerebral autosomal dominant arteriopathy with subcortical infarctsand leucencephalopathy), which is characterized by relapsing subcortical ischemia.The condition is caused by a mutation in the notch3-gene on chromosome 19 [7, 8].MELAS (mitochondrial encephalopathy, lactate acidosis, stroke-like episodes) is an-other rare familial stroke disorder, characterized by migraine, grand-mal seizures,and recurrent cortical infarctions. The condition is caused by a range of mutationsin the mitochondrial genome, which explains its maternal inheritance. The severityof the phenotype depends mainly on the mutation locus and on the proportion ofmutated mitochondrial genomes in the cerebral mosaic (for review see [9]). How-ever, family and twin studies suggest that genetic factors are likely to be involved incommon types of ischemic stroke, not just in rare and well defined familial strokedisorders [10–13].A locus on chromosome 5q12 described for the Icelandic population seemed to

increase stroke susceptibility significantly [14]. The responsible gene has beenidentified as phosphodiesterase 4D (PDE4D). The risk-conveying polymorphismsare located in the gene-regulatory part. This hints at a defect in the regulation ofthe encoded cAMP degrading protein [15]. Phosphodiesterase 4D belongs to agroup of proteins that are drug targets in the treatment of asthma, erectile dysfunc-tion and inflammation.Symptoms of focal cerebral ischemia depend on the individual affected vessel

with its typical supply territory. Table 1 summarizes the frequency of affection[16] and the typical clinical syndromes for the main brain arteries. Stroke mortalityis rated between 20 and 30%. Unfavorable prognostic factors are old age, initialcoma, papillary asymmetry, coronary heart disease and heart failure. Pyrexia andinfections (pneumonia in particular) deteriorate the prognosis particularly during

4 1 Stroke

the initial phase (see below). Of the surviving patients, one third improves within aweek, 40% remain unchanged in their disability, and 20% deteriorate further dur-ing the first week [17].

Table 1.1 Territorial stroke syndromes.

Anterior territoryMCA (z60%) hemiparesis mainly arm, hemihypaesthesia, hemianopia,

dysphasia or neglectACA (4%) hemiparesis mainly leg, urinary incontinence, apraxiaAchA (8%) hemiparesis, hemihypaesthesia, hemianopia

Posterior territoryVA/BA (10%) vertigo, diplopia, bilateral hypaesthesia and paresis, crossed

syndromes, amaurosis, ataxia, headache, comaCerebellum(PICA/ AICA)

(7%) headache, ataxia, vertigo, gaze palsy, facial weakness, deafness

PCA (9%) hemianopia, dyslexia, visual agnosia

A great deal of our knowledge about the pathophysiology of stroke comes fromexperimental research. The experimental animal models follow two main para-digms: One is the model of focal cerebral ischemia as a model of ischemic stroke.The other is the model of global cerebral ischemia as model of circulatory arrest.For obvious reasons, we will focus mainly at studies of focal cerebral ischemia.The bulk of experimental studies has been carried out with rats and mice, althoughsome primates have been used. In rodents, ischemia is mostly induced by intravas-cular occlusion of the MCA, using a monofilament just exceeding the critical vesseldiameter. Depending on the duration of occlusion, we distinguish permanent fromtransient ischemia models. The latter are also used as models for spontaneousreperfusion or for the state after successful lysis therapy with recombinant tissueplasminogen activator (rtPA) in humans, respectively.In models of cerebral ischemia perfusion to the brain is impeded by some form

of manipulation. Since the essence of ischemic stroke is vascular occlusion, it isnot surprising that in these models the extent of the eventual infarction can be pre-dicted quite accurately from the degree of reduction in regional cerebral blood(rCBF) flow. Thus, if the rCBF is reduced to less than 25% of normal the likelihoodof infarction in a given volume of brain tissue greater than 95%. In contrast, thelikelihood of infarction is less than 5% if rCBF does not fall below 50% of normal.These thresholds have been established in comparative human PET and MRI stud-ies [18] and correspond with experimental data from animal models [19]. Thus, theinitial reduced rCBF determines the extent of the anticipated infarction. This, how-ever, holds only with the provision that a spontaneous or therapeutic reperfusiondoes not take place.We now look at ischemic infarction as the result of a complex and prolonged pro-

cess of infarct maturation rather than a simple result of reduced regional perfu-

51.1 Introduction

sion. As brain tissue has a high demand for oxygen and glucose, a disruption ofperfusion leads to substrate depletion within few minutes, while toxic metabolitesaccumulate. The ensuing cellular energy deficit leads to a collapse of the estab-lished ion gradients and the membrane potential. Neurons and glial cells depolar-ize. Depending on the extent and the duration of this energy deficiency, the cellswill suffer not only a functional but also a structural breakdown. The highly com-plex sequence of events within the ischemic area follows a fairly well definedstereotypic spatio-temporal pattern, which we will discuss below in more detail.The concept of the ischemic penumbra is crucial to the understanding of these me-chanisms. The cascade of ischemic damage begins with excitotoxicity, the forma-tion of reactive oxygen free radicals, the increasing tissue acidosis and the occur-rence of periinfarct-depolarizations. It is followed by the stages of inflammationand programmed cell death (apoptosis). This is associated with DNA damagethat, in turn, induces DNA repair programs. Although the process is not yetfully understood, we know that chromatin re-modeling, i. e. epigenetic mechan-isms, and the activation of transcription factors induce complex gene programs.These changes initiate the expression of destructive proteins involved in inflamma-tion and apoptosis, as well as a host of protective genes that help repair the ische-mic damage. It is the activation of these protective genes that builds up ischemictolerance. Among the many newly discovered protective mechanisms, endogenousand exogenous cell replacement have met with a lot of interest. Apart from theseautochthonous mechanisms of the brain tissue, there are other mechanisms on asystemic scale, which bear important clinical significance. For instance, the phe-nomenon of stroke-induced immunodepression can help to understand why strokepatients are at such high risk of contracting serious bacterial infections. Neuropro-tective treatment must be based on an understanding of these mechanisms.

1.2The Penumbra Concept

In the ischemic brain, we commonly distinguish two tissue volumes – the core ofthe infarction and the surrounding zone, known as ischemic penumbra [20] – theunderperfused and metabolically compromised margin surrounding the irrevoc-ably damaged core. Core and penumbra are characterized by two different kindsof cell death: necrosis and apoptosis (which is also called programmed cell deathor delayed neuronal cell death). The severe perfusion deficit in the core causes abreakdown of metabolic processes, cellular energy supply and ion homeostasis,which causes the cells to lose their integrity within minutes. Thus, acute necrosisof cell and tissue prevails in the core. In the penumbra, some residual perfusion ismaintained by collateral vessels, which may be unable to maintain the full func-tional metabolism, but prevents immediate structural disintegration. However,over time, the alteration of cellular homeostasis causes more and more cells todie, and the volume of the infarction increases. The penumbra has thus to beconsidered as tissue at risk during the maturation of the infarct. In this region,

6 1 Stroke

apoptosis and inflammatory signaling cascades play an important role. It mayinitially constitute 50% of the volume that will end up as infarction. Themechanisms that lead to delayed cell death within the penumbra are the subjectof intense research, as they provide targets for a specific neuroprotective therapyin brain regions challenged by ischemia, but which are still viable (for a reviewsee [19, 21]).

1.3Excitotoxicity

The depolarization of neurons and glia due to local energy deficit causes the acti-vation of voltage-gated calcium channels and the release of excitatory amino acidsinto the extracellular space. Glutamate in particular, which, under normal cellularenergy conditions, would be immediately taken up pre-synaptically or through as-trocytes, now remains in the extracellular space where it accumulates dramatically.Through the activation of glutamate receptors NMDA and AMPA, the intracellularCa2þ level rises. Furthermore, metabotropic glutamate receptors are activatedthrough the induction of phospholipase C (PLC) and inositol triphosphate (IP3),and calcium is mobilized from intracellular stores.Furthermore, the over-activation of AMPA receptors causes a rise of sodium and

chloride concentrations. Altogether, the result is a massive disturbance of ionhomeostasis, accompanied by passive water influx and cell edema. Ultimately,these massive changes in cell volume account for osmotic cell lysis. This lytictype of cell death, also referred to as necrosis, is primarily observed in the coreof the infarction. Cells that escape this most dramatic form of disintegration, asthey can not found in the core but in the penumbra where excitotoxicity may bean initiator of molecular events that lead to apoptosis and inflammation (for anoverview see [21, 22]).

1.4Oxygen Free Radicals

As a consequence of ischemia and particularly of reperfusion, reactive oxygen freeradicals such as superoxide, hydrogen peroxide and hydroxyl radicals are gener-ated. Nitric oxide is generated via the activation of the calcium-calmodulin-depen-dent nitric oxide synthase (NOS); it reacts with superoxide radicals and forms thusthe highly reactive peroxy-nitrite radical. Further sources of oxygen free radicals inthe damaged brain tissue are the breakdown products of the adenosine phosphates,which contribute to radical production via xanthine oxidase and the iron-catalyzedHaber-Weiss reaction. The many different radical species that are thus formed canreact with virtually any cellular components (carbohydrates, amino acids, DNA,phospholipids) and damage them. The peroxidation of membrane lipids releasesfurther radicals – and further glutamate. Oxygen free radicals gain even more

71.4 Oxygen Free Radicals

significance when new oxygen reaches the damaged tissue by virtue of reperfu-sion, or in the penumbra where oxygen supply has not ceased entirely (overviewin [21, 22]).Hypoxia itself as well as the elevated intracellular concentration of calcium ions

and free radicals disrupt the function of neuronal mitochondria. Consequently, aso-called mitochondrial permeability transition pore (MPT) in the mitochondrialmembrane may form. Besides impeding ATP production through loss of mito-chondrial potential, the MPT leads to mitochondrial swelling, a burst of free oxy-gen radicals, and the release of pro-apoptotic molecules. Thus a vicious cycle offurther disintegration is fuelled (see below, for review [21, 23]). This vicious cycleis counterbalanced in part by anti-oxidative enzymes like the manganese-superox-ide dismutase (Mn-SOD) and the cytosolic forms of the copper-zinc superoxide dis-mutase (CuZn-SOD). These may prevent the breakdown of the mitochondrialmembrane and, thus, the release of cytochrome C, which would be the triggerof apoptosis (see below and for review [24]).

1.5Tissue Acidosis

In the context of stroke pathophysiology, the proton balance is intimately linkedwith the glucose metabolism. With reduced oxygen availability, anaerobic glycolysisas only remaining source of ATP production leads to tissue acidosis. It has longbeen assumed that this acidosis was one of the main noxious mechanisms in is-chemic stroke. This so-called ‘lactate-acidosis-hypothesis’ is often quoted as expla-nation for the “glucose paradox” of cerebral ischemia. This paradox refers to theobservation that excessive supply of glucose, the most important source of energyof the brain, during focal cerebral ischemia does not reduce tissue damage as oneshould think but, instead, augments it [25]. However, by which mechanism thishappens and, in fact, whether levels of acidosis reached in brain ischemia can gen-erate brain tissue damage at all is still far from being clear. Possibly, the pH depen-dent transition of Fe(III) to Fe(II) and the release of iron from molecular storageslead to a facilitation of the Haber-Weiss reaction that forms toxic free oxygen radicalspecies (see above). Besides the production of different species of oxygen free radi-cals, acidosis also interferes with intracellular protein synthesis. However, the lac-tate-acidosis-hypothesis is not unchallenged. Particularly the fact that acidosisblocks the NMDA-receptor and thus has an anti-excitotoxic effect indicates thecomplexity of the role that acidosis plays in cerebral ischemia (for a review [21]).A similarly hotly debated topic are the findings on hyperglycemia during stroke.

Experimental data from animal models show a detrimental effect of hyperglycemiaduring focal cerebral ischemia [26–28]. Clinical data also suggest that hyperglyce-mia during the acute phase of stroke worsens prognosis [29–31]. Persisting hyper-glycemia beyond the acute phase is also an independent prognostic factor for largerinfarct volume and poorer functional outcome in stroke patients [32]. It is a matterof debate whether these observations are due to a causal relationship or if hypergly-

8 1 Stroke

cemia is a mere epiphenomenon, possibly due to a stress reaction (i. e. an effect ofconcomitant catecholamine and glucocorticoid release proportional to initial tissuedamage) [34, 34]. The controversial role of an associated lactate-acidosis has alreadybeen discussed. An alternative concept is the glucocorticoid hypothesis [35]. It sug-gests that hyperglycemia enhances the release of glucocorticoids, which are also re-leased during ischemia in the sense of a stress reaction. Glucocorticoids have beenshown to have a direct cytotoxic effect [36] and the blockade of glucocorticoid recep-tors does reduce the damaging effect of hyperglycemia [35]. On the other handhyperglycemia and acidosis have a protective effect in vitro [37, 38], and some ex-perimental studies demonstrate that hyperglycemia can also have a beneficial effectunder special circumstances [39, 40].However, a recent study on humans seems to shed light on this controversy. Per-

fusion and diffusion MR imaging and MR spectroscopy shows that stroke patientswith hyperglycemia have a higher lactate accumulation in the penumbra. Whilethe perfusion-diffusion-mismatch (as a measure of the size relation of penumbraand core) was similar in the normoglycemic and hyperglycemic groups, the even-tual infarct size and disability scores were significantly higher in the hyperglycemicgroup [41]. While it could still be argued that initially hyperglycemic patients werealso likely to be longstanding or latent diabetics with an overall worse microvascu-lar state, we take this study- which links in with much of the experimental animaldata – as strong indicator for the vulnerability of the penumbra to hyperglycemia.Indirectly, it also corroborates the lactacidosis theory, which, however, still lacksdirect proof.

1.6Peri-infarct Depolarizations

The anoxic as well as excitotoxic depolarization of neurons and glia increases theconcentration of glutamate and potassium in the ischemic core. These diffuseinto the penumbral zone where they induce the electrochemical depolarizationof more neuronal and glial cells that are “just managing” under the circumstancesof oligemia. Depolarization in this context is not to be confused with short-livedaction potentials, but means sustained depolarization of the entire cell membrane,a breakdown of the membrane potential that has to be actively rectified. Depolar-ization also causes a further significant increase in extracellular glutamate and po-tassium. Thus, waves of sustained electrochemical depolarization spread from thecore of infarction into the penumbra, bearing similarity with Leao’s spreading de-pression. These waves are called peri-infarct depolarizations [42–44]. Peri-infarctdepolarizations occur with a frequency of 1–4 per hour and have been detectedby functional MRI [45] and near infrared spectroscopy [46]. Since active rectifica-tion of the membrane potential requires energy, these peri-infarct depolarizationscontribute further to the metabolic compromise and each wave turns a larger areaof the penumbra towards irrevocable infarction [43, 44, 47, 48]. In addition, peri-infarct depolarizations seem to compromise the impaired microcirculation within

91.6 Peri-infarct Depolarizations

the penumbra even further [49]. Peri-infarct depolarizations have been detected inpatients with head trauma [50].

1.7Inflammation

The early stage of inflammation, which starts a few hours after the onset of ische-mia, is characterized by the expression of adhesion molecules in the vascularendothelium as well as on circulating leukocytes. Thus, leukocytes adhere to theendothelium and transmigrate from the blood into the brain parenchyma, whichis of decisive importance for stroke induced brain inflammation (for review see[51–53]).Newly expressed endothelial adhesion molecules like ICAM-1 and VCAM-1 facil-

itate the interaction between endothelium and leucocytes as first step of trans-migration of white blood cells from blood into the tissue. They interact with theb2-integrins CD11b/CD18 (Mac-1) and CD11a/CD18 (LFA-1) that are, in turn,expressed on leucocytes [54–56]. Neutrophil granulocytes accumulate in the capil-laries of the penumbra and interfere with the already impaired microcirculation[57]. Blocking this interaction with CD18, CD11, or ICAM-1 antibodies reducesnot only the number of leucocytes but also the size of infarction [58–60]. ICAM-1knockouts have also smaller infarcts than wild type mice [61]. Next to the b2-in-tegrins, other integrins that are more specific for lymphocytes like VLA-4(a4 b1)come into action, too. Consequently, blocking the a4 -component reduces also thevolume of infarction [62].A major part in inflammation is ascribed to the population of microglial cells.

These cells are the primary immunoeffectors of the CNS and make up about20% of the entire glial mass, about the same proportion as neurons. Activationof microglia is a striking feature in the penumbra. The cells change their shapefrom highly ramified cells to a more plump and amoeba like form, and they pro-liferate. They also express a number of characteristic surface molecules (for in-stance MHC I and MHC II) and are capable of antigen presentation [63–66]. Simi-lar to leukocytes, activated microglia cells are able to produce a variety of pro-in-flammatory cytokines (but also cytotoxic metabolites (especially oxygen free radicalssuch as peroxy-nitrite and superoxide) and enzymes (for instance cathepsins). Inaddition, microglia is phagocytically active. The inhibition of microglial activationturns out to be protective in experimental stroke models [67]. On the other hand,activated microglia have also been shown to release anti-inflammatory cytokines(such as TGF-b1) and growth factors [68]. Because of the Janus-faced nature of mi-croglial products (destructive, e. g. free radicals, vs. protective, e. g. growth factors)the overall role of microglia in cerebral ischemia is not clear at present. It is verylikely that microglia play different roles at different times, with protective or regen-erative activities occurring days or even weeks after the onset of ischemia.Activated leukocytes (granulocytes, monocytes/macrophages, lymphocytes) as

well as neurons and glial cells (astrocytes, microglia) produce cytokines and che-

10 1 Stroke

mokines [69, 70]. In particular, pro-inflammatory cytokines like TNFa, IL-1 andIL-6 play a major role as mediators of the inflammatory response. They are respon-sible for the transition from the early, excitotoxic phase to the inflammation phase(for a review: [51–53, 71]. Their regulation depends on transcription factors likeNFkB, which, in turn, are activated by oxygen free radicals [72, 73]. Cytokine-mRNA and also the protein of TNFa and IL-1 can be detected only few hoursafter induction of an experimental focal ischemia [71]. Expression of IL-6 followsonly after about 24 hours [74]. These cytokines are mainly released by microglialcells and macrophages [69, 70]. In stroke patients, it could be shown that the in-trathecal concentration of IL-1 and IL-6 correlated with the infarct size [75]. Cyto-kine receptor antagonists reduce the infarct volume in animal models. For exam-ple, blockade of TNFa by TNF-binding proteins reduces brain injury after focal cer-ebral ischemia [76]. The role of TNFa is not entirely clear, however, as mice lackingthe TNFa-receptor (TNFR2) have larger infarctions [77]. In part, these contrastingresults may reflect a difference in signal transduction cascades activated by the twoTNFa-receptors TNFR1 and TNFR2 [78]. It is also arguable that an increased IL-6induction may be responsible, since IL-6 knockouts do not have smaller infarctionsthan their litter mates [79]. Further to the induction of adhesion molecules, theabove-mentioned cytokines also render the blood brain barrier more permeableand induce pro-thrombotic functions of the endothelium (for a review [80]).Not only can the pro-inflammatory cytokines be observed, but anti-inflammatory

cytokines such as TGF-1b or IL-10 also induced. By down-regulating the inflamma-tion, these cytokines have a protective effect in the context of cerebral ischemia[81–84]. The neuroprotective action of TGF-b1 is well known (for review see[85]). Interestingly, this cytokine seems to play an important part in immunologicaltolerance. As in ischemic tolerance (see below), it is possible to induce a tolerantstate towards ischemia. This protection is induced by immunogenetic proteins(for instance myelin basic protein, MBP) and mediated by a certain lymphocyte po-pulation expressing TGF-b1 [86–88]. In contrast to ischemic tolerance, immunolo-gical tolerance is longer lasting. Protection lasts for months, whereas in ischemictolerance it lasts no longer than 7 days.Apart from cytokines, there are two other proteins that dominate the inflam-

matory phase: the inducible nitric oxide synthase (iNOS) and cyclo-oxygenase 2(COX-2). Focal cerebral ischemia models have shown that iNOS is induced onthe mRNA as well as on the protein level [89]. Protein expression reaches itsheight after 24 hrs. By inhibiting the expression of iNOS the size of experimentalinfarcts can be reduced by about 30%, even if treatment is undertaken only 24hours after the onset of ischemia [90, 91]. Nitric oxide (NO), which is producedin much greater amounts by iNOS than by the constitutional NO synthases, actsvia the formation of peroxynitrite highly cytotoxic [92]. NO triggers apoptosis notonly through peroxynitrite but also via the induction of p53, which probably causesdirect DNA damage [93]. COX-2 is mainly generated in the penumbra [94]. The en-zyme has a destructive effect on the penumbral tissue, mainly through producingoxygen free radicals and toxic prostanoids (for review see [51, 53]). This seems alikely hypothesis, as both genetic and pharmacological inhibition of COX-2 have

111.7 Inflammation

a protective effect [95, 96]. COX-2 as well as iNOS are promising targets for thetreatment of stroke patients because blocking them is still effective even 6 to24 hours after ischemia [91, 96].Not only the activated residual microglia, but also monocytes from the circulat-

ing blood migrate into the affected brain tissue. This transmigration is mediated bya chemokine, the monocyte-attractant protein 1 (MCP-1), which is induced 1 to2 days after ischemia and causes immigration of monocytes [97]. Moreover,about a third of these transmigrated cells differentiate within 14 days post strokeinto microglial cells – virtually indistinguishable from autochthonous microglia[98]. A trans-differentiation into astrocytes has also been observed [99, 100], butour group has not been able to reproduce this finding [101]. Irrespective of that,the important role of astrocytes during inflammation has been recognized; produc-ing both pro-inflammatory as well as neuroprotective factors such as erythropoietin(EPO), TGF-b1, or metallothionein 2 [85, 102, 103].In general, much further study needs to be conducted before the complex role of

inflammation in cerebral ischemia is sufficiently understood to be able to specifi-cally block the damaging proteins and foster protective elements during the in-flammation process in stroke.

1.8Damage to the Blood–Brain-Barrier

The function of the blood-brain-barrier (BBB) depends on the integrity of its cellu-lar matrix which consists of endothelial cells, basal lamina, and astrocytes. Cerebralischemia causes damage to this matrix and its intercellular signal exchange. Cen-tral to these damaging processes are proteases like cathepsins, plasminogen activa-tors (PA), and the matrix metalloproteinases (MMP). The action of the MMPs, agroup of more than 20 zinc-endopeptidases, has been extensively studied. Proteo-lytic destruction of the basal membrane by the MMPs permits the immigration ofleukocytes and promotes vasogenic edema. MMP-2 and MMP-9 are expressedwithin 1 to 3 hours after cerebral ischemia; the level of their expression correlateswith the level of BBB-breakdown, the risk of hemorrhagic transformation and theextent of eventual neuronal damage [104–106]. Pharmacological inhibition of theMMPs prevents this breakdown and reduces stroke volumes. Mice with target dis-ruption of MMP9 are partly protected from the effects of ischemia [107, 108].Of the plasminogen activators, tPA is best known. In the brain, tPA is expressed

mainly in neurons, astrocytes and microglia. Studies of mainly primary cortex neu-rons have shown that tPA acts also as neurotoxin [109], and experiments on tPA-knockouts have produced conflicting data [110, 111]. Since recombinant tPA isthe sole licensed drug for thrombolytic therapy, these findings reflect also on cur-rent therapeutic practice. tPA has furthermore been shown to induce MMP-9 and,thus, – at least indirectly – damaging the BBB [112]. The same holds probably forthe recombinant tPA. This may explain why such a high number of secondaryhemorrhages have been observed after thrombolytic stroke therapy [106, 113,

12 1 Stroke

114]. Animal experiments showed that delayed rtPA treatment – mimicking real-life delays in treatment – lead to significantly greater damage to the BBB and amajor increase in mortality The pharmacological inhibition of MMPs can reversethis effect of rtPA [115]. In addition, tPA enhances NMDA-mediated calcium sig-naling and subsequent excitotoxic neuronal injury [116]. It is a matter of contro-versy, whether this detrimental effect is due to a cleavage of the NR1 subunit ofthe NMDA receptor by tPA [117].The example of rtPA demonstrates once again that in order to refine treatment of

ischemia, a thorough understanding of the complex mechanisms is indispensable.The development of effective and low-risk thrombolytic treatment is, of course, stillone of the main objectives of pharmacological stroke research [113]. One promis-ing approach is the development of novel plasminogen activators, such as recom-binant desmodus rotundus salivary plasminogen activator (DSPA-a1; Desmoteplase).Experimental evidence indicates that Desmoteplase has pharmacological and toxi-cological properties superior to rtPA [118].

1.9Programmed Cell Death and Apoptosis

Apoptosis follows a similar time schedule as inflammation. It sets in after a fewhours and continues over days. It is the commonest form of cell damage in thepenumbra and is very distinct from the rapid neuronal and glial death in the ad-jacent ischemic core, where pan-necrosis is complete within few hours. In necro-sis, it is the cellular edema that leads to osmolysis (see above). The released cellcontents are a potent inflammatory stimulus as discussed above.Caspases, a group of proteases, has a pivotal role in apoptosis. They facilitate the

“orderly” destruction of the cell, which is characterized by a condensation of bothnucleus and the cell as a whole. The cell contents are fragmented into so-calledapoptotic bodies. Endonucleases fragment the cell’s DNA into typical strands of180 bp length. Other morphologic criteria of apoptosis are membrane blebbingand shrunk cytoplasm with morphologically intact organelles. While the terms“delayed neuronal death” (DND) and “programmed cell death” (PCD) are oftenused synonymously with “apoptosis”, the two former may well occur without thetypical features of the latter. Although most experimental studies show some bio-chemical evidence for apoptosis to have taken place, histological features are oftenmissing to substantiate the claim. This has led some researchers to question apop-tosis as an actual feature of ischemia. A number of reviews have dealt extensivelywith the significance of apoptosis in the pathophysiology of ischemia (for example[21, 22, 119–122]). This is why we shall focus here on some of its mechanisticaspects in this context.Essentially, apoptosis can be triggered in two ways, by intrinsic (i. e. basically

mitochondrial) or extrinsic activation. Binding to the Fas-receptor, for exampleby TNF-a, triggers the extrinsic pathway. Intrinsic triggers of apoptosis in the con-text of ischemia are elevated concentrations of intracellular calcium, reactive oxy-

131.9 Programmed Cell Death and Apoptosis

gen species, and glutamate as well as increased DNA damage. Through damage tothe mitochondrial membranes, both pathways directly or indirectly lead to theactivation of caspases.The caspases are a hierarchic group of at least 14 cysteine-dependent and aspar-

tate-specific proteases. Their inactive precursor proteins exist in every cell and canbe activated through cleavage. Apoptosis is essentially an active and in itself energydependent cellular deconstruction process. In cerebral ischemia, the caspases 1, 3,8, and 9 are involved, of which the caspases 8 and 9 are the initiators of a signalcascade that activates the so-called execution caspase 3. Caspase 1 is predominantlyengaged in cytokine activation (for review see [119, 121]). Caspase 3 is of predomi-nant importance not only in the context of cerebral ischemia [122], and its geneticor pharmacological inhibition is not only in this context neuroprotective [124, 125].The substrates degraded by caspase 3 are DNA repair enzymes like poly-(ADP-ribose)-polymerase (PARP) and the DNA dependent protein kinase (DNA-PK). Thelatter is important for the repair of double strand breaks. PARP recognizes singlestrand disruptions and mediates self-repair of the DNA through swift auto-modifi-cation with synthesis of highly negatively charged, long and branched ADP-ribosemolecules. This process utilizes ATP and thus taps the cellular energy pool.Although not universally accepted, this ATP consumption by PARP is widely be-lieved to be in itself cytotoxic. PARP is also capable of inducing caspase-indepen-dent cell suicide via the so-called apoptosis-inducing factor (AIF; see below;[126]). This helps to explain also the partial failure of caspase inhibitors in somemodels of cerebral ischemia. The genetic or pharmacological blockade of PARPhas, however, been demonstrated to be neuroprotective in experimental stroke[127–129]. Of significance is the cleavage by caspase 3 of the inhibitor of cas-pase-activated DNase (ICAD). ICAD is thus activated and causes the characteristicfeature of “laddering” that has been described in all stroke models (for review see[130]). The term laddering draws on the regular appearance of those DNA-frag-ments as ladder-like bands on a gel electrophoretic separation. The length of theresulting DNA fragments reflects the complex structure of eukaryotic genetic infor-mation in the chromatin. Chromatin is a structure, in which most of the DNA isenclosed by specific proteins, the histones (see epigenetics), leaving only few posi-tions accessible for endonucleases. Complete apoptotic laddering leaves DNA-oligo-mers of about 140 to 180 base pairs in length, which do no longer contain any use-ful genetic information. Apart from destroying genetic information, caspases alsodegrade structure proteins of the nucleus and cell, such as laminin, actin, andgelsolin (for a review [120]).Cytochrome C is not only an essential part of the mitochondrial electron transfer

chain and hence for the energy production in the cell, but also a crucial mediator inthe caspase activation cascade [131, 132]. On the extrinsic pathway of apoptosis,stimulation of the Fas-receptor leads to the formation of the “death-inducing sig-naling complex” (DISC). DISC activates caspase 8 from its precursor protein,which, in turn, activates “Bid”, a pro-apoptotic protein of the Bcl-2 family (seebelow). Bid then induces the liberation of cytochrome C from the mitochondrialmembrane. Other members of the pro-apoptotic Bcl-2 family are Bad and Bax, im-

14 1 Stroke

portant on the intrinsic pathway. Their activation leads to their translocation intothe outer mitochondrial membrane and to the establishment of the mitochondrialpermeability transition pore (MPTP), leading to the release of cytochrome C. An-other way of its release from mitochondria is obviously direct membrane damageby free radicals (for review see [130, 132]). Cytochrome C, together with the cyto-solic protein Apaf-1, forms the caspase precursor protein, and, using ATP, theso-called apoptosome. The apoptosome activates caspase 9 through cleavagefrom its precursor [133]. Whether cytochrome C is eventually released depends,however, on the balance of pro-apoptotic members of the Bcl-2 family like Bid,Bax, Bad, and Bag and their anti-apoptotic counterparts Bcl-2 and Bcl-xL (for reviewsee [134]). Prevailing expression of Bcl-2 and Bcl-xL results in clearly smaller infarc-tion after focal cerebral ischemia [135–137]. Treatment with a Bcl-xL fusion proteincontaining the TAT domain of the human immunodeficiency virus (HIV) for cellpenetration acts also neuroprotectively in models of ischemia [138, 139]. Bcl-2and Bcl-xL seem to have a stabilizing effect on the mitochondrial membrane;they prevent the formation of MPTP, the ensuing cytochrome C liberation, andthe formation of AIF (for review see [132]). Not only the reinforcement of anti-apoptotic factors but also the attenuation of pro-apoptotic proteins may be neuro-protective in cerebral ischemia. For instance, the phosphorylation of Bad leads toits sequestration by the protein 14-3-3 and thereby to its inactivation [140]. ThePI3K/Akt-pathway, utilized by many neurotrophic factors such as erythropoietin(see below: endogenous neuroprotection), plays an essential role in this [102,141]. It has been reported that Akt has been activated during cerebral ischemia[142]. Whether this has a functional impact in stroke remains unclear at the mo-ment.Another important protein group involved in stroke are the members of the IAP-

group (inhibitors of apoptosis proteins). The over-expression of the X-chromosome-linked IAP (XIAP) has been shown to be protective in stroke models [143]. IAPs acteither indirectly by inhibiting the apoptosome or directly inhibiting caspase 3 (forreview see [132]), but can themselves be blocked by other proteins such as DIABLO(in mice, the human homologue is Smac) that is physiologically localized in mito-chondria. DIABLO is released after focal ischemia and binds to XIAP. It is yetunclear what its function is during focal ischemia [144]. The data discussedshow, however, that apoptosis underlies a complex regulation process which iscontrolled at many levels.

1.10Ischemia-induced DNA Damage, DNA Repair, and p53 as Genotoxic Sensor

The emergence of DNA damage and its repair functions in the context of focalcerebral ischemia are particularly interesting. DNA damage is caused by oxygenfree radicals (see above). Particularly the interaction of hydroxyl radicals withDNA results in strand breaks and base alterations. Hydroxyl radicals have a veryshort half-life – with the exception of peroxynitrite, formed from NO and super-

151.10 Ischemia-induced DNA Damage, DNA Repair

oxide – and are unlikely to reach the nuclear DNA from outside the cell. They aremore likely to be produced in close proximity to the nucleus as a homolyticcleavage product of peroxynitrite, which has the ability to diffuse over longerdistances.Under physiological circumstances, DNA lesions are generated with a frequency

of about 10,000 events each day [145]. They are recognized and repaired efficiently,provided the repair enzymes are intact and not overloaded. In the wake of ischemiaand reperfusion, the tissue is flooded with oxygen free radicals causing abundantDNA damage. Alterations that are typically caused by free radicals [146] have beenfound in the mouse model of transient focal ischemia, where a massive increase inDNA damage was observed 10 to 20 minutes after reperfusion. Foremost, 8-hy-droxy-deoxy guanosine becomes much more prevalent in both nuclear as well asmitochondrial DNA after ischemia [147–150]. These alterations are repaired par-tially within 4 to 6 hours, and the increase in repair activity is proportional tothe degree of DNA damage [148, 151].There are three basic DNA repair mechanisms – base excision repair (BER), nu-

cleotide excision repair (NER), and what is known as mismatch repair. BER seemsto be of particular significance in ischemia. During BER, DNA glycosylases excisedamaged or modified bases. The thus generated purine and pyrimidine-free sites(AP-sites) are eliminated from the damaged DNA strand by AP-endonucleases(apurinic/apyrimidinic endonucleases) and replaced by DNA-polymerase b andDNA-ligase with a short strand of 1 to 4 nucleotides [147] based on the complemen-tary blueprint of the remaining undamaged strand.Apart from oxidative DNA damage, deamination of cytosine may produce uracil.

In BER, such alterations are sorted out by uracil-DNA-glycosylase (UDG). Wefound that knockout mice deficient in UDG have substantially larger infarctionsthan the wild-type controls (Endres, Meisel, Dirnagl, Jaenisch et al. unpublished).DNA repair is neither inexhaustible nor infallible. Persisting mutations are the

result of erroneous DNA repair. In models of transient focal ischemia, a dramaticincrease in the frequency of mutations has been observed. Half of them were tran-sitions and transversions. The other half were deletions and occasional insertions,mostly resulting in a frame shift (frame shift mutations) of the open reading frame(ORF). It seems that not only the exhaustion of the repair capacity, but also animpaired proofreading ability in the exonuclease-negative DNA polymerase b ex-pressed in the brain is to blame for these mutations. The major role of DNA repairis to ensure transcription of mRNA from intact genes. An obvious consequence ofmutations is instability of the genetic information that may lead to faulty proteins,eventually resulting in either cellular malfunction, disintegration, or spawning oftumors [147, 150].The maintenance of swift and effective DNA repair is therefore essential in post-

mitotic cells like neurons. In order to adapt to an increased demand, a cell may tryto increase its repair efficiency [151]. Apoptosis is another way of avoiding erro-neous protein synthesis and the potential emergence of tumors (see above).Both mechanisms are found in focal cerebral ischemia, and it seems to be the ex-tent of the damage that determines which scenario will prevail [147, 149].

16 1 Stroke

We have several reasons to believe that the tumor suppressor protein p53 repre-sents the molecular switch on the crossroads of attempted repair and apoptosis[152]. (1) p53 is mainly induced by DNA damage. (2) p53 can halt the cell cycle,and thus prevent the division of non-neuronal cells before complete repair. (3)p53 can induce apoptosis if the damage is extensive [153]. In most models of ische-mia, it could be shown that p53 had been induced. [154, 155]. The fact that p53knockouts have significantly lower infarct volumes accords with p53’s pro-apopto-tic function. However, it seems paradoxical that heterozygous p53þ/– mice showeven smaller infarctions [156]. An attenuated expression of p53 seems to give betterneuroprotection than a total lack of it, which hints at the dual function of the gene.It would be interesting to find out if longer-term neoplasias could be the downsideof this short-term effect.

1.11Epigenetics

Experimental work of the last years has shown that, apart from DNA repair, epige-netic mechanisms like DNA methylation and histone modification play also an im-portant role in ischemic stroke. The transcriptional gene expression is not onlyregulated by transcription factors, but depends also on epigenetic mechanismsthat mainly modify the chromatin structure.In the mammalian genome, DNA methylation is a covalent, postreplicative mod-

ification at the carbon 5 position of the pyrimidine ring of cytosine bases (m5C); ittypically occurs in CpG dinucleotides. DNA methylation is catalyzed by DNAmethyltransferases (DNMT1, DNMT3a, DNMT3b). The main function of DNAmethylation lies in the regulation of gene transcription. Methylated genes arebarred from expression in a process called “gene silencing” (for a review [157].First of all, many transcription activators are unable to bind to methylated loci atall. Furthermore, there is a group of methyl-CpG binding domain proteins(MBD [158]) that form a complex with histone deacetylases (HDAC), which, inturn, by deacetylation of histones, cause the chromatin structure to condense.Through this condensation, the gene regulatory loci become inaccessible for tran-scription factors. Five proteins of the MBD family are known to date: MBD1through 4 and MeCP2. HDACs deacetylate predominantly histones H3 and H4.Conversely, acetylated histones, open up the chromatin structure, permitting thegene regulatory regions (promoters, enhancers) to become active again, and, ulti-mately, enabling gene transcription through binding of transcriptional activatorproteins. While the state of deacetylation or acetylation of a CpG hypomethylatedpromoter or enhancer region is comparatively volatile, DNA hypermethylation ofthese regulatory regions results in a more stable repression [159, 160].Initially, DNA methylation was thought to be of biological significance only for

fundamental phenomena of cell biology like embryogenesis and differentiation,in X-chromosome inactivation, genomic imprinting, and in several neoplastic dis-eases. In more recent years, its pathophysiological role in a number of further con-

171.11 Epigenetics

ditions has been recognized, such as fragile-X-syndrome, ICF (Immunodeficiency,Centromere instability, Facial anomalies) syndrome, or Rett syndrome (for a review[157, 161]).For the first time, we were able to detect the involvement of DNA methyltrans-

ferase 1 and DNA methylation in an acute disease, such as ischemic stroke [162–164]. In addition, a complex pattern of an alterated expression of MBD-family pro-teins has been described in a model of transient cerebral ischemia [165]. The inhi-bition of DNA methylation had a protective effect, as had manipulations to theacetylation state of histones. Pharmacological inhibition of a histone acetyltransfer-ase with trichostatin A (TSA) caused enhanced histone acetylation in vitro as wellas in vivo and proved to be neuroprotective ([162, Meisel et al. unpublished results).A similar effect was shown recently in a mouse model of Huntington’s disease[166]. Thus, the epigenetic mechanisms of gene regulation may provide yet anotheraxis of neuroprotective therapeutic intervention.

1.12Gene Expression

The role of transcription factors such as AP-1, HIF-1 and NF-kB in the context ofcerebral ischemia has been recognized for much longer [167–169]. They initiate thetranscriptional activation of complex programs of gene expression (for a review[170, 171]). The whole hierarchical system of such programs has been unraveledover more than 15 years. Within few hours the transcription of immediate earlygenes such as c-fos is activated. Some of these, in turn, form transcription factorssuch as AP-1. They induce a variety of genes that may have protective or destructiveeffects. Cells that survive the ensuing biochemical negotiation finally express genesthat contribute to a structural re-organization and thus to greater plasticity. A greatmany of the genes involved have been subject to pathophysiological studies instroke. Their functional significance has been studied either in descriptive ways,i. e. by describing candidate genes on RNA or protein level or using transgenictechnology. Especially knockout mice have been extremely useful for this approach.The genetic elimination of specific proteins in mice used for these experiments hasbeen particularly enlightening for the untangling of the signaling cascades in-volved in and after stroke.In recent years, the advent of genetic screening methods has allowed for the sys-

tematic investigation of complex gene expression programs that are initiated bycerebral ischemia. In contrast to candidate-gene research, they permit a more un-biased approach. They are based on the assumption that focal cerebral ischemiadoes incite complex changes of gene expression. Not just a single, specifically tar-geted gene is examined for its role in the scenario of stroke but – ideally – all genes.For this new research paradigm terms have been coined like discovery science,genomics, transcriptomics, proteomics, and cellomics. So-called high throughputtechnologies as well as a highly evolved bioinformatic methodology are a prerequi-site for acquiring and managing the vast bulk of information involved. DNA micro

18 1 Stroke

arrays or SAGE technology facilitate the examination of global changes in gene ex-pression on mRNA level (transcriptomics). Global changes on the level of proteinsynthesis can be described with a combination of two-dimensional gel electrophor-esis and identification techniques (proteomics). The first results of a number ofstudies using these techniques in the context of cerebral ischemia are now avail-able [103, 172–174].What do these results imply? First of all, they confirm that the assumption of the

existence of complex stroke-induced programs of gene expression is correct. Sec-ondly, they demonstrate that they follow a robust temporal and spatial pattern.Not only do the expression patterns differ between core and penumbra, but alsowithin the penumbra, there are many spatial and temporal differentiations. Thishas even given rise to a concept of “multiple penumbras” [175]. Thirdly, thisapproach has not only confirmed many findings of candidate gene research butalso underpinned our pathophysiological understanding. Thus, a number of heatshock proteins, anti-oxidative enzymes, growth factors, and pro- as well as anti-in-flammatory proteins have been detected. Not just transcription factors, but a varietyof proteins have been have been found to play essential roles in the metabolism ofRNA and DNA. Finally, proteins that contribute to structural and functional reor-ganization [103, 176–178] are thought to establish the molecular base of the func-tional healing potential after strokes. Not surprisingly, a number of hitherto uni-dentified genes have been brought forth by discovery science.Although the results of these screening methods give a great boost to stroke

research and open doors to new insights, a full functional characterization of thedifferentially expressed genes will take many more years. Our understanding ofthe emerging network of gene expression increasingly depends on not only biolo-gical but also mathematical models. However, there are things we do know already;for instance that protein modifications are of great functional importance. Thus,the functional value of numerous proteins depends on their phosphorylationstate. For example, the specific phosphorylation of the usually pro-apoptotic proteinBad prevents apoptosis in the ischemic penumbra. The differential function equalsthus not simply differential gene expression, and genetic screening methods willnever be sufficient to paint the full picture of stroke-induced molecular pathways.The same holds for proteomics, which, just like transcriptomics, neglects the func-tional importance of RNAs that do not code for proteins.These molecular pathways, however, potentially lead to new, neuroprotective

therapies and open up new perspectives for fundamental issues like neuronal plas-ticity and stem cell research.

1.13Cell Replacement

In recent years, it has been shown in different models that ischemic brain injuryinduces neuron formation from neuronal stem cells. Such cells reside in the sub-granular zone of the dentate gyrus, in the subventricular zone below the lateral

191.13 Cell Replacement

ventricles, and periventricularly in the hippocampus. They can migrate into partsof ischemic lesions within the striatum, the dentate nucleus, and the hippocampalCA1 region [179]. It is still a matter of debate to what extent a similar neuroneogen-esis also takes place within the penumbra in cortical regions. Neither is it entirelyunderstood to what extent these migrating cells, that doubtlessly express neuronalmarkers, become functional. Observations so far indicate that it is not a very effec-tive process and that most of the migrating cells perish. However, if these neuronalcells could be enabled or encouraged to assume the function of damaged neurons,this would imply another new therapeutic paradigm (for a review [180, 181]). In-deed, intrathecal infusion of growth factors like FGF-2 and EGF after global cere-bral ischemia did not only stimulate the migration and the survival of these cellsbut resulted also in improved functional results both on cellular and on neurobe-havioral level [182]. Astrocytes are newly formed in a similar fashion [183]. For fu-ture therapeutic evaluation, it should be kept in mind not only that growth factorssuch as erythropoietin, FGF-2, EGF, VEGF, SCF, and BDNF have this beneficialeffect [182, 184, 185], but also that their expression depends on the stimulationof the NMDA receptor by glutamate [186]. A pharmacological blockade of thisreceptor with the intention of inducing neuroprotection could interfere with thiseffect and thus be counterproductive. This is yet another example of an incompleteunderstanding of the signaling mechanisms, leading to a paradox in the interpre-tation of our neuroprotective data. Bearing this in mind, it is needless to say thatthe ongoing experimental studies of cell replacement are still far from clinical im-plementation. However, the available experimental observations are very exciting:For instance, neural and also haematopoietic pluripotent cells differentiate intoneuronal and glial cells when injected into infarction sites (for a review [187]).On the other hand, this potentially helpful strategy may have detrimental effectstoo. In a mouse model of stroke, for example, murine embryonic stem cells didnot migrate into the damaged tissue, but produced highly malignant teratocarcino-mas at the site of implantation, independent of whether or not they had been pre-differentiated in vitro into neural progenitor cells [188]. As far as safety is con-cerned, an endogenous pathway of cell regeneration may be more promising.Bone-marrow-derived cells (BMDC) introduced into the brain possess a differentia-tion potential capable not only of forming microglia (see above; [98]), but also neu-rons. This apparent plasticity has been attributed to transdifferentiation [189].However, more recent data suggest in the case of neurons that BMDCs not differ-entiate into but rather fuse with Purkinje cells in the brain. These observationssuggest that cell fusion contributes to the maintenance of neurons [190]. One po-tential use for haematopoietic cells migrating from blood into ischemic brain tis-sue could be the therapeutic transfer of protective gene products such as growthfactors [98].

20 1 Stroke

1.14Endogenous Neuroprotection – Ischemic Tolerance

In the wake of ischemic stroke, not only cascades of destruction are induced, butalso endogenous protective mechanisms are activated. These mechanisms arecurrently studied intensively. Central to this research are models of ischemicpreconditioning. The idea behind ischemic preconditioning is to achieve aprotected state of a cell, tissue, or whole organism by subliminal exposure to a nox-ious stimulus (trigger), which is applied just below the level at which damagewould occur. The stimulus induces a protective state against insults that would nor-mally be lethal. Subliminal ischemia (‘ischemic tolerance’), hypoxia, reactive oxy-gen free radicals, inflammation, etc. can serve as tolerance-inducing stimuli.After induction, there are two time windows within which ischemic tolerancecan be observed. Early tolerance occurs after about 5 to 120 minutes followingstimulation (classical preconditioning), while delayed preconditioning sets inafter a latency of about 24 to 72 hours. Seven days after induction, the state of pro-tection is no longer detectable [191, 192]. Studies of patients who had transient is-chemic attacks (TIA) before suffering a persistent stroke indicate that ischemic tol-erance may also exist in humans. Subsequent to a TIA, brain infarcts were smallerin comparison with patients without preceding TIA. As a TIA, by definition, doesnot lead to infarction, it may thus be considered as a stimulus that induces toler-ance [193, 194].Unlike severe ischemia, leading to stroke, a preconditioning stimulus induces

exclusively protective, but not destructive signaling cascades. There are basicallythree stages of protective signaling that can be distinguished in delayed precondi-tioning. In the induction phase molecular sensors (mostly receptors, channels andregulators) are activated by transcription factors. During the transduction phasecertain protein kinases, transcription factors and paracrine and autocrine media-tors, such as growth factors, amplify the signal and thereby prepare the effectorphase. In this phase, proteins with a direct protective impact (anti-apoptotic,anti-inflammatory, anti-oxidative) are switched on. These mechanisms are the mo-lecular basis for ischemic tolerance and have been reviewed in detail elsewhere (fora review [191–193]). Here, we want to give only a brief outline of the concept. Onthe basis of animal models in mice and rats as well as cell culture experiments, wewere able to describe the following signaling cascade: In astrocytes, a tolerance-in-ducing hypoxic stimulus induces the transcription factor HIF-1 (hypoxia-inducible-factor-1). HIF-1 is one of the central sensors of hypoxia and induces the expressionof erythropoietin (EPO; [102, 168, 196]). Erythropoietin is released by astrocytes andbinds to the neuronal EPO-receptor. Via the activation of a protein kinase cascade,in particular phosphoinositol-3 kinase, the pro-apoptotic protein BAD becomesphosphorylated and thus inactivated [102]. This neuroprotective signaling cascademediates approximately 50% of the protection observed in our model [197]. Theremaining 50% of protection it does not account for suggest the presence offurther mechanisms in ischemic preconditioning. As a matter of fact, beside thisparacrine cascade, which is mediated by astrocytes and neurons, a number of

211.14 Endogenous Neuroprotection – Ischemic Tolerance

other mechanisms have been described. A host of anti-excitotoxic, anti-inflamma-tory and anti-apoptotic mechanisms, including programs for regeneration andrepair, have been implicated so far (for a review [191–193]).Understanding these mechanisms may enable us in the future to induce or

boost endogenous protection in patients as a novel strategy to safeguard thebrain against hypoxic/ischemic damage. For example, EPO is a promising drugin stroke therapy. There is solid evidence that exogenously applied EPO is neuro-protective in vivo and in vitro models of cerebral ischemia [102, 198, 199]. Clinicalphase I and II trials suggest that EPO is not only safe, but also beneficial in thetherapy for acute stroke [200]. In addition, pharmacological inducers of ischemictolerance in brain could become a promising tool in circumstances when ischemiacan be anticipated, e. g. before cardiac or carotid surgery. For example, desferriox-amine, an approved drug for the treatment of hemosiderosis, induces robustischemic tolerance by inducing erythropoietin in a HIF-1 dependent manner [197].Combined, the above-mentioned results demonstrate that brain cells are not only

challenged by deleterious mechanisms, but also activate innate programs to protectthemselves from ischemia. However, cerebral ischemia does not only affectthe brain parenchyma, but also other vital organs, in particular the immunesystem.

1.15Stroke Induced Immune Depression (SIDS)

Within three days after stroke, up to 61% of the stroke patients develop fever [201–205]. Findings from animal experiments show that pyrexia of more than 1hC doesnot only lead to a dramatic increase of the cytotoxic neurotransmitter glutamate[206], but also to a significantly larger volume of the infarction [207, 208]. Clinicaltrials demonstrated that both mortality and morbidity rise with the occurrence offever in stroke [209, 210]. The most common cause of fever in the acute courseof stroke is infection [201, 205, 211–213]. Infectious complications occur in 21 to65% of stroke patients [214]. Pneumonia is the most frequent infection in strokepatients [215] and a major mortality risk factor [216, 217].Stroke patients have a much higher risk of developing severe bacterial infections.

It is obvious that factors such as immobilization, reduced bulbar reflexes, drow-siness and, subsequently, a higher risk of aspiration promote pulmonary infec-tions. Although these are risk factors for bacterial infection, they cannot fullyexplain the high risk of infection observed. A large meta-analysis showed thatother factors had to be equally important. In particular, a reduced function ofthe immune system was postulated [218–220]. Very recently, we were able toprove that immunodepression following stroke causes severe bacterial infections[221]. In a mouse model we demonstrated that, three days after stroke, severe bac-terial infections (mostly pneumonia and sepsis) develop spontaneously. Stroke in-duces an over-activation of the sympathetic nervous system and leads to a rapid,severe, and sustained lymphopenia as well as to a disturbed function of lympho-

22 1 Stroke

cytes and monocytes. Interestingly, there is also evidence for an over-activation ofthe sympathetic nervous system within the first two days after stroke in humans[222]. Furthermore, we were able to show that a disturbed secretion of inter-feron-g from T cells and NK cells is central to these severe infections. Both a cel-lular reconstitution of the immune system and the application of interferon-g pre-vented infections. In addition, pharmacological inhibition of the sympathetic ner-vous system with the b-receptor-blocker propranolol not only prevented infections,but also dramatically reduced the high rate of mortality in this model [221]. Ourstudies are the first to provide a mechanistic explanation for the clinical phenom-enon of increased susceptibility to infections after stroke [223].Approaches to prevent infection after stroke seem to be desirable. Preventive

anti-infective therapy is expected to reduce a number of negative prognostic factorssuch as fever, infection-induced arterial hypotension, and the systemic release ofpro-inflammatory cytokines. In addition, with the prevention of severe infections,an earlier mobilization and rehabilitation becomes feasible. In an experimentalstroke model, we were able to demonstrate that preventive anti-infective therapywith an antibiotic after focal cerebral ischemia not only reduces mortality andthe infarct sizes, but also decreases the functional deficit [224]. These experimentalresults lead to the initiation of a Phase IIb trial for preventive antibacterial short-term therapy in patients with acute ischemic media-infarction (Preventive ANti-in-fective THERapy In Stroke; PANTHERIS, Berlin). Other groups initiated similartrials (A. Chamorro, Spain; M. Hennerici, Germany; personal communications).

1.16Conclusion

Due to intensive experimental and clinical research, our pathophysiological under-standing of cerebral ischemia has greatly improved over the last 15 years. Complexpathways involving excitotoxicity, oxidative and nitrosative stress, peri-infarct depo-larizations, inflammation, and apoptosis-like mechanisms have been described incell culture as well as animal models of cerebral ischemia. Despite impressive re-sults in animal experiments on neuroprotection, the transfer of preclinical resultsinto human stroke therapy has been so far been unsuccessful. Possible reasons forthis dilemma include problems with the models used, the design of the clinicaltrials as well as the drugs, and they have been discussed intensively [21, 225,226]. However, we believe that although we have often not been able to transferour experimental results into human stroke therapy, some cautious optimism isjustified. More and more experimental data demonstrate that the division of molec-ular mechanisms in “good” and “bad guys” is an improper simplification, and weneed to gather more data in order to identify more suitable molecular targets. Sincein stroke there are multiple pathways that cause cell death, it seems reasonable todevelop combination therapies targeting all these mechanisms. In addition, the ad-vances in understanding endogenous neuroprotective mechanisms provide newtherapeutic options to improve tissue survival after stroke. Targeting systemic com-

231.16 Conclusion

plications like severe bacterial infections in a preventive manner may also improveoutcome. Thus, an ever more refined understanding of complex damaging cas-cades as well as the development of more complex therapeutic strategies will con-tinue to be the basis for more successful therapies of stroke in the future.

24 1 Stroke

References

1 Sacco RL, Neurology 1997, 49:S39–44.2 Bonita R, Lancet 1992, 339:342–344.3 Taylor TN, Davis PH, Torner JC,

Holmes J, Meyer JW, Jacobson MF,Stroke 1996, 27:1459–1466.

4 Youman P, Wilson K, Harraf F,Kalra L, Pharmacoeconomics 2003; 21Suppl 1:43–50.

5 Adams HP Jr, Bendixen BH, KappelleLJ, Biller J, Love BB, Gordon DL,Marsh EE, Stroke 1993, 24:35–41.

6 Caplan, L.R, Caplan’s Stroke: A Clini-cal Approach. Butterworth-Heine-mann, Boston, 2000.

7 Joutel A, Corpechot C, Ducros A,Vahedi K, Chabriat H, Mouton P,Alamowitch S, Domenga V, Cecil-lion M, Marechal E, Maciazek J,Vayssiere C, Cruaud C, Cabanis EA,Ruchoux MM, Weissenbach J, BachJF, Bousser MG, Tournier-Lasserve E,Nature 1996, 383:707–710.

8 Tournier-Lasserve E, Joutel A, Melki J,Weissenbach J, Lathrop GM, Chab-riat H, Mas JL, Cabanis EA, Baudri-mont M, Maciazek J, et al, Nat. Genet.1993, 3:256–259.

9 Schmiedel J, Jackson S, Schafer J,Reichmann H, J. Neurol. 2003,250:267–277.

10 Brass LM, Isaacsohn JL, MerikangasKR, Robinette CD, Stroke 1992,23:221–223.

11 Bak S, Gaist D, Sindrup SH,Skytthe A, Christensen K, Stroke2002, 33:769–774.

12 Kiely DK, Wolf PA, Cupples LA,Beiser AS, Myers RH, Stroke 1993,24:1366–1371.

13 Liao D, Myers R, Hunt S, Shahar E,Paton C, Burke G, Province M,Heiss G, Stroke 1997, 28:1908–1912.

14 Gretarsdottir S, Sveinbjornsdottir S,Jonsson HH, Jakobsson F, Einars-dottir E, Agnarsson U, Shkolny D,Einarsson G, Gudjonsdottir HM, Val-dimarsson EM, Einarsson OB, Thor-geirsson G, Hadzic R, Jonsdottir S,Reynisdottir ST, Bjarnadottir SM,Gudmundsdottir T, GudlaugsdottirGJ, Gill R, Lindpaintner K, Sainz J,Hannesson HH, Sigurdsson GT,Frigge ML, Kong A, Gudnason V,Stefansson K, Gulcher JR, Am. J.Hum. Genet. 2002, 70:593–603.

15 Gretarsdottir S, Thorleifsson G, Rey-nisdottir ST, Manolescu A, Jonsdot-tir S, Jonsdottir T, Gudmundsdottir T,Bjarnadottir SM, Einarsson OB,Gudjonsdottir HM, Hawkins M,Gudmundsson G, GudmundsdottirH, Andrason H, Gudmundsdottir AS,Sigurdardottir M, Chou TT, Nah-mias J, Goss S, Sveinbjornsdottir S,Valdimarsson EM, Jakobsson F,Agnarsson U, Gudnason V, Thor-geirsson G, Fingerle J, Gurney M,Gudbjartsson D, Frigge ML, Kong A,Stefansson K, Gulcher JR, Nat.Genet.. 2003, 35:131–138.

16 Moulin T, Tatu L, Vuillier F, Berger E,Chavot D, Rumbach L, Cerebrovasc.Dis. 2000, 10:261–71.

17 Hankey GJ, Cerebrovasc. Dis. 2003, 16Suppl 1:14–19.

18 Heiss WD, Kracht LW, Thiel A, GrondM, Pawlik G, Brain 2001, 124:20–29.

19 Ginsberg MD, Stroke 2003, 34:214–223.

20 Astrup J, Siesjo BK, Symon L, Stroke1981, 12:723–725.

21 Dirnagl U, Iadecola C, MoskowitzMA, Trends. Neurosci. 1999, 22:391–397.

25References

22 Lo EH, Dalkara T, Moskowitz MA,Nat. Rev. Neurosci. 2003, 4:399–415.

23 Sims NR, Anderson MF, Neurochem.Int. 2002, 40:511–526.

24 Chan PH, J. Cereb. Blood Flow. Metab.2001, 21:2–14.

25 Siesjo BK, Neurochem. Pathol. 1988,9:31–88.

26 Myers RE, Yamagutchi M, Arch.Neurol. 1977, 34: 65–74.

27 Pulsinelli WA, Waldman S, Rawlin-son D, Plum F, Neurology 1982,32:1239–1246.

28 Jorgensen HS, Nakayama H,Raaschou HO, Olsen TS, Lancet 1994,344:156–159.

29 Weir CJ, Murray GD, Dyker AG,Lees KR, BMJ 1997, 314:1303–1306.

30 Bruno A, Biller J, Adams HP Jr,Clarke WR, Woolson RF, WilliamsLS, Hansen MD, Neurology 1999,52:280–284.

31 Bruno A, Levine SR, Frankel MR,Brott TG, Lin Y, Tilley BC, Lyden PD,Broderick JP, Kwiatkowski TG, Fine-berg SE, Neurology 2002, 59:669–674.

32 Baird TA, Parsons MW, Phanh T,Butcher KS, Desmond PM, Tress BM,Colman PG, Chambers BR, DavisSM, Stroke 2003, 34:2208–2214.

33 Tracey F, Stout RW, Stroke 1994,25:524–525.

34 Counsell C, McDowall M, Dennis M,BMJ 1997, 315:810.

35 Schurr A, Curr. Opin. Clin. Nutr.Metab. Care 2001, 4:287–292.

36 Smith-Swintosky VL, Pettigrew LC,Sapolsky RM, Phares C, CraddockSD, Brooke SM, Mattson MP,J. Cereb. Blood Flow Metab. 1996,16:585–598.

37 Tian GF, Baker AJ, J. Neurophysiol.2002, 88:236–248.

38 Seo SY, Kim EY, Kim H, Gwag BJ,J. Neurosci. 1999, 19: 8849–8855.

39 Ginsberg MD, Prado R, Dietrich WD,Busto R, Watson BD, Stroke 1987,18:570–574.

40 Zasslow MA, Pearl RG, Shuer LM,Steinberg GK, Lieberson RE, LarsonCP Jr, Stroke 1989, 20:519–523.

41 Parsons MW, Barber PA, DesmondPM, Baird TA, Darby DG, Byrnes G,

Tress BM, Davis SM, Ann. Neurol.2002, 52:20–28.

42 Nedergaard M, Hansen AJ, J. Cereb.Blood. Flow Metab.. 1993, 13:568–574.

43 Back T, Kohno K, Hossmann KA,J. Cereb. Blood. Flow Metab. 1994,14:12–19.

44 Back T, Ginsberg MD, Dietrich WD,Watson BD, J. Cereb. Blood FlowMetab. 1996, 16:202–213.

45 Busch E, Gyngell ML, Eis M, Hoehn-Berlage M, Hossmann KA, J. Cereb.Blood Flow Metab. 1996, 16:1090–1099.

46 Wolf T, Lindauer U, Reuter U, Back T,Villringer A, Einhaupl K, Dirnagl U,J. Cereb. Blood Flow Metab. 1997,17:950–954.

47 Iijima T, Mies G, Hossmann KA,J. Cereb. Blood. Flow Metab. 1992,12:727–733.

48 Hoehn-Berlage M, NMR Biomed.1995, 8:345–358.

49 Pinard E, Nallet H, MacKenzie ET,Seylaz J, Roussel S, Stroke 2002,33:606–612.

50 Strong AJ, Fabricius M, Boutelle MG,Hibbins SJ, Hopwood SE, Jones R,Parkin MC, Lauritzen M, Stroke 2002,33:2738–2743.

51 Emsley HC, Tyrrell PJ, J. Cereb. BloodFlow Metab. 2002, 22:1399–1419.

52 Becker KJ, Curr. Opin. Neurol. 2001,14:349–353.

53 del Zoppo G, Ginis I, Hallenbeck JM,Iadecola C, Wang X, Feuerstein GZ,Brain. Pathol. 2000, 10:95–112.

54 Zhang RL, Chopp M, Zaloga C,Zhang ZG, Jiang N, Gautam SC,Tang WX, Tsang W, Anderson DC,Manning AM, Brain Res. 1995,682:182–188.

55 Lindsberg PJ, Carpen O, Paetau A,Karjalainen-Lindsberg ML, Kaste M,Circulation. 1996, 94:939–945.

56 Fiszer U, Korczak-Kowalska G, Pala-sik W, Korlak J, Gorski A, Czlon-kowska A, Acta. Neuro. Scand. 1998,97:221–224.

57 del Zoppo GJ, Schmid-SchonbeinGW, Mori E, Copeland BR, ChangCM, Stroke 1991, 22:1276–1283.

26 1 Stroke

58 Mori E, del Zoppo GJ, Chambers JD,Copeland BR, Arfors KE, Stroke 1992,23:712–718.

59 Chen H, Chopp M, Zhang RL, Bod-zin G, Chen Q, Rusche JR, Todd RF,Ann. Neurol. 1994, 35:458–463.

60 Zhang RL, Chopp M, Jiang N, TangWX, Prostak J, Manning AM, Ander-son DC, Stroke 1995, 26:1438–1442.

61 Soriano SG, Lipton SA, Wang YF,Xiao M, Springer TA, Gutierrez-Ramos JC, Hickey PR, Ann. Neurol.1996, 39:618–624.

62 Becker K, Kindrick D, Relton J, Har-lan J, Winn R, Stroke 2001, 32:206–211.

63 Korematsu K, Goto S, Nagahiro S,Ushio Y, J. Cereb. Blood Flow Metab.1994, 14:825–830.

64 Ueyama T, Ren Y, Sakai N, TakahashiM, Ono Y, Kondoh T, Tamaki N,Saito N, J. Neurochem. 2001,79:903–913.

65 Marks L, Carswell HV, Peters EE,Graham DI, Patterson J, DominiczakAF, Macrae IM, Hypertension 2001,38:116–122.

66 Tarozzo G, Campanella M, Ghiani M,Bulfone A, Beltramo M, Eur. J. Neu-rosci. 2002, 15:1663–1668.

67 Yrj�nheikki J, Tikka T, Keinanen R,Goldsteins G, Chan PH, Koistinaho J,Proc. Natl. Acad. Sci. U S A. 1999,96:13496–13500.

68 Gehrmann J, Matsumoto Y, Kreutz-berg GW, Brain Res. Brain. Res. Rev.1995, 20:269–287.

69 Davies CA, Loddick SA, Toulmond S,Stroemer RP, Hunt J, Rothwell NJ,J. Cereb. Blood Flow Metab. 1999,19:87–98.

70 Gregersen R, Lambertsen K, Fin-sen B, J. Cereb. Blood. Flow Metab.2000, 20:53–65.

71 Stoll G, Rev. Neurol. (Paris) 2002,158:887–891.

72 O’Neill LA, Kaltschmidt C, TrendsNeurosci. 1997, 20:252–258.

73 Iadecola C, Alexander M, Curr. Opin.Neurol. 2001, 14:89–94.

74 Block F, Peters M, Nolden-Koch M,Neuroreport 2000, 11:963–967.

75 Tarkowski E, Rosengren L, Blom-strand C, Wikkelso C, Jensen C,

Ekholm S, Tarkowski A., Stroke 1995,26:1393–1398.

76 Nawashiro H, Tasaki K, Ruetzler CA,Hallenbeck JM, J. Cereb. Blood FlowMetab. 1997, 17:483–390.

77 Bruce AJ, Boling W, Kindy MS,Peschon J, Kraemer PJ, CarpenterMK, Holtsberg FW, Mattson MP, Nat.Med. 1996, 2:788–794.

78 Fontaine V, Mohand-Said S, Hano-teau N, Fuchs C, Pfizenmaier K,Eisel U, J. Neurosci. 2002, 22:RC216.

79 Clark WM, Rinker LG, Lessov NS,Hazel K, Hill JK, Stenzel-Poore M,Eckenstein F, Stroke 2000, 31:1715–1720.

80 Priller J, Dirnagl U, Ernst ScheringRes. Found Workshop 2002, 39:133–157.

81 Bogdan C, Paik J, Vodovotz Y,Nathan C, J. Biol. Chem. 1992,267:23301–23308.

82 Strle K, Zhou JH, Shen WH, Brous-sard SR, Johnson RW, Freund GG,Dantzer R, Kelley KW, Crit. Rev. Im-munol. 2001, 21:427–449.

83 Prehn JH, Backhauss C, Krieglstein J,J. Cereb. Blood Flow Metab. 1993,13:521–525.

84 Spera PA, Ellison JA, Feuerstein GZ,Barone FC, Neurosci. Lett. 1998,251:189–192.

85 Letterio JJ, Cytokine Growth Factor Rev.2000, 11:81–87.

86 Becker KJ, McCarron RM, Ruetzler C,Laban O, Sternberg E, Flanders KC,Hallenbeck JM, Proc. Natl. Acad Sci.U S A. 1997, 94:10873–19878.

87 Becker K, Kindrick D, McCarron R,Hallenbeck J, Winn R, Stroke 2003,34:1809–1815.

88 Takeda H, Spatz M, Ruetzler C,McCarron R, Becker K, Hallenbeck J,Stroke 2002, 33:2156–2163.

89 Iadecola C, Zhang F, Casey R, ClarkHB, Ross ME, Stroke 1996, 27:1373–1380.

90 Iadecola C, Zhang F, Casey R,Nagayama M, Ross ME, J. Neurosci.1997, 17:9157–9164.

91 Zhang F, Casey RM, Ross ME, Iade-cola C, Stroke 1996, 27:317–323.

92 Crow JP, Beckman JS, Curr. TopMicrobiol. Immunol. 1995, 196:57–73.

27References

93 Kolb JP, Leukemia 2000, 14:1685–1694.

94 Sairanen T, Ristimaki A, Karjalainen-Lindsberg ML, Paetau A, Kaste M,Lindsberg PJ, Ann. Neurol. 1998,43:738–747.

95 Iadecola C, Niwa K, Nogawa S,Zhao X, Nagayama M, Araki E, Mor-ham S, Ross ME, Proc. Natl. Acad. Sci.U S A. 2001, 98:1294–1299.

96 Sugimoto K, Iadecola C, Brain Res.2003, 960:273–276.

97 Gourmala NG, Buttini M, Limonta S,Sauter A, Boddeke HW, J. Neuroim-munol. 1997, 74:35–44.

98 Priller J, Flugel A, Wehner T, Boen-tert M, Haas CA, Prinz M, Fernan-dez-Klett F, Prass K, Bechmann I, deBoer BA, Frotscher M, KreutzbergGW, Persons DA, Dirnagl U, Nat.Med. 2001, 7:1356–1361.

99 Eglitis MA, Mezey E, Proc. Natl. Acad.Sci. U S A. 1997, 94:4080–4085.

100 Eglitis MA, Dawson D, Park KW,Mouradian MM, Neuroreport 1999,10:1289–1292.

101 Wehner T, Bontert M, Eyupoglu I,Prass K, Prinz M, Klett FF, Heinze M,Bechmann I, Nitsch R, Kirchhoff F,Kettenmann H, Dirnagl U, Priller J,J. Neurosci. 2003, 23:5004–5011.

102 Ruscher K, Freyer D, Karsch M,Isaev N, Megow D, Sawitzki B, PrillerJ, Dirnagl U, Meisel A, J. Neurosci.2002, 22:10291–10301.

103 Trendelenburg G, Prass K, Priller J,Kapinya K, Polley A, Muselmann C,Ruscher K, Kannbley U, Schmitt AO,Castell S, Wiegand F, Meisel A,Rosenthal A, Dirnagl U, J. Neurosci.2002, 22:5879–5888.

104 Heo, J. H. et al, J. Cereb. Blood FlowMetab. 1999, 19:624–633.

105 Gasche Y, Fujimura M, Morita-Fuji-mura Y, Copin JC, Kawase M, Mas-sengale J, Chan PH, J Cereb. BloodFlow Metab. 1999; 19:1020–1028.

106 Lapchak PA, Chapman DF, Zivin JA,Stroke 2000, 31:3034–3040.

107 Asahi, M. et al, J. Neurosci. 2001,21:7724–7732.

108 Rosenberg, G. A., Estrada, E. Y. &Dencoff J. E, Stroke 1998, 29:2189–2195.

109 Nagai, N. Yamamoto S, Tsuboi T,Ihara H, Urano T, Takada Y, Tera-kawa S, Takada A, J. Cereb. Blood FlowMetab. 2002, 21:631–634.

110 Wang YF, Tsirka SE, Strickland S,Stieg PE, Soriano SG, Lipton SA, Nat.Med. 1998, 4:228–231.

111 Tabrizi P, Wang L, Seeds N, McCombJG, Yamada S, Griffin JH, Carme-liet P, Weiss MH, Zlokovic BV,Arterioscler. Thromb. Vasc. Biol. 1999,19:2801–2806.

112 Wang X, Lee SR, Arai K, Lee SR,Tsuji K, Rebeck GW, Lo EH, Nat.Med. 2003, 9:1313–1317.

113 Lapchak PA, Expert. Opin. Investig.Drugs 2002, 11:1623–1632.

114 Dijkhuizen RM, Asahi M, Wu O,Rosen BR, Lo EH, Stroke 2002,33:2100–2104.

115 Pfefferkorn T, Rosenberg GA, Stroke2003, 34:2025–2030.

116 Nicole O, Docagne F, Ali C, Mar-gaill I, Carmeliet P, MacKenzie ET,Vivien D, Buisson A, Nat. Med. 2001,7:59–64.

117 Matys T, Strickland S, Nat. Med. 2003,9:371–372.

118 Liberatore GT, Samson A, Bladin C,Schleuning WD, Medcalf RL, Stroke2003, 34:537–543.

119 Friedlander RM, N. Engl. J. Med.2003, 348:1365–1375.

120 Endres M, Dirnagl U, Adv. Exp. Med.Biol. 2002, 513:455–473.

121 Ferrer I, Planas AM, J. Neuropathol.Exp. Neurol. 2003, 62:329–339.

122 Mattson MP, Duan W, Pedersen WA,Culmsee C, Apoptosis. 2001,6:69–81.

123 Namura S, Zhu J, Fink K, Endres M,Srinivasan A, Tomaselli KJ, Yuan J,Moskowitz MA, J. Neurosci. 1998,18:3659–3668.

124 Le DA, Wu Y, Huang Z, Matsushita K,Plesnila N, Augustinack JC, HymanBT, Yuan J, Kuida K, Flavell RA,Moskowitz MA. Proc. Natl. Acad. Sci.U S A. 2002, 99:15188–193.

125 Hara H, Friedlander RM, GagliardiniV, Ayata C, Fink K, Huang Z, Shi-mizu-Sasamata M, Yuan J, MoskowitzMA, Proc. Natl. Acad. Sci. U S A.1997, 94:2007–2012.

28 1 Stroke

126 Yu SW, Wang H, Poitras MF,Coombs C, Bowers WJ, Federoff HJ,Poirier GG, Dawson TM, Dawson VL,Science 2002, 297:259–263.

127 Endres, M., Wang, Z.–Q., Namura, S.,Waeber, C. & Moskowitz, M. A,J. Cereb. Blood Flow Metab. 1997,17:1143–1151.

128 Eliasson MJ, Sampei K, Mandir AS,Hurn PD, Traystman RJ, Bao J, Pie-per A, Wang ZQ, Dawson TM, SnyderSH, Dawson VL, Nat. Med. 1997,3:1089–1095.

129 Ying W, Garnier P, Swanson RA,Biochem. Biophys. Res. Commun. 2003,308:809–813.

130 Love S, Prog. Neuropsychopharmacol.Biol. Psychiatry 2003, 27:267–282.

131 Mattson MP, Liu D, Biochem. Biophys.Res. Commun. 2003, 304:539–549.

132 van Gurp M, Festjens N, van Loo G,Saelens X, Vandenabeele P, Biochem.Biophys. Res. Commun. 2003,304:487–497.

133 Li P, Nijhawan D, Budihardjo I, Sri-nivasula SM, Ahmad M, Alnemri ES,Wang X, Cell 1997, 91:479–489.

134 Hengartner MO, Nature 2000,407:770–776.

135 Linnik MD, Zahos P, Geschwind MD,Federoff HJ, Stroke 1995, 26:1670–1674.

136 Martinou, J. C., Dubois-Dauphin, M.,Staple, J. K., Rodriguez, P., Fran-kowski, H., Missotten, M., Alber-tini, P., Talabot, D., Catsicas, S.,Pietra, C. & Huarte, J, Neuron 1994,13:1017–1030.

137 Wiessner C, Allegrini PR, Rupalla K,Sauer D, Oltersdorf T, McGregor AL,Bischoff S, Bottiger BW, van der Put-ten H, Neurosci. Lett. 1999, 268:119–122.

138 Cao G, Pei W, Ge H, Liang Q, Luo Y,Sharp FR, Lu A, Ran R, Graham SH,Chen J, J. Neurosci. 2002, 22:5423–5431

139 Kilic E, Dietz GP, Hermann DM,Bahr M, Ann. Neurol. 2002, 52:617–622.

140 Datta SR, Katsov A, Hu L, Petros A,Fesik SW, Yaffe MB, Greenberg ME,Mol. Cell. 2000, 6:41–51.

141 Datta SR, Brunet A, Greenberg ME,Genes. Dev. 1999, 13:2905–2927.

142 Noshita N, Lewen A, Sugawara T,Chan PH, J. Cereb. Blood. Flow. Metab.2001, 21:1442–1450.

143 Trapp T, Korhonen L, Besselmann M,Martinez R, Mercer EA, Lindholm D,Mol. Cell. Neurosci. 2003, 23:302–313.

144 Saito A, Hayashi T, Okuno S, Fer-rand-Drake M, Chan PH, J. Cereb.Blood. Flow. Metab. 2003, 23:1010–1019.

145 Lindahl T, Nature 1993, 362:709–715.146 Jaruga P, Dizdaroglu M, Nucleic Acids

Res. 1996, 24:1389–1394.147 Liu PK, Grossman RG, Hsu CY,

Robertson CS, Trends. Neurosci. 2001,24:581–588.

148 Liu, P. K., Hsu, C. Y., Dizdaroglu, M.,Floyd, R. A., Kow, Y. W., Karakaya, A.,Rabow, L. E. & Cui, J.–K, J. Neurosci.1996, 16:6795–6806.

149 Englander EW, Hu Z, Sharma A, LeeHM, Wu ZH, Greeley GH, J. Neuro-chem. 2002, 83:1471–1480.

150 Chopp M, Chan PH, Hsu CY,Cheung ME, Jacobs TP, Stroke 1996,27:363–369.

151 Lan J, Li W, Zhang F, Sun FY, Na-gayama T, O’Horo C, Chen J, J. Cereb.Blood. Flow Metab. 2003, 23:1324–1339.

152 Oren M, Cell. Death. Differ. 2003,10:431–442.

153 Jacks T, Weinberg RA, Nature 1996381:643–644.

154 Chopp M, Li Y, Zhang ZG, FreytagSO, Bioche. Biophy. Re. Commun.1992, 182:1201–2107.

155 Li Y, Chopp M, Zhang ZG, Zaloga C,Niewenhuis L, Gautam S, Stroke1994, 25:849–855.

156 Crumrine RC, Thomas AL, MorganPF, J. Cereb. Blood. Flow. Metab. 1994,14:887–891.

157 Jaenisch R, Bird A, Nat. Genet. 2003,33 Suppl:245–254.

158 Wade PA, Bioessays 2001, 23:1131–1137.

159 Jenuwein T, Allis CD, Science 2001,293:1074–1080.

160 Bird A, Genes Dev. 200, 16:6–21.161 Robertson KD, Wolffe AP, Nat. Rev.

Genet. 2000, 1:11–9.

29References

162 Endres M, Meisel A, Biniskiewicz D,Namura S, Prass K, Ruscher K, Lip-ski A, Jaenisch R, Moskowitz MA,Dirnagl U, J. Neurosci. 2000,20:3175–3181.

163 Endres M, Fan G, Meisel A, Dir-nagl U, Jaenisch R, Neuro Report2001, 12:3763–3766.

164 Meisel A, U Dirnagl, M Endres,Maturation phenomenon in cerebralischemia, Springer-Verlag Berlin Hei-delberg, 2001.

165 Jung BP, Zhang G, Ho W, Francis J,Eubanks JH, Neuroscience2002;115:515–524.

166 Hockly E, Richon VM, Woodman B,Smith DL, Zhou X, Rosa E, Sathasi-vam K, Ghazi-Noori S, Mahal A,Lowden PA, Steffan JS, Marsh JL,Thompson LM, Lewis CM, Marks PA,Bates GP, Proc. Natl. Acad. Sci. U S A2003, 100:2041–2046.

167 Kapinya KF, Penzel R, Sommer C,Kiessling M, Brain Res. 2000,872:282–293.

168 Ruscher K, Isaev N, Trendelenburg G,WeihM, Iurato L, Meisel A, Dirnagl U,Neurosci. Lett. 1998, 254:117–120.

169 O’Neill LA, Kaltschmidt C, Trends.Neurosci. 1997, 20:252–258.

170 Carmichael ST, Curr. Opin Neurol.2003, 16:699–704.

171 Read SJ, Parsons AA, Harrison DC,Philpott K, Kabnick K, O’Brien S,Clark S, Brawner M, Bates S, Gloger I,Legos JJ, Barone FC, J. Cereb. BloodFlow Metab. 2001, 21:755–778.

172 Lu A, Tang Y, Ran R, Clark JF, Aro-now BJ, Sharp FR, J. Cereb. BloodFlow Metab. 2003, 23:786–810.

173 Keyvani K, Witte OW, Paulus W,J. Cereb. Blood Flow Metab. 2002,22:153–160.

174 Kim YD, Sohn NW, Kang C, Soh Y,Brain Res. Bull. 2002, 58:491–498.

175 Sharp FR, Lu A, Tang Y, Millhorn DE,J. Cereb. Blood Flow Metab. 2000,20:1011–1032.

176 Wang H, Zhan Y, Xu L, FeuersteinGZ, Wang X, Stroke 2001, 32:1020–1027.

177 Raghavendra Rao VL, Bowen KK,Dhodda VK, Song G, Franklin JL,

Gavva NR, Dempsey RJ, J. Neuro-chem. 2002, 83:1072–1086.

178 Jin K, Mao XO, Eshoo MW, Na-gayama T, Minami M, Simon RP,Greenberg DA, Ann. Neurol. 2001,50:93–103.

179 Kokaia Z, Lindvall O, Curr. Opin.Neurobiol. 2003, 13:127–132.

180 Moskowitz MA, Lo EH, Stroke 2003,34:324–326.

181 Nadareishvili Z, Hallenbeck J,N. Engl.J Med. 2003, 348:2355–2356.

182 Nakatomi H, Kuriu T, Okabe S,Yamamoto S, Hatano O, Kawahara N,Tamura A, Kirino T, Nakafuku M, Cell2002, 11:429–441.

183 Sharp FR, Liu J, Bernabeu R, BrainRes. Dev. Brain Res. 2002, 134:23–30.

184 Shingo T, Sorokan ST, Shimazaki T,Weiss S, J. Neurosci. 2001, 21:9733–43.

185 Jin K, Mao XO, Sun Y, Xie L, Green-berg DA, J. Clin. Invest. 2002,110:311–319.

186 Arvidsson A, Collin T, Kirik D,Kokaia Z, Lindvall O, Nat.Med. 2002,8:963–970.

187 Savitz SI, Rosenbaum DM, DinsmoreJH, Wechsler LR, Caplan LR, Ann.Neurol. 2002, 52:266–275.

188 Erdo F, Buhrle C, Blunk J, Hoehn M,Xia Y, Fleischmann B, Focking M,Kustermann E, Kolossov E, Hesche-ler J, Hossmann KA, Trapp T, J. Cereb.Blood Flow Metab. 2003, 23:780–785.

189 Priller J, Persons DA, Klett FF, Kem-permann G, Kreutzberg GW, Dir-nagl U, J. Cell Biol. 2001, 155:733–8.

190 Alvarez-Dolado M, Pardal R, Garcia-Verdugo JM, Fike JR, Lee HO, Pfef-fer K, Lois C, Morrison SJ, Alvarez-Buylla A, Nature 2003, 425:968–973.

191 Dirnagl U, Simon RP, HallenbeckJM, Trends.Neurosci. 2003, 26:248–254.

192 Weih M, Prass K, Ruscher K, Tren-delenburg G, Dirnagl U, Riepe MW,Meisel A, Nervenarzt 2001, 72:255–260.

193 Kirino T, J. Cereb. Blood Flow Metab.2002, 22:1283–1296.

194 Weih M, Kallenberg K, Bergk A,Dirnagl U, Harms L, Wernecke KD,

30 1 Stroke

Einhaupl KM, Stroke 1999, 30:1851–1854.

195 Moncayo J, de Freitas GR, Bogous-slavsky J, Altieri M, van Melle G,Neurology 2000 Jun, 54:2089–94.

196 Prass K, Ruscher K, Karsch M,Isaev N, Megow D, Priller J, ScharffA, Dirnagl U, Meisel A, J. Cereb. BloodFlow Metab. 2002, 22:520–525.

197 Prass K, Scharff A, Ruscher K,Lowl D, Muselmann C, Victorov I,Kapinya K, Dirnagl U, Meisel A,Stroke 2003, 34:1981–1986.

198 Brines ML, Ghezzi P, Keenan S,Agnello D, de Lanerolle NC, Cera-mi C, Itri LM, Cerami A, Proc. Natl.Acad. Sci. U S A. 2000, 97:10526–31.

199 Sakanaka M, Wen TC, Matsuda S,Masuda S, Morishita E, Nagao M,Sasaki R, Proc. Natl. Acad. Sci. U S A1998, 95:4635–4640.

200 Ehrenreich H, Hasselblatt M, Dem-bowski C, Cepek L, Lewczuk P, Stie-fel M, Rustenbeck HH, Breiter N,Jacob S, Knerlich F, Bohn M, PoserW, Ruther E, Kochen M, Gefeller O,Gleiter C, Wessel TC, De Ryck M,Itri L, Prange H, Cerami A, Brines M,Siren AL, Mol. Med. 2002, 8:495–505.

201 Castillo J, Davalos A, Marrugat J,Noya M, Stroke 1998, 29:2455–2460.

202 Georgilis K, PlomaritoglouA, Dafni U,Bassiakos Y, Vemmos K, Journal ofInternal Medicine 1999, 246:203–209.

203 Johnston KC, Li JY, Lyden PD, Han-son SK, Feasby TE, Adams RJ, FaughtRE Jr, Haley EC Jr, Stroke 1998,29:447–453.

204 Boysen G, Christensen H, Stroke2001, 32:413–417.

205 Reith J, Jorgensen HS, Pedersen PM,Nakayama H, Raaschou HO, Jeppe-sen LL, Olsen TS, Lancet 1996,347:422–425.

206 Takagi K, Ginsberg MD, Globus MY,Martinez E, Busto R, Am. J. Physiol.1994, 267:1770–1776.

207 Wass CT, Lanier WL, Hofer RE,Scheithauer BW, Andrews AG,Anesthesiology 1995, 83:325–335.

208 Kim Y, Busto R, Dietrich WD, Kray-dieh S, Ginsberg MD, Stroke 1996,27:2274–2280.

209 Hajat, C., S. Hajat, and P. Sharma,Stroke 2000, 31:410–414.

210 Kammersgaard LP, Jorgensen HS,Rungby JA, Reith J, Nakayama H,Weber UJ, Houth J, Olsen TS, Stroke2002, 33:1759–1762.

211 Hilker R, Poetter C, Findeisen N,Sobesky J, Jacobs A, Neveling M,Heiss WD, Stroke 2003, 34:975–981.

212 Hindfelt B. Acta. Neurol. Scand. 1976,53:72–79.

213 Grau AJ, Buggle F, Schnitzler P, SpielM, Lichy C, Hacke W, J. Neurol. Sci.1999, 171:115–120.

214 Prass K, Meisel C, Wolf T, Volk HD,Dirnagl U, Meisel A, Pharmacology ofcerebral ischemia. medpharm, Stutt-gart 2002.

215 Davenport RJ, Dennis MS, Well-wood I, Warlow CP, Stroke 1996,27:415–420.

216 Katzan IL, Cebul RD, Husak SH,Dawson NV, Baker DW, Neurology2003, 60:620–625.

217 Vernino S, Brown RD Jr, Sejvar JJ,Sicks JD, Petty GW, O’Fallon WM,Stroke 2003, 34:1828–1832.

218 Perry, L., and C. P. Love, Dysphagia2001, 16:7–18.

219 Marik PE, N. Engl. J. Med. 2001,344:665–671.

220 Finegold SM, Rev. Infect. Dis. 1991,13:S737–S742.

221 Prass K, Meisel C, Hoflich C, Braun J,Halle E, Wolf T, Ruscher K, Victo-rov IV, Priller J, Dirnagl U, Volk HD,Meisel A, J. Exp. Med. 2003, 198:725–736.

222 Meyer S, Fischer C, Treib J, Georg TStrittmatter M, Akt. Neurologie 2001,28:170–175.

223 Davenport RJ, SAGE KE 2003 (36):nw 124–124.

224 Meisel C, Prass K, Braun J, Victorov I,Wolf T, Megow D, Halle E, Volk HD,Dirnagl U, Meisel A, Stroke 2004,35:2–6.

225 Gladstone DJ, Black SE, Hakim AM,Stroke 2002, 33:2123–2136.

226 del Zoppo GJ, Neurology 1998,51(Suppl 3):S59–61.