Cerebral ischemia studied with positron emission tomography and microdialysis

Full text

(1)

(2)

(3)

(4)

(5)

(6)

(7) .

(8)

(9) .

(10)

(11)

(12)

(13)

(14) !. "#$#% %!&'(). *$* +,-#%,$*$, "*),#+, ""*)* .//..

(15) Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in Neurosurgery presented at Uppsala University in 2002 Abstract Frykholm, P 2002. Cerebral ischemia studied with PET and microdialysis. Acta Universitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1157. 72 pp. Uppsala. ISBN 91-554-5319-8 Stroke is the third leading cause of morbidity and mortality in the industrialized world. Subarachnoid hemorrhage (SAH), the least common form of stroke, is one of the most demanding diseases treated in neurointensive care units. Cerebral ischemia may develop rapidly, and has a major influence on outcome.To be able to save parts of the brain that are at risk for ischemic brain damage, there is a need for reliable monitoring techniques.Understanding the pathophysiology of cerebral ischemia is a prerequisite both for the correct treatment of these diseases and for the development of new monitoring techniques and treatment modalities. The main aim of this thesis was to gain insight into the mechanisms of cerebral ischemia by studying early hemodynamic and metabolic changes with positron emission tomography and neurochemical changes with microdialysis. A secondary aim was to evaluate the potential of these techniques for detecting ischemia and predicting the degree of reversibility of ischemic changes. Early changes in cerebral blood flow (CBF) and metabolism (CMRO2) were studied with repeated positron emission tomography in an experimental model (MCAO) of transient focal ischemia, and in SAH patients. CMRO 2 was superior to CBF in discriminating between tissue with irreversible damage and tissue with the potential for survival in the experimental model. A metabolic threshold of ischemia was found. Neurochemical changes in the ischemic regions were studied simultaneously with microdialysis. Extracellular concentrations of glucose, lactate, hypoxanthine, glutamate and glycerol were measured, and the lactate/pyruvate (LP) and lactate/glucose ratios were calculated. Changes in all the microdialysis parameters were related to the degree of ischemia (severe ischemia or penumbra). Especially the LP ratio and glycerol were found to be robust markers of the degree of ischemic brain damage. In the patients, hemodynamic and metabolic changes were common, but diverse in the acute phase of SAH, and it was suggested that these changes may contribute to an increased vulnerability for secondary events and the development of secondary ischemic brain damage. Key Words: Cerebral ischemia, penumbra, positron emission tomography, microdialysis, middle cerebral artery occlusion, subarachnoid hemorrhage. Peter Frykholm, Department of Neuroscience, Section of Neurosurgery, Uppsala University Hospital, SE-751 85 Uppsala, Sweden, Peter.Frykholm@neurokir.uu.se. © Peter Frykholm 2002 ISSN 0282-7476 ISBN 91-554-5319-8 Printed in Sweden by Fyris-Tryck AB, Uppsala, 2002. 2.

(16) PUBLICATIONS This thesis is based on the following papers, which will be referred to in the text by their Roman numerals. I. Peter Frykholm, Jesper L. R. Andersson, Johann Valtysson, Hans C:son Silander, Lars Hillered, Lennart Persson, Yngve Olsson, Wen Ru Yu, Gunnar Westerberg, Yasuyoshi Watanabe, Bengt Långström, Per Enblad. A metabolic threshold of irreversible ischemia demonstrated by PET in a middle cerebral artery occlusion-reperfusion model. Acta Neurol Scand 2000; 102: 18-26. II. Per Enblad, Peter Frykholm, Johann Valtysson, Hans C:son Silander, Jesper L. R. Andersson, Karl-Johan Fasth, Yasuyoshi Watanabe, Bengt Långström, Lars Hillered, Lennart Persson. Middle cerebral artery occlusion and reperfusion in primates monitored by microdialysis and sequential positron emission tomography. Stroke 2001; 32: 1574-1580. III. Peter Frykholm, Lars Hillered, Bengt Långström, Lennart Persson, Johann Valtysson, Yasuyoshi Watanabe, Per Enblad. Increase of interstitial glycerol reflects the degree of ischaemic brain damage: a PET and microdialysis study in a middle cerebral artery occlusion-reperfusion primate model. J Neurol Neurosurg Psychiatry 2001; 71: 455-461. IV. Peter Frykholm, Lars Hillered, Bengt Långström, Lennart Persson, Johann Valtysson, Per Enblad. The relation between CBF and oxygen metabolism and the extracellular glucose and lactate concentrations during middle cerebral artery occlusion and reperfusion. A microdialysis and PET study in primates. Submitted. V. Peter Frykholm, Jesper L. R. Andersson, Bengt Långström, Lennart Persson, Per Enblad. Hemodynamic and metabolic disturbances in the acute stage of subarachnoid hemorrhage demonstrated by PET. Submitted. Reprints were made with the permission of the publishers.. 3.

(17) CONTENTS ABBREVIATIONS. 6. INTRODUCTION. 8. REVIEW OF THE LITERATURE The concept of the ischemic penumbra Flow thresholds Metabolic thresholds Uncoupling between flow and metabolism The no-reflow phenomenon and delayed hypoperfusion Cerebral energy metabolism in normal physiology Energy failure The role of calcium Glutamate and excitotoxicity Radical oxygen species (ROS) Mitochondrial dysfunction Cell membrane phospholipid breakdown Microdialysis studies of cerebral ischemia Cerebral ischemia in the pathophysiology of subarachnoid hemorrhage. 9 9 10 11 11 12 12 14 14 15 16 17 17 18 19. AIMS. 21. METHODS Experimental animal model MCAO with reperfusion Patients Positron emission tomography Microdialysis Statistical methods Histopathological methods. 22 22 23 23 29 32 32. RESULTS Paper I: A metabolic threshold of ischemia Papers II IV: Neurochemical changes in cerebral ischemia Paper V: Early PET changes in SAH. 34 34 38 46. DISCUSSION A quest for thresholds Microdialysis as an instrument for clinical monitoring of cerebral ischemia Microdialysis in elucidating the pathophysiology of cerebral ischemia Primary and secondary ischemia in SAH. 49 51 54 58 60. 4.

(18) CONCLUSIONS. 60. ACKNOWLEDGEMENTS. 61. REFERENCES. 63. 5.

(19) ABBREVIATIONS AA ACA AComA AMPA CBF/rCBF CBV CMRO2 CMRG CSF CT DAG EAA eNOS GOS ICH ICA ICP IP3 iNOS IR LG LP MCA MCAO MD MR mM NAD NMDA nNOS NO OER PAF PET PICA PL PLA PLC PR ROS TBI SAH. arachidonic acid anterior cerebral artery anterior communicating artery α-amino-3-hydroxy-5-methyl-5-isoxazoleproprionate cerebral blood flow* cerebral blood volume cerebral metabolic rate of oxygen cerebral metabolic rate of glucose cerebrospinal fluid computed tomography diacylglycerol excitatory amino acid endothelial nitric oxide Glasgow outcome scale intracerebral hematoma internal carotid artery intracranial pressure inositol-3-phosphate inducible nitric oxide synthase infarction region lactate glucose (ratio) lactate pyruvate (ratio) middle cerebral artery middle cerebral artery occlusion microdialysis magnetic resonance millimoles per litre nicotine adenine dinucleotide N-methyl-D-aspartate neural nitric oxide synthase nitric oxide oxygen extraction ratio platelet activating factor positron emission tomography posterior inferior cerebellar artery phospholipid phospholipase A phopholipase C penumbra region radical oxygen species traumatic brain injury subarachnoid hemorrhage. 6.

(20) SD TCA WFNS XO. standard deviation tricarboxylic acid (cycle) World Federation of Neurological Surgeons xanthine oxidase. The suffix r stresses the difference between regional and global values. However, it is usually implicit, and therefore omitted.. 7.

(21) INTRODUCTION Stroke is one of the leading causes of morbidity and mortality in Sweden. The major categories of stroke are ischemic stroke (85%), intracerebral hemorrhage (10%) and subarachnoid hemorrhage (2 5%). The incidence of stroke in the Swedish population is 30 000 cases per year; two thirds are first-ever stroke. The risk of death within one month is approximately 20%, and one third of survivors are dependent on care-givers for daily activities. Stroke is the leading diagnosis for in-patient hospital care patients with stroke consume one million hospital-care days in Sweden per year. The economical burden on society is heavy appraised to be ten billion Swedish kronor per year.1 Morbidity and mortality in stroke is being reduced. This has been attributed to better acute care and better rehabilitation (especially the so-called Stroke-units) but may also reflect a shift to less serious forms of the disease.2 In the future, we can expect that new treatment modalities, such as neuroprotective agents, may improve prognosis further, even though numerous studies of pharmacological treatment as yet have been fruitless, or have at best yielded conflicting results.3 On the other hand, thrombolysis with tissue plasmin activator in ischemic stroke is becoming an established therapy after promising results.4 It has become evident that treatment is beneficial and cost-effective especially in certain sub-groups of patients, while signs of hemorrhage or large ischemic areas on diagnostic CT dramatically increase the complication rate. In this setting, there is a need for more efficient diagnostic modalities that facilitate choosing patients suitable for treatment. In this thesis, the potential of positron emission tomography (PET) in predicting reversible and irreversible ischemia is investigated. The prognosis of subarachnoid hemorrhage (SAH) has improved with modern neurointensive care, but still a significant proportion of the patients have a poor clinical outcome.5, 6 The disease process is complex, and secondary ischemia in the subacute phase is a major cause of morbidity. Monitoring for signs of ischemia in the neurointensive care unit is a prerequisite for rapid treatment. The ideal monitoring technique is minimally invasive, continuous, sensitive and specific. Part of this thesis aims to evaluate and validate microdialysis as a tool for monitoring different degrees of ischemia. Finally, to improve the treatment of critically ill patients with ischemic stroke and subarachnoid hemorrhage, in which primary and secondary ischemia are the dominating causes of morbidity and mortality, we need to widen our understanding of the pathophysiology of cerebral ischemia. This understanding is the basis for the correct application as well as the development of new monitoring technologies. Furthermore, the design of new treatment modalities involving neuroprotective drugs is based on the elucidation of the mechanisms of ischemia. The experimental model with PET and microdialysis that was used in this thesis was well suited to gain insights into different aspects of cerebral ischemia. We set out to explore the effects of cerebral ischemia on the hemodynamic, metabolic and chemical levels.. 8.

(22) REVIEW OF THE LITERATURE The aim of this review is to clarify certain aspects of cerebral ischemia that are relevant to the studies in the thesis, and it has no ambition of being complete. A few published reviews and especially relevant or representative studies are referred to. At first, the hemodynamics and the metabolic consequences of cerebral ischemia are outlined, concentrating on evidence from PET studies. The evolution of the concept of the ischemic penumbra is described. Evidence for thresholds of ischemia is reviewed, and different states of uncoupling between flow and metabolism are discussed. Second, a description of the main mechanisms of ischemia on the cellular level is attempted. Current concepts of cerebral energy metabolism under normal and altered circumstances are outlined. The significance of calcium, excitatory amino acids (EAAs), mitochondrial damage and membrane breakdown is described. Third, a short summary of the role of microdialysis in experimental and clinical research on cerebral ischemia is presented. Finally, the pathophysiology of subarachnoid hemorrhage with focus on cerebral ischemia is summarized.. The concept of the ischemic penumbra The occlusion of a major cerebral artery by an embolus, a local thrombosis or by an experimental intervention causes ischemia in the region that was perfused through the vessel. The degree of ischemia in the acute phase varies topographically within the region, in a typical pattern: the deep, subcortical structures with poor collateral blood supply forms the core of infarction, with markedly reduced perfusion causing irreversible ischemia. Typically surrounding the core and including the cortex, a region with less profound ischemia is found. This latter region is called the ischemic penumbra.7 The hallmark of the penumbra is its potential for survival. Thus, it is the target of numerous studies of cerebral ischemia with the aims of diagnostic imaging, of elucidating mechanisms for ischemia, of pharmacological intervention, etc. Astrup et al., studying reductions of cerebral blood flow (CBF) to levels around the loss of the somatosensory response originally used the term ischemic penumbra for describing a zone with functional inactivation, but not yet cell death. Their observation was that the critical ischemic threshold for electrical failure was higher than the threshold for structural integrity (loss of ion gradients).7, 8 Since then, many definitions of the penumbra have been used. In the early 1980:s, the emphasis was on determining the exact flow threshold for viability, and further on the importance of preserved metabolism as well as the duration of ischemia was stressed.9-11 The dynamic nature of the penumbra, i.e. that infarction may expand from the ischemic core into parts or all of the penumbra has been ascribed to biochemical mechanisms such as the influx of calcium, the production of free radicals, mitochondrial damage, the inhibition of protein synthesis and initiation of apoptosis.12 Recently, the term has been expanded to molecular biology, and zones of different gene expression or multiple molecular penumbras have been described.13. 9.

(23) Flow thresholds Early work on focal cerebral ischemia demonstrated a flow threshold for electrophysiological function (flattening of the EEG or abolished sensory evoked potentials).14, 15 A systematic study of neurological function in reversible or permanent middle cerebral artery occlusion (MCAO) in awake monkeys was performed by Jones et al.16 A negative correlation was found between residual blood flow (measured with a hydrogen clearance technique) and the severity of symptoms, and thresholds for functional integrity and infarction were found. Furthermore, the authors suggested that the development of infarction seems to be a function of intensity and duration of ischemia. This was later confirmed by PET studies of transient MCAO.17 Hossmann summarized the results from a large number of studies in different species, and expanded the threshold concept to include a full spectrum of pathophysiological changes (figure 1).11 Thus, protein synthesis begins to be inhibited at flow rates below 55 ml 100 ml1 min-1, glucose utilization and lactate accumulation increase at flow rates below 35 ml 100 ml-1min-1, while phosphocreatine and ATP levels decline with flow rates below 26 ml 100 ml1 min-1, as acidosis becomes severe.18 At around 23 ml 100 ml-1min-1, neurological dysfunction and suppression of EEG activity and evoked potentials appear. Finally, irreversible hemiparesis and infarction concomitant with terminal depolarization and potassium efflux and calcium influx ensue at flow rates between 5 and 18 ml 100 ml-1min-1 (depending on duration of ischemia and the species studied). In humans, it has not been possible to confirm this sequence of pathophysiological events with declining flow rates, for obvious reasons. Nor has the exact duration of critical hypoperfusion for irreversible ischemia been established. Therefore, these threshold values based on data from experiments with focal ischemia in animals should not be directly extrapolated to human physiology. But there is a growing body of evidence from PET studies in humans, that regions with blood flow below 12 ml 100 ml-1min-1 or 60 % of contralateral values usually evolve into infarction.19-21 Furthermore, in stroke patients,. CBF (ml 100 ml-1min-1). 60. Protein synthesis, selective gene expression. 50 40 30. Lactate production, increased glucose utilization Oligemia Acidosis EEG suppression, Neurol function. 20 Penumbra. Anoxic depolarization, ion gradients cease. 10 Infarction 0. Figure 1. Pathophysiological events with decreasing blood flow. (Modified from Hossmann11). 10.

(24) regions with CBF between 12 and 22 ml 100 ml-1min-1 usually display intact metabolism and will survive or become infarcted to a varying extent, thus fulfilling the criteria for penumbra.22, 23. Metabolic thresholds Normal brain function relies almost exclusively on the oxidative metabolism of glucose for energy production. As it became possible to study the metabolic rate of oxygen in conjunction with blood flow with PET, it was a natural step to examine minimum levels of oxyen consumption required for function and viability. Early studies in humans showed that the cerebral metabolic rate of oxygen (CMRO2) was below 56 60 mmol 100 ml-1min-1 or below 40 - 45 % of the contralateral region in infarction regions.24-26. These studies were performed with low-resolution PET scanners and in the subacute or chronic stages of stroke, but were confirmed by a later study.21 Powers and coworkers analyzed blood flow and metabolism in normal subjects and patients with vaso-occlusive carotid disease without infarction and found that CMRO2 in viable tissue was always above 58 65 mmol 100 ml-1min-1.26, 27 Ackerman et al. observed that regions with CMRO2 values < 67 mmol 100 ml-1min-1 always showed infarction on a late computed tomography (CT) scan. Touzani et al. addressed the question of the evolution of hypometabolic tissue in a MCAO primate model, and found that tissue with CMRO2 below 65 mmol 100 ml-1min-1 or below 40 50 % of contralateral values expanded with time, and ovestimated infarction size as compared with histopathological analysis.28 In the latter study, the 40% level showed the best correlation with histopathology. To summarize, there is as yet no conclusive data for an unequivocal absolute CMRO2 threshold of viability/ irreversible ischemia. From the studies reviewed above, we can infer that a threshold value of 58 67 mmol 100 ml-1min-1 is reasonably valid in the subacute phase, but may be higher in the acute phase. A corresponding relative threshold value, based on the ratio between ipsiand contralateral values could be 40 - 45 % in the subacute phase, but is perhaps higher in the acute phase. Uncoupling between flow and metabolism The oxygen steady-state method allows for nearly simultaneous measurements of CBF, CMRO2 and oxygen extraction ratio (OER). Early PET studies of cerebral ischemia made clear that uncoupling of CBF and CMRO2, resulting in either increased or decreased OER, is a common phenomenon. 24, 25 Baron et al. introduced the term misery perfusion for a focal mismatch resulting in an increased OER, and this term was in many studies used as a synonym for penumbra.29 The outcome of regions with maximal OER was variable in sequential PET studies with MCAO in baboons, but often a decrease in CMRO2 and OER with time was observed.30, 31 Similar findings have been reported in human stroke studies.21 The opposite phenomenon of unproportionally increased blood flow resulting in a reduced OER is also a common finding in cerebral ischemia.25 The presumed mechanism is a vasodilation due to acidosis or perturbed autoregulation after spontaneous or induced. 11.

(25) reperfusion. The luxury perfusion syndrome was originally described by Lassen in a case study,32 and this term has sometimes been used as a synonym for the phenomenon of postischemic hyperperfusion. There has been some controversy in the literature as to whether hyperperfusion is beneficial, or predicts a poor outcome. In general terms, postischemic hyperperfusion is the hallmark of effective recanalization in interventional studies, and as such a sign of the desired effect of therapy. On the other hand, postischemic hyperperfusion can cause critical brain swelling, especially in small animals such as cats, and has been shown to be associated with poor or variable outcome.17 Early postischemic hyperperfusion has been reviewed by Marchal et al.33 On the cellular level, reperfusion may have deleterious effect on the mitochondria, as is discussed below. The no-reflow phenomenon and delayed hypoperfusion In animal models of transient ischemia, many workers have observed that blood flow is sometimes not restituted following removal of the occluding intervention.This was originally described as the no-reflow phenomenon by Ames et al.34 The evolution of concepts of noreflow and delayed hypoperfusion have been reviewed by Ginsberg.35 The regions affected with no-reflow are often described as having a markedly heterogeneous distribution of blood flow, and there may be concurrent or preceding hyperperfusion. The distribution between animals may be somewhat random and unpredictable. The mechanism was first thought to be compression of the capillaries by edematous endothelial and glia cells, on the basis of electron microscopic studies. However, subsequent studies have shown that vessel occlusion by stagnant blood or overt thrombosis may be a more likely cause of no-reflow. Several workers have found evidence of increased platelet activity and infiltration of polymorphonuclear leukocytes in postischemic small vessels. In addition, pretreatment with indomethacin, heparin or hemodilution all reduced no-reflow and infarction volume. Calcium channel blockers such as nimodipine also reduced postischemic hypoperfusion. Electron microscopy has been used to show the occurrence of microvilli protruding from the endothelial surface in ischemic vessels. Available evidence thus suggests a prominent role for an inflammatory reaction triggered by cells damaged by ischemia. Cerebral energy metabolism in normal physiology Until recently, the widely held view was that the sole energy substrate for neurons is glucose, that crosses the blood-brain barrier to reach neurons through a membrane transporter. Glucose is metabolized to pyruvate in the process called glycolysis, yielding two ATP molecules. Pyruvate enters the tricarboxylic acid (TCA) cycle in the mitochondria, and an additional 34 ATP molecules are regenerated as the end product of oxidative phosphorylation. The driving force for the latter is the electrochemical gradient over the inner mitochondrial membrane produced by the electron transport chain when oxygen is reduced. ATP provides the energy for most of the active processes in the cell. The sodium-potassium ATPase activity that restores the cell membrane potential after depolarization is quantitatively the most important of these.36 Small quantitities of glycogen and phospocreatine may store substrate and energy within the cell, but these stores last only a few minutes of ischemia. 12.

(26) However, Fox and coworkers observed that the brain seems to resort to glycolysis rather than increased oxidative phosphorylation when challenged with functional activation.37, 38 Several studies have since confirmed this observation, and have added to our understanding of cerebral energy metabolism. We now know that lactate is an obligatory aerobic energy substrate in neurons,39 and a new concept of substrate flow has been proposed and substantiated by Magistretti and Pellerin.40, 41 Briefly, astrocytes sense synaptic activity when glutamate is released into the synaptic space (figure 2). Glutamate and sodium are taken up into the astrocyte, which stimulates uptake of glucose via the astrocyte´s podocytes on the local capillary. Glucose is processed glycolytically, yielding lactate as a substrate for oxidative phosphorylation in nearby neurons. Lactate may be shuttled directly to the neuron via monocarboxylate transporters that have been shown to be present in mouse brain. Evidence is accumulating that astrocytes are the source of lactate consumed by neurons.42 In addition, brain glycogen stores may be limited to perivascular astrocytes, supporting the concept of their nutritional role for neurons.43 However, this view has been challenged by Gjedde et al., who argue for a more complex model in which glycolysis and oxidative metabolism, as well as blood flow, play different parts depending on the level of functional activity in the brain.44. Neuron. Astrocyte. 4. 7&$F\FOH. 3. Lactate. *O\FRO\VLV. 5HVSLUDWRU\FKDLQ. Lactate + 2 ATP. 34 ATP. Synaptic cleft. Capillary. Gln. Gln. Glu. Glu. Glu. Glu. 2. Glucose. 1D.$73DVH. 1. Figure 2. Coupling of glutamate release to glucose metabolism in astrocytes and neurons. When the neuron depolarizes, glutamate is released into the synaptic cleft (1). This is rapidly cleared of transmittor via the efficient glutamate transporter of the adjacent astrocyte, driven by the electrochemical gradient of Na+. Glutamate (Glu) is converted to glutamine (Gln) in the astrocyte, released and taken up by the neuron to be recycled into glutamate (2). The uptake of glutamate into the astrocyte stimulates glucose uptake via glucose transporters on the end-feet covering the capillary (3). Each glucose molecule is converted to lactate through astrocytic glycolysis, yielding 2 ATP to drive the Na+-K+ -ATPase and glutamine synthase activity. Lactate is transported into the neuron, and used as a substrate in the tricarboxylic acid cycle (TCA), yielding 34 ATP after aerobic metabolism in the respiratory chain (4).. 13.

(27) Energy failure When blood flow is interrupted, the supply of glucose and oxygen (the substrates of aerobic metabolism) ceases. Stores of glucose and glycogen are scant in the brain, but residual glucose is metabolized anaerobically in both astrocytes and neurons, resulting in lactate accumulation and acidosis. Since lactate cannot be metabolized further due to lack of oxygen, energy failure and an altered redox state with accumulation of NADH ensue. The acidosis can be detrimental in itself, as it alters cell metabolism and mitochondrial membrane function. Pyruvate dehydrogenase is inhibited in mitochondria that have been ischemic, which contributes to prolonged acidosis, even if ischemia is transient. Interestingly, the pre-ischemic plasma glucose level is coupled to the degree of acidosis in ischemia.45 Preischemic hyperglycemia aggravates brain damage after transient ischemia, and the suggested mechanism is acidosis-mediated increase in free radical production.12, 35 On the other hand, in vitro recovery from transient ischemia is dependent on adequate glucose concentration in the perfusate.46 This may not be contradictory, but probably only reflects the fact that the post-ischemic cells need energy substrate to restitute membrane function. A correlate of the increased glycolysis in normal functional activation may be the PET finding of hyperglycolysis following traumatic brain injury.47 The role of calcium Calcium accumulates intracellularly because of membrane depolarization and ionotropic NMDA-receptor activation. The increase in the intracellular calcium level can be dramatic (from 0.1 to 30 mM in hippocampal CA1 neurons), and has a central role in both acute cell dysfunction or destruction and secondary brain damage, as has been reviewed by Siesjö and Siesjö.12 Lysis of membrane phospholipids (PLs) is stimulated, which leads to free fatty acid (FFA), especially arachidonic acid (AA) liberation and leucotriene production (figure 3). This attracts leucocytes, triggering an inflammatory response that may cause vascular occlusion. An array of calcium-dependent enzymes are activated, among them the nitric oxide synthases (iNOS, nNOS and eNOS). The production of nitric oxide may have beneficial effects, but in severe ischemia, when intrinsic scavengers such as superoxide dismutase are saturated, several radical oxygen species (ROS) may be unleashed, causing lipid peroxidation and damage to the cell membrane, mitochondria and DNA.48, 49 Calcium also induces the production of ROS by conversion of xanthine dehydrogenase to xanthine oxidase (XO).12 Protein phosphorylation is altered by increased calcium levels, and this and the stimulation of growth factors may lead to changes in gene expression and protein synthesis that finally result in apoptosis (programmed cell death).12 Mitochondrial calcium overload may cause the expulsion of cytochrome c, which ultimately causes DNA damage via interaction with caspases. Finally, intracellular calcium accumulation induces disaggregation of microtubuli and induction of the proteolytic enzyme calpain which acts on spectrin, a component of the cytoskeleton. The result is an inhibition of axonal transport, which contributes to calcium overload, mitochondrial damage and apoptosis.12, 50 14.

(28) Lipolysis. Enzyme activity. Gene expression. Proteolysis. FFAs, PLs, AA. XO, nNOS, iNOS,. Caspases. Calpain. ROS. DNA damage. Spectrin breakdown. PAF, Leucotrienes Mitochondrial damage. Vascular dysfunction, edema. Membrane dysfunction Cellular swelling. Apoptosis. Cytoskeleton damage. Inhibition of axonal transport. Figure 3. Calcium triggers multiple mechanisms leading to cell damage. (Modified from refs 12 and 49) See text for details and abbreviations.. Glutamate and excitotoxicity Glutamate is the principal fast neurotransmittor in the brain, acting on ionotropic receptors (the NMDA, AMPA, kainate and AP4 receptors). It is also involved in learning and memory through slow synaptic responses and modulation via metabotropic receptors. Aspartate, and perhaps other peptides may also act on some of the glutamate receptors. The ionotropic receptors incorporate ion channels, thus glutamate stimulation results in influx into the cell of sodium via the AMPA receptor channels, and calcium via the NMDA receptor channels. The metabotropic receptors, in a G-protein mediated second messenger system, elicit increased levels of inositol 3-phosphate (IP3) and diacylglycerol (DAG).51, 52 In ischemia, cellular depolarization and inhibited (energy-dependent) reuptake mechanisms result in a large increase in the extracellular glutamate concentration. This elicits all of the above mechanisms, and it is a key factor in the early increase of intracellular calcium concentration. In addition, an increased metabolic demand on the already compromised cell is created with the increased membrane permeability and the stimulation of intracellular ATP-dependent processes. The role of glutamate (and aspartate, both called excitatory amino acids) in ischemic brain damage is called excitotoxicity based on the early work by Olney, who observed neuronal damage after application of glutamate to mouse brain slices and brain damage caused by oral intake of glutamate.53, 54 Further evidence for the detrimental effect of excitotoxicity comes from experiments with NMDA antagonists and lesioning of excitatory pathways. EAAs also have an important role in the initiation of epileptic activity, which may further aggravate ischemic brain damage. Increased levels of glutamate have been observed in numerous studies of human cerebral ischemia using microdialysis.55-60. 15.

(29) Radical oxygen species (ROS) Oxygen (free) radicals are extremely reactive molecules implicated in ischemic brain damage, expecially reperfusion injury. In normal physiology, they are continuously formed in small amounts as biproducts of oxidative metabolism. Free radicals are also used by macrophages, T-killer cells and other leukocytes in inflammation and combatting infection. A carefully balanced system of prooxidants and antioxidants has been found at work in cerebral tissue, probably controlling the potential harmful effects in normal function. Manipulation with e.g. the level of an endogeneous antioxidant such as superoxide-dismutase has been found to influence the extent of cell death in experimental models of ischemia, recently by using transgenic mice, see review by Chan.49 In the classical view of cerebral ischemia, reperfusion of ischemic tissue leads to grossly enhanced ROS production (triggered by intracellular calcium via conversion of xanthine dehydrogenase to xanthine oxidase and activation of phospholipases). This causes activation of arachidonic acid cascades and lipid peroxidation, protein oxidation and DNA damage, leading to cell membrane damage, swelling and necrosis.12, 61 Recent studies have provided evidence that ROS also are involved in redox signal transduction pathways that cause cellular damage and apoptosis. One example of this is that nNOS may via peroxynitrite formation (or directly) signal the mitochondrial release of cytochrome c, activating caspases that cleave nuclear DNA, resulting in apoptosis.49. $SRSWRVLV Ca2+ taken up by intact mitochondria. Lipid peroxidation. ROS. Ca2+ induced expulsion of cytochrome c. ROS-mediated damage to the electron transport chain. Ca2+. 0LWRFKRQGULD Ca2+ overload causes membrane permeability transition. ROS production in disrupted electron tranport chain. Ca2+ released by damaged mitochondria. 1HFURVLV. Figure 4. The mitochondria and the double vicious circles of calcium and ROS accumulation in ischemia. In ischemia, the electron transport chain is inhibited. With reperfusion, or incomplete ischemia, the partially inhibited electron transport chain produces increasing amounts of ROS, that may severely damage membranes causing necrosis, or contribute to delayed mitochondrial damage and subsequent apoptosis. An ischemic increase in the intracellular calcium concentration is initially buffered by the mitochondrion. Accumulation of Ca2+ in the mitochondrion may induce the expulsion of cytochrome c, eventually resulting in apoptosis. With severe mitochondrial calcium overload, the mitochondrial membrane undergoes a permeability transition, and large amounts of Ca2+ may be released to the cytosol. This is taken up by remaining functional mitochondria that may subsequently be damaged, until total energy failure ensues, leading to necrosis.. 16.

(30) Mitochondrial dysfunction Mitochondria are thus both a target for, and a contributor to the production of ROS as well as intracellular calcium, in a double vicious circle (figure 4). The central role of mitochondria in cerebral ischemia has been elucidated by Siesjö and Siesjö, and more recently reviewed by Fiskum, Murphy & Beal.12, 50, 62 Brain mitochondria are sensitive to ischemic injury, and exhibit signs of impaired function after periods of moderately reduced blood flow, even when ATP and phosphocreatine levels are not significantly affected. In fact, incomplete ischemia may result in greater mitochondrial damage than complete ischemia, due to the immediate action of ROS. The consequences of mitochondrial injury after cerebral ischemia thus depend on the severity of ischemia and reperfusion, and include cellular energy failure (abolished ATP production), oxidative stress (ROS), exacerbation of excitotoxicity through impaired calcium buffering (mitochondria are normally an important part of regulating intracellular calcium homeostasis). Furthermore, calcium overload in mitochondria can cause mitochondrial swelling or opening of a non-specific proteinacious pore (the mechanisms have not been completely clarified) and membrane permeability increase, or if less severe, expulsion of cytochrome C, which interacts with caspases and results in apoptosis. With complete ischemia, different patterns of pathophysiological events have been shown to occur depending on the duration of ischemia. With short duration (up to 20 minutes), energy failure is reversible, but reperfusion triggers a cascade of events that lead to delayed cell death. In this setting, there is a differentiation of brain cells as to the sensitivity to ischemia, reflected in the tendency to develop apoptosis. With up to a few hours of complete ischemia, partial irreversible injury to the electron transport chain is seen, exacerbated by reperfusion. With longer duration of severe ischemia, mitochondrial function is completely disrupted, and cells rapidly develop necrosis. Cell membrane phospholipid breakdown Degradation of cell membrane phospholipids is one of the earliest biochemical events that occur as a consequence of cerebral ischemia.63 Free fatty acids, lysophospholipids and DAG accumulate, while the total phospholipid content of the ischemic brain decreases.52, 63, 64 Diacylglycerol is further degraded to glycerol and arachidonic acid. 65 The cause of phospholipid degradation in ischemia is the glutamate-mediated activation of phospholipases, via pathways that normally are important in transmembrane signal processing via metabotropic glutamate receptors ( figure 5). One effect of this activation in ischemia is an augmentation of intracellular calcium release from intracellular stores. Another effect is the production of ROS by oxidation of arachidonic acid.61 Arachidonic acid also inhibits presynaptic glutamate uptake, creating a vicious circle of excitotoxicity. Paschen, et al. (1986) suggested that glycerol could be used as a postischemic indicator of the severity of brain damage.65 Subsequent studies have corroborated the hypothesis that glycerol reliably signals phospholipid degradation, but yielded partly conflicting conclusions about the reliability of glycerol as a marker of ischemia.66-68. 17.

(31) Phosphatidyl - X 3/$ 3/$ (activated by intracellular Ca2+ or PAF). 3/&. /\VRSKRVSKROLSDVH. Glycerolphosphoryl - X. Diacylglycerol. *O\FHUROSKRVSKRU\OGLHVWHUDVH. Glycerol - 3 - phosphate 3KRVSKDWDVHV. Glycerol. Glycerol. Figure 5. Pathways in phospholipid metabolism. The two parallel biochemical pathways in which membrane phospholipids (Phosphatidyl-X) are degraded to glycerol . Phospholipids are liberated from the cell membrane through the action of phospholipases in normal receptor signal processing. Further degradation to glycerol occurs in the cytosol. In ischemia, these processes are augmented by e.g. Ca2+ overload. (Modified from ref 65.) PLA1 - phospholipase A1, PLA2 - phospholipase A2, PAF - platelet activating factor, PLC - phospholipase C. Microdialysis studies of cerebral ischemia There has been an increasing number of studies of different aspects of cerebral ischemia since 1984, when the first microdialysis paper on experimental cerebral ischemia was published.69 Microdialysis has been used for studying the mechanisms of ischemia, such as excitotoxicity, free radical damage, and membrane degradation.69-72 New pharmacological treatment modalities have been evaluated in several experimental models of ischemia.73-75 Microdialysis, with the advantages of sampling in vivo and continuously, has been beneficial in the understanding of altered glucose metabolism and purine metabolism.76-79 Finally, the potential of microdialysis for monitoring cerebral ischemia has been investigated in MCAO and TBI models.67, 77, 80 Microdialysis was first used in the human brain during neurosurgical procedures, demonstrating the feasibility and indicating the potential of the technique.59, 81 The use of microdialysis in humans has been reviewed by Hillered and Persson, and by Hamani et al.82, 83 Persson and Hillered first applied microdialysis for long-term monitoring of the brain in neurointensive care, and reported fluctuations of energy-related metabolites and EAAs that parallelled secondary ischemic events.60 This finding was corroborated by several studies in head injured patients. Bullock et al. found marked glutamate increases that were transient or prolonged, depending on the occurrence of secondary ischemic events.84 In addition, extracellular potassium levels increased in patients with very high glutamate levels, suggesting ion flux due to membrane damage. Goodman et al. found that increased lactate and decreased glucose levels, as an indicator of accelerated glycolysis, was associated with poor outcome in a larger study of head injured patients.85 Numerous studies in SAH patients have confirmed 18.

(32) that increases in energy-related substances and EAAs are associated with ischemic events. Säveland et al. demonstrated increases in EAAs in localized regions with vasospasm.86 Enblad et al. correlated changes in energy-related metabolites and glutamate to CBF and CMRO2 determined by PET.55 Hillered et al. introduced glycerol as a marker of phospholipid degradation in SAH patients.68 Schultz et al. and Reinstrup et al. have attempted to determine reference values of glucose, lactate, pyruvate, glycerol and glutamate for normal and severely ischemic metabolism.87, 88 Cerebral ischemia in the pathophysiology of aneurysmal subarachnoid hemorrhage Subarachnoid hemorrhage is a complex disease, in which cerebral ischemia is a dominating problem. SAH is usually caused by the rupture of an arterial aneurysm in the cerebral circulation. The initial rupture is a catastrophic event, leading to immediate death in perhaps 10 %, prolonged drowsiness or focal deficits in 37 %, and a brief period unconsciousness in 32 % of cases.89 The hemorrhage causes an acute rise in intracranial pressure (ICP), which results in transient or prolonged global ischemia due to a decrease in cerebral perfusion pressure. The degree of altered consciousness and focal deficits seems to influence outcome, which is reflected in the Hunt & Hess and World Federation of Neurological Surgeons (WFNS) grading systems for SAH.90, 91 Survivors of the primary ischemic event still suffer a substantial risk of morbidity and mortality due to secondary pathophysiological events or complications that may occur in the weeks following hemorrhage. Rebleeding is the most feared complication in the course of the disease. It occurs in 20 % of unoperated patients, and it is frequently fatal.92 Cerebral edema may contribute to ICP elevations and hemispheric shifts. Clinically significant hydrocephalus due to increased resistance to outflow in the ventricular system often complicates the further clinical course (15 20% of cases). Seizures are common in the early phase of SAH (10 26% of cases). Epileptic activity may aggravate ischemia by increasing metabolic demand in a state of compromized blood flow.92 Ischemic brain damage is a major factor influencing the morbidity and mortality of SAH. The conventional view is that secondary focal ischemia is caused by vasosopasm, which is a common complication to SAH. It typically occurs within the first two weeks, but rarely in the first days of the disease. Transient or permanent focal neurological deficits may occur in conjunction with vasospasm, but there is no evidence of a causal relationship, and these symptoms may occur without radiological signs of vasospasm. Thus, it has been suggested that secondary events such as hypotension, hypoxia, hypercarbia and intracranial hypertension may be as important in the pathophysiology of secondary ischemia in SAH.93 The concept of avoidable factors was introduced more than twenty years ago to emphasize that secondary insults after head injury were common and had a devastating influence on outcome.94 Enblad and Persson suggested that the concept also should apply to subarachnoid hemorrhage.93 Thus, factors such as hypoxemia or systemic hypotension (see Table 1) are often avoidable, and influence outcome in SAH. Results from animal experiments support. 19.

(33) this concept.80 This is the essential reason for treating SAH patients in a dedicated neurointensive care unit, beginning as early as possible after the primary ictus.95 In fact, a major avoidable factor may be the delayed referral of SAH patients to such a unit.96 However, it remains to be shown why SAH patients are exceptionally prone to develop secondary ischemic brain damage. Although several studies have addressed the problem in relation to vasospasm,97-100 the early changes in blood flow and metabolism occurring in the acute phase of SAH have not been described, with the exception of a few studies.99, 101 We thought that a PET study of such early changes could give important insights into the pathophysiology of SAH, with focus on conditions that may contribute to the development of secondary ischemia.. Table 1. Avoidable factors after SAH Hypoxemia Hypercapnea Systemic hypotension Hyperpyrexia Hyponatremia Epileptic activity ICP elevation due to ICH ICP elevation due to hydrocephalus. 20.

(34) AIMS General aim The main objective of this thesis is to gain insight into the pathophysiology of cerebral ischemia by studying early hemodynamic and metabolic changes with PET and neurochemical changes with microdialysis. Specific aims To describe the early hemodynamic and metabolic changes due to transient focal ischemia (MCAO reperfusion). (Paper I) To evaluate if early PET measurements of CBF, CMRO2 and OER can predict the fate of ischemic brain tissue. (Paper I) To relate changes in known neurochemical ischemic markers - the lactate/pyruvate (LP) ratio, hypoxanthine and glutamate directly to the degree of ischemia during transient MCAO as defined by PET measurements of CBF, CMRO2 and OER. (Paper II) To evaluate extracellular glycerol as a marker of ischemia by studying the changes in glycerol in direct relation to the changes in PET measurements of CBF, CMRO2 and OER and other microdialysis parameters (LP ratio and glutamate) during transient MCAO. (Paper III) To study changes in extracellular glucose in relation to extracellular lactate and PET measurements of CBF and CMRO2 during ischemia. (Paper IV) To validate microdialysis as an instrument for monitoring of impending cerebral ischemia. (Papers II, III and IV) To investigate the occurrence of early hemodynamic and metabolic changes in the acute stage of subarachnoid hemorrhage. (Paper V). 21.

(35) METHODS Experimental animal model - middle cerebral artery occlusion (MCAO) with reperfusion Eight adult Macaca Mulatta monkeys, weighing between 4.4 and 12.8 kg, were used for the MCAO studies (papers I IV). The reasons for using monkeys instead of smaller animals were threefold. First, an important goal when conceiving the studies was devising an experimental model in which comparisons could be made between data from irreversibly ischemic tissue with data from penumbra. The resolution of the PET scanner, as well as the need for placing two microdialysis (MD) probes into anatomically distinct regions of each brain thus demanded the larger size of the primate brain. Second, a few studies of PET changes after permanent MCAO in monkeys had been published, but there was a need for elucidating changes occurring also in the reperfusion phase, as the concept of reperfusion injury had become the focus of much current research. Third, the model was designed to validate microdialysis for monitoring of the human ischemic brain.60 The study protocol was approved by the local ethical committee for animal research (permissions c 38/96 and c 37/ 97). The animals were sedated with intramuscular ketamine before transport to the PET centre. Intravenous (i.v.) lines were placed and 5 ml/kg of dextran solution was administered for volume expansion. Anesthesia was induced with propofol, and maintained with a continuous infusion of morphine, midazolam and atracurium. During surgery, additional doses of the midazolam/morphine solution were injected when needed. Total intravenous anesthesia was chosen instead of inhalational agents, since the latter may cause an uncoupling between CBF and CMRO2 at high doses. Basal glucose, electrolyte and fluid requirements were maintained. A femoral artery was catheterized for blood pressure monitoring and blood sampling. Continuous arterial blood pressure, electrocardiography, pulse-oximetry, end-tidal-CO2, intracranial pressure (ICP) and rectal temperature were monitored, and blood gases were analysed intermittently. A controlled heating mattress and a humidifier in the ventilatory circuit were used to maintain normothermia. The animals were killed at the end of the experiments by an i.v. injection of KCl solution. The surgical procedure started with the insertion of two MD probes into the right hemisphere, aiming for the deep and superficial middle cerebral artery (MCA) territory, respectively, and using a stereotaxic technique. The objective was to place one probe in the expected core of infarction, and one in the cortical penumbra. A right sided transorbital MCAO was performed as described by others.102, 103 A Mayfield aneurysm clip was used for occluding the artery, since it has a reduced closing force and smooth blades which makes damage to the vessel wall unlikely. After two hours of MCAO, the clip was removed to permit reperfusion. The duration of MCAO was based on previous experience from studies in primates and our pilot experiments.16, 104 This model was also chosen since it resembles stroke of a few hours duration, with spontaneous or therapeutic reperfusion.. 22.

(36) Patients Eleven patients (two men and nine women) with subarachnoid hemorrhage were studied in Paper V (table 2). They were selected from a larger group of SAH patients that had been studied with PET at different stages in the disease process. Included were patients in which a complete PET investigation comprising cerebral blood volume (CBV), CBF, CMRO2 and OER had been performed in the acute phase within two days of admittance to our institution. The initial neurological condition of the patients varied from alert and without focal neurological deficits (WFNS, grade 1) to unconscious with hemiparesis (WFNS grade 5). The initial CT scans were graded according to Fisher105 as grade 2 4, i.e. from a small rim of subarachnoid blood to large hemorrhages with intraventricular blood. The patients thus represent the varying degrees of severity with which SAH patients present (with the exception of death or coma with dilated pupils). The patients were managed according to standard neurointensive care principles for SAH patients.95 PET scanning was completed within 10 40 hours of admittance and within 22 to 53 hours of ictus. Surgical aneurysm clipping was performed in the early stage in nine patients, in the late stage in one patient, and not at all in one patient who was in a poor clinical condition. The PET procedure always preceded the surgical procedure. Four of the patients were sedated with propofol during PET examination, the rest were unsedated. Six unsedated healthy male subjects served as controls. Positron emission tomography Positron emission tomography is a non-invasive and quantitative technique that permits the measurement of radioactive tracer concentrations regionally. The underlying principle for measurement is that with the decay of radioactive tracer, a positron is emitted; it travels an infinitesimal distance through tissue to collide with an electron. The collision results in the production of two photons of the same wavelength. These leave the tissue at opposite directions, and are simultaneously detected by two of a set of detectors arranged as a ring around the subject. A computerised procedure calculates a value that represents the number of photons emitted from each pixel (a pixel is essentially a volume unit) of the investigated region, and an array of pixel data is produced (emitted photons per second and unit volume, or activity concentration). Depending on by what physiological principle the tracer is distributed in the tissue, this data array will after modification represent e.g. regional blood flow. However, a mathematical model is necessary to translate the pixel data into physiological data. In this thesis, 15Oxygen (15O), with a half-life of 123 seconds, is the tracer that was used. It was incorporated into carbon monoxide (C15O), carbon dioxide (C15O2), water (H215O), or molecular oxygen (15O2) for studies of cerebral blood volume, cerebral blood flow, cerebral oxygen metabolism and the oxygen extraction ratio. The main mathematical model used was the oxygen steady-state method. The 15O steady state method was conceived in 1969 by TerPogossian, et al.106, 107 After PET technology had become available in 1978, the method was refined and adapted for quantitative measurements of CBF, OER and CMRO2 by workers at Hammersmith Hospital in London, England.108-111 In paper V, the steady-state measurement of CBF is exchanged for the autoradiographic method to reduce the total radioactivity dose. The latter PET method was devised by Herscovitch et al, who modified the classic autoradiographic method of Kety et al.112, 113 In the following sections, the objective is to 23.

(37) 2 2 1 1 2 2 4 4 4 2 5. 4 3 4 2 4 4 4 4 4 4 4. Aneurysm location 2 2 3 7 2 3. Aneurysm repair (d). 3 2 2 16. R ACA*+L MCA L ICA R ICA + R MCA Basilar R MCA AComA Not diagnosed AComA+L MCA AComA L ICA PICA+R ICA. Table 2. Patient characteristics. 51 46 64 37 49 59 69 54 70 49 66. Case Age WFNS Fisher grade grade 1 2 3 4 5 6 7 8 9 10 11. yes. yes yes yes. GR GR GR GR GR MD SD SD GR GR D. PET time Sedation GOS1 (h from ictus) 24 53 48 22 30 48 23 36 32 32 31. Follow-up CT Infarction location None L frontal + thalamic, lacunar Missing CT L basal ganglia, lacunar R temporal R frontal + L frontal None R temporal + L frontal R parietofrontal + L frontal None Dead. * Pericallosal artery 1 GOS - Glasgow Outcome Scale. GR - good recovery, MD - moderate disability, SD - severe disability, D -dead. GOS evaluated at 3 months after SAH.. 24.

(38) describe the practical procedure for a complete PET session as performed in Papers I and V, as well as to describe the physiologic principles involved and finally, to summarize the theory for analysis to give an idea of the calculations that are performed. rCBV (Papers I V) The main reason for wanting to measure cerebral blood volume is that it is used for correcting subsequent 15O2 measurements for the presence of activity bound in blood. Without this correction rCMRO2 is severely overestimated in areas close to large vessels. The basic principle for measurement of cerebral blood volume is that inhaled carbon monoxide rapidly equilibrates in the total erythrocyte pool of the blood, since it binds tightly to hemoglobin. This allows for a simple model and procedure. C15O is administered through the endotracheal tube during a few breath cycles (a maximum of 30 seconds, typical dose 50 100 mCi) until the count-rate of the scanner is 30-50 kilocounts per second, after which the administration is discontinued. After a delay of two minutes for equilibration, the scanner is started and data collected for five 1-minute frames. The five frames are decay corrected and averaged. Two arterial blood samples are taken 1.5 and 3.5 minutes after the start of scanning, and are decay corrected. The activity concentration of the blood samples is measured in a well-counter that has previously been cross-calibrated against the scanner. The blood volume is calculated as described by Martin and coworkers114 using the following model: Ct = rCBV 兰Ca(t)dt R ρa in which Ct is the activity concentration in each pixel during the scan, Ca is the blood activity integrated over the time of the scan, R is the ratio of cerebral small vessel hematocrit to large vessel hematocrit and ρa is the density of blood. The value of R is assumed to be 0.85, ρa is 1.05. rCBF measured with the steady-state method (Papers I-IV) Trace amounts of C15O2 are continuously delivered through the breathing system. In the presence of carbonic anhydrase in the pulmonary capillaries, the 15O label is transferred to water. H215O is thus distributed throughout the systemic circulation, and equilibrates dynamically with water in the cerebral tissues after three or four half-lives of 15O, i.e. within approximately six to eight minutes. At steady state, the supply of H215O from arterial blood is balanced by losses through radioactive decay and washout with venous blood. Thus, if rCBF is the regional blood flow, Cwa and Cwt represent the concentrations of H215O in arterial blood and brain tissue, respectively, and λ is the decay, rCBF Cwa = λ wt + rCBF Cwt. Rearranging, rCBF(Cwa Cwt) = λ Cwt, or rCBF = λ / (Cwa / Cwt 1). Cwt in each pixel is measured by the PET detectors, and Cwa in blood samples from an arterial 25.

(39) line is measured in a well counter. rCBF measured with the water bolus method (Paper V) This method is based on a modification of the tissue autoradiographic method for the measurement of rCBF, in which the equation below describes the relation between tissue activity concentration and CBF. Ct is the calculated tissue activity concentration, and Cwa is the measured arterial blood activity concentration. P is the blood brain partition coefficient for water. All activity values are decay corrected to a common time. Ct = 兰C(t)dt = rCBF兰Cwa(t) e-(rCBF/P)t dt A bolus of H215O in 4 ml of saline is injected i.v. as the PET scan starts. Arterial blood is sampled automatically (frequency 1 Hz), and Cwa is measured.115 The calculated tissue activity concentration according to the equation above, may be plotted against time for inserted values of rCBF. The next step is to plot the calculated activity for an integration time of 45 seconds against rCBF. The true rCBF value for each pixel may then be obtained by inserting the measured tissue activity concentration (Cwt, from the PET scan data array) into the latter plot. Luckily, these three operations are performed by the computer. rOER and rCMRO2 (Papers I V) The breathing system is fed with 15O2. Inhaled labeled oxygen binds to hemoglobin in the pulmonary capillaries and is extracted to varying degrees in the peripheral tissues. Oxygen functions as an electron acceptor in the mitochondrial respiratory system, and is metabolized to water. The concentration of H215O in the cerebral tissue (Ct) is thus a function of the uptake of 15O from arterial blood, which can be expressed as OER. Steady-state is achieved after about three half-lives, i.e. 6 minutes. At steady state, input again balances output (Coa is the concentration of 15O in arterial blood): rOER rCBF Coa = λ Cot + rCBF Cot in which rCBF is obtained from the preceding scan. However, a significant amount of H215O is produced by metabolism in other parts of the body, and is recirculated to the brain, and taken up into cerebral tissue as above. This recirculated tracer normally contributes to up to 30% of the count rate in the scanner. In addition, the 15O in the cerebral blood vessels emits some positrons that falsely contribute to Cot, as measured by the scanner. The previous CBV scan is used for correction for this intravascular 15O. After correction for recirculated H215O and intravascular 15O and a fair amount of algebra (see references 108 110): rOER = [(Cot/Cwt . (Cwa/Cop Cwa/Cwp) / (Coa/Cop Cwa/Cwp) X] / (1 X) in which the factor X is introduced for the correction for intravascular oxygen. X = (rCBF/ 26.

(40) 100 + λ) / (rCBF/(R rCBV) + λ). C signifies radioactivity concentration during CBF scan (w subscript) or oxygen inhalation scan (o subscript). The subscripts t, a and p designate activity in brain tissue, arterial blood or plasma, respectively. If the above equations are scrutinized, it is realized that several blood and plasma samples from the consecutive PET procedures (as well as a rigorous sampling procedure) are required to correctly calculate rOER. Finally, and to the relief of all, rCMRO2 is obtained by simply multiplying rOER with rCBF and the arterial oxygen content. Patient safety There are potential hazards to patient safety with the oxygen steady-state method. Since labelled carbon monoxide is used for obtaining measurements of CBV, the risk of CO toxicity must be taken into account. With the steady-state method, as long as inhalation time is limited to seven minutes, which is the maximum required for equilibrium and data acquisition, blood levels are well below toxic doses. In our studies, a single breath technique was used to limit exposure to CO further.114 Another potential risk is tissue exposure to the radioactivity of the tracers. This problem has been addressed by Deloar and coworkers, who estimated radioactivity uptake in various organs, and found (not surprisingly) that the critical organ in terms of radioactivity load is the airway, which typically is exposed to less than 0.03 mGy in a complete PET session with inhalation of C15O, C15O2 and O2.116 This is well below doses considered to be mutagenic, and again exposure was less in the patient study since the single breath method for CBV and the autoradiographic method for CBF were used. However, strict routines have to be used to minimize the exposure to both patients and personnel from other sources such as transmission scan radioactivity and contamination of the ventilator circuit (e.g. water in humidifier), apart from practising an efficient emission scan protocol. Technical limitations and potential errors The limitations of various PET methods used in cerebral ischemia research has been reviewed by Baron et al.117 For CBV measurement, it is important to allow time for collection of adequate number of counts, since the distribution volume is small. The CBV value is dependent on an assumed value for the ratio of regional cerebral to large vessel hematocrit. This ratio is likely to be variable especially in ischemic regions, an issue that has yet to be resolved, since there is no method to assess the true value. Furthermore, mostly due to partial volume effects, rCBV values in the vicinity of the large cerebral arteries and sinuses will be overestimated. In paper V, the vascular region VOIs were adjusted to avoid this effect. We have avoided basing conclusions on the analysis of absolute CBV values in all our studies, since all the errors described above can cause spurious regional values. On the other hand, in correcting for intravascular activity in calculations of OER and CMRO2, these errors become insignificant.109 For measurement of CBF with the steady-state method, the main limitation is that the relationship to tissue radioactivity is non-linear, which causes an underestimation of rCBF in high flow regions, as well as mean values from heterogeneous tissue. In low flow states, however, the model is more accurate. It is important to deliver tracer at a constant rate to 27.

(41) achieve steady-state rapidly, which can be a problem with spontatneous breathing. However, this problem was overcome in the animal experiment by using controlled ventilation, and in the patient study an autoradiographic method was used. This latter method is sensitive to time shifts between brain and blood curves and dispersion of cerebral input function. All uptake curves were evaluated for these errors before accepting the CBF scans. A major limitation in the calculations of OER and CMRO2 is that the complete procedure takes up to one hour. It is of course important that the actual CBF and CBV levels are stable during the procedure, since corrections for recirculated water and intravascular oxygen are based on the latter values. Meticulous care must be taken to avoid fluctuations in pCO2, for instance. Furthermore, errors in CBF measurement will inevitably be propagated into the calculation of CMRO2, which will be underestimated in regions with high blood flow and heterogeneous regions, as shown above. Multiple blood and plasma sampling is required for calculation of rOER, and this implies a high risk of measurement error or failure. Finally, since calculations are performed pixel by pixel, head movements during the procedure are not tolerated. In practice, movements can be compensated for by using an extension of the method described below. Problems that arise with intraindividual comparisons In Papers I - IV, the regional evolution of ischemia was followed in multiple PET sessions. It was difficult to place the animals head in exactly the same position in the scanner after it had been moved for a surgical procedure. The animal was moved twice after the baseline run: for the MCAO surgical procedure, and for removal of the clip. Thus two reorientations had to be performed, using a method for computerized realigning of consecutive PET scans that has been developed at the Uppsala University PET centre.118 In this method, an array of maximal pixel variations using a fraction of the total volume of data is created for the reference image, typically the CBF scan from the first session after clip removal. A similar array is created for the baseline session, compared to the reference array, and a reorientation matrix is created. This is subsequently used for mathematical reorientation and reslicing of all the scans of the session. An identical procedure is performed for the occlusion session. The resliced images are compared manually in three dimensions to check if the reorientation procedure was successful. PET data analysis in the MCAO model (ROI strategy) A penumbra region was delineated on the MCAO PET scan. The penumbra was operationally defined as a region in which all pixels in the OER image were more than 125 % of the contralateral hemisphere mean value and CMRO2 was more than 45 % of the contralateral value.21, 55 A probable infarction region was delineated in the last PET session based on the criterium of CMRO2 < 45 % of the contralateral hemisphere mean value.21, 55 At the end of the experiment, the microdialysis probes were perfused with a radioactive medium. Circular ROIs were delineated around the probes in the most basal slice showing activity. All ROIs were mirrored into the corresponding contralateral region, so that ipsi/contralateral ratios of the PET parameters could be calculated.. 28.

(42) PET data analysis in the patient study Standardized regions of interest (ROIs) with the dimensions 10 x 26 mm were outlined in the different vascular territories. To obtain more representative values, volumes of interest (VOIs) were produced by stacking ROIs from two adjacent slices in each vascular territory. Global cortical values of CBF, CMRO2, and OER were obtained by calculating the mean values of all cortical volume of interests (VOIs) in the respective scans. The vascular territory VOIs were categorized according to their pathophysiologic characteristics modified from Baron and Astrup et al.8, 119 First, regions were categorized as irreversible ischemia if the mean CMRO2 value was less than the established metabolic threshold of 60 mmol 100ml-1min-1.9, 120, 121 Second, an operational definition for penumbra was set up, based on criteria for each of the three PET parameters: 10 < CBF < 22 ml 100ml-1min-1 and 60 < CMRO2 < 146 mmol 100 ml-1min-1 and OER > 51.2%. The flow limits were based on the classic functional threshold values9, 120 and the latter two criteria were based on the controls´ mean values of CMRO2 3 SD and OER + 3 SD, respectively, i.e. a significantly increased OER and significantly reduced CMRO2, but above the metabolic threshold for irreversible ischemia. Third, oligemia was defined as CBF < 38.2 ml 100 ml-1min-1 (the controls´ mean CBF value 3 SD). Fourth, hyperperfusion was defined as CBF > 71.2 ml 100ml-1min-1, i.e. above the controls´ CBF mean + 3 SD. The remaining VOIs were considered normal. Infarction ROIs were transferred from the 3 month follow-up CT to each corresponding slice on the PET images for analysis of the outcome of the vascular region VOIs. Tissue outcome in the vascular region VOIs was defined as normal (completely separated from the CT infarction VOI), mixed infarction (partially overlapping the infarction VOI or infarction (within the CT infarction VOI). Microdialysis Microdialysis is a technique to monitor the chemistry of the extracellular space (i.e. the interstitial fluid) in living tissue. The principle was originally described by Bito et al. and it was developed by Delgado et al, but its modern application has primarily been refined by Ungerstedt and colleagues.122-127 The understanding of the kinetics in microdialysis probes has been advanced by Lindefors and coworkers.128, 129 Theoretical and practical aspects as well as limitations of the technique have been reviewed by Benveniste.130 A double-lumen probe with a semi-permeable membrane is implanted into the cerebral tissue. When a physiological salt solution is slowly pumped through the microdialysis probe the solution equilibrates with the surrounding extracellular fluid (see figure 6). The equilibrated solution is then pumped back to collecting vials which are changed regularly, e.g. every fifteen minutes. The fluid in the vials can be frozen for later analysis. The insertion of the MD probe implies a risk of tissue trauma. The risk of hemorrhage or infarctions around the MD probe is very low, and microdialysis can be regarded as a safe technique.82 Nevertheless, a disruption of the blood brain barrier is unavoidable following insertion, and the first samples collected will be unreliable due to the contamination of the interstitial fluid with blood, as well as by transient ischemia in the local tissue.130 In our studies, baseline values were calculated from the samples collected two hours immediately 29.

(43) MD perfusion fluid Dialysate to collecting vial. Shaft. Membrane. Figure 6. Schematic drawing of a microdialysis probe. The probe is constructed as a concentric tube in which the perfusion fluid enters through an inner tube; flows to its distal end; exits the tube and enters the space between the inner tube and the outer dialysis membrane. This is where the dialysis takes place, i.e. the diffusion of molecules between the extracellular fluid and the perfusion fluid. The dialysate is transported within the outer tube and is deposited into collecting vials for later analysis. Shaft length can be varied according to how deep into the brain the probe is to be inserted. No exchange of substances takes place in the shaft, which thus contributes to dead space.. preceding experimental MCAO. Since the surgical procedure started with probe insertion, more than one hour had always passed before collection of baseline samples. This allowed for normalization of the implantation artifact.59 Microdialysis does not provide true samples of the extracellular fluid. The concentration of a solute in the dialysate is a fraction of the true interstitial concentration, unless a very large membrane surface and/or a very slow perfusion flow is being used. Therefore, the probe has to be calibrated before use in vivo. The recovery is defined as the ratio between the concentration of a solute in the outflow solution and its concentration outside the probe. Recovery is measured by placing the probe in a solution with known concentrations of the substances for which it is to be calibrated. The probe is perfused at a constant flow with an identical solution except for the calibration substances which, of course, are omitted. The absolute recovery (mol/time unit) of a substance from the tissue depends mainly on the cut off of the dialysis membrane, the length of the membrane, the flow rate of the perfusion fluid and the diffusion coefficient of the compound through the extracellular fluid. The relative recovery in vitro (RRin vitro) is the ratio of the concentrations in the outflow (Cout) and the calibration solution (Ccs), respectively. RRin vitro = Cout/Ccs After calibration and start of in vivo microdialysis, the true interstitial concentration (Cis) of the substance can be calculated by dividing the outflow concentration with the recovery in vitro. Cis = Cout / RRin vitro 30.

(44) Table 3. Factors affecting recovery Flow rate Time after start of perfusion Diffusion coefficient Temperature Membrane area Composition of perfusion fluid Substance concentration in tissue (only influences the absolute recovery). However, the relative recovery in vivo is influenced by several additional factors summarized in table 3. It is important to consider factors such as flow rate and membrane characteristics when comparing absolute values of brain interstitial composition between different studies. Ideally, the MD probe should also be calibrated in vivo, since the diffusion characteristics of many substances will depend on the composition of the fluid that is investigated. In addition, this composition is bound to change with time, especially in extreme pathophysiological situations such as cerebral ischemia or traumatic brain injury. Unfortunately, there is as yet very limited data available on the diffusion characteristics of cerebral tissue in different pathophysiological states. Previously, there was no available method for in vivo calibration except for the zero net flux method.131 This method can only be used for steady-state measurements. Recently, promising results with simultaneous measurements of urea in microdialysis samples and subcutaneous tissue were reported.132 This method was not available at the time of the study. In our investigations, we have used a commercially available probe (CMA/10, CMA/ Microdialysis, Stockholm, Sweden) with a cut off level of 20 000 Daltons, and a membrane length of 4 mm. The flow rate was 2 µl per minute and the perfusion medium was designed to mimic normal cerebrospinal fluid (CSF, concentrations of Na+, 148 mM; Ca2+, 1.2 mM; Mg2+, 0.9 mM; K+, 2.7 mM and Cl-, 155 mM). Probe shaft length was 50 mm and 20 mm, which resulted in dead space volumes of 13.6 ml and 7.6 ml for the deep and superficial probes, corresponding to a time lag of seven and four minutes, respectively. Microdialysis sampling and data analysis Sampling in fifteen minute fractions started after probe insertion, typically three to four hours before MCAO. After the occlusion period and approximately six hours of reperfusion, samples were collected in 60 minute periods. The changes in the absolute values of the lactate/pyruvate (LP) ratio, hypoxanthine and glutamate with time were studied in paper II. In paper III, the absolute changes in glycerol was studied, and the temporal patterns observed were compared to those of the LP ratio and glutamate. In paper IV, the changes of glucose, lactate and the lactate/glucose (LG) ratio were analyzed relative baseline values. For comparisons with baseline, and for the statistical analyses, mean values from two hour periods throughout the experimental period were calculated. The precise location of each MD probe in the PET scans was identified by perfusing the probe with a radioactive solution at the end of the experiment, and running an emission scan. The microdialysis data was grouped into the categories of severe ischemia and penumbra, depending on if CMRO2 in the region surrounding the MD probe was below or above the metabolic threshold found in Paper I. In addition, the 31.

No results found